Compare commits

63 Commits

Author SHA1 Message Date
mix 2f888dc7be ci: drop per-arch GHCR tags via GitHub Packages API
test / test (push) Has been cancelled
imagetools has no rm subcommand; the merge cleanup had been failing silently. Document Gitea-to-GitHub tag mirror pitfalls in development.md. CHANGELOG [1.3.0] updated.

Co-Authored-By: Cursor <cursoragent@cursor.com>
2026-08-15 22:57:34 +03:00
mix 5598b62e4e release: 1.3.0
test / test (push) Has been cancelled
Close Unreleased for security, tests, docs, and release CI; pin compose and docs to 1.3.0.

Co-authored-by: Cursor <cursoragent@cursor.com>
2026-08-14 21:51:29 +03:00
mix 32f5085c48 docs: record release CI workflow changes in [Unreleased]
The three post-1.2.5 CI commits that tie image publish to a published GitHub
Release and trim per-arch GHCR tags were missing from the changelog.

Co-authored-by: Cursor <cursoragent@cursor.com>
2026-08-14 21:47:09 +03:00
mix 5601f73622 ci: consolidate release workflow and trim GHCR tags
test / test (push) Has been cancelled
Run release.yml directly on release:published instead of a dispatcher
workflow. Remove per-arch tags from GHCR after the manifest merge so only
X.Y.Z remains visible to operators.

Co-authored-by: Cursor <cursoragent@cursor.com>
2026-08-14 21:44:33 +03:00
mix e17b1680bd ci: dispatch release build from release-on-publish workflow
test / test (push) Has been cancelled
GitHub recorded release published at 16:44:21Z but no release.yml run
started. A thin release-on-publish workflow listens for publish and starts
release.yml via workflow_dispatch (always creates a run). release.yml now
checks out vX.Y.Z from the version input, not main HEAD.

Co-Authored-By: Cursor <cursoragent@cursor.com>
2026-08-14 19:48:20 +03:00
mix 5b63da0e49 ci: trigger release workflow on published GitHub Release
test / test (push) Has been cancelled
Bare git tag pushes no longer start the image build; publishing a GitHub
Release does. workflow_dispatch still requires an explicit X.Y.Z version.
Docs explain draft vs published releases and that deleting a tag on GitHub
converts a published release back to draft.

Co-Authored-By: Cursor <cursoragent@cursor.com>
2026-08-14 19:43:06 +03:00
mix e9aaed1c7b docs: reconcile [Unreleased] with full git history since 1.2.5
test / test (push) Has been cancelled
Expand the doc-alignment entry to cover guide RBAC and restore Resync
wording; trim the P2 entry of claims superseded by that pass while keeping
its /license route-table addition with a cross-reference.

Co-authored-by: Cursor <cursoragent@cursor.com>
2026-08-14 19:40:46 +03:00
mix 1bd71c10d0 docs: align operator docs with routes, sessions, and restore
test / test (push) Has been cancelled
Fix architecture route RBAC, restore Resync, and session/password wording;
correct guide restore/session/PTR claims and README port-587 text; point
the CSRF ADR at authz.go for global route gating.

Co-authored-by: Cursor <cursoragent@cursor.com>
2026-08-14 19:36:24 +03:00
mix 74acf09212 docs: verify [Unreleased] against git since 1.2.5
test / test (push) Has been cancelled
Move restore Resync under Fixed to match e70ba90, and extend the restore
test bullet with the resync boot step and TestResyncAfterRestoreHealsDriftedMaps.

Co-authored-by: Cursor <cursoragent@cursor.com>
2026-08-14 19:30:12 +03:00
mix ba07decbb0 docs: bring [Unreleased] up to date with guide and review closure
Record the operator-guide restructure, merge the stale code-review plan
entry with its completion note, and align the restore Resync bullet with
the current backup package comment.

Co-authored-by: Cursor <cursoragent@cursor.com>
2026-08-14 19:24:29 +03:00
mix 4fc5cff266 docs(guide): add worked restore commands to Full backup and restore
test / test (push) Has been cancelled
The section explained what a restore means conceptually but gave no
runnable commands. Add a step-by-step in-place restore, a
move-to-a-new-host variant, and the encrypted-backup decrypt-first
sequence, plus the actual version-mismatch error text so operators
know what to expect.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
2026-08-14 19:18:54 +03:00
mix ac8c2af500 docs(guide,architecture): move internal env vars out of the operator guide
test / test (push) Has been cancelled
Internal (non-operator) env vars were listed in guide.md's Environment
variables section even though the guide's own text says operators
aren't meant to touch them. They now live in architecture.md §
Configuration, the as-built/internals doc; guide.md keeps a one-line
pointer.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
2026-08-14 19:13:33 +03:00
mix 9e3f0d0fad docs(guide): reorder Installation and add a Full deployment walkthrough
test / test (push) Has been cancelled
Installation now reads Ports, Local trial, Initial setup, Full
deployment (with Fixed image tag nested under it), Environment
variables, Reverse proxy. The step-by-step production deploy and the
per-proxy TLS commands move from README's "Reference deploy" into the
operator guide; README keeps a short pointer instead of duplicating
the walkthrough.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
2026-08-14 15:49:58 +03:00
mix 403c76d17d docs(guide): restructure into Installation / Instance admin / Domain admin
test / test (push) Has been cancelled
Splits mixed DNS-setup, Operations, Rate-limiting, and Backup sections
along the instance/domain boundary and fixes README anchors that
pointed at the old #dns-setup heading.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
2026-08-14 15:45:38 +03:00
mix d9ab7bfa45 docs: retire the completed full-tree review plan
test / test (push) Has been cancelled
Every phase (P0-P7) is closed, so the plan file goes per its own exit criteria; history lives in git and the CHANGELOG. The roadmap's recommended order returns to queue-retries and then inbound-relay, and the order notes in both plans stop pointing at the deleted file, as does the 1.2.5 changelog entry (same treatment as logrotate-mode.md).

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-14 15:23:52 +03:00
mix 18bfd148ec docs: close P7 — security review of the P0–P1 diff, no findings
The Fable review of the send-log authorization, tryAdmit, session create and app-delete ordering against security.md found no further issues: every send-log query path states its scope and an empty scope matches nothing, rejected rows cannot consume rate-limit budget, the milter's nil-registry fail-open is test-only, and each failure residue of the reordered delete fails safe. Outcome recorded in security.md's header and the plan; nothing added to the accepted risks.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-14 15:02:07 +03:00
mix e70ba9046e fix(panel): resync mail-path maps once after restore
When CheckRestore accepts a backup manifest, the panel re-derives OpenDKIM tables and the Postfix sender map from SQLite on that first boot and reloads both daemons, so archive/database drift is healed before mail flows.

Co-authored-by: Cursor <cursoragent@cursor.com>
2026-08-14 14:58:08 +03:00
mix 02afa0fa80 fix(panel): code-review P6 optional cleanup
test / test (push) Has been cancelled
Close the remaining low-risk items from the full-tree review: rename the settings handler, query assigned domains in SQL, bound the login limiter map, collapse panel.js show/hide helpers, and soften DMARC copy that promised a future in-panel receiver.

Co-authored-by: Cursor <cursoragent@cursor.com>
2026-08-14 14:48:43 +03:00
mix 97a0c7e508 licence,ci: ship OFL with Plex fonts and guard release dispatch (P5)
test / test (push) Has been cancelled
IBM Plex WOFF2 files were shipped without the SIL OFL 1.1 text, NOTICE
told modifiers to edit layout.html for a Source URL that lives in
legal.go, and workflow_dispatch took GITHUB_REF_NAME as the version so a
run from main would publish ghcr.io/...:main.

Closes P5 of docs/plans/code-review.md. CHANGELOG updated.

Co-Authored-By: Cursor Grok 4.6 <noreply@anthropic.com>
Co-authored-by: Cursor <cursoragent@cursor.com>
2026-08-14 09:47:33 +03:00
mix efdc2d63fb docs: agree queue-retries (Postfix retry policy in the panel)
test / test (push) Has been cancelled
Roadmap item and plan: show first retry delay, backoff cap and queue
lifetime on Mail queue and the delivery page, reading postconf -h once
at panel start so a manual override is visible. Explanation only; not
implemented yet. CHANGELOG updated.

Co-Authored-By: Cursor Grok 4.6 <noreply@anthropic.com>
Co-authored-by: Cursor <cursoragent@cursor.com>
2026-08-13 23:51:13 +03:00
mix 8dfb483244 test: boot a panel on a data directory restored from its own backup (P4)
test / test (push) Has been cancelled
Restore is not a code path in the panel — the operator extracts the archive
onto /data and starts the image — so it had no test. cmd/panel/restore_test.go
now performs that path in process: download a backup from a running panel
through POST /backup (plain and encrypted), unpack it the way tar -xzf does,
and boot a second panel on the result through run()'s own startup order
(CheckRestore, store.Open, newPanel, Start).

Covered: the restored panel shows the domain and journal the archive carried
and finds the DKIM key, sasldb2 and Postfix sender map where its configuration
says they are; the setup link is not reopened by a restore; a session that
predates the backup still works, as the guide documents; an encrypted download
restores identically; a data directory from another version is refused with
both versions named and the manifest kept for the retry.

serveHTTP is split so the composition it performs (newPanel) can be started
without binding a port. No behaviour change.

Closes the optional P4 item in docs/plans/code-review.md, and with it the
"HandleBackup POST untested" gap from the review's test section. CHANGELOG
updated.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Co-authored-by: Cursor <cursoragent@cursor.com>
2026-08-13 23:46:01 +03:00
mix e6af1a7866 test: cover the auth, sign-in and RBAC surfaces P0 shipped through (P4)
test / test (push) Has been cancelled
- internal/web/auth/ratelimit_test.go: the login/setup limiter's ceiling,
  per-address scope, window reset and the sweep that keeps finished
  buckets out of memory.
- internal/web/auth/handlers_test.go: sign-in (session issued, refusals
  that do not reveal which usernames exist, a lockout a correct password
  cannot bypass, the setup hint before the first administrator) and the
  one-time setup link (creates the first global administrator and then
  closes, rejects a wrong or expired token, refuses credentials the panel
  would not accept later, rate-limited).
- internal/web/handlers/authz_test.go: every global-only route answers a
  domain administrator — and a request with no principal — with 404, with
  a positive control so the table cannot pass on a handler that always
  404s. This is the check that would have caught the send-log leak.
- test/e2e: CoreDNS pinned to 1.14.6 instead of latest; the level-1
  failure message quoted RATE_LIMIT_MESSAGES_PER_IP=5 while the stand
  sets 50.

docs/plans/code-review.md P4 checked off except the optional backup-boot
test; CHANGELOG updated. No production code changed.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Co-authored-by: Cursor <cursoragent@cursor.com>
2026-08-13 23:35:46 +03:00
mix aab35f0679 fix(panel): P3 optional items — urlquery, aria-live, confirm-without-JS note
- Deliveries fragment/pagination links urlquery-encode the domain/app
  filters instead of splicing them into the query string raw.
- The four polled regions (deliveries rows, status, mail queue, system
  log) carry aria-live="polite" so a screen reader announces refreshes.
- security.md accepted risks note that data-confirm prompts are
  JavaScript-only and why that is acceptable.

docs/plans/code-review.md P3 fully checked off; CHANGELOG updated.

Co-authored-by: Cursor <cursoragent@cursor.com>
2026-08-13 22:54:31 +03:00
mix 0570608738 fix(panel): GUI defects from the 1.2.x layout pass (P3)
- .flash.error now renders on the danger surface instead of the success
  one; RateLimitErr previously showed as green with red text.
- User delete goes through a confirmation page (GET/POST
  /users/{uid}/delete), matching the domain-delete pattern, instead of a
  plain submit button next to Save with no confirmation.
- Extracted the repeated DNS Host/Type/Value markup on a domain's page and
  the duplicated Settings credentials form into shared partials. No
  behaviour change.

docs/plans/code-review.md P3 checked off; CHANGELOG updated.

Co-authored-by: Cursor <cursoragent@cursor.com>
2026-08-13 21:50:42 +03:00
mix 1cf8bfcbe2 docs: rewrite security/operator docs for the multi-user panel (P2)
test / test (push) Has been cancelled
The CSRF ADR, guide.md, architecture.md route table, the internal/backup
package comment, roadmap/dmarc-reports admin.dmarc_report_email references,
the e2e coverage summary, and the setupManager/store package comments still
described the pre-1.2.0 single-administrator panel or stale implementation
details. Closes P2 of docs/plans/code-review.md; no behaviour change.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Co-authored-by: Cursor <cursoragent@cursor.com>
2026-08-13 16:15:14 +03:00
mix 4761991dd5 panel,mail: fail closed on the rate-limit race, session create and app delete
test / test (push) Has been cancelled
The level-2 limiter counted stored plus in-flight messages and reserved its own slot in two critical sections, so SMTP sessions that overlapped could each take the last free slot; tryAdmit now does both under one lock. A session that cannot be written no longer yields a cookie the browser would carry while every request bounced to /login. Deleting an application clears its SASL account before its registry row, matching domain delete, so a saslpasswd2 failure leaves a retryable application rather than an account that still authenticates.

Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
Co-authored-by: Cursor <cursoragent@cursor.com>
2026-08-13 14:50:04 +03:00
mix 8355479e03 docs: note e2e send-log badge scraper fix in Unreleased
test / test (push) Has been cancelled
The release gate timed out waiting for status=sent because its HTML scrapers had drifted from the panel's badge markup; record the fix and the handler regression test.

Co-authored-by: Cursor <cursoragent@cursor.com>
2026-08-13 14:36:11 +03:00
mix 6216e430e9 test(e2e): follow send-log status badge markup in scrapers
test / test (push) Has been cancelled
The deliveries_rows template wraps status in badge spans, but the e2e gate still looked for bare td cells, so send_verify_dkim_and_status timed out even when mail was delivered and logged. Add a handler regression test so go test ./... catches this drift without Docker.

Co-authored-by: Cursor <cursoragent@cursor.com>
2026-08-13 14:19:31 +03:00
mix a1e7e05d26 ci: gofmt files that failed the formatting workflow check
test / test (push) Has been cancelled
Whitespace-only alignment from gofmt; removes a stray BOM in web.go.

Co-authored-by: Cursor <cursoragent@cursor.com>
2026-08-13 12:24:34 +03:00
mix fae0068c4a panel: scope the send log to a domain admin's own domains
test / test (push) Has been cancelled
The Deliveries list narrowed the journal only when exactly one domain was assigned, so an administrator with none or with two or more read every domain's rows. The domain scope is now an IN constraint the store query carries, a filter that states no scope matches nothing, and the domain/app query parameters are validated against the principal before the query runs.

Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
Co-authored-by: Cursor <cursoragent@cursor.com>
2026-08-13 12:20:51 +03:00
mix b829625066 docs: record full-tree review and 2.x schema squash
test / test (push) Has been cancelled
Capture the 2026-08-13 review as the active plan, queue it ahead of inbound-relay, and defer SQLite migration squash to 2.x with a user_version gate.

Co-authored-by: Cursor <cursoragent@cursor.com>
2026-08-13 01:36:08 +03:00
mix 71048922b9 release: 1.2.5
test / test (push) Has been cancelled
Close Unreleased for the rate-limit form polish; pin compose and docs to 1.2.5.

Co-authored-by: Cursor <cursoragent@cursor.com>
2026-08-13 01:04:19 +03:00
mix 5caca024c3 panel: restore natural Export and Danger card layout
test / test (push) Has been cancelled
Drop the shared-baseline flex pin on .split cards; mismatched body
lengths made Export / Danger look worse than the plain stack.

Co-authored-by: Cursor <cursoragent@cursor.com>
2026-08-13 01:02:15 +03:00
mix 8289ed3916 panel: align paired domain settings and export actions
test / test (push) Has been cancelled
Subgrid Domain settings rows, match Address mode and trusted-IP controls,
and pin Export/Danger buttons to a shared baseline in split cards.

Co-authored-by: Cursor <cursoragent@cursor.com>
2026-08-13 00:39:55 +03:00
mix 958d07fcdd panel: polish rate-limit form copy and field alignment
test / test (push) Has been cancelled
Drop the app L1 backstop line, move trusted-IP help under the IP field,
shorten the domain rate-limit title, and align field-pair labels with DMARC.

Co-authored-by: Cursor <cursoragent@cursor.com>
2026-08-13 00:26:17 +03:00
mix f0cd804cb3 panel: tighten rate-limit and address-mode copy
Restore muted L1 backstop line, move domain limit help above Save, and drop
the duplicate Address mode field label.

Co-authored-by: Cursor <cursoragent@cursor.com>
2026-08-13 00:13:34 +03:00
mix 3950bd333b docs: put rate-limit UI polish back under Unreleased
test / test (push) Has been cancelled
1.2.4 stays as cut at the release commit; the L1 display and badge work
belongs in Unreleased until the next version cut (development.md).

Co-authored-by: Cursor <cursoragent@cursor.com>
2026-08-12 23:40:33 +03:00
mix 584f8cbc1c panel: clearer L1 rate-limit display and status badges
test / test (push) Has been cancelled
release / prepare (push) Has been cancelled
release / build (amd64, ubuntu-latest) (push) Has been cancelled
release / build (arm64, ubuntu-24.04-arm) (push) Has been cancelled
release / merge (push) Has been cancelled
Show the level-1 backstop as its own line and code-row, use st badges for
active/inactive domain and app limits, and fold the polish into 1.2.4.

Co-authored-by: Cursor <cursoragent@cursor.com>
2026-08-12 23:37:12 +03:00
mix 6671bd3393 release: 1.2.4
test / test (push) Has been cancelled
release / prepare (push) Has been cancelled
release / build (amd64, ubuntu-latest) (push) Has been cancelled
release / build (arm64, ubuntu-24.04-arm) (push) Has been cancelled
release / merge (push) Has been cancelled
Close Unreleased for the level-2 rate-limit semantics change and DNS Type
field height fix; pin compose and docs to 1.2.4.

Co-authored-by: Cursor <cursoragent@cursor.com>
2026-08-12 23:21:28 +03:00
mix 00e36df553 rate limit: domain ceiling for all IPs, trusted app override
Invert level-2 semantics so domain limits apply to every client IP and
application limits with trusted IPs raise the ceiling above the domain
(still capped by level 1). Panel shows L1, validates maxima, and documents
the model on Settings.

Co-authored-by: Cursor <cursoragent@cursor.com>
2026-08-12 23:19:51 +03:00
mix b0ebe061b5 panel: match DNS Type TXT height to Host fields
Co-authored-by: Cursor <cursoragent@cursor.com>
2026-08-12 23:06:09 +03:00
mix 601e183e0c release: 1.2.3
Close CHANGELOG [Unreleased] as 1.2.3; bump the compose pin and image references. Includes domain detail layout polish and Domains add-row.

Co-authored-by: Cursor <cursoragent@cursor.com>
2026-08-12 22:52:19 +03:00
mix f9e259a66d panel: refine domain detail layout and drop section index
test / test (push) Has been cancelled
Co-authored-by: Cursor <cursoragent@cursor.com>
2026-08-12 22:16:03 +03:00
mix 44e79c4120 panel: pair domain detail cards like Status (SPF | DMARC)
test / test (push) Has been cancelled
Co-authored-by: Cursor <cursoragent@cursor.com>
2026-08-12 15:33:41 +03:00
mix 1d30605533 panel: align page URLs, titles, and headings
test / test (push) Has been cancelled
Settings moves to /settings with a 308 redirect from /account; domains, Status, Users, and user forms get matching browser titles and h1 text; backup page title reflects domain import.

Co-Authored-By: Composer <noreply@cursor.com>
Co-authored-by: Cursor <cursoragent@cursor.com>
2026-08-12 15:24:14 +03:00
mix b87baa6dd6 fix: pin logrotate config mode in image and fail loud on bad permissions
test / test (push) Has been cancelled
COPY --chmod makes /etc/logrotate.d/mail 0644 regardless of build context
file modes (Windows tar sync). logrotate-loop preflight exits non-zero when
logrotate would ignore the config. E2e covers mode, forced rotation, and a
group-writable context build.

Co-Authored-By: Composer 2.5 <noreply@cursor.com>
Co-authored-by: Cursor <cursoragent@cursor.com>
2026-08-12 00:51:37 +03:00
mix 9420e30d6f release: 1.2.2
test / test (push) Has been cancelled
Close CHANGELOG [Unreleased] as 1.2.2; bump the compose pin and image references.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

Co-authored-by: Cursor <cursoragent@cursor.com>
2026-08-12 00:34:09 +03:00
mix 93cf1de3b7 panel: denser Status layout with true two-column cards
test / test (push) Has been cancelled
Pair Machine|Processes, queue|cert, and sockets|hostname; fix .split
auto-margins so cards fill half the row; trim Status prose and machine
details; drop the page section index; note panel-docs on the roadmap.

Co-authored-by: Cursor <cursoragent@cursor.com>
2026-08-12 00:29:34 +03:00
mix 6b8658602d panel: place mail queue and TLS cert side by side on Status
test / test (push) Has been cancelled
Widen the status page and wrap the short queue/certificate cards in
.split so the stack is shorter; other checks stay full width.

Co-authored-by: Cursor <cursoragent@cursor.com>
2026-08-11 23:24:24 +03:00
mix 3a520d9ec7 panel: fix Assigned domains checkbox layout on user form
test / test (push) Has been cancelled
Form-wide block labels and full-width inputs stacked the box above the
domain name; checkbox rows now share label.check and fieldset spacing.

Co-authored-by: Cursor <cursoragent@cursor.com>
2026-08-11 23:15:37 +03:00
mix 4015ecd70d release: 1.2.1
test / test (push) Has been cancelled
Close CHANGELOG [Unreleased] as 1.2.1; bump the compose pin and image references.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Co-authored-by: Cursor <cursoragent@cursor.com>
2026-08-11 23:03:54 +03:00
mix ad014ecdc1 panel: centre status badge text optically
test / test (push) Has been cancelled
IBM Plex Mono sits low in its em square; top-heavy badge padding left
ok/warn sunk below the adjacent heading. Bottom padding is now heavier.

Co-authored-by: Cursor <cursoragent@cursor.com>
2026-08-11 22:58:45 +03:00
mix de6f7a88da panel: disable demote and delete for the only global admin
test / test (push) Has been cancelled
Co-authored-by: Cursor <cursoragent@cursor.com>
2026-08-11 22:42:01 +03:00
mix 5985977f21 panel: hide domain pick for global administrators on user form
test / test (push) Has been cancelled
Global administrators manage every domain; the assignment checkboxes are
now hidden when that role is selected.

Co-authored-by: Cursor <cursoragent@cursor.com>
2026-08-11 22:37:04 +03:00
mix 9ff7514621 panel: tidy the Users nav icon
test / test (push) Has been cancelled
Two full silhouettes with staggered baselines read lopsided at 16 px; the
rear person is now a right-side crescent behind an aligned front silhouette.

Co-authored-by: Cursor <cursoragent@cursor.com>
2026-08-11 22:31:07 +03:00
mix ae7f5258a6 panel: use a gear for the Settings nav icon
The icon was a sun-with-rays (circle plus spokes), not a cog, despite the
1.2.0 note that Settings used a gear.

Co-Authored-By: Cursor <cursoragent@cursor.com>
2026-08-11 22:13:29 +03:00
mix 5bd632c8e0 panel: place drill-down back links under the heading
test / test (push) Has been cancelled
The user form had Back to users at the bottom of the card; all drill-down
pages now use a shared back_link template with a structural test.

Co-Authored-By: Claude <Composer 2.5> <noreply@anthropic.com>
Co-authored-by: Cursor <cursoragent@cursor.com>
2026-08-11 21:58:08 +03:00
mix f1074eb526 panel: show Settings credentials and DMARC side by side
test / test (push) Has been cancelled
Global administrators get two cards in the .split layout; domain-scoped
users keep the single narrow form.

Co-Authored-By: Claude <Composer 2.5> <noreply@anthropic.com>
Co-authored-by: Cursor <cursoragent@cursor.com>
2026-08-11 18:40:59 +03:00
mix bc005dea85 release: 1.2.0
test / test (push) Has been cancelled
Close CHANGELOG [Unreleased] as 1.2.0; correct missing entries; bump the
compose pin and image references.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Co-authored-by: Cursor <cursoragent@cursor.com>
2026-08-11 17:39:11 +03:00
mix 4b2f3880ef panel: align heading with narrow cards on signed-in pages
Settings and the user form centred the title on the 48rem measure while
.card.narrow centred on 24rem, so their left edges drifted by 12rem. Cap every
direct child of main when a narrow card is present, without shrinking the
column, so navigation stays put and login/setup stay unchanged.

Co-Authored-By: Cursor <cursoragent@cursor.com>
2026-08-11 17:34:04 +03:00
mix fae930a37a panel: give Settings, Users, and the session user distinct nav icons
test / test (push) Has been cancelled
Settings had been sharing the account silhouette with Users and the signed-in user label, so three different concepts read as the same icon. Settings now uses a gear, the user line the single-user icon, and Users a two-person group mark.

Co-Authored-By: Cursor <cursoragent@cursor.com>
2026-08-11 16:06:38 +03:00
mix c443149623 docs: plan the narrow pages' heading alignment
test / test (push) Has been cancelled
Settings and the user form leave their heading at the left of the reading
measure while the card floats 12rem to the right of it. main > * caps every
child at 48rem and centres it; .card.narrow overrides the max-width down to
24rem but not the auto margins it never mentions, so the two centre on
different measures and half the difference is the offset.

Records what the question was actually about: not what is wrong with Settings
but what it shares with exactly one other page. Four templates use the narrow
card and eleven use a full-width one, which sorts every page into three groups
— full-width card, narrow card without navigation, narrow card with it — and
only the last has the problem. Settings and the user form are the only pages
that are both signed-in and built from a single narrow card; every other page
has one of those properties, never both.

Also records why the obvious fix is unavailable, so it is not tried a fourth
time: the shell centres the navigation and the page as a pair, so anything that
changes the column width is visible from the navigation (296px, measured). The
fix has to work inside the column.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
2026-08-11 10:19:11 +03:00
mix 5eaf665c01 docs: remove completed plans and trim the open-work tracker
test / test (push) Has been cancelled
Co-authored-by: Cursor <cursoragent@cursor.com>
2026-08-11 01:34:04 +03:00
100 changed files with 5328 additions and 2431 deletions
+63 -13
View File
@@ -1,11 +1,9 @@
name: release
# Publishes an immutable, version-tagged image on ghcr.io (spec 10.1).
# Ordinary commits do not publish anything — only a pushed tag matching
# vX.Y.Z does. That tag is the single source the version comes from: it goes
# into both the image tag and the panel binary's -ldflags version, so the two
# can never drift apart (the invariant restore's version check in spec 7.5.A
# depends on).
# Ordinary commits and bare git tag pushes do not publish anything. A published
# GitHub Release runs this workflow directly (same pattern as gosentry / imap-scrub).
# You can also run it manually via workflow_dispatch with an explicit SemVer X.Y.Z.
#
# Native per-architecture builds (see docs/development.md), not qemu:
# running the full Postfix/OpenDKIM stack under emulation for the e2e gate
@@ -13,14 +11,18 @@ name: release
# tag on its own native runner; a merge job then combines them into the one
# manifest tag documented in deploy/docker-compose.yml. "test, then push" (not
# push-by-digest then test) is deliberate: it means the bytes that get tagged
# are exactly the bytes that passed e2e, at the cost of per-arch tags lingering
# in the registry as a side effect (harmless — the version tag's immutability,
# spec 10.1, is about that tag, not these).
# are exactly the bytes that passed e2e. Per-arch tags are pushed only so
# imagetools can assemble the multi-arch manifest; merge removes them from GHCR
# so operators see a single version tag (spec 10.1).
on:
push:
tags:
- "v[0-9]+.[0-9]+.[0-9]+"
release:
types: [published]
workflow_dispatch:
inputs:
version:
description: "Image version as X.Y.Z (no v prefix)."
required: true
type: string
permissions:
contents: read
@@ -32,9 +34,25 @@ jobs:
outputs:
version: ${{ steps.version.outputs.version }}
steps:
- name: Derive version from tag
- name: Derive a SemVer X.Y.Z version
id: version
run: echo "version=${GITHUB_REF_NAME#v}" >> "$GITHUB_OUTPUT"
env:
INPUT_VERSION: ${{ github.event.inputs.version }}
RELEASE_TAG: ${{ github.event.release.tag_name }}
EVENT_NAME: ${{ github.event_name }}
run: |
set -euo pipefail
if [ "$EVENT_NAME" = "release" ]; then
raw="${RELEASE_TAG:-}"
else
raw="${INPUT_VERSION:-}"
fi
raw="${raw#v}"
if ! [[ "$raw" =~ ^[0-9]+\.[0-9]+\.[0-9]+$ ]]; then
echo "::error::release version is not SemVer X.Y.Z (event=${EVENT_NAME}; tag=${RELEASE_TAG:-} input=${INPUT_VERSION:-}). Refusing to publish ghcr.io/${{ github.repository }}:${raw:-?}"
exit 1
fi
echo "version=${raw}" >> "$GITHUB_OUTPUT"
build:
needs: prepare
@@ -49,6 +67,8 @@ jobs:
runs-on: ${{ matrix.runner }}
steps:
- uses: actions/checkout@v4
with:
ref: v${{ needs.prepare.outputs.version }}
- uses: docker/setup-buildx-action@v3
@@ -113,6 +133,10 @@ jobs:
needs: [prepare, build]
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v4
with:
ref: v${{ needs.prepare.outputs.version }}
- uses: docker/setup-buildx-action@v3
- name: Log in to ghcr.io
@@ -152,3 +176,29 @@ jobs:
attempt=$((attempt + 1))
backoff=$((backoff * 2))
done
- name: Remove per-arch tags from GHCR
# Side-effect tags for imagetools assembly only — not part of the public
# version surface (deploy/docker-compose.yml pins X.Y.Z, not X.Y.Z-amd64).
# imagetools has no "rm" subcommand; delete via the GitHub Packages API.
env:
GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
run: |
set -euo pipefail
version="${{ needs.prepare.outputs.version }}"
owner="${{ github.repository_owner }}"
pkg="${{ github.event.repository.name }}"
api="/users/${owner}/packages/container/${pkg}/versions"
for suffix in amd64 arm64; do
tag="${version}-${suffix}"
mapfile -t ids < <(gh api "$api" --paginate \
--jq ".[] | select([.metadata.container.tags[]] | index(\"${tag}\")) | .id")
if [ "${#ids[@]}" -eq 0 ]; then
echo "no GHCR package version for tag ${tag}"
continue
fi
for id in "${ids[@]}"; do
echo "deleting GHCR package version ${id} (tag ${tag})"
gh api -X DELETE "${api}/${id}"
done
done
+406 -8
View File
@@ -5,6 +5,383 @@ Format follows [Keep a Changelog](https://keepachangelog.com/en/1.1.0/); version
## [Unreleased]
## [1.3.0] - 2026-08-14
Security and quality after 1.2.5: domain-admin send-log authorization,
fail-closed sign-in and application delete, level-2 rate-limit race fix,
restore Resync, expanded tests, operator docs, release CI, and OFL for IBM
Plex. Upgrading from 1.2.x is a tag bump; no migration.
### Added
- licence: the SIL Open Font License 1.1 text now travels with the IBM Plex
WOFF2 files (`internal/web/view/static/OFL.txt`). The image copies it next
to LICENSE and NOTICE under `/usr/share/doc/selfpost/`; the panel serves it
at `/static/OFL.txt`. OFL requires the licence to accompany the font.
- docs: agreed roadmap item **queue-retries** — show this Postfix's retry
policy (first delay, backoff cap, queue lifetime) on Mail queue and on a
delivery's history, reading `postconf -h` once at panel start so a manual
override is visible. Plan: [docs/plans/queue-retries.md](docs/plans/queue-retries.md).
Explanation only; no attempt counter and no panel knobs. Not yet
implemented.
### Security
- The independent security review of the send-log authorization and
fail-closed fixes below (code-review plan § P7; reviewer model ≠ author
model) found no further issues: the domain scope holds on every query path,
a rate-limit refusal cannot consume window budget, and each failure residue
of the reordered application delete fails safe. Nothing was added to
[docs/security.md](docs/security.md) § Accepted risks; the review is
recorded in that file's header.
### Fixed
- test (e2e): send-log status scrapers follow the badge markup in
`deliveries_rows`. The release gate still looked for bare `<td>sent</td>`
after the panel started rendering status as `<span class="st st-*">` badges,
so `send_verify_dkim_and_status` timed out even when mail was delivered and
logged. A handler regression test catches this drift in `go test ./...`
without Docker.
- security (panel): the Deliveries list is scoped to a domain administrator's
assigned domains for every number of assignments. Previously the send log was
narrowed only when exactly one domain was assigned, so an administrator with
none or with two or more read every domain's rows (sender, recipient, subject)
on the list and its polled fragment. The domain scope is now an `IN`
constraint carried by the store query — a filter that states no scope returns
nothing — and the `domain` and `app` query parameters are checked against the
principal's own domains and applications before the query runs, so a
hand-written URL cannot widen the scope. Global administrators are unaffected.
- mail (level-2 rate limit): the ceiling is no longer overshot by messages that
arrive at the same instant. The milter counted the stored and in-flight
messages and reserved its own slot in two separate steps, so several SMTP
sessions could pass the same check before any of them had reserved. Counting
and reserving now happen as one operation, and the ceiling is handed out
exactly as many times as configured. Postfix's level-1 limit remains the
backstop and the level-2 check stays fail-open on store errors.
- panel (sign-in): a session that cannot be written to the database no longer
produces a session cookie. The login used to log the failure, set the cookie
and redirect to the dashboard, leaving the browser looking signed in while
every request bounced back to `/login`; it now fails closed with an error on
the sign-in page.
- panel (applications): deleting an application removes its SASL credentials
before its registry row. If `saslpasswd2` fails, the application stays listed
and the delete can be retried, instead of leaving a hidden account that could
still authenticate to Postfix. This matches the order domain deletion already
used.
- panel (GUI): a rejected rate-limit change on a domain's page now renders on
the danger surface (`.flash.error`) instead of the success one — it was
green with red text, reading as good news. Deleting a panel user now goes
through a confirmation page, the same pattern as domain deletion, instead of
a plain submit button next to Save with no confirmation at all.
- panel (restore): after a backup is extracted and the version guard passes,
the panel runs one mail-path Resync on the first boot — OpenDKIM's tables
and Postfix's sender map are re-derived from SQLite and the daemons are
reloaded, so drift between the archive and the database is healed before
mail flows. Later starts skip that step; the Status page Reload button runs
the same Resync on demand. The `internal/backup` package comment now matches
this behaviour.
- ci (GHCR): per-arch package tags (`X.Y.Z-amd64`, `X.Y.Z-arm64`) are dropped
after the manifest merge via the GitHub Packages API. The merge job had called
`docker buildx imagetools rm`, which is not a valid subcommand — cleanup failed
with a warning and the side-effect tags stayed in the registry.
### Changed
- docs: operator and as-built docs aligned with the code after a full
pass — [architecture.md](docs/architecture.md) route table now marks
**global** routes (404 for domain administrators) and documents the
one-time restore Resync in Persistence; session/password and restore-session
wording corrected in [guide.md](docs/guide.md) and architecture (own-password
change vs admin reset, no "logout everywhere", immediate session restore on
the next request, PTR cache ≈1 min, decrypt has no version check, restore
Resync on first boot, domain add/delete/import and `POST /reload` global-only,
Settings DMARC global-only); [README.md](README.md) port-587 and quick-start
volume wording fixed; [security.md](docs/security.md) CSRF ADR points at
`authz.go` for route gating. No behaviour change.
- docs: [guide.md](docs/guide.md) reorganised into **Installation**, **Instance
administration**, and **Domain administration** — DNS setup, operations,
rate limiting, and backup sections follow the instance/domain boundary
instead of mixing them. **Installation** now reads Ports → Local trial →
Initial setup → Full deployment (with the fixed image tag nested under it) →
Environment variables → Reverse proxy; the step-by-step production deploy and
per-proxy TLS commands move here from README's "Reference deploy" (README
keeps a short pointer). Internal (non-operator) environment variables move
to [architecture.md](docs/architecture.md) § Configuration; the guide keeps
a one-line pointer. **Full backup and restore** gains worked commands for
in-place restore, move-to-a-new-host, and encrypted-backup decrypt-first,
plus the version-mismatch error text. README anchors updated for the new
headings. No behaviour change.
- docs: the 2026-08-13 full-tree review plan is complete — every phase (P0P7)
is closed — and `docs/plans/code-review.md` is deleted per its own exit
criteria (history in git and in this file). The plan covered architecture,
quality, GUI, tests, and licence work; P0 was domain-admin send-log
authorization. The [roadmap](docs/roadmap.md)'s recommended order returns to
**queue-retries** and then **inbound-relay**; it still records
**schema-squash** (replace the 1.x SQLite migration chain with a 2.x baseline;
not a reason to cut a major on its own).
- licence: [NOTICE](NOTICE) tells modifiers to update `SourceURL` in
`internal/legal/legal.go` (the value the panel footer actually injects), not
`layout.html`. Per-file `SPDX-License-Identifier` headers on the two command
packages were dropped so the tree is consistent; AGPL-3.0 does not require
them ([development.md](docs/development.md) § External libraries). Deleted the
completed `docs/plans/logrotate-mode.md` (history in git and
[1.2.3](#123---2026-08-12)).
- ci: the release image is published only for a **published** GitHub Release
(`vX.Y.Z`) or a manual `workflow_dispatch` with an explicit SemVer version — a
bare git tag push no longer starts the build. `release.yml` listens for
`release: published`, checks out that tag (not `main` HEAD), e2e-gates each
native arch build, merges `X.Y.Z-amd64` and `X.Y.Z-arm64` into one manifest,
then removes the per-arch tags from GHCR via the GitHub Packages API so
operators see only `ghcr.io/mixeme/selfpost:X.Y.Z` (what
`deploy/docker-compose.yml` pins). A dispatch whose version input is missing
or not `X.Y.Z` fails in `prepare`. [development.md](docs/development.md)
documents draft vs published releases, why deleting a release tag converts
it back to draft, and Gitea → GitHub tag-mirror pitfalls (do not prune release
tags on GitHub; a mirrored `v1.0.0` still runs that tag's `on: push: tags`
workflow).
- test: the authorization and sign-in surfaces that had no tests now have them.
The login limiter is covered for its ceiling, its per-address scope, the reset
at the end of a window and the sweep that keeps finished buckets out of
memory; sign-in for a successful session, for refusals that do not reveal
which usernames exist, and for a lockout that a correct password cannot
bypass; the one-time setup link for creating the first administrator, closing
afterwards, rejecting a wrong or expired token, and refusing credentials the
panel would not accept later. Every global-only route (`/users`, `/backup`,
domain import, `/status`, `/mail-queue`, `/system-log`, domain add and delete,
reload) is checked to answer a domain administrator — and a request with no
principal — with 404, the check that would have caught the send-log leak.
- test: restore is covered as the operator performs it, in process. A backup is
downloaded from a running panel through `POST /backup` (plain and encrypted),
unpacked the way `tar -xzf` unpacks it onto the `/data` bind mount, and a
second panel is booted on the result through the startup order the panel
itself uses — version guard, database, one Resync when restoring, then
services and the HTTP application. The restored panel shows the domain and
journal the archive carried, finds the DKIM key, SASL database and Postfix
sender map where its configuration says they are, does not reopen the
one-time setup link, and still honours a session that predates the backup.
Drifted on-disk maps are healed by that Resync step
(`TestResyncAfterRestoreHealsDriftedMaps`). A data directory left by another
version is refused with both versions named and the manifest kept. `serveHTTP`
is split in two so that composition can be started without binding a port; no
behaviour change.
- test (e2e): the CoreDNS image is pinned to `1.14.6` instead of `latest`, so
the release gate cannot change under a commit between two runs. The level-1
rate-limit failure message quoted `RATE_LIMIT_MESSAGES_PER_IP=5` while the
stand sets `50`.
- ci: gofmt on eight files that failed the formatting workflow check (panel
config, DNS check, domain transfer export, rate-limit tests, auth principal,
domain and delivery handlers, web package doc comment).
- panel (templates): the repeated Host/Type/Value DNS record markup on a
domain's page and the duplicated credentials form on Settings are now
shared partials (`host_type`, `host_type_copy`, `field_value`,
`field_values`, `credentials_fields`) instead of copy-pasted blocks. No
behaviour or visible change.
- panel (GUI, accessibility): the Deliveries fragment's `hx-get` and pagination
links now `urlquery`-encode the `domain`/`app` filters instead of splicing
them into the query string raw. The four polled regions (deliveries rows,
status, mail queue, system log) carry `aria-live="polite"` so a screen
reader announces the refreshed content.
- docs: [security.md](docs/security.md) accepted risks now note that
`data-confirm` prompts on destructive forms are JavaScript-only — with
JavaScript disabled the form submits immediately, the same as before the
prompts existed — and why that is acceptable (the prompt is a mis-click
guard, not an authorization boundary).
- docs: security and operator docs updated for the panel that has shipped
global administrators and domain-admins since 1.2.0. The CSRF ADR in
[security.md](docs/security.md) no longer argues from "single-user"; it now
states that cross-user CSRF between panel roles is not the threat the origin
check defends against, and gives a new revisit trigger. Dropped the
unimplemented "or argon2" alternative for the password hash.
[guide.md](docs/guide.md) documents the Users page and the two roles,
level-2 rate limiting's fail-open behaviour, and that a domain-admin can
export working SASL passwords for domains assigned to them.
[architecture.md](docs/architecture.md) gains `/license` and the
`/account``/settings` redirect in the route table (later expanded for
RBAC in the doc-alignment pass above). Corrected stale
`admin.dmarc_report_email` references in
[roadmap.md](docs/roadmap.md) and
[docs/plans/dmarc-reports.md](docs/plans/dmarc-reports.md) to the setting's
actual home after migration `0005`. No behaviour change.
- panel: code-review P6 cleanup — the unused `auth.RequireGlobal` middleware is
gone (handlers already call `requireGlobal`); the settings route handler is
named `HandleSettings` in `handlers_settings.go`; domain lists for a
domain-admin now come from `ListDomainsForUser` in SQL instead of loading
every domain and filtering in Go; the login and setup rate limiters sweep
expired buckets on a timer and cap the map at 4096 keys; the five
show/hide field helpers in `panel.js` are one rule table; DMARC copy no
longer promises in-panel report reception in a future release — SelfPost
does not receive inbound mail.
## [1.2.5] - 2026-08-13
Rate-limit form polish after 1.2.4. Upgrading is a tag bump; no migration.
### Changed
- panel: rate-limit forms refined — level-1 backstop as muted copy on the
domain settings form (`N messages / Ws — Settings`) and in message-limit
labels (`max N`); Settings shows the L1 ceiling in a code-row. Domain,
application, and Edit-toggle limit state use the shared `st` badge instead of
bold text or parenthetical copy. Address mode and trusted-IP override columns
carry muted leads and matched control height; trusted-IP help sits under the
IP field. Domain settings pairs DMARC reports with the level-2 rate limit
using CSS subgrid so titles, fields, and Save / Remove buttons line up across
columns.
## [1.2.4] - 2026-08-12
Level-2 rate-limit semantics inverted after 1.2.3, plus a small DNS field
height fix. Upgrading is a tag bump; no migration.
### Changed
- rate limiting (level 2): domain ceilings apply to every client IP (no IP
allowlist). An application ceiling with trusted IPs is an override
**above** the domain limit (still ≤ level 1) and skips the domain check for
those IPs; without IPs the application override is inactive. When no domain
ceiling is set, non-privileged senders use level 1 only. The panel shows the
level-1 backstop on domain/application forms and Settings, rejects maxima
above level 1, and requires an application override to exceed the domain
maximum. Operator guide and architecture updated.
### Fixed
- panel: on the domain DNS status grid, the Type (TXT) field height matches
the Host fields.
## [1.2.3] - 2026-08-12
Domain detail layout and panel polish after 1.2.2. Upgrading is a tag bump; no
migration.
### Changed
- panel: the domain detail page is wide with paired cards (DKIM+SPF ‖ DMARC;
connection settings ‖ add application; export ‖ danger). DNS status,
Applications and Domain settings are full-width. DNS status is two rows
(DKIM ‖ SPF, DMARC ‖ report authorization) with Host ‖ Type (narrow TXT)
and a Value label when records are present. Domain settings pairs DMARC
report mode with the optional level-2 domain rate limit; application Edit
opens address mode and an optional level-2 application rate limit side by
side (with a note that domain level-2 and global level-1 still apply); the
custom `rua=` address field is shown only for Custom address. The in-nav
“On this page” section index is removed. Shorter blurbs; *Sending server
settings* renamed **Connection settings**.
- panel: Domains list — **Add domain** sits beside the domain field; the
lead blurb under the form is dropped.
- panel: page URLs, browser titles, and headings are aligned — **Settings** is
now `/settings` (legacy `/account` redirects with 308); the domains list title
is `SelfPost — domains`; Status, Users, and user create/edit titles match their
nav labels and `<h1>` text; the backup page title is `SelfPost — backup &
migration` to reflect domain import as well as full backup. Operator guide and
architecture route tables updated.
### Fixed
- image: `mail.log` rotation no longer silently stops when the build context
ships `logrotate-mail.conf` with group/other write bits (common after a
Windows checkout sync). Runtime `COPY --chmod` pins config and script modes
in the Dockerfile; `logrotate-loop.sh` refuses a config logrotate would
ignore. E2e checks mode `644`, forced rotation, and a group-writable context
build.
## [1.2.2] - 2026-08-12
Status page layout after 1.2.1: paired cards in a wide column, denser machine
and check copy, and a user-form checkbox fix. Upgrading is a tag bump; no
migration.
### Changed
- panel: **Status** is wide again so paired cards fill the column. Layout:
Overall; Machine ‖ Processes; Mail queue ‖ TLS certificate; Milter sockets ‖
Hostname / reverse DNS; Configuration. Dropped lead blurbs on Machine, TLS
certificate, and Hostname (and Hostname's trailing detail line); milter
socket paths omitted from the table; queue link reads **View queue**; milter
ok detail is `Listening` without a trailing period and sits in its own
Detail column beside the status badge; CPU detail is only core and thread
counts (no load average); memory detail is `N used of M` without the
«available to new work» clause; network detail lists per-interface totals
only (rates stay in the Usage column). No «On this page» section index —
the paired layout is short enough. Hostname stays in the polled fragment so
the pair survives HTMX refresh. In-panel docs for the removed blurbs filed
as roadmap `panel-docs`.
### Fixed
- panel: cards inside `.split` used `margin: 0 auto`, which in a CSS grid
shrinks each card to its content and centres it in the track instead of
filling half the row. Side margins are cancelled for `.split > .card`
(Status, Settings, and a delivery's message/history).
- panel: on the user create/edit form, **Assigned domains** checkboxes stacked
the box above the domain name (and stretched it full-width) because the form's
block-label and full-width input rules applied to them. Checkbox rows now use
the shared `label.check` layout; the fieldset has matching spacing.
## [1.2.1] - 2026-08-11
Panel refinements after 1.2.0: navigation icons, status-badge centreing,
drill-down back-link placement, and safeguards for the sole global
administrator. Upgrading is a tag bump; no migration.
### Fixed
- panel: status badge text sat low in the box (and below the heading or label
beside it). IBM Plex Mono sits low in its em square; the previous top-heavy
padding made that worse. Bottom padding is now heavier so the word centres
optically.
- panel: the **Users** navigation icon was two full silhouettes with staggered
baselines, so the pair looked lopsided at 16 px. The rear person is now a
right-side crescent (head + shoulder) behind a full front silhouette aligned
with `icon-account`.
- panel: the **Settings** navigation icon was a sun-with-rays (circle plus
spokes), not a gear. It now uses a toothed cog so it matches the other
session icons and the 1.2.0 release note.
- panel: the user create/edit form placed «Back to users» at the bottom of the
card instead of under the heading like the delivery, domain, and domain-delete
pages. A shared `back_link` template now renders every drill-down up-link, and
`TestDrillDownPagesPlaceBackLinkAboveContent` guards its position.
### Changed
- panel: the user edit form disables role change and delete for the only global
administrator, with a short note, instead of allowing the action and showing
an error on submit.
- panel: the user create/edit form hides **Assigned domains** when the role is
global administrator, since that role manages every domain anyway.
- panel: **Settings** shows panel credentials and DMARC aggregate reports side
by side for global administrators (the same `.split` layout as a delivery's
message and history). Domain-scoped users keep the single narrow card.
## [1.2.0] - 2026-08-11
The second MINOR after 1.0.0: domain administrators with per-domain scope, a
panel visual refresh on the SelfPost palette, and refinements to navigation and
the send log. Upgrading runs one SQLite migration (`0005_panel_users`); the
single administrator becomes a global user.
### Added
- panel: **domain-admin role** — global administrators manage panel users and
@@ -15,28 +392,49 @@ Format follows [Keep a Changelog](https://keepachangelog.com/en/1.1.0/); version
administrator into a global user; sessions and full backup restore carry users
and domain bindings.
### Fixed
- Docker build: `LICENSE` is no longer excluded by `.dockerignore`, so the
runtime image can copy it into `/usr/share/doc/selfpost/` as AGPL requires. A
clean build failed once the cached layer was invalidated.
- panel: on signed-in pages with a narrow card (**Settings**, the user form)
the heading, flash, card and footer now share one left edge. `.card.narrow`
had overridden only `max-width` while `main > *` still centred siblings on
the 48rem measure, so the card floated 12rem to the right of the title.
`main:has(> .card.narrow) > *` keeps the stack on 24rem without narrowing
the column, so the navigation stays put; login/setup are unchanged.
### Changed
- panel: navigation session icons are distinct — **Settings** uses a gear,
the signed-in user line carries the single-user icon, and **Users** a
two-person group mark instead of the same account silhouette for all three.
- panel: `/` redirects domain administrators to `/domains`; global users still
land on `/status`. Navigation hides global-only sections for domain
administrators.
- panel: **Sign out** in the navigation column uses the same type size and
weight as the page entries above it; only the red tint and border mark it as
destructive.
- panel: **visual style** brought in line with the SelfPost mark — brick accent
and warm paper in place of the blue-on-cool-grey defaults, IBM Plex Sans and
IBM Plex Mono served by the panel itself, squarer corners, and column
headings, status badges and small labels set in the mono face. Light and dark
schemes both keep their contrast; no page, control or workflow changed. A page
whose only card is the narrow one now takes its heading and footer down to the
card's width instead of splitting them across the wider measure, and the send
log stops breaking `Details` and `deferred` across two lines when a row is
tight.
schemes both keep their contrast; no page, control or workflow changed. The
send log stops breaking `Details` and `deferred` across two lines when a row
is tight. Badge padding and line-height are tuned so lowercase labels sit
centred in the box. The three WOFF2 files add ~76 KB to the image and are
served from the panel's own origin, so the Content-Security-Policy is
unchanged (`default-src 'self'`).
- panel: the send log's **status is a badge**, in the same ok/warn/error/unknown
colours the status page and the DNS checks use, instead of the one place in
the panel where a status was bare text. The mapping is the one the delivery
page already applied — `sent` is ok, `deferred` a warning, `bounced` and
`rejected` errors, `queued` unknown because nothing has gone wrong yet.
The three
WOFF2 files add ~76 KB to the image and are served from the panel's own origin,
so the Content-Security-Policy is unchanged (`default-src 'self'`).
- docs: [roadmap.md](docs/roadmap.md) and [product.md](docs/product.md) no
longer list domain-admin or visual-style as open work — both ship in this line.
Completed plan files (`domain-admin`, `visual-style`, `web-split`,
`narrow-page-alignment`) are removed; history stays in git and the entries
above. Inbound relay is the main agreed 1.x+ item left on the roadmap.
## [1.1.0] - 2026-08-10
+5 -2
View File
@@ -10,8 +10,8 @@ Corresponding Source for the published releases is at:
If you modify SelfPost and let users interact with it over a network,
AGPL-3.0 §13 requires you to offer them the Corresponding Source of
your modified version. Update the Source link in the panel footer
(internal/web/view/templates/layout.html) so it points at your sources.
your modified version. Update SourceURL in internal/legal/legal.go
so the panel footer points at your sources.
Third-party software included in this distribution
----------------------------------------------------
@@ -24,6 +24,9 @@ htmx 2.0.4 (internal/web/view/static/htmx.min.js)
IBM Plex Sans / IBM Plex Mono
Copyright © 2017 IBM Corp., with Reserved Font Name "Plex"
Licence: SIL Open Font License 1.1
Full text: internal/web/view/static/OFL.txt
(also /usr/share/doc/selfpost/OFL.txt in the image, and
/static/OFL.txt from the control panel)
https://github.com/IBM/plex
Distributed unmodified as three latin-subset WOFF2 files served by the
control panel:
+16 -134
View File
@@ -24,9 +24,9 @@ send log and DNS checks in the panel, encrypted backups.
- Web panel — domains, applications, deliveries, mail queue, system log, backup
- Multi-domain relay — each SASL application is bound to one sending domain
- DNS status checks (PTR, SPF, DKIM, DMARC) with in-panel re-check
- Two-level rate limiting — IP backstop (Postfix) and per-domain/per-app limits
- Two-level rate limiting — IP backstop (Postfix), per-domain ceilings, and trusted-IP app overrides
- Full-server backup and single-domain export/import (optional password encryption)
- Single Docker image; data in a `./data` bind mount
- Single Docker image; production data in a `./data` bind mount (the quick start below uses a named Docker volume instead)
## Documentation
@@ -76,14 +76,15 @@ For every domain you add in the panel:
- [ ] DKIM TXT record (value shown on the domain page)
- [ ] DMARC `_dmarc` TXT record
See [DNS setup](docs/guide.md#dns-setup) in the operator guide.
See [Domain-level DNS](docs/guide.md#domain-level-dns-spf-dkim-dmarc) in the operator guide.
## Quick start
> **First boot — create the admin account.** On a fresh container SelfPost prints
> a **one-time setup URL** (valid ten minutes). Open it in a browser to choose
> the administrator username and password. Until you do, the panel has no login.
> Production deploy: [step 3](#3-start-selfpost).
> Production deploy: [Full deployment](docs/guide.md#full-deployment) in the
> operator guide.
One container, panel at `http://127.0.0.1:8080` — no reverse proxy, no TLS
files, no compose files. Good for clicking through the UI on your machine;
@@ -95,7 +96,7 @@ docker run --rm -d --name selfpost-try \
-e SELFPOST_HOSTNAME=mail.local.test \
-e PANEL_COOKIE_SECURE=false \
-v selfpost-try-data:/data \
ghcr.io/mixeme/selfpost:1.1.0
ghcr.io/mixeme/selfpost:1.3.0
```
**Get the setup URL** (pick one):
@@ -129,136 +130,17 @@ TLS PEM files at `./certs` (read by Postfix on 465/587). The panel is reached
only through a reverse proxy on 443 — port 8080 is bound to localhost in the
default compose file.
| Artefact | Path |
|---|---|
| Compose file (fixed image tag) | [deploy/docker-compose.yml](deploy/docker-compose.yml) |
| Environment template | [deploy/.env.example](deploy/.env.example) |
| Apache vhost (recommended) | [deploy/apache/selfpost-vhost.conf](deploy/apache/selfpost-vhost.conf) |
| nginx | [deploy/nginx/](deploy/nginx/) |
| Caddy | [deploy/caddy/](deploy/caddy/) |
| Traefik | [deploy/traefik/](deploy/traefik/) |
Full walkthrough — fetching the base files, setting up a reverse proxy and
TLS (Apache/nginx/Caddy/Traefik), starting the container, and wiring up
DNS — lives in the operator guide's [Full
deployment](docs/guide.md#full-deployment) section, with proxy-specific
commands under [Reverse proxy](docs/guide.md#reverse-proxy-mandatory).
### 1. Fetch the base files
```sh
mkdir -p selfpost/data selfpost/certs && cd selfpost
curl -O https://raw.githubusercontent.com/mixeme/selfpost/main/deploy/docker-compose.yml
curl -O https://raw.githubusercontent.com/mixeme/selfpost/main/deploy/.env.example
cp .env.example .env
```
Edit `.env` — at minimum set `SELFPOST_HOSTNAME` to your mail hostname (bare
FQDN, e.g. `mail.example.com`). It must match the PTR record you request from
your provider and the certificate your proxy will obtain.
### 2. Reverse proxy and TLS
Pick one proxy. In every case the proxy terminates HTTPS for the panel; the
same certificate must end up under `./certs` as `fullchain.pem` and
`privkey.pem` so Postfix can serve it on 465 (and 587 if enabled). The proxy
must **pass the original `Host` header** — details and rationale:
[Reverse proxy](docs/guide.md#reverse-proxy-mandatory).
**Apache (recommended, on the host).** Install Apache with `ssl`, `proxy`, and
`proxy_http` enabled. Copy
[deploy/apache/selfpost-vhost.conf](deploy/apache/selfpost-vhost.conf) into your
vhost directory, replace `mail.example.com` with your hostname, enable the site,
then issue a certificate:
```sh
sudo certbot --apache -d mail.example.com
```
Point `./certs` at the PEM files certbot wrote (symlink is fine):
```sh
ln -s /etc/letsencrypt/live/mail.example.com certs
```
**nginx (containerised).** From the `deploy/` directory, merge the nginx
fragment and issue the first certificate before nginx can serve HTTPS:
```sh
docker compose -f docker-compose.yml -f nginx/docker-compose.nginx.yml \
run --rm certbot certonly --webroot -w /var/www/certbot \
-d mail.example.com --email you@example.com --agree-tos --no-eff-email
docker compose -f docker-compose.yml -f nginx/docker-compose.nginx.yml up -d
```
Edit [deploy/nginx/nginx.conf.example](deploy/nginx/nginx.conf.example) and
replace `mail.example.com` first. The fragment bind-mounts certbot's output into
both nginx and SelfPost.
**Caddy (containerised, automatic ACME).** Edit
[deploy/caddy/Caddyfile](deploy/caddy/Caddyfile) and the `<hostname>` placeholders
in [deploy/caddy/docker-compose.caddy.yml](deploy/caddy/docker-compose.caddy.yml),
then:
```sh
docker compose -f docker-compose.yml -f caddy/docker-compose.caddy.yml up -d
```
Verify Caddy's on-disk cert path for your version before relying on the
default mount — see the comment at the top of the Caddy compose fragment.
**Traefik (containerised).** Edit the `Host(...)` label and ACME email in
[deploy/traefik/docker-compose.traefik.yml](deploy/traefik/docker-compose.traefik.yml),
start the stack, then extract PEM files for Postfix whenever Traefik issues or
renews a certificate:
```sh
docker compose -f docker-compose.yml -f traefik/docker-compose.traefik.yml up -d
./traefik/extract-cert.sh ./traefik/letsencrypt/acme.json mail.example.com ./traefik/extracted-certs
```
Schedule `extract-cert.sh` (cron or a timer) alongside Traefik's renewals.
### 3. Start SelfPost
If you used Apache on the host (step 2, first option), start only the base
compose file from your `selfpost/` directory:
```sh
docker compose up -d
```
The nginx/Caddy/Traefik fragments from step 2 already include `docker compose up
-d` — skip this if you ran one of those.
**Get the setup URL** — open it in a browser to create the admin account
([first boot](#quick-start)):
```sh
docker compose logs selfpost 2>&1 | grep -m1 'http'
```
```sh
cat ./data/setup-token
```
The file is deleted as soon as setup completes. If logs are shipped to a
central aggregator, prefer `cat ./data/setup-token` so the bearer token does
not enter the log pipeline.
### 4. DNS and sending
Before sending real mail:
1. Confirm PTR/rDNS for the server IP points at `SELFPOST_HOSTNAME` (Status
page → *Re-check*).
2. For each domain you add in the panel, publish SPF, DKIM, and DMARC at the
same time ([DNS setup](docs/guide.md#dns-setup)).
3. Warm up a new IP gradually ([IP warmup](docs/guide.md#ip-warmup)).
### Ports and upgrades
The compose file maps **465** (always) and **587** (when
`SUBMISSION_ENABLE=true`). Bump the pinned image tag deliberately when
upgrading — never use `:latest` ([why](docs/guide.md#fixed-image-tag)).
Optional variables (`TRUSTED_PROXY_CIDR`, rate limits, retention): see
[Environment variables](docs/guide.md#environment-variables).
The compose file always publishes **465** and **587**; Postfix listens on 587
only when `SUBMISSION_ENABLE=true` (see [Ports](docs/guide.md#ports)). Bump the
pinned image tag deliberately when upgrading, never `:latest` ([why](docs/guide.md#fixed-image-tag)). Optional
variables (`TRUSTED_PROXY_CIDR`, rate limits, retention): see [Environment
variables](docs/guide.md#environment-variables).
## License
+18 -14
View File
@@ -87,21 +87,25 @@ COPY --from=build /out/panel /usr/local/bin/panel
COPY --from=build /out/selfpost-backup /usr/local/bin/selfpost-backup
# Licence text shipped with the image (AGPL-3.0 conveyance). The panel also
# serves the same text at /license from an embedded copy.
COPY LICENSE NOTICE /usr/share/doc/selfpost/
# serves the same text at /license from an embedded copy. OFL.txt is the SIL
# Open Font License 1.1 that must travel with the IBM Plex WOFF2 files
# embedded in the panel binary (NOTICE).
COPY --chmod=0644 LICENSE NOTICE /usr/share/doc/selfpost/
COPY --chmod=0644 internal/web/view/static/OFL.txt /usr/share/doc/selfpost/OFL.txt
COPY build/opendkim.conf /etc/opendkim.conf
COPY build/logrotate-mail.conf /etc/logrotate.d/mail
COPY build/postfix-wrapper.sh /usr/local/bin/postfix-wrapper.sh
COPY build/postfix-config.sh /usr/local/bin/postfix-config.sh
COPY build/postfix-cert-reload.sh /usr/local/bin/postfix-cert-reload.sh
COPY build/logrotate-loop.sh /usr/local/bin/logrotate-loop.sh
COPY build/crashexit.py /usr/local/bin/crashexit.py
COPY build/entrypoint.sh /usr/local/bin/entrypoint.sh
COPY build/supervisord.conf /etc/supervisor/supervisord.conf
RUN chmod +x /usr/local/bin/postfix-wrapper.sh /usr/local/bin/postfix-config.sh \
/usr/local/bin/postfix-cert-reload.sh /usr/local/bin/logrotate-loop.sh \
/usr/local/bin/crashexit.py /usr/local/bin/entrypoint.sh
# File modes are pinned here so a build context copied from a checkout without
# POSIX permissions (e.g. Windows tar sync) cannot land group-writable config
# that logrotate would silently ignore — see docs/development.md § Building
# binaries and the image.
COPY --chmod=0644 build/opendkim.conf /etc/opendkim.conf
COPY --chmod=0644 build/logrotate-mail.conf /etc/logrotate.d/mail
COPY --chmod=0755 build/postfix-wrapper.sh /usr/local/bin/postfix-wrapper.sh
COPY --chmod=0755 build/postfix-config.sh /usr/local/bin/postfix-config.sh
COPY --chmod=0755 build/postfix-cert-reload.sh /usr/local/bin/postfix-cert-reload.sh
COPY --chmod=0755 build/logrotate-loop.sh /usr/local/bin/logrotate-loop.sh
COPY --chmod=0755 build/crashexit.py /usr/local/bin/crashexit.py
COPY --chmod=0755 build/entrypoint.sh /usr/local/bin/entrypoint.sh
COPY --chmod=0644 build/supervisord.conf /etc/supervisor/supervisord.conf
# Published submission ports: 465 (smtps, primary) and 587 (submission, optional)
# plus the panel on 8080. Outbound delivery dials remote MXs on 25 as a client,
+40 -1
View File
@@ -14,10 +14,49 @@
# often than daily — polling merely bounds how late a legitimate rotation runs.
set -eu
CONFIG=/etc/logrotate.d/mail
INTERVAL="${LOGROTATE_INTERVAL_SECONDS:-21600}"
# logrotate refuses configs writable by group or others and exits 0 while
# ignoring them — fail here so supervisord reports the fault.
logrotate_config_ok() {
mode=$(stat -c '%a' "$CONFIG")
mode=${mode#0}
grp=$(( (mode / 10) % 10 ))
oth=$(( mode % 10 ))
case $grp in 2|3|6|7) return 1 ;; esac
case $oth in 2|3|6|7) return 1 ;; esac
return 0
}
logrotate_config_fatal() {
echo "logrotate-loop: refusing to run: $CONFIG mode $(stat -c '%a' "$CONFIG") is writable by group or others" >&2
exit 1
}
if ! logrotate_config_ok; then
logrotate_config_fatal
fi
run_logrotate() {
out=$(logrotate "$CONFIG" 2>&1) || {
echo "$out" >&2
return 1
}
case "$out" in
*Ignoring*|*Potentially\ dangerous\ mode*)
echo "$out" >&2
logrotate_config_fatal
;;
esac
return 0
}
while true; do
if logrotate /etc/logrotate.d/mail; then
if ! logrotate_config_ok; then
logrotate_config_fatal
fi
if run_logrotate; then
:
else
echo "logrotate-loop: logrotate failed, will retry after ${INTERVAL}s" >&2
+3 -2
View File
@@ -104,8 +104,9 @@ stderr_logfile=/dev/stderr
stderr_logfile_maxbytes=0
; Periodic logrotate for /data/log/mail.log (spec 9, 10: daily, 7-14 files kept
; in the image). Runs as root so logrotate can read/rotate the log; never exits
; non-zero, so it neither trips the crashexit listener nor needs restarting.
; in the image). Runs as root so logrotate can read/rotate the log. Exits
; non-zero when the config is group/other-writable (logrotate would ignore it
; silently); autorestart surfaces BACKOFF on the Status page.
[program:logrotate]
command=/usr/local/bin/logrotate-loop.sh
priority=400
+1 -1
View File
@@ -17,7 +17,7 @@ var documentedPublic = []string{
"TRUSTED_PROXY_CIDR",
}
// documentedInternal matches guide.md "Internal variables (not part of the operator interface)".
// documentedInternal matches architecture.md § Configuration "Internal env vars".
var documentedInternal = []string{
"SELFPOST_DATA_DIR",
"SELFPOST_DB_PATH",
+60 -10
View File
@@ -3,6 +3,7 @@ package main
import (
"context"
"errors"
"fmt"
"log"
"net/http"
"time"
@@ -15,17 +16,57 @@ import (
"github.com/mixeme/selfpost/internal/web"
)
// serveHTTP runs the control-panel HTTP server until ctx is cancelled, using
// the database handle shared by all roles: setup, login and the authenticated
// panel surface (security.md).
func serveHTTP(ctx context.Context, cfg config, st *store.Store) error {
// Applications own the SASL accounts and the Postfix sender map; the domain
// service delegates to them when a domain (and its applications) is deleted.
pf := postfix.New(cfg.postfixDir)
apps := app.NewService(st, app.NewSASLDB(cfg.saslDBPath, cfg.saslRealm), pf)
domains := domain.NewService(st, domain.NewOpenDKIM(cfg.opendkimDir), apps, cfg.dkimSelectorDef)
// mailStack is the panel's domain and application services plus the on-disk
// mail-path adapters they write through.
type mailStack struct {
Domains *domain.Service
Apps *app.Service
pf *postfix.Postfix
odk *domain.OpenDKIM
}
srvApp, err := web.New(st, domains, apps, web.Config{
func newMailStack(cfg config, st *store.Store) *mailStack {
pf := postfix.New(cfg.postfixDir)
odk := domain.NewOpenDKIM(cfg.opendkimDir)
apps := app.NewService(st, app.NewSASLDB(cfg.saslDBPath, cfg.saslRealm), pf)
domains := domain.NewService(st, odk, apps, cfg.dkimSelectorDef)
return &mailStack{Domains: domains, Apps: apps, pf: pf, odk: odk}
}
// Resync rebuilds OpenDKIM's tables and Postfix's sender map from SQLite and
// reloads both daemons — the same work as the Status page's Reload button.
func (m *mailStack) Resync() error {
if err := m.Domains.Resync(); err != nil {
return fmt.Errorf("opendkim resync: %w", err)
}
if err := m.Apps.Resync(); err != nil {
return fmt.Errorf("postfix resync: %w", err)
}
return nil
}
func (m *mailStack) skipReloadForTest() {
m.pf.SetReloadHook(func() error { return nil })
m.odk.SetReloadHook(func() error { return nil })
}
// resyncAfterRestore runs one mail-path Resync on the first boot after a
// backup restore. testNoReload skips the supervisord reload step so restore
// tests can verify file regeneration without a running mail stack.
func resyncAfterRestore(cfg config, st *store.Store, testNoReload bool) error {
ms := newMailStack(cfg, st)
if testNoReload {
ms.skipReloadForTest()
}
return ms.Resync()
}
// newPanel wires the panel's services over the shared database handle and
// builds the HTTP application from cfg. It is the composition of the panel as
// the environment describes it, with nothing bound to a port yet.
func newPanel(cfg config, st *store.Store) (*web.Server, error) {
ms := newMailStack(cfg, st)
return web.New(st, ms.Domains, ms.Apps, web.Config{
Hostname: cfg.hostname,
CookieSecure: cfg.cookieSecure,
SubmissionEnabled: cfg.submissionEnabled,
@@ -39,7 +80,16 @@ func serveHTTP(ctx context.Context, cfg config, st *store.Store) error {
JournalSocket: cfg.journalSocket,
SessionIdleDays: cfg.sessionIdleDays,
DNSResolvers: cfg.dnsResolvers,
RateLimitMessagesPerIP: cfg.rateLimitMessagesPerIP,
RateLimitWindowSeconds: cfg.rateLimitWindowSeconds,
}, cfg.setupTokenPath)
}
// serveHTTP runs the control-panel HTTP server until ctx is cancelled, using
// the database handle shared by all roles: setup, login and the authenticated
// panel surface (security.md).
func serveHTTP(ctx context.Context, cfg config, st *store.Store) error {
srvApp, err := newPanel(cfg, st)
if err != nil {
return err
}
+15 -2
View File
@@ -4,7 +4,6 @@
// the rate-limit checks.
//
// Copyright (C) 2026 Mikhail Yenuchenko
// SPDX-License-Identifier: AGPL-3.0-only
package main
import (
@@ -63,6 +62,8 @@ type config struct {
trustedProxies []*net.IPNet
sessionIdleDays int
dnsResolvers []string
rateLimitMessagesPerIP int
rateLimitWindowSeconds int
// Read-only inputs to the panel's status page: the certificate Postfix
// serves and the two milter sockets it connects to. The defaults mirror
@@ -117,6 +118,10 @@ func loadConfig() config {
// means dnscheck's public defaults; a closed network names its own here.
dnsResolvers: dnscheck.ParseResolvers(os.Getenv("SELFPOST_DNS_RESOLVERS")),
// Level-1 anvil defaults match build/postfix-config.sh / guide.md.
rateLimitMessagesPerIP: envInt("RATE_LIMIT_MESSAGES_PER_IP", 100),
rateLimitWindowSeconds: envInt("RATE_LIMIT_WINDOW_SECONDS", 3600),
tlsCertFile: envDefault("TLS_CERT_FILE", "/etc/postfix/tls/fullchain.pem"),
opendkimSocket: envDefault("OPENDKIM_SOCKET", "/run/opendkim/opendkim.sock"),
@@ -217,7 +222,8 @@ func run() error {
// touch the database, so schema/format skew between versions cannot corrupt
// the restored state. A match consumes the manifest; its absence is the
// normal (non-restore) case.
if err := backup.CheckRestore(cfg.manifestPath, buildinfo.Version); err != nil {
restored, err := backup.CheckRestore(cfg.manifestPath, buildinfo.Version)
if err != nil {
return err
}
@@ -230,6 +236,13 @@ func run() error {
}
defer st.Close()
if restored {
log.Printf("restore manifest accepted; regenerating mail-path maps from SQLite")
if err := resyncAfterRestore(cfg, st, false); err != nil {
return err
}
}
var wg sync.WaitGroup
errc := make(chan error, 3)
+456
View File
@@ -0,0 +1,456 @@
package main
import (
"archive/tar"
"bytes"
"compress/gzip"
"encoding/json"
"io"
"net/http"
"net/http/httptest"
"net/url"
"os"
"path/filepath"
"strings"
"testing"
"github.com/mixeme/selfpost/internal/backup"
"github.com/mixeme/selfpost/internal/buildinfo"
"github.com/mixeme/selfpost/internal/secretfile"
"github.com/mixeme/selfpost/internal/store"
"golang.org/x/crypto/bcrypt"
)
// Restoring a SelfPost backup is not a code path in the panel: the operator
// extracts the archive into the /data bind mount and starts the image, and the
// panel is expected to come up on it (architecture.md § Persistence). Nothing
// below stubs that story out — the archive is downloaded from a running panel
// through /backup, unpacked the way `tar -xzf` unpacks it, and a second panel
// is started on the result through the same startup sequence run() uses:
// CheckRestore, store.Open, newPanel, Start.
const (
restorePassword = "correct-horse-battery"
restoreDomain = "bs.example.ru"
restoreSubject = "Order confirmation"
)
// restored is the outcome of a full backup-and-restore round trip.
type restored struct {
panel http.Handler // panel booted on the restored data directory
dataDir string // the restored /data
session *http.Cookie // a session opened before the backup was taken
}
// restoreFromOwnBackup runs the operator's path end to end: seed a panel that
// has been in use, sign in, download a backup from it, extract that archive
// into an empty directory and boot a second panel there. A non-empty password
// takes the encrypted download and decrypts it on the way in, which is what an
// operator does with a .spbk file.
func restoreFromOwnBackup(t *testing.T, password string) restored {
t.Helper()
live := seedPanelData(t)
panel := bootPanel(t, live)
session := signIn(t, panel)
archive := downloadBackup(t, panel, session, password)
target := t.TempDir()
extract(t, archive, target)
return restored{panel: bootPanel(t, target), dataDir: target, session: session}
}
// The panel has to come up on the restored directory and show the state that
// was in the archive, without the operator touching anything else: the domain
// and its journal are in the database the archive carried, and the credentials
// that worked before the restore still work after it.
func TestPanelBootsOnADataDirectoryRestoredFromItsOwnBackup(t *testing.T) {
r := restoreFromOwnBackup(t, "")
body := getPage(t, r.panel, "/deliveries", signIn(t, r.panel))
for _, want := range []string{restoreDomain, restoreSubject} {
if !strings.Contains(body, want) {
t.Errorf("the restored panel's send log does not show %q:\n%s", want, body)
}
}
// The archive carries the daemon files, and the first boot after restore
// re-derives the maps from SQLite so they stay aligned with the database.
for path, want := range map[string]string{
filepath.Join("opendkim", "keys", restoreDomain, "selfpost.private"): "PRIVATE KEY",
filepath.Join("sasl", "sasldb2"): "SASLDB",
filepath.Join("postfix", "sender_login_maps"): "@" + restoreDomain + " shop\n",
} {
got, err := os.ReadFile(filepath.Join(r.dataDir, path))
if err != nil {
t.Errorf("the restored data directory has no %s: %v", path, err)
continue
}
if string(got) != want {
t.Errorf("%s = %q, want %q", path, got, want)
}
}
}
// The one-time setup link is closed by the presence of a panel user, and the
// restored database has one. A restore that reopened it would publish a link
// that creates a second global administrator on a server holding live mail
// credentials (security.md).
func TestARestoreDoesNotReopenTheSetupLink(t *testing.T) {
r := restoreFromOwnBackup(t, "")
if _, err := os.Stat(filepath.Join(r.dataDir, "setup-token")); !os.IsNotExist(err) {
t.Errorf("the restored panel wrote a setup token (stat err = %v)", err)
}
body := getPage(t, r.panel, "/login", nil)
if strings.Contains(body, "No administrator has been created yet") {
t.Errorf("the restored panel offers first-run setup:\n%s", body)
}
}
// Sessions live in the database, so they travel in the archive: a cookie that
// was valid when the backup was taken is valid again on the restored panel.
// That is the documented consequence of restoring an older backup (guide §
// Backup and restore) — stated here so it cannot change by accident.
func TestARestoredPanelHonoursSessionsFromTheArchive(t *testing.T) {
r := restoreFromOwnBackup(t, "")
rec := request(t, r.panel, http.MethodGet, "/deliveries", nil, r.session)
if rec.Code != http.StatusOK {
t.Errorf("a session from before the backup = %d on the restored panel, want 200", rec.Code)
}
}
// An encrypted download is the same archive inside an envelope, so it restores
// the same way once the password is supplied. The archive is never written to
// disk in the clear by the panel, so this is the only place the two paths can
// be shown to agree.
func TestAnEncryptedBackupRestoresTheSameWay(t *testing.T) {
r := restoreFromOwnBackup(t, "a-long-enough-password")
body := getPage(t, r.panel, "/deliveries", signIn(t, r.panel))
if !strings.Contains(body, restoreSubject) {
t.Errorf("the panel restored from an encrypted backup lost the send log:\n%s", body)
}
}
// A restore boot runs one Resync from SQLite. If the archive's Postfix map
// drifted from the database, that step puts it back before mail flows.
func TestResyncAfterRestoreHealsDriftedMaps(t *testing.T) {
dataDir := seedPanelData(t)
cfg := panelConfig(t, dataDir)
mapPath := filepath.Join(dataDir, "postfix", "sender_login_maps")
if err := os.WriteFile(mapPath, []byte("stale map\n"), 0o640); err != nil {
t.Fatalf("write stale map: %v", err)
}
manifest, err := json.Marshal(backup.Manifest{
Format: backup.FormatFull,
Version: buildinfo.Version,
CreatedAt: "2026-08-14T00:00:00Z",
})
if err != nil {
t.Fatalf("marshal manifest: %v", err)
}
if err := os.WriteFile(cfg.manifestPath, manifest, 0o644); err != nil {
t.Fatalf("write manifest: %v", err)
}
restored, err := backup.CheckRestore(cfg.manifestPath, buildinfo.Version)
if err != nil {
t.Fatalf("CheckRestore: %v", err)
}
if !restored {
t.Fatal("CheckRestore did not report a restore")
}
st, err := store.Open(cfg.dbPath)
if err != nil {
t.Fatalf("open store: %v", err)
}
defer st.Close()
if err := resyncAfterRestore(cfg, st, true); err != nil {
t.Fatalf("resync after restore: %v", err)
}
got, err := os.ReadFile(mapPath)
if err != nil {
t.Fatalf("read sender map: %v", err)
}
want := "@" + restoreDomain + " shop\n"
if string(got) != want {
t.Errorf("sender map = %q, want %q", got, want)
}
}
// The version guard is what stops a restore from being silently corrupted by
// schema skew, and it runs before anything opens the database. The manifest
// stays put on a mismatch: the operator's next move is to start the image the
// backup names, and it has to be there when they do.
func TestPanelRefusesADataDirectoryRestoredFromAnotherVersion(t *testing.T) {
live := seedPanelData(t)
var archive bytes.Buffer
if err := backup.Create(&archive, backup.Params{
DataDir: live,
DBPath: filepath.Join(live, "selfpost.db"),
Version: "9.9.9",
}); err != nil {
t.Fatalf("create backup: %v", err)
}
target := t.TempDir()
extract(t, archive.Bytes(), target)
cfg := panelConfig(t, target)
_, err := backup.CheckRestore(cfg.manifestPath, buildinfo.Version)
if err == nil {
t.Fatal("the panel booted on a data directory left by another version")
}
for _, want := range []string{"9.9.9", buildinfo.Version} {
if !strings.Contains(err.Error(), want) {
t.Errorf("the refusal does not name %q, so the operator cannot tell which image to run: %v", want, err)
}
}
if _, statErr := os.Stat(cfg.manifestPath); statErr != nil {
t.Errorf("the manifest was consumed by a refused restore: %v", statErr)
}
}
// seedPanelData builds the /data tree of a panel that has been in use: an
// administrator, a sending domain with an application and one logged message,
// and the daemon state the mail path needs (a DKIM key, the SASL database and
// Postfix's sender map).
func seedPanelData(t *testing.T) string {
t.Helper()
dataDir := t.TempDir()
st, err := store.Open(filepath.Join(dataDir, "selfpost.db"))
if err != nil {
t.Fatalf("open store: %v", err)
}
hash, err := bcrypt.GenerateFromPassword([]byte(restorePassword), bcrypt.MinCost)
if err != nil {
t.Fatalf("hash password: %v", err)
}
if err := st.CreateGlobalUser("admin", string(hash)); err != nil {
t.Fatalf("create user: %v", err)
}
dom, err := st.AddDomain(restoreDomain, "selfpost")
if err != nil {
t.Fatalf("add domain: %v", err)
}
if _, err := st.AddApplication(dom.ID, "shop", store.AddressModeWildcard, nil); err != nil {
t.Fatalf("add application: %v", err)
}
if err := st.InsertQueued(store.SendLogEntry{
QueueID: "4A1B2C3D", Domain: restoreDomain, AppLogin: "shop",
From: "noreply@" + restoreDomain, To: "customer@example.net", Subject: restoreSubject,
}); err != nil {
t.Fatalf("insert send-log row: %v", err)
}
if err := st.Close(); err != nil {
t.Fatalf("close store: %v", err)
}
for path, content := range map[string]string{
filepath.Join("opendkim", "keys", restoreDomain, "selfpost.private"): "PRIVATE KEY",
filepath.Join("sasl", "sasldb2"): "SASLDB",
filepath.Join("postfix", "sender_login_maps"): "@" + restoreDomain + " shop\n",
filepath.Join("log", "mail.log"): "postfix/smtp[1]: 4A1B2C3D: status=sent",
} {
full := filepath.Join(dataDir, path)
if err := os.MkdirAll(filepath.Dir(full), 0o750); err != nil {
t.Fatalf("mkdir %s: %v", full, err)
}
if err := os.WriteFile(full, []byte(content), 0o640); err != nil {
t.Fatalf("write %s: %v", full, err)
}
}
return dataDir
}
// panelConfig resolves the panel's own configuration for a data directory, so
// the test finds the files where the running binary would look for them rather
// than where it put them. Cookies are marked insecure for the same reason the
// e2e stand does it: the test client speaks plain HTTP.
func panelConfig(t *testing.T, dataDir string) config {
t.Helper()
t.Setenv("SELFPOST_DATA_DIR", dataDir)
t.Setenv("PANEL_COOKIE_SECURE", "false")
t.Setenv("SELFPOST_HOSTNAME", "mail.example.ru")
// MAIL_LOG's default is an absolute path, not one derived from the data
// directory; without this the panel would read the host's /data.
t.Setenv("MAIL_LOG", filepath.Join(dataDir, "log", "mail.log"))
return loadConfig()
}
// bootPanel performs the startup sequence run() performs, in the same order,
// and returns the panel's HTTP handler.
func bootPanel(t *testing.T, dataDir string) http.Handler {
t.Helper()
cfg := panelConfig(t, dataDir)
restored, err := backup.CheckRestore(cfg.manifestPath, buildinfo.Version)
if err != nil {
t.Fatalf("the panel refused to start on %s: %v", dataDir, err)
}
if restored {
if _, err := os.Stat(cfg.manifestPath); err == nil {
t.Errorf("the restore manifest was not consumed, so the next start is gated by it too")
}
}
st, err := store.Open(cfg.dbPath)
if err != nil {
t.Fatalf("open the restored database: %v", err)
}
t.Cleanup(func() { _ = st.Close() })
if restored {
if err := resyncAfterRestore(cfg, st, true); err != nil {
t.Fatalf("resync after restore: %v", err)
}
}
panel, err := newPanel(cfg, st)
if err != nil {
t.Fatalf("build the panel: %v", err)
}
if err := panel.Start(); err != nil {
t.Fatalf("start the panel: %v", err)
}
return panel.Handler()
}
// signIn signs in as the seeded administrator and returns the session cookie.
func signIn(t *testing.T, h http.Handler) *http.Cookie {
t.Helper()
form := url.Values{"username": {"admin"}, "password": {restorePassword}}
rec := request(t, h, http.MethodPost, "/login", strings.NewReader(form.Encode()), nil)
if rec.Code != http.StatusSeeOther {
t.Fatalf("sign in = %d, want 303:\n%s", rec.Code, rec.Body.String())
}
cookies := rec.Result().Cookies()
if len(cookies) == 0 {
t.Fatal("sign in issued no session cookie")
}
return cookies[0]
}
// downloadBackup takes a backup through the panel's own /backup route, the way
// the operator does. An empty password downloads the plain archive; otherwise
// the response is a .spbk envelope, which is decrypted here.
func downloadBackup(t *testing.T, h http.Handler, session *http.Cookie, password string) []byte {
t.Helper()
form := url.Values{}
if password != "" {
form.Set("encrypt", "1")
form.Set("password", password)
form.Set("password_confirm", password)
}
rec := request(t, h, http.MethodPost, "/backup", strings.NewReader(form.Encode()), session)
if rec.Code != http.StatusOK {
t.Fatalf("download a backup = %d, want 200:\n%s", rec.Code, rec.Body.String())
}
if got := rec.Header().Get("Cache-Control"); got != "no-store" {
t.Errorf("Cache-Control = %q; an archive of every secret on the server must not be cached", got)
}
body := rec.Body.Bytes()
if password == "" {
if secretfile.HasMagic(body) {
t.Fatal("an unencrypted download came back as an envelope")
}
return body
}
if !secretfile.HasMagic(body) {
t.Fatal("the download is not an encrypted envelope, so the archive left the panel in the clear")
}
r, err := secretfile.NewReader(bytes.NewReader(body), password)
if err != nil {
t.Fatalf("open the encrypted backup: %v", err)
}
plain, err := io.ReadAll(r)
if err != nil {
t.Fatalf("decrypt the backup: %v", err)
}
return plain
}
// extract unpacks a backup archive into dir, as `tar -xzf` does onto the /data
// bind mount before the image is started.
func extract(t *testing.T, archive []byte, dir string) {
t.Helper()
gz, err := gzip.NewReader(bytes.NewReader(archive))
if err != nil {
t.Fatalf("the download is not a gzip stream: %v", err)
}
tr := tar.NewReader(gz)
for {
hdr, err := tr.Next()
if err == io.EOF {
break
}
if err != nil {
t.Fatalf("read the archive: %v", err)
}
name := filepath.Clean(filepath.FromSlash(hdr.Name))
if strings.HasPrefix(name, "..") || filepath.IsAbs(name) {
t.Fatalf("the archive escapes the directory it is extracted into: %q", hdr.Name)
}
path := filepath.Join(dir, name)
switch hdr.Typeflag {
case tar.TypeDir:
if err := os.MkdirAll(path, hdr.FileInfo().Mode().Perm()); err != nil {
t.Fatalf("mkdir %s: %v", path, err)
}
case tar.TypeReg:
if err := os.MkdirAll(filepath.Dir(path), 0o750); err != nil {
t.Fatalf("mkdir %s: %v", filepath.Dir(path), err)
}
f, err := os.OpenFile(path, os.O_CREATE|os.O_WRONLY|os.O_TRUNC, hdr.FileInfo().Mode().Perm())
if err != nil {
t.Fatalf("create %s: %v", path, err)
}
if _, err := io.Copy(f, tr); err != nil {
f.Close()
t.Fatalf("write %s: %v", path, err)
}
if err := f.Close(); err != nil {
t.Fatalf("close %s: %v", path, err)
}
}
}
}
// getPage performs a GET and returns the body, failing on any non-200.
func getPage(t *testing.T, h http.Handler, target string, session *http.Cookie) string {
t.Helper()
rec := request(t, h, http.MethodGet, target, nil, session)
if rec.Code != http.StatusOK {
t.Fatalf("GET %s = %d, want 200:\n%s", target, rec.Code, rec.Body.String())
}
return rec.Body.String()
}
// request drives the panel's real handler chain, including the origin check,
// with the headers a browser on the panel's own page would send.
func request(t *testing.T, h http.Handler, method, target string, body io.Reader, session *http.Cookie) *httptest.ResponseRecorder {
t.Helper()
req := httptest.NewRequest(method, "http://mail.example.ru"+target, body)
req.Host = "mail.example.ru"
if method == http.MethodPost {
req.Header.Set("Content-Type", "application/x-www-form-urlencoded")
req.Header.Set("Sec-Fetch-Site", "same-origin")
}
if session != nil {
req.AddCookie(session)
}
rec := httptest.NewRecorder()
h.ServeHTTP(rec, req)
return rec
}
-1
View File
@@ -18,7 +18,6 @@
// docker exec -i <container> selfpost-backup -decrypt < backup.spbk > backup.tar.gz
//
// Copyright (C) 2026 Mikhail Yenuchenko
// SPDX-License-Identifier: AGPL-3.0-only
package main
import (
+1 -1
View File
@@ -22,7 +22,7 @@
services:
selfpost:
image: ghcr.io/mixeme/selfpost:1.1.0
image: ghcr.io/mixeme/selfpost:1.3.0
restart: unless-stopped
environment:
SELFPOST_HOSTNAME: "${SELFPOST_HOSTNAME:?set the mail/panel hostname, e.g. mail.example.com}"
+68 -25
View File
@@ -81,11 +81,13 @@ One process, three roles:
proxy only.
2. **journal-milter** — unix socket `JOURNAL_MILTER_SOCKET`; records From/To/
Subject/SASL user at DATA; enforces level-2 rate limits; **fail-open**
(`default_action=accept`) so milter failure does not stop mail. The level-2
count is the stored send-log rows plus the messages this process has admitted
but not yet written (`internal/milter/inflight.go`), so concurrent sessions
cannot each spend the same last slot; a reservation is released at
end-of-message, on ABORT, or after a 10-minute TTL.
(`default_action=accept`) so milter failure does not stop mail. Domain
ceilings apply to every client IP; an application ceiling with trusted IPs
raises the limit for those IPs only and skips the domain check (guide § Rate
limiting). The level-2 count is the stored send-log rows plus the messages
this process has admitted but not yet written (`internal/milter/inflight.go`),
so concurrent sessions cannot each spend the same last slot; a reservation
is released at end-of-message, on ABORT, or after a 10-minute TTL.
3. **log-tailer** — follows `MAIL_LOG`, updates send-log delivery status by
queue-id. Send-log `queued → sent` transitions depend on this goroutine alone
(`UpdateStatus` is only called from [internal/logtail](../internal/logtail/logtail.go)).
@@ -152,25 +154,31 @@ state for an older message and the page reports it as such, not as a failure.
## Panel HTTP surface
Canonical routes: [internal/web/web.go](../internal/web/web.go). Authenticated
unless noted. The table below is a summary — HTMX fragment endpoints
unless noted. Routes marked **global** return **404** for domain administrators
(`requireGlobal()` in
[internal/web/handlers/authz.go](../internal/web/handlers/authz.go)). The table
below is a summary — HTMX fragment endpoints
(`/status/fragment`, `/deliveries/rows`, `/mail-queue/body`,
`/system-log/body`, …) and every POST variant live in `web.go`.
| Route | Purpose |
|---|---|
| `/healthz` | Liveness (no auth) |
| `/license` | Embedded `LICENSE` text (no auth) |
| `/setup/*` | One-time admin bootstrap |
| `/login`, `/logout` | Session auth |
| `/status` | Process, cert, socket, PTR checks; machine CPU/memory/network |
| `/domains`, `/domains/*` | Domain and application CRUD, DKIM, L2 limits |
| `/domains/import` | Domain import (`POST`; form on the Backup page) |
| `/deliveries` | Send log with filters |
| `/deliveries/{id}` | One send-log row in full, with its `mail.log` lines |
| `/mail-queue` | Postfix queue view |
| `/system-log` | `mail.log` tail |
| `/reload` | Reload OpenDKIM + Postfix maps |
| `/backup` | Full backup download (page also hosts the import form) |
| `/account` | Admin username/password |
| `/account` | 308 redirect to `/settings` (pre-1.2.3 route, kept as a compat shim) |
| `/status`, `/status/*` | **Global.** Process, cert, socket, PTR checks; machine CPU/memory/network |
| `/domains` | Domain list; `POST /domains` (add domain) is **global** |
| `/domains/{id}`, `/domains/{id}/*` | Assigned-domain detail for domain-admins; delete domain is **global** |
| `/domains/import` | **Global.** Domain import (`POST`; form on the Backup page) |
| `/deliveries`, `/deliveries/{id}` | Send log with filters; scoped to assigned domains for domain-admins |
| `/mail-queue`, `/mail-queue/*` | **Global.** Postfix queue view |
| `/system-log`, `/system-log/*` | **Global.** `mail.log` tail |
| `/reload` | **Global.** `POST` — reload OpenDKIM + Postfix maps |
| `/backup`, `/backup/*` | **Global.** Full backup download (page also hosts the import form) |
| `/settings` | Username/password for any user; DMARC report default is **global** only |
| `/users`, `/users/*` | **Global.** Panel user CRUD |
HTMX polling refreshes monitoring fragments (5 s while the operator is active on
the page, 30 s when the tab is visible but idle, none when hidden — scheduled in
@@ -189,13 +197,16 @@ holds the cookie works after process restart, redeploy, or full backup restore.
absolute cap (regular use keeps the session alive indefinitely).
- **Renewal** — DB `last_seen` and cookie `Max-Age` update at most once per hour
(`renewThreshold` in [internal/web/auth/session.go](../internal/web/auth/session.go)).
- **Password change** — all other sessions are deleted; the current session stays
active ([internal/store/sessions.go](../internal/store/sessions.go),
[handlers_account.go](../internal/web/handlers/handlers_account.go)).
- **Password change on `/settings`** — changing your own password deletes
every other session for that user; the current session stays active
([internal/store/sessions.go](../internal/store/sessions.go),
[handlers_settings.go](../internal/web/handlers/handlers_settings.go)).
A global administrator resetting another user's password on `/users` updates
the hash but does not delete that user's existing sessions.
Restoring an **older** backup also restores session rows: a session invalidated
after that backup was taken can become valid again if the browser still has the
cookie and idle timeout has not expired.
Restoring an **older** backup also restores session rows: a session removed
after that backup was taken can become valid again if the browser still holds
the cookie and the restored row's `expires_at` has not passed.
---
@@ -290,9 +301,13 @@ Not in `/data`: TLS certificates (reverse-proxy mount), Postfix queue
`mail.log` via logrotate (14 rotated files, check every 6h, rename +
`postfix reload` in `postrotate` — see § Log tailer above).
**Backup:** panel button or `selfpost-backup` CLI — SQLite snapshot + tar of
**Restore:** panel button or `selfpost-backup` CLI — SQLite snapshot + tar of
`/data` tree, minus `log/`, the setup token and any `tls/`; version check on
restore. Stopped-container `tar` of `./data` is safe (see guide).
restore. On the first successful boot after restore, the panel runs one
**Resync** — OpenDKIM's tables and Postfix's sender map are re-derived from
SQLite and both daemons are reloaded, so drift between the extracted archive
and the database is healed before mail flows (same step as `POST /reload` on
demand). Stopped-container `tar` of `./data` is safe (see guide).
**Optional encryption** of the two secret-bearing downloads
([internal/secretfile](../internal/secretfile/secretfile.go)): password →
@@ -314,5 +329,33 @@ origin check, no CSRF tokens) are documented there separately.
## Configuration
Public and internal env vars: [guide § Environment variables](guide.md#environment-variables).
Public env vars: [guide § Environment variables](guide.md#environment-variables).
Regression test: [cmd/panel/envdoc_test.go](../cmd/panel/envdoc_test.go).
**Internal env vars.** The following are read by the panel or startup scripts
but are not part of the operator interface — not meant to be changed in a
normal deployment; documented here so an accidental override reads as
unsupported rather than as a missing doc:
- **Panel paths and tuning:** `SELFPOST_DATA_DIR` (`/data`), `SELFPOST_DB_PATH`
(`/data/selfpost.db`), `SELFPOST_SETUP_TOKEN_FILE`
(`/data/setup-token`), `PANEL_HTTP_ADDR` (`:8080`),
`JOURNAL_MILTER_SOCKET` (`/run/selfpost/journal.sock`), `MAIL_LOG`
(`/data/log/mail.log` — read by the panel and written by Postfix, so a change
here has to be matched in `build/postfix-config.sh`),
`PANEL_COOKIE_SECURE` (`true`), `OPENDKIM_SOCKET`
(`/run/opendkim/opendkim.sock`), `OPENDKIM_DIR` (`/data/opendkim`),
`DKIM_SELECTOR_DEFAULT` (`selfpost`), `SASL_DB_PATH`
(`/data/sasl/sasldb2`), `SASL_REALM` (defaults to `SELFPOST_HOSTNAME`),
`POSTFIX_DIR` (`/data/postfix`), `POSTFIX_SENDER_LOGIN_MAPS`
(`/data/postfix/sender_login_maps` — read by Postfix config only; the panel
always writes `<POSTFIX_DIR>/sender_login_maps`, so overriding this env alone
desyncs the map Postfix reads from the file the panel maintains).
- **Milter and Postfix startup:** `MILTER_CONNECT_TIMEOUT` (`15s`),
`MILTER_COMMAND_TIMEOUT` (`15s`), `MILTER_CONTENT_TIMEOUT` (`30s`),
`MILTER_WAIT_TIMEOUT` (`30` seconds).
- **Background maintenance:** `TLS_RELOAD_INTERVAL_SECONDS` (`86400` — daily
`postfix reload` to pick up renewed certificates),
`LOGROTATE_INTERVAL_SECONDS` (`21600` — check `mail.log` rotation every six
hours; logrotate keeps 14 rotated files on a daily schedule, and each
rotation triggers `postfix reload`).
+76 -17
View File
@@ -93,8 +93,9 @@ mechanics → Haiku. Reviewers must not be the author of the code under review.
## External libraries
The project is **AGPL-3.0** ([LICENSE](../LICENSE)). Copyright holder and
third-party notices: [NOTICE](../NOTICE). New Go dependencies must be
permissive or GPL-family (see
third-party notices: [NOTICE](../NOTICE). The tree does not use per-file
`SPDX-License-Identifier` headers; AGPL-3.0 does not require them. New Go
dependencies must be permissive or GPL-family (see
[.cursor/rules/agent-rules.mdc](../.cursor/rules/agent-rules.mdc)).
### Main module (`go.mod`)
@@ -113,6 +114,7 @@ the tree are AGPL-3.0-compatible.
| Asset | Version | Repository | License |
|---|---|---|---|
| `internal/web/view/static/htmx.min.js` | 2.0.4 | <https://github.com/bigskysoftware/htmx> | 0BSD |
| `internal/web/view/static/ibm-plex-*.woff2` | latin subset | <https://github.com/IBM/plex> | SIL OFL 1.1 (`OFL.txt` beside the files) |
### E2e module (`test/e2e/go.mod`)
@@ -128,8 +130,10 @@ the image.
Postfix, OpenDKIM, `supervisord`, `sasl2-bin`, `logrotate`, and others come
from Debian bookworm repositories; licenses are in each package's `copyright`
file on <https://packages.debian.org/bookworm/>.
The image also ships [LICENSE](../LICENSE) and [NOTICE](../NOTICE) under
`/usr/share/doc/selfpost/`. The panel serves the AGPL text at `/license`.
The image also ships [LICENSE](../LICENSE), [NOTICE](../NOTICE), and the IBM
Plex [OFL.txt](../internal/web/view/static/OFL.txt) under
`/usr/share/doc/selfpost/`. The panel serves the AGPL text at `/license` and
the OFL text at `/static/OFL.txt`.
---
@@ -141,7 +145,7 @@ Requires Go 1.26+ and `CGO_ENABLED=0`.
```sh
make build # bin/panel, bin/selfpost-backup (VERSION=dev by default)
make build VERSION=1.1.0
make build VERSION=1.2.3
```
Or directly:
@@ -162,8 +166,10 @@ docker build -f build/Dockerfile -t selfpost:dev --build-arg VERSION=dev .
```
The Dockerfile has a build stage (`go vet`, `go build` with `VERSION`) and a
runtime stage (Debian + mail stack). See [architecture.md](architecture.md) §
Image and processes.
runtime stage (Debian + mail stack). Runtime config and scripts use `COPY
--chmod` so file modes in the image do not depend on how the build context was
synced (e.g. a Windows checkout widening permissions on `logrotate-mail.conf`).
See [architecture.md](architecture.md) § Image and processes.
---
@@ -185,9 +191,12 @@ tag / push only on explicit request (see `release.yml`).
### Release image
The release image is published **only on tag** `vX.Y.Z` (not on every push to
`main`). The tag is the single source of version: it drives the image tag and
`-ldflags` in the binaries so they cannot drift apart.
The release image is published **only** for a SemVer version `X.Y.Z`: a
**published** GitHub Release whose tag is `vX.Y.Z`, or a `workflow_dispatch`
that supplies that version. Pushing a git tag alone does not publish. Ordinary
commits, and a dispatch from `main` without a version input, do not publish.
The version is the single source that drives the image tag and `-ldflags` in
the binaries so they cannot drift apart.
**Steps (on explicit request):**
@@ -195,8 +204,42 @@ The release image is published **only on tag** `vX.Y.Z` (not on every push to
tag in [deploy/docker-compose.yml](../deploy/docker-compose.yml) (and any
local-trial image references) in the **same** release commit.
2. Create and push git tag `vX.Y.Z` on that commit.
3. Workflow [release.yml](../.github/workflows/release.yml) builds, e2e-gates,
and publishes `ghcr.io/mixeme/selfpost:X.Y.Z`.
3. Publish the GitHub Release for `vX.Y.Z` (not a draft).
4. Workflow [release.yml](../.github/workflows/release.yml) builds, e2e-gates,
and publishes `ghcr.io/mixeme/selfpost:X.Y.Z` (checks out tag `vX.Y.Z`).
**GitHub Release vs GHCR.** The public [Releases](https://github.com/mixeme/selfpost/releases)
page lists only **published** releases. A draft is visible to maintainers only —
it looks like “no releases” to everyone else. CI does not create or publish the
GitHub Release; you do that in the UI. Deleting a releases git tag on GitHub
(or re-pushing tags while cleaning the registry) converts a published release
back into a **draft** — that matches “I published three times and it keeps
disappearing”. After publish, leave the tag on GitHub; clean up only unwanted
GHCR package versions, not the git tag.
Push workflow and source changes to **github.com/mixeme/selfpost** before
publishing — Actions reads that repo, not Gitea.
**Gitea → GitHub tag mirror.** If every tag push from Gitea is mirrored to
GitHub, two things follow:
1. **GitHub Release tags must not be deleted on GitHub.** Many mirror setups
prune remote tags that are absent on Gitea (or re-push with `--force` /
`--prune`). Deleting `v1.0.0` / `v1.3.0` on GitHub converts a published
Release back to draft. Mirror **branches and new tags forward**; do not
delete release tags on the GitHub side. GHCR cleanup is package versions in
the UI — not `git push github --delete` and not tag prune on the mirror.
2. **Tag push runs the workflow file at that tag's commit**, not `main`. `v1.0.0`
still points at a commit whose `release.yml` has `on: push: tags` and no
per-arch GHCR cleanup — every mirror (re)push of that tag can republish
`1.0.0-amd64` / `1.0.0-arm64`. Tags from `v1.3.0` onward only run
`release.yml` on **Publish release** (`release: published`), so mirroring
those tags alone does not start the image build.
Safe mirror: push tags to GitHub without deleting existing ones; keep release
tags on Gitea; publish the GitHub Release on github.com after the mirror has
the tag.
Ordinary commits **do not** publish an image. The compose pin and the git tag
must match (`1.0.0` / `v1.0.0` for the first published release). Intermediate
@@ -263,7 +306,11 @@ operator would actually use.
**Coverage (summary):** bootstrap → SMTP AUTH → delivery → DKIM verify →
send-log `queued → sent`; negatives (no AUTH, relay, sender/login mismatch,
L1/L2 limits, milter fail-open, bad `SELFPOST_HOSTNAME`, session survives
`docker restart`). Polling with timeouts only — no fixed `sleep`.
`docker restart`); startup checks that supervisord actually brought up
OpenDKIM, the panel, and Postfix (`checkSupervisorProcesses`), plus logrotate
config-mode and forced-rotation checks (`checkLogrotateConfigMode`,
`checkLogrotateRotation` — [test/e2e/logrotate_check.go](../test/e2e/logrotate_check.go)).
Polling with timeouts only — no fixed `sleep`.
Requires **Docker + Compose v2** on the machine running the suite.
@@ -278,21 +325,33 @@ Workflows in [.github/workflows/](../.github/workflows/). What each job runs —
`gofmt -l``go vet ./...``go test ./...` (main module, no e2e).
### `release.yml` — push of tag `vX.Y.Z` or `workflow_dispatch`
### `release.yml` — published GitHub Release, or `workflow_dispatch` with SemVer
Publishing a GitHub Release runs `release.yml` directly (`release: published`,
same pattern as gosentry / imap-scrub). You can also run it manually via
`workflow_dispatch` with an explicit `X.Y.Z` input. A bare git tag push does not
run the workflow. The build always checks out `vX.Y.Z`, not `main` HEAD.
`prepare` takes the version from `github.event.release.tag_name` on a release
event, or from the `workflow_dispatch` `version` input. A dispatch whose input
is missing or not `X.Y.Z` fails in `prepare`.
```
prepare (version from tag)
release: published
prepare (version from release tag or workflow_dispatch input; checkout vX.Y.Z)
→ build [matrix: ubuntu-latest / ubuntu-24.04-arm]
→ docker build --load (VERSION from tag)
→ docker build --load (VERSION from prepare)
→ e2e (test/e2e)
→ push ghcr.io/...:X.Y.Z-amd64 | X.Y.Z-arm64
→ merge
→ docker buildx imagetools create → unified manifest X.Y.Z
→ GitHub Packages API → drop X.Y.Z-amd64 and X.Y.Z-arm64 from GHCR
```
Native per-arch matrix (no QEMU): running the full Postfix/OpenDKIM stack under
emulation for e2e is impractical. E2e first, then push — the registry receives
the bytes that passed the gate.
the bytes that passed the gate. Only `ghcr.io/mixeme/selfpost:X.Y.Z` remains
tagged in GHCR; per-arch names exist briefly during the merge job.
A failed e2e **blocks** image publication.
+531 -237
View File
@@ -3,55 +3,53 @@
Detailed install, configuration, and day-to-day operations. For a short
overview and quick start, see [README.md](../README.md).
This guide has three parts: **[Installation](#installation)** (getting a
container running with a working reverse proxy and TLS), **[Instance
administration](#instance-administration)** (running and maintaining the
SelfPost server itself — status, backups, users, upgrades), and **[Domain
administration](#domain-administration)** (day-to-day work on the sending
domains hosted on that instance — DNS, deliveries, rate limits, applications).
## Table of contents
- [Reverse proxy (mandatory)](#reverse-proxy-mandatory)
- [Local trial](#local-trial)
- [Environment variables](#environment-variables)
- [DNS setup](#dns-setup)
- [IP warmup](#ip-warmup)
- [Operations](#operations)
- [Rate limiting](#rate-limiting)
- [Backup, restore, and moving a single domain](#backup-restore-and-moving-a-single-domain)
- [Installation](#installation)
- [Ports](#ports)
- [Local trial](#local-trial)
- [Initial setup](#initial-setup)
- [Full deployment](#full-deployment)
- [Fixed image tag](#fixed-image-tag)
- [Environment variables](#environment-variables)
- [Reverse proxy (mandatory)](#reverse-proxy-mandatory)
- [Instance administration](#instance-administration)
- [Status](#status)
- [Mail queue and System log](#mail-queue-and-system-log)
- [Settings](#settings)
- [Users](#users)
- [Sessions](#sessions)
- [Upgrading](#upgrading)
- [Container health](#container-health)
- [Server-level DNS (PTR/rDNS)](#server-level-dns-ptrrdns)
- [Rate limiting — level 1 (IP backstop)](#rate-limiting--level-1-ip-backstop)
- [Full backup and restore](#full-backup-and-restore)
- [Encrypting a backup or export](#encrypting-a-backup-or-export)
- [Published ports](#published-ports)
- [Fixed image tag](#fixed-image-tag)
- [Domain administration](#domain-administration)
- [Domains page](#domains-page)
- [Domain-level DNS (SPF, DKIM, DMARC)](#domain-level-dns-spf-dkim-dmarc)
- [IP warmup](#ip-warmup)
- [Rate limiting — level 2 (domain and application)](#rate-limiting--level-2-domain-and-application)
- [Deliveries](#deliveries)
- [Exporting and importing a single domain](#exporting-and-importing-a-single-domain)
## Reverse proxy (mandatory)
## Installation
SelfPost's panel speaks plain HTTP and never terminates TLS itself — a reverse
proxy in front of it is not optional. The proxy is also the project's only
source of TLS certificates: whatever it obtains via ACME/Let's Encrypt gets
bind-mounted **read-only** into the SelfPost container, and Postfix uses those
same PEM files for TLS on 465 (and 587, if enabled). If the panel and the mail
service share one hostname — the common case — it's genuinely one certificate
serving both.
### Ports
SelfPost isn't tied to a specific proxy; pick whichever fits your host:
`deploy/docker-compose.yml` maps **465** and **587** to the host. Port 465
(smtps) is always active. Port **587** is published even when
`SUBMISSION_ENABLE=false`; nothing listens until you set it to `true` — harmless,
but it can look like an open port in external scans.
| Proxy | Where certs live | Fragment |
|---|---|---|
| **Apache** (default/recommended) | Host disk, via the certbot Apache plugin — PEM files ready to bind-mount, no extraction step. | [deploy/apache/selfpost-vhost.conf](../deploy/apache/selfpost-vhost.conf) |
| nginx | Host disk, via a certbot sidecar container — same PEM-ready shape as Apache. | [deploy/nginx/](../deploy/nginx/) |
| Caddy | Automatic ACME, zero extra containers — simplest, but its on-disk cert path is versioned internal layout, not a stable API; verify it for the Caddy version you run. | [deploy/caddy/](../deploy/caddy/) |
| Traefik | Bundled inside `acme.json` — needs a small extraction script to produce standalone PEM files. | [deploy/traefik/](../deploy/traefik/) |
Apache is the recommended default because the certbot Apache plugin already
writes plain `fullchain.pem`/`privkey.pem` files to a predictable path with no
extra moving parts between "certificate issued" and "Postfix can read it."
**The proxy needs no security configuration of its own.** The panel emits its
own `Content-Security-Policy`, `Strict-Transport-Security`, `X-Frame-Options`,
`X-Content-Type-Options` and `Referrer-Policy` — deliberately, so the part
that's easy to get wrong lives in the service rather than in a config file
somebody edits under pressure. There is exactly one thing the proxy must do:
**pass the original `Host` header through**. All four fragments above already
do (Apache `ProxyPreserveHost On`, nginx `proxy_set_header Host $host`, Caddy
and Traefik by default). A proxy that rewrites `Host` instead makes the panel
reject every form submission as cross-origin — the log says so explicitly,
printing the `Origin` and `Host` it compared.
## Local trial
### Local trial
The [README quick start](../README.md#quick-start) runs a single container
with `PANEL_COOKIE_SECURE=false` and port 8080 published on localhost. No
@@ -85,7 +83,99 @@ Add `-v /tmp/selfpost-certs:/etc/postfix/tls:ro` to the `docker run` command
(and keep `SELFPOST_HOSTNAME=mail.local.test` so it matches the certificate CN).
Clients must skip TLS verification — the cert is not from a public CA.
## Environment variables
### Initial setup
On first start the one-time setup URL is printed in the container log
(`docker compose logs -f`) and written to `/data/setup-token` inside the
container — `./data/setup-token` on the host, mode `0600` — then deleted when
setup completes. The link is `https://<SELFPOST_HOSTNAME>/setup/<token>` (path
token, not a query string), valid for ten minutes. Open it to choose the
administrator username and password — until then the panel has no login. If
this host ships container logs to a central aggregator, prefer reading the
file:
```sh
docker compose exec selfpost cat /data/setup-token
```
### Full deployment
Production layout: one `docker-compose.yml`, a `.env`, persistent `./data`, and
TLS PEM files at `./certs` (read by Postfix on 465/587). The panel is reached
only through a [reverse proxy](#reverse-proxy-mandatory) on 443 — port 8080 is
bound to localhost in the default compose file.
| Artefact | Path |
|---|---|
| Compose file (fixed image tag) | [deploy/docker-compose.yml](../deploy/docker-compose.yml) |
| Environment template | [deploy/.env.example](../deploy/.env.example) |
| Apache vhost (recommended) | [deploy/apache/selfpost-vhost.conf](../deploy/apache/selfpost-vhost.conf) |
| nginx | [deploy/nginx/](../deploy/nginx/) |
| Caddy | [deploy/caddy/](../deploy/caddy/) |
| Traefik | [deploy/traefik/](../deploy/traefik/) |
**1. Fetch the base files.**
```sh
mkdir -p selfpost/data selfpost/certs && cd selfpost
curl -O https://raw.githubusercontent.com/mixeme/selfpost/main/deploy/docker-compose.yml
curl -O https://raw.githubusercontent.com/mixeme/selfpost/main/deploy/.env.example
cp .env.example .env
```
Edit `.env` — at minimum set `SELFPOST_HOSTNAME` to your mail hostname (bare
FQDN, e.g. `mail.example.com`). It must match the PTR record you request from
your provider and the certificate your proxy will obtain. See [Environment
variables](#environment-variables) for the full list.
**2. Reverse proxy and TLS.** Pick and set up one proxy — see [Reverse proxy
(mandatory)](#reverse-proxy-mandatory) for the per-proxy commands. The same
certificate must end up under `./certs` as `fullchain.pem` and `privkey.pem`
so Postfix can serve it on 465 (and 587 if enabled).
**3. Start SelfPost.** If you used Apache on the host (the recommended
option), start only the base compose file from your `selfpost/` directory:
```sh
docker compose up -d
```
The nginx/Caddy/Traefik fragments already include `docker compose up -d`
skip this if you ran one of those.
**Get the setup URL** — open it in a browser to create the admin account (see
[Initial setup](#initial-setup)):
```sh
docker compose logs selfpost 2>&1 | grep -m1 'http'
```
```sh
cat ./data/setup-token
```
**4. DNS and sending.** Before sending real mail:
1. Confirm PTR/rDNS for the server IP points at `SELFPOST_HOSTNAME` (Status
page → *Re-check*) — see [Server-level DNS](#server-level-dns-ptrrdns).
2. For each domain you add in the panel, publish SPF, DKIM, and DMARC at the
same time ([Domain-level DNS](#domain-level-dns-spf-dkim-dmarc)).
3. Warm up a new IP gradually ([IP warmup](#ip-warmup)).
#### Fixed image tag
`deploy/docker-compose.yml` pins an explicit version (`ghcr.io/mixeme/selfpost:X.Y.Z`),
deliberately never `:latest`. The current pin is `1.3.0`. Intermediate
CHANGELOG sections (`0.2.0``0.6.0`) record development cuts from before that
image was published. Pinning matters because of the backup version check (see
[Full backup and restore](#full-backup-and-restore)): the panel binary's
embedded version and the image tag that produced it are the same value by
construction (the release CI stamps both from one git tag — see
`.github/workflows/release.yml`), so the pin is what makes "restore into the
same version" a checkable fact rather than a guess. Upgrade by bumping the tag
deliberately, not by riding a moving target — see [Upgrading](#upgrading).
### Environment variables
Copy [deploy/.env.example](../deploy/.env.example) to `.env` next to your
`docker-compose.yml`. The table below lists every variable an operator is
@@ -95,7 +185,7 @@ expected to set; defaults match the code exactly.
|---|---|---|---|
| `SELFPOST_HOSTNAME` | Mail-server identity: Postfix HELO/EHLO, SASL realm, certificate CN/SAN, and the hostname the PTR check expects. Bare FQDN only — no scheme or port. | *(required)* | `.env` |
| `SUBMISSION_ENABLE` | When `true`, also listen on port 587 with STARTTLS (RFC 6409 submission) alongside the primary 465/smtps listener. | `false` | `.env` |
| `RATE_LIMIT_MESSAGES_PER_IP` | Level-1 backstop: maximum messages one client IP may submit per window (Postfix `smtpd_client_message_rate_limit`). See [Rate limiting](#rate-limiting). | `100` | `.env` |
| `RATE_LIMIT_MESSAGES_PER_IP` | Level-1 backstop: maximum messages one client IP may submit per window (Postfix `smtpd_client_message_rate_limit`). See [Rate limiting — level 1](#rate-limiting--level-1-ip-backstop). | `100` | `.env` |
| `RATE_LIMIT_WINDOW_SECONDS` | Level-1 window length in seconds (Postfix `anvil_rate_time_unit`). | `3600` | `.env` |
| `SEND_LOG_RETENTION_DAYS` | Days of send-log history kept before the background sweep deletes rows — the main driver of `/data` growth over time. | `90` | `.env` |
| `PANEL_SESSION_IDLE_DAYS` | Sliding idle timeout for the panel login session, in days. There is no absolute cap: an admin who keeps coming back stays signed in indefinitely. | `7` | `.env` |
@@ -106,118 +196,128 @@ TLS certificate paths (`TLS_CERT_FILE`, `TLS_KEY_FILE`) are fixed in
[deploy/docker-compose.yml](../deploy/docker-compose.yml) to match the `./certs`
bind mount — configure the mount, not these variables.
**Internal variables (not part of the operator interface).** The following are
read by the panel or startup scripts but are not meant to be changed in a
normal deployment; documenting them here avoids treating accidental overrides as
supported configuration:
The image also reads a number of internal, non-operator env vars (paths,
timeouts, tuning) — not part of this interface; see
[architecture.md § Configuration](architecture.md#configuration) if you need
them.
- **Panel paths and tuning:** `SELFPOST_DATA_DIR` (`/data`), `SELFPOST_DB_PATH`
(`/data/selfpost.db`), `SELFPOST_SETUP_TOKEN_FILE`
(`/data/setup-token`), `PANEL_HTTP_ADDR` (`:8080`),
`JOURNAL_MILTER_SOCKET` (`/run/selfpost/journal.sock`), `MAIL_LOG`
(`/data/log/mail.log` — read by the panel and written by Postfix, so a change
here has to be matched in `build/postfix-config.sh`),
`PANEL_COOKIE_SECURE` (`true`), `OPENDKIM_SOCKET`
(`/run/opendkim/opendkim.sock`), `OPENDKIM_DIR` (`/data/opendkim`),
`DKIM_SELECTOR_DEFAULT` (`selfpost`), `SASL_DB_PATH`
(`/data/sasl/sasldb2`), `SASL_REALM` (defaults to `SELFPOST_HOSTNAME`),
`POSTFIX_DIR` (`/data/postfix`), `POSTFIX_SENDER_LOGIN_MAPS`
(`/data/postfix/sender_login_maps` — read by Postfix config only; the panel
always writes `<POSTFIX_DIR>/sender_login_maps`, so overriding this env alone
desyncs the map Postfix reads from the file the panel maintains).
- **Milter and Postfix startup:** `MILTER_CONNECT_TIMEOUT` (`15s`),
`MILTER_COMMAND_TIMEOUT` (`15s`), `MILTER_CONTENT_TIMEOUT` (`30s`),
`MILTER_WAIT_TIMEOUT` (`30` seconds).
- **Background maintenance:** `TLS_RELOAD_INTERVAL_SECONDS` (`86400` — daily
`postfix reload` to pick up renewed certificates),
`LOGROTATE_INTERVAL_SECONDS` (`21600` — check `mail.log` rotation every six
hours; logrotate keeps 14 rotated files on a daily schedule, and each
rotation triggers `postfix reload`).
### Reverse proxy (mandatory)
## DNS setup
SelfPost's panel speaks plain HTTP and never terminates TLS itself — a reverse
proxy in front of it is not optional. The proxy is also the project's only
source of TLS certificates: whatever it obtains via ACME/Let's Encrypt gets
bind-mounted **read-only** into the SelfPost container, and Postfix uses those
same PEM files for TLS on 465 (and 587, if enabled). If the panel and the mail
service share one hostname — the common case — it's genuinely one certificate
serving both.
Two different scopes — don't confuse them:
SelfPost isn't tied to a specific proxy; pick whichever fits your host:
**Server level (once, for the machine itself):**
- **PTR/rDNS** for the server's IP, pointing at its mail hostname. Most
receiving mail servers weigh this heavily; get it from whoever assigns the IP
(hosting provider's panel/support), not from your own DNS zone.
| Proxy | Where certs live | Fragment |
|---|---|---|
| **Apache** (default/recommended) | Host disk, via the certbot Apache plugin — PEM files ready to bind-mount, no extraction step. | [deploy/apache/selfpost-vhost.conf](../deploy/apache/selfpost-vhost.conf) |
| nginx | Host disk, via a certbot sidecar container — same PEM-ready shape as Apache. | [deploy/nginx/](../deploy/nginx/) |
| Caddy | Automatic ACME, zero extra containers — simplest, but its on-disk cert path is versioned internal layout, not a stable API; verify it for the Caddy version you run. | [deploy/caddy/](../deploy/caddy/) |
| Traefik | Bundled inside `acme.json` — needs a small extraction script to produce standalone PEM files. | [deploy/traefik/](../deploy/traefik/) |
**Domain level (for *every* sending domain you add in the panel):**
- **SPF** — a TXT record on the domain authorizing this server to send on its
behalf (e.g. `v=spf1 a mx ip4:<server IP> -all`, adjusted to your setup).
- **DKIM** — a TXT record with the exact value the panel shows on that
domain's page (`domain page → DKIM TXT record`), one selector per domain.
- **DMARC** — a `_dmarc` TXT record. The panel suggests `p=none` (monitoring
only, safe to publish immediately). On a send-only relay the sending domain
often has no inbox, so `rua=` is optional — configure a default report address
in *Settings* or per domain when you have a mailbox that receives inbound mail
elsewhere. If `rua=` points at another domain, publish `_report._dmarc` on that
hub domain too; the panel checks it. Public mail hosts (Gmail, Outlook, …)
cannot be used as external report destinations.
Apache is the recommended default because the certbot Apache plugin already
writes plain `fullchain.pem`/`privkey.pem` files to a predictable path with no
extra moving parts between "certificate issued" and "Postfix can read it."
Skipping any of the three per-domain records is the single most common reason
mail lands in spam even though SelfPost delivered it correctly — DKIM passing
doesn't help if SPF/DMARC are absent. **Whenever you add a new domain in the
panel, add its DNS records at the same time**, not later.
**The proxy needs no security configuration of its own.** The panel emits its
own `Content-Security-Policy`, `Strict-Transport-Security`, `X-Frame-Options`,
`X-Content-Type-Options` and `Referrer-Policy` — deliberately, so the part
that's easy to get wrong lives in the service rather than in a config file
somebody edits under pressure. There is exactly one thing the proxy must do:
**pass the original `Host` header through**. All four fragments above already
do (Apache `ProxyPreserveHost On`, nginx `proxy_set_header Host $host`, Caddy
and Traefik by default). A proxy that rewrites `Host` instead makes the panel
reject every form submission as cross-origin — the log says so explicitly,
printing the `Origin` and `Host` it compared.
The panel checks both scopes for you and tells you what is actually published:
the *Status* page verifies the server's hostname and its reverse record
(forward-confirmed reverse DNS), and each domain's page shows a *DNS status*
card comparing the published DKIM record against the key this server signs with,
plus the domain's SPF, DMARC, and (when configured) DMARC report-authorisation
records. Results are cached for a few minutes;
use *Re-check* right after publishing a record. The SPF check is deliberately
shallow — it looks for a mechanism that literally covers this server's address
and does not follow `include:` or `redirect=`, so a record that authorizes the
server through an include is reported as "cannot tell" rather than as a failure.
In every case the proxy terminates HTTPS for the panel; the resulting
certificate must end up under `./certs` as `fullchain.pem` and `privkey.pem`.
## IP warmup
**Apache (recommended, on the host).** Install Apache with `ssl`, `proxy`, and
`proxy_http` enabled. Copy
[deploy/apache/selfpost-vhost.conf](../deploy/apache/selfpost-vhost.conf) into your
vhost directory, replace `mail.example.com` with your hostname, enable the site,
then issue a certificate:
A brand-new IP has no sending history, so receiving servers are cautious with
it regardless of how correct your DKIM/SPF/DMARC are. Start with low volume to
a domain, increase gradually over days/weeks rather than sending everything on
day one, and check the IP against major blocklists (Spamhaus and similar)
before and during warmup. This is inherent to how mail reputation works on the
public internet, not something SelfPost's configuration can shortcut.
```sh
sudo certbot --apache -d mail.example.com
```
## Operations
Point `./certs` at the PEM files certbot wrote (symlink is fine):
```sh
ln -s /etc/letsencrypt/live/mail.example.com certs
```
**nginx (containerised).** From the `deploy/` directory, merge the nginx
fragment and issue the first certificate before nginx can serve HTTPS:
```sh
docker compose -f docker-compose.yml -f nginx/docker-compose.nginx.yml \
run --rm certbot certonly --webroot -w /var/www/certbot \
-d mail.example.com --email you@example.com --agree-tos --no-eff-email
docker compose -f docker-compose.yml -f nginx/docker-compose.nginx.yml up -d
```
Edit [deploy/nginx/nginx.conf.example](../deploy/nginx/nginx.conf.example) and
replace `mail.example.com` first. The fragment bind-mounts certbot's output into
both nginx and SelfPost.
**Caddy (containerised, automatic ACME).** Edit
[deploy/caddy/Caddyfile](../deploy/caddy/Caddyfile) and the `<hostname>` placeholders
in [deploy/caddy/docker-compose.caddy.yml](../deploy/caddy/docker-compose.caddy.yml),
then:
```sh
docker compose -f docker-compose.yml -f caddy/docker-compose.caddy.yml up -d
```
Verify Caddy's on-disk cert path for your version before relying on the
default mount — see the comment at the top of the Caddy compose fragment.
**Traefik (containerised).** Edit the `Host(...)` label and ACME email in
[deploy/traefik/docker-compose.traefik.yml](../deploy/traefik/docker-compose.traefik.yml),
start the stack, then extract PEM files for Postfix whenever Traefik issues or
renews a certificate:
```sh
docker compose -f docker-compose.yml -f traefik/docker-compose.traefik.yml up -d
./traefik/extract-cert.sh ./traefik/letsencrypt/acme.json mail.example.com ./traefik/extracted-certs
```
Schedule `extract-cert.sh` (cron or a timer) alongside Traefik's renewals.
## Instance administration
After sign-in the panel opens on **Status** — the place to answer "is the
service healthy and will mail be accepted?"
- **Status** (`/status`) — supervised processes (Postfix, OpenDKIM, panel),
TLS certificate validity and expiry, milter socket presence, and a short
Postfix queue summary. The **Machine** card adds the resource usage of the
host underneath — processor (with the load average), memory and swap, and
per-interface network throughput and totals — read from the kernel's
counters; CPU and throughput are measured between refreshes, so they appear
one refresh after the page opens. A fully busy processor or a machine out of
memory is a warning here, because both delay or kill the mail path;
throughput is only reported. The hostname block compares `SELFPOST_HOSTNAME`
against the PTR record the internet publishes for this server's IP
(forward-confirmed reverse DNS); use *Re-check* after changing DNS. The
**Reload configuration** button re-applies OpenDKIM tables and the Postfix
sender map from the database — use it if daemons drifted from what the panel
shows after manual edits under `/data`.
- **Domains** (`/domains`) — add sending domains, inspect each domain's DKIM
TXT value, SPF/DMARC checks, and SASL applications. Per-domain rate limits
(level 2) are configured here. *Export domain* writes a single-domain archive;
*Import a domain* on the Backup page reads one back in.
- **Deliveries** (`/deliveries`) — searchable send log with server-side filters
by domain and application. A row identifies its message and nothing more —
time, sender, recipient, subject and status `queued` (accepted, not yet
delivered), `sent` (handed off successfully), `deferred` (Postfix is retrying),
`bounced` (final failure), or `rejected` (refused — for example by a level-2
rate limit); *Details* opens that row's own page
(`/deliveries/{id}`). That page carries the sending domain, the application it
was submitted under, the Postfix queue id and the journal id, beside the
message's history — when it was accepted and what Postfix later reported for
the recipient — and, under both, the `mail.log` lines for its queue id: the
connection to the receiving server, the server's reply, and the status that
reply was filed as. Rows outlive `mail.log`, so an older message's lines may
have rotated away; the page says so. Retention is controlled by
`SEND_LOG_RETENTION_DAYS`.
### Status
`/status` shows supervised processes (Postfix, OpenDKIM, panel), TLS
certificate validity and expiry, milter socket presence, and a short Postfix
queue summary. The **Machine** card adds the resource usage of the host
underneath — processor (core and thread counts), memory and swap, and
per-interface network throughput and totals — read from the kernel's
counters; CPU and throughput are measured between refreshes, so they appear
one refresh after the page opens. A fully busy processor or a machine out of
memory is a warning here, because both delay or kill the mail path;
throughput is only reported. The hostname block compares `SELFPOST_HOSTNAME`
against the PTR record the internet publishes for this server's IP
(forward-confirmed reverse DNS) — see
[Server-level DNS](#server-level-dns-ptrrdns); use *Re-check* after changing
DNS. The **Reload configuration** button re-applies OpenDKIM tables and the
Postfix sender map from the database — use it if daemons drifted from what
the panel shows after manual edits under `/data`.
### Mail queue and System log
- **Mail queue** (`/mail-queue`) — live view of messages Postfix is still
trying to deliver or deferring.
- **System log** (`/system-log`) — tail of `/data/log/mail.log` (Postfix and
@@ -226,112 +326,206 @@ service healthy and will mail be accepted?"
hours. It lives in the data volume, so it survives a container recreate along
with the rest of the state — `./data/log/` on the host — but it is *not*
included in backups: it is diagnostics, not state.
- **Backup** (`/backup`) — download a full-server backup; the same page hosts
the domain-import form (`POST /domains/import`). See
[Backup, restore, and moving a single domain](#backup-restore-and-moving-a-single-domain).
- **Settings** (`/account`) — change the administrator username and/or password.
Application SASL logins are separate and are not changed here.
**Sessions.** A login survives a container restart: sessions live in SQLite, not
in memory. Expiry is a sliding idle window (`PANEL_SESSION_IDLE_DAYS`, default
### Settings
`/settings` changes the signed-in user's username and/or password. **Global
administrators** also set the panel-wide default DMARC report address (`rua=`)
offered when a domain doesn't set its own — see
[Domain-level DNS](#domain-level-dns-spf-dkim-dmarc). Application SASL logins
are separate and are not changed here.
### Users
`/users` (global administrator only) creates, edits, and deletes panel users.
There are two roles:
- **Global administrator** — full access to every page and every domain,
including Users, Backup, Status, Mail queue, and System log.
- **Domain-admin** — scoped to one or more domains assigned by a global
administrator. Sees only those domains' pages, applications, and
Deliveries rows; cannot add or delete domains. `/users`, `/backup`,
`/status`, `/mail-queue`, `/system-log`, and `POST /reload` are not
reachable (404). A domain-admin can *export* the
domains assigned to them — see
[Exporting and importing a single domain](#exporting-and-importing-a-single-domain).
The panel refuses to remove or demote the **last** global administrator, so
it can never end up with none.
### Sessions
A login survives a container restart: sessions live in SQLite, not in
memory. Expiry is a sliding idle window (`PANEL_SESSION_IDLE_DAYS`, default
seven days) with no absolute lifetime cap — an admin who keeps using the panel
stays signed in indefinitely. HTMX polling on the monitoring screens
(Deliveries, Mail queue, System log, and the Status health fragment) does
**not** count as activity, so an auto-refreshing tab left open will not keep a
session alive forever. Changing the password signs out every other session but
leaves the current browser signed in.
session alive forever. Changing **your own** password on `/settings` signs out
every other session for that user but leaves the current browser signed in.
Signing out (`POST /logout`) ends only the current session — other browsers or
tabs for the same user keep working until their session rows expire.
**Upgrading.** Bump the pinned image tag in `docker-compose.yml` to the target
release, then `docker compose up -d`. The backup version check requires the
running image to match the version that created a full backup — see [Fixed image
### Upgrading
Bump the pinned image tag in `docker-compose.yml` to the target release, then
`docker compose up -d`. The backup version check requires the running image
to match the version that created a full backup — see [Fixed image
tag](#fixed-image-tag).
**Container health.** The image declares a Docker `HEALTHCHECK` that probes
`GET /healthz` on port 8080 (unauthenticated). It returns `200 ok` when
OpenDKIM, the panel, and Postfix are all `RUNNING` under supervisord;
otherwise `503 unhealthy`. This catches a dead mail path that would still leave
the HTTP server up, but it does **not** verify TLS certificates, DNS records,
or end-to-end delivery — use the authenticated **Status** page for that. External
monitoring can use the same endpoint through the reverse proxy if you expose it,
or poll `docker inspect` health state on the host.
### Container health
**First-time setup link.** On first start the one-time setup URL is printed in
the container log (`docker compose logs -f`) and written to `/data/setup-token`
inside the container — `./data/setup-token` on the host, mode `0600` — then
deleted when setup completes. The link is
`https://<SELFPOST_HOSTNAME>/setup/<token>` (path token, not a query string),
valid for ten minutes. If this host ships container logs to a central
aggregator, prefer reading the file:
The image declares a Docker `HEALTHCHECK` that probes `GET /healthz` on port
8080 (unauthenticated). It returns `200 ok` when OpenDKIM, the panel, and
Postfix are all `RUNNING` under supervisord; otherwise `503 unhealthy`. This
catches a dead mail path that would still leave the HTTP server up, but it
does **not** verify TLS certificates, DNS records, or end-to-end delivery —
use the authenticated [Status](#status) page for that. External monitoring
can use the same endpoint through the reverse proxy if you expose it, or poll
`docker inspect` health state on the host.
```sh
docker compose exec selfpost cat /data/setup-token
```
### Server-level DNS (PTR/rDNS)
## Rate limiting
Once, for the machine itself: **PTR/rDNS** for the server's IP, pointing at
its mail hostname. Most receiving mail servers weigh this heavily; get it
from whoever assigns the IP (hosting provider's panel/support), not from
your own DNS zone.
SelfPost applies two independent limits; both can refuse a submission, but only
level 2 writes a `rejected` row in the send log.
The [Status](#status) page verifies the server's hostname against this
record (forward-confirmed reverse DNS). Results are cached for about one
minute; use *Re-check* right after publishing a record.
**Level 1 (IP backstop)** — always on, configured via `.env`:
Per-domain DNS (SPF, DKIM, DMARC) is a separate scope — see
[Domain-level DNS](#domain-level-dns-spf-dkim-dmarc).
### Rate limiting — level 1 (IP backstop)
SelfPost applies two independent layers of rate limiting; both can refuse a
submission, but only level 2 (domain/application, see
[Domain administration](#rate-limiting--level-2-domain-and-application))
writes a `rejected` row in the send log. Level-2 ceilings set in the panel
cannot exceed level 1 (the panel shows the level-1 values and rejects higher
numbers).
Level 1 is always on, configured via `.env`:
- `RATE_LIMIT_MESSAGES_PER_IP` → Postfix `smtpd_client_message_rate_limit`
- `RATE_LIMIT_WINDOW_SECONDS` → Postfix `anvil_rate_time_unit`
This is an anvil limit per connecting client IP. It keeps working even if the
journal-milter (level 2) is down.
journal-milter (level 2) is down. There is no per-IP bypass.
**Level 2 (per domain / per application)** — optional, configured in the panel
on each domain's page or on an individual application. You set a message
ceiling, a time window, and optionally restrict the limit to specific client
IPs; an empty IP list means the differentiated limit does not apply. When
exceeded, Postfix returns a 4xx and the refusal is recorded in Deliveries as
`rejected`.
### Full backup and restore
## Backup, restore, and moving a single domain
**Full backup** (whole `/data` except `log/`: SQLite, all domains' DKIM keys,
all applications' SASL credentials, `manifest.json` with the version that
created it): panel button (*Backup* → *Full backup*), or from the host:
Two related but distinct operations
([architecture.md](architecture.md) § Persistence):
```sh
docker exec <container> selfpost-backup > selfpost-backup.tar.gz
```
- **Full backup** (whole `/data` except `log/`: SQLite, all domains' DKIM keys,
all applications' SASL credentials, `manifest.json` with the version that
created it): panel button (*Backup* → *Full backup*), or from the
host:
```sh
docker exec <container> selfpost-backup > selfpost-backup.tar.gz
```
**Restore** means unpacking that archive into a fresh `/data` bind mount and
starting a container of the **exact same image version** that created it —
SelfPost refuses to start otherwise and tells you which tag to use. On the
first successful start after restore, `manifest.json` from the archive is
**deleted** — it guards only that one boot, so a later in-place upgrade is
not blocked. This is why the compose file pins a fixed tag rather than
`:latest`: without a known version, there'd be no way to tell which image
restoring a given backup actually requires.
**Restore** means unpacking that archive into a fresh `/data` bind mount and
starting a container of the **exact same image version** that created it —
SelfPost refuses to start otherwise and tells you which tag to use. On the
first successful start after restore, `manifest.json` from the archive is
**deleted** — it guards only that one boot, so a later in-place upgrade is
not blocked. On that same first boot the panel also runs one **Resync**
OpenDKIM's tables and Postfix's sender map are re-derived from SQLite and both
daemons are reloaded, healing any drift between the extracted files and the
database (the Status page's *Reload configuration* button runs the same step
on demand). This is why the compose file pins a fixed tag rather than
`:latest`: without a known version, there'd be no way to tell which image
restoring a given backup actually requires (see [Fixed image
tag](#fixed-image-tag)).
**Alternative: archive `./data` while stopped.** If the service can be taken
offline, `docker compose down` then `tar czf selfpost-data.tar.gz ./data` on
the host is safe — nothing is writing to SQLite. Unlike the panel/CLI backup
this sweeps in `./data/log/` too, which is Postfix's raw log and usually the
bulk of the archive; add `--exclude=./data/log` if you only want the state.
Do **not** tar `./data` while
the container is running: the database uses WAL mode and a naive copy can
capture an inconsistent snapshot. The panel/CLI backup remains preferable when
you cannot afford downtime because it takes a consistent SQLite snapshot via
the Backup API on a live container.
**Restoring in place** (same host — recovering from data loss, or rolling
back after a bad change):
- **Export/import a single domain** (domain page → *Export domain* to write the
file, *Backup**Import a domain* to read it back in): moves one domain — its DKIM key and its applications' **working**
SASL passwords — to a different SelfPost instance without regenerating
anything, so DNS (the DKIM TXT record) doesn't need to change. Unlike a full
restore, this works across different hostnames/instances.
```sh
# 1. Stop the instance being replaced
docker compose down
Both files are **secrets** — they contain the admin password hash (full
backup) or working application credentials (domain export) in the clear or in
directly reversible form. Treat them like any other credential material:
restrict who can read them, don't email them around — and encrypt them, which
SelfPost can do for you.
# 2. Move the current /data aside rather than deleting it, start from empty
mv ./data ./data.before-restore
mkdir ./data
### Encrypting a backup or export
# 3. Unpack the backup into the fresh directory
tar xzf selfpost-backup.tar.gz -C ./data
# 4. docker-compose.yml must already pin the exact tag the backup was made
# with — check the archive's manifest if unsure:
tar xzf selfpost-backup.tar.gz -O manifest.json
# 5. Start it and watch the boot
docker compose up -d
docker compose logs -f selfpost
```
A version mismatch at step 5 refuses to start and leaves `/data` untouched —
the panel exits with a message naming the tag to use, e.g.:
```
backup: this backup was created by SelfPost 1.2.3 but this image is 1.3.0 — restore into the matching image (selfpost:1.2.3)
```
Fix the tag in `docker-compose.yml`, `docker compose pull && docker compose up
-d` again — the manifest is still there because the failed boot never got to
delete it.
**Moving to a different host** is the same five steps, just starting cold:
bring the compose files and the correct pinned image tag to the new host
(step 1 of [Full deployment](#full-deployment)), put the backup archive in
place of step 3 above, then redo the reverse-proxy/TLS and DNS steps of a
[Full deployment](#full-deployment) — the PTR record and the certificate both
belong to the old IP/host and have to be reissued for the new one; nothing in
the backup carries them.
**Restoring an encrypted (`.spbk`) backup** needs a running container to
decrypt it first — any container with the `selfpost-backup` CLI works; decryption
does not read `/data` and performs no version check. Start one normally
(step 5, but on an empty `/data` you haven't unpacked yet), then:
```sh
docker exec -i <container> selfpost-backup -decrypt < backup.spbk > selfpost-backup.tar.gz
```
Stop it, wipe `/data` again, and continue from step 2 above with the
resulting `.tar.gz` — see [Encrypting a backup or
export](#encrypting-a-backup-or-export) for the decrypt command's password
options.
Restoring an archive taken **before** a session row was removed can bring
that session back: session rows travel with the backup, and a browser that
still holds the matching cookie is signed in again on the next request if the
restored row's idle expiry has not passed. `POST /logout` removes only the
current session; there is no "logout everywhere". Changing your own password
on `/settings` deletes your other sessions, but a global administrator
resetting another user's password on `/users` does not invalidate that user's
existing sessions.
**Alternative: archive `./data` while stopped.** If the service can be taken
offline, `docker compose down` then `tar czf selfpost-data.tar.gz ./data` on
the host is safe — nothing is writing to SQLite. Unlike the panel/CLI backup
this sweeps in `./data/log/` too, which is Postfix's raw log and usually the
bulk of the archive; add `--exclude=./data/log` if you only want the state.
Do **not** tar `./data` while the container is running: the database uses
WAL mode and a naive copy can capture an inconsistent snapshot. The
panel/CLI backup remains preferable when you cannot afford downtime because
it takes a consistent SQLite snapshot via the Backup API on a live
container.
See also [Exporting and importing a single
domain](#exporting-and-importing-a-single-domain) — a different, domain-scoped
operation that also lives on the *Backup* page (`/backup`).
Both a full backup and a domain export are **secrets** — they contain the
admin password hash (full backup) or working application credentials (domain
export) in the clear or in directly reversible form. Treat them like any
other credential material: restrict who can read them, don't email them
around — and encrypt them, which SelfPost can do for you.
#### Encrypting a backup or export
Both download forms carry an **Encrypt with a password** checkbox. Ticked, the
file that comes down is an encrypted envelope instead of the plain archive:
@@ -371,21 +565,121 @@ docker exec -e SELFPOST_BACKUP_PASSWORD="$PW" <container> selfpost-backup > back
With no password set, the CLI keeps writing the plain `.tar.gz` it always has.
## Published ports
## Domain administration
`deploy/docker-compose.yml` maps **465** and **587** to the host. Port 465
(smtps) is always active. Port **587** is published even when
`SUBMISSION_ENABLE=false`; nothing listens until you set it to `true` — harmless,
but it can look like an open port in external scans.
### Domains page
## Fixed image tag
`/domains` lists sending domains and hosts the add-domain form (**global
administrator only**). Domain administrators see only domains assigned to
them. Each row shows its DKIM TXT value, SPF/DMARC checks, and SASL
applications. Per-domain rate limits (level 2) and trusted-IP application
overrides are configured here — see [Rate limiting —
level 2](#rate-limiting--level-2-domain-and-application). *Export domain*
writes a single-domain archive; *Import a domain* on the Backup page reads
one back in (**global administrator only**) — see [Exporting and importing a single
domain](#exporting-and-importing-a-single-domain).
`deploy/docker-compose.yml` pins an explicit version (`ghcr.io/mixeme/selfpost:X.Y.Z`),
deliberately never `:latest`. The current pin is `1.1.0`. Intermediate
CHANGELOG sections (`0.2.0``0.6.0`) record development cuts from before that
image was published. Pinning matters because of the backup version check above:
the panel binary's embedded version and the image tag that produced it are the
same value by construction (the release CI stamps both from one git tag — see
`.github/workflows/release.yml`), so the pin is what makes "restore into the
same version" a checkable fact rather than a guess. Upgrade by bumping the tag
deliberately, not by riding a moving target.
### Domain-level DNS (SPF, DKIM, DMARC)
For *every* sending domain you add in the panel:
- **SPF** — a TXT record on the domain authorizing this server to send on its
behalf (e.g. `v=spf1 a mx ip4:<server IP> -all`, adjusted to your setup).
- **DKIM** — a TXT record with the exact value the panel shows on that
domain's page (`domain page → DKIM TXT record`), one selector per domain.
- **DMARC** — a `_dmarc` TXT record. The panel suggests `p=none` (monitoring
only, safe to publish immediately). On a send-only relay the sending domain
often has no inbox, so `rua=` is optional — configure a default report address
in *Settings* (see [Settings](#settings)) or per domain when you have a
mailbox that receives inbound mail elsewhere. If `rua=` points at another
domain, publish `_report._dmarc` on that hub domain too; the panel checks
it. Public mail hosts (Gmail, Outlook, …) cannot be used as external
report destinations.
Skipping any of the three records is the single most common reason mail
lands in spam even though SelfPost delivered it correctly — DKIM passing
doesn't help if SPF/DMARC are absent. **Whenever you add a new domain in the
panel, add its DNS records at the same time**, not later.
Each domain's page shows a *DNS status* card comparing the published DKIM
record against the key this server signs with, plus the domain's SPF,
DMARC, and (when configured) DMARC report-authorisation records. Results are
cached for a few minutes; use *Re-check* right after publishing a record.
The SPF check is deliberately shallow — it looks for a mechanism that
literally covers this server's address and does not follow `include:` or
`redirect=`, so a record that authorizes the server through an include is
reported as "cannot tell" rather than as a failure.
Server-level DNS (the PTR/rDNS record) is a separate, once-per-machine scope
— see [Server-level DNS](#server-level-dns-ptrrdns).
### IP warmup
A brand-new IP has no sending history, so receiving servers are cautious with
it regardless of how correct your DKIM/SPF/DMARC are. Start with low volume to
a domain, increase gradually over days/weeks rather than sending everything on
day one, and check the IP against major blocklists (Spamhaus and similar)
before and during warmup. This is inherent to how mail reputation works on the
public internet, not something SelfPost's configuration can shortcut.
### Rate limiting — level 2 (domain and application)
Level 2 is optional, configured on each domain's page, and layers on top of
the always-on [level-1 IP backstop](#rate-limiting--level-1-ip-backstop).
Level-2 ceilings cannot exceed level 1 (the panel shows the level-1 values
and rejects higher numbers). When a level-2 ceiling is exceeded, Postfix
returns a 4xx and the refusal is recorded in [Deliveries](#deliveries) as
`rejected`.
**Level 2 — domain** — a message ceiling and window for **every** client IP
sending as that domain. When unset, only level 1 applies for non-privileged
senders.
**Level 2 — application (trusted IPs)** — optional override on an
application: list one or more client IPs and a ceiling **strictly above**
the domain limit (still ≤ level 1). Connections from those IPs use the
application ceiling and skip the domain check. Other IPs stay under the
domain limit (or level 1 alone). An application override without trusted
IPs is inactive.
**Level 2 is best-effort, not a guarantee.** It runs inside the
journal-milter and is deliberately fail-open: if the rate-limit lookup hits
a store error, or the connecting client's IP is not available to the
milter, level 2 is skipped and the message is accepted rather than held up.
Level 1 is the backstop that keeps working even when level 2 cannot run.
### Deliveries
`/deliveries` is a searchable send log with server-side filters by domain
and application. A row identifies its message and nothing more — time,
sender, recipient, subject and status `queued` (accepted, not yet
delivered), `sent` (handed off successfully), `deferred` (Postfix is
retrying), `bounced` (final failure), or `rejected` (refused — for example
by a [level-2 rate limit](#rate-limiting--level-2-domain-and-application));
*Details* opens that row's own page (`/deliveries/{id}`). That page carries
the sending domain, the application it was submitted under, the Postfix
queue id and the journal id, beside the message's history — when it was
accepted and what Postfix later reported for the recipient — and, under
both, the `mail.log` lines for its queue id: the connection to the
receiving server, the server's reply, and the status that reply was filed
as. Rows outlive `mail.log`, so an older message's lines may have rotated
away; the page says so. Retention is controlled by
`SEND_LOG_RETENTION_DAYS`.
### Exporting and importing a single domain
Domain page → *Export domain* to write the file, *Backup* → *Import a
domain* to read it back in. This moves one domain — its DKIM key and its
applications' **working** SASL passwords — to a different SelfPost instance
without regenerating anything, so DNS (the DKIM TXT record) doesn't need to
change. Unlike a full restore (see [Full backup and
restore](#full-backup-and-restore)), this works across different
hostnames/instances. *Import* is global-administrator only; *export* is
available to any user who can access the domain, **including a domain-admin**
for a domain assigned to them — so a domain-admin can walk away with that
domain's working SASL passwords in the clear. Weigh that when deciding which
domains to assign to a domain-admin account.
A domain export is a secret in the same way a full backup is, and can be
encrypted the same way — see [Encrypting a backup or
export](#encrypting-a-backup-or-export).
+2 -1
View File
@@ -18,7 +18,8 @@ operator.
- gzip + XML aggregate parsing → SQLite summaries per sending domain.
- Panel page and/or per-domain section: recent reports, third-party senders,
delivery health of report ingestion.
- Reuse `admin.dmarc_report_email` and `domains.dmarc_rua` for DNS templates;
- Reuse the `dmarc_report_email` setting (moved off the old `admin` table into
`settings` by migration `0005`) and `domains.dmarc_rua` for DNS templates;
when enabled, suggest a SelfPost-hosted report address.
**Out:**
-259
View File
@@ -1,259 +0,0 @@
# Plan: domain-admin (domain administrator role)
**Status:** agreed
**Version:** target bump **1.x** MINOR, given a compatible migration of the
current administrator into a global one.
**Order:** recommended after [web-split](web-split.md) (done), before
[inbound-relay](inbound-relay.md).
---
## What this is
Today the panel has exactly one subject: `RequireAuth` is a boolean gate, not a
role ([web.go](../../internal/web/web.go) — the
`mux.Handle("/", s.auth.RequireAuth(authed))` wrapper), and the session carries
nothing beyond the fact of being signed in.
Two panel roles:
| Role | Scope |
|------|-------|
| **global** | Full panel except nothing new — same powers as today's single admin |
| **domain_admin** | Only **assigned** domains (one or several; list set by global admin) |
For each assigned domain, a domain-admin can:
- applications (create, sender mode, password regeneration, delete, L2 limit);
- DKIM/DNS status and recheck;
- per-domain DMARC `rua=` (inherit / none / custom) — full control on the
domain page;
- send log filtered to assigned domains;
- domain export (encrypted `.spde` optional, same as today);
- domain-level L2 rate limit.
What stays **global-only** (domain-admin gets 404 or redirect):
- adding and removing domains;
- domain import;
- creating/editing/deleting panel users and assigning domains;
- `/reload`;
- full backup (`/backup` — all of `/data` including every domain's `sasldb2`);
- mail queue (`/mail-queue*`);
- system log tail (`/system-log*`);
- status page (`/status*`) — server-wide health, queue summary, reload, DNS
recheck of the **hostname**; same treatment as queue and system log.
Domain-admin **self-service** on `/account`: username and password only (not
global DMARC report email).
## Why this extends v1.0
[product.md](../product.md) puts "multiple panel users, roles" out of scope
(one administrator). A second subject is a deliberate widening of the project's
boundary, as inbound-relay is.
The cost is phase-sized, not patch-sized:
- a users table and their binding to domains;
- the role in the session;
- authorisation in every handler (not only on the route — today `{id}`/`{aid}`
are checked for nothing beyond existence);
- reworking first-run setup and password change for several users;
- accounting for the new subject in backup and domain export.
*(The earlier wording of this item — "2FA and multiple administrators" — has
been replaced: 2FA is off the table, and "multiple administrators" is narrowed
to one specific role, because what is needed is not a second all-powerful admin
but limited access for the owner of one or several domains, with the list set
by the global administrator.)*
---
## Schema and migration
**New migration** `0005_panel_users.sql`:
```sql
CREATE TABLE users (
id INTEGER PRIMARY KEY AUTOINCREMENT,
username TEXT NOT NULL UNIQUE,
password_hash TEXT NOT NULL,
role TEXT NOT NULL CHECK (role IN ('global', 'domain_admin')),
dmarc_report_email TEXT NOT NULL DEFAULT '',
created_at TEXT NOT NULL
);
CREATE TABLE user_domains (
user_id INTEGER NOT NULL REFERENCES users(id) ON DELETE CASCADE,
domain_id INTEGER NOT NULL REFERENCES domains(id) ON DELETE CASCADE,
PRIMARY KEY (user_id, domain_id)
);
-- Migrate existing administrator → global user (idempotent guard via admin count).
INSERT INTO users (username, password_hash, role, dmarc_report_email, created_at)
SELECT username, password_hash, 'global', dmarc_report_email, created_at
FROM admin WHERE id = 1;
DROP TABLE admin;
```
**Sessions:** keep `sessions.username` (no schema change). On login and
`RequireAuth`, resolve username → `User` row (role + domain IDs). Stale session
after username change behaves as today (`Lookup` fails → redirect login).
**Backup/restore:** full backup already snapshots `selfpost.db` via
`VACUUM INTO`; users and bindings restore with the DB. No manifest format change
required (same `selfpost-full-backup`).
**DMARC two levels:**
- **Global** `users.dmarc_report_email` — only on global user's `/account`;
default `rua=` when a domain uses *inherit*.
- **Per-domain** `domains.dmarc_rua` — domain-admin edits on the domain page
(existing handler); domain-admin never sees the global default field.
---
## Principal model
Request context carries a `Principal` (in `internal/web/auth`):
```go
type Role string // "global" | "domain_admin"
type Principal struct {
ID int64
Username string
Role Role
Domains []int64 // assigned domain IDs; empty for global (meaning "all")
}
```
Helpers:
- `CurrentPrincipal(r)` — from context;
- `IsGlobal(p)``p.Role == "global"`;
- `CanAccessDomain(p, domainID)` — global or `domainID` in `p.Domains`;
- `CanAccessApp(p, app)``CanAccessDomain(p, app.DomainID)`.
`lookupDomain` / `lookupApplication` in handlers call `CanAccess*` after
existence check; return 404 (not 403) to avoid leaking IDs.
---
## Route matrix
| Method | Path | global | domain_admin |
|--------|------|--------|--------------|
| GET | `/` | → `/status` | → `/domains` |
| GET | `/status`, `/status/fragment` | yes | **no** (404) |
| POST | `/status/recheck` | yes | **no** |
| GET | `/domains` | all domains | assigned only |
| POST | `/domains` | yes | **no** |
| POST | `/domains/import` | yes | **no** |
| GET | `/domains/{id}` | yes | assigned |
| POST | `/domains/{id}/dns-recheck` | yes | assigned |
| GET/POST | `/domains/{id}/delete` | yes | **no** |
| POST | `/domains/{id}/applications` | yes | assigned |
| POST | `/domains/{id}/ratelimit` | yes | assigned |
| POST | `/domains/{id}/dmarc` | yes | assigned |
| POST | `/domains/{id}/export` | yes | assigned |
| POST | `/applications/{aid}/*` | yes | if app in assigned domain |
| POST | `/reload` | yes | **no** |
| GET/POST | `/account` | username, password, global DMARC email | username, password only |
| GET/POST | `/backup` | yes | **no** |
| GET | `/deliveries*` | all (optional filter) | clamped to assigned domains |
| GET | `/mail-queue*` | yes | **no** |
| GET | `/system-log*` | yes | **no** |
| GET | `/users` | list users | **no** |
| GET/POST | `/users/new` | create user | **no** |
| GET/POST | `/users/{uid}` | edit/delete user | **no** |
| POST | `/logout` | yes | yes |
**Deliveries:** for domain-admin, `sendLogData` forces filter to assigned
domain set; dropdowns list only assigned domains/apps; reject `domain` query
param outside assignment; `HandleDelivery` checks log row's `domain` field.
---
## User management UI (global only)
New routes under `/users`:
- **List** — username, role, assigned domain names (or "all" for global).
- **Create** — username, password, role (`domain_admin` default), multi-select
domains (required when role is `domain_admin`).
- **Edit** — change password (optional), reassign domains, delete user.
- **Guards:** cannot delete the last `global` user; cannot demote self to
`domain_admin` without another global user; domain-admin role cannot access
these routes.
Templates: `users.html`, `user_form.html`; nav link visible only for global
users.
---
## Auth / setup / sessions
- **Setup** (`/setup/{token}`): unchanged semantics — creates first **global**
user via `CreateGlobalUser`; `AdminExists``UserExists`.
- **Login:** authenticate against `users` by username + bcrypt.
- **Password change:** per-user `UpdateUser`; domain-admin cannot change
another user's password.
- **Session rename / destroy others:** unchanged behaviour keyed by username.
---
## Navigation
[layout.html](../../internal/web/view/templates/layout.html) `nav` template:
- **global:** all items today (status, domains, deliveries, mail queue, system
log, backup, settings) + **Users**.
- **domain_admin:** domains, deliveries, settings only.
Pass `IsGlobal` (or `Principal`) into every rendered page.
---
## Security
- **CSRF:** keep origin-check-only for now ([security.md](../security.md) ADR);
note in CHANGELOG that multi-user panel reopens the ADR — no CSRF tokens in
this phase.
- **Export encryption:** optional password on domain export remains; full backup
encryption trigger ("second administrator") is satisfied by domain-admin
existing — no change required.
- **Authorization tests:** table-driven tests for global vs domain-admin on
representative handlers; explicit `{aid}` cross-domain mutation blocked.
---
## Implementation order
1. Migration `0005_panel_users.sql` + `store/users.go` (CRUD, domain bindings).
2. Auth: login against `users`, `Principal` in context, setup creates global user.
3. `CanAccessDomain` / `CanAccessApp`; harden `lookupDomain` / `lookupApplication`.
4. Route guards: global-only middleware or per-handler checks.
5. Filter lists: dashboard, deliveries, domain detail DMARC inherit source.
6. User management handlers + templates.
7. Nav visibility + default redirect (`/`).
8. Tests + `go build` / `go vet` / `go test`; CHANGELOG `[Unreleased]`.
---
## Done when
- A global administrator and a domain-admin with different rights both work
through the panel; the domain-admin cannot reach past the **assigned**
domains;
- the current single admin migrates into a global one without losing access;
- backup/restore accounts for users and their bindings;
- `build`/`vet`/`test`/image green.
## Risks
- An incomplete `{id}`/`{aid}` check in a handler — access leaking to someone
else's domain;
- breaking setup or backup — that would be a semver major, not 1.x.
+5 -2
View File
@@ -3,8 +3,11 @@
**Status:** agreed
**Version:** target bump **1.x** MINOR; **`2.x` possible** — to be settled once
the implementation lands (do not fix a major in advance).
**Order:** recommended after [web-split](web-split.md) and
[domain-admin](domain-admin.md).
**Order:** the 2026-08-13 full-tree review follow-ups are closed (CHANGELOG
`[Unreleased]`), so this is next up as the largest remaining 1.x+ **feature**.
The domain-admin role
and `internal/web` split have shipped — see [CHANGELOG](../../CHANGELOG.md)
`[1.2.0]`.
---
-89
View File
@@ -1,89 +0,0 @@
# Plan: logrotate-mode (mail.log stops rotating in some images)
**Status:** candidate
**Version:** patch; no schema, no configuration surface.
**Order:** independent. Worth doing before anything that lets an instance run
unattended for months.
---
## What was observed
The container log carries, on every start:
```
warning: Potentially dangerous mode on /etc/logrotate.d/mail: 0664
error: Ignoring /etc/logrotate.d/mail because it is writable by group or others.
```
logrotate refuses a configuration file that group or others may write, so
`mail.log` is never rotated in an image with that mode. It grows until the
volume does.
Measured across three images built on the same host from the same commit range:
| Image built from | Mode of `/etc/logrotate.d/mail` |
|---|---|
| a sync made two days earlier | `0644` — works |
| a later sync (`tar -czf -` pipe from a Windows checkout) | `0666` |
| a later sync (`git archive` from the same checkout) | `0664` |
So the file is fine in the repository (git records `100644`) and is spoiled on
the way into the build context. `COPY build/logrotate-mail.conf
/etc/logrotate.d/mail` ([build/Dockerfile](../../build/Dockerfile)) takes the
mode from the context as it finds it, and an archive produced from a checkout
without POSIX permissions carries the umask-widened mode instead of the one git
recorded.
**Not established:** whether images built by the release workflow are affected.
They are built from a checkout on Linux, where the mode should survive as
`0644`, but the published image could not be pulled to check. Confirm before
concluding that only locally built images have this.
## Why it deserves a plan rather than a one-line fix
Three separate things are wrong, and fixing only the visible one leaves the
other two.
1. **The image trusts the build context's file modes.** Every `COPY` in the
Dockerfile has this property, not just this one; the scripts happen to be
`chmod +x`-ed afterwards, which is why they were never noticed.
2. **The failure is silent.** `logrotate-loop.sh` runs
`logrotate /etc/logrotate.d/mail` and only reports a failure on a non-zero
exit — but logrotate *ignores* the file and exits 0, so the loop reports
nothing and the operator's only clue is a warning printed once at start.
3. **Nothing checks the outcome.** No test or health check notices that
`mail.log` has not rotated, and the panel's Status page has no view of it.
## Directions to weigh
- `COPY --chmod=0644` on the configuration files (and an explicit mode on the
scripts instead of the later `chmod +x`), which makes the image's file modes a
property of the Dockerfile rather than of whoever built it. Needs a check of
the minimum BuildKit version the project is willing to require.
- Or an explicit `chmod` in the same `RUN` that already fixes the scripts —
cruder, no build-time requirement.
- Make `logrotate-loop.sh` fail loudly: `logrotate` has `--debug`-free ways to
be told to care, but the simplest reliable check is that the loop verifies
the configuration is readable-and-not-writable before entering the loop, and
exits non-zero so supervisord reports it.
- Consider whether the e2e stack should assert that a rotation actually happens
(it can run with a short `LOGROTATE_INTERVAL_SECONDS`).
## Done when
- An image built from a Windows checkout and one built by the release workflow
both carry `0644`, and rotation runs in both.
- A configuration logrotate would ignore makes the container say so in a way an
operator will see, rather than exiting 0.
- The dev loop's sync step cannot silently widen file modes again, or the image
no longer cares if it does.
## Risks
- Low blast radius, but it touches the image's startup path — a mistake here is
a container that will not start rather than a log that does not rotate.
- The `create 0640 postfix selfpost` line in the rotate configuration is load
bearing (see the comment in `logrotate-loop.sh`: a postlogd-triggered recreate
lands the file unreadable by the unprivileged panel). Any rework of the
configuration must keep it.
+141
View File
@@ -0,0 +1,141 @@
# Plan: queue-retries (Postfix retry policy in the panel)
**Status:** agreed
**Date:** 2026-08-13
**Version:** patch; no schema, no configuration surface.
**Order:** small panel item; does not wait on inbound-relay.
---
## Goal
Show the operator how Postfix retries deferred mail: first retry delay, later
backoff cap, and how long a message stays in the queue before it bounces. The
numbers come from this container's effective Postfix config, not from
hard-coded copy.
## Scope
**In:**
- A static «How delivery retries work» card on `/mail-queue` (global
administrator). Not inside the HTMX poll fragment — the snapshot is taken at
panel start.
- The same human-readable intervals in `/deliveries/{id}` history for
`deferred` and `bounced` (domain administrators never see Mail queue).
- Operator docs: [guide.md](../guide.md) Mail queue bullet;
[architecture.md](../architecture.md) notes the one-shot `postconf -h` at
panel start. [CHANGELOG.md](../../CHANGELOG.md) `### Added`.
**Out:**
- Changing Postfix retry parameters, or exposing them as panel settings
([product.md](../product.md): Postfix is used as-is).
- An `attempts` column on `send_log`, or «attempt 3 of N» — Postfix has no
attempt budget; it is time-based.
- Counting `status=deferred` lines in `mail.log` or reading `postcat`. Each
attempt is already on the delivery page's log table; the journal stores only
the last status.
- Duplicating the card on Status (summary + link to Mail queue already exist).
- Re-reading `postconf` on every HTTP request.
## Architecture
Acceptance is still synchronous SMTP. Delivery stays in Postfix's on-disk
queue. SelfPost does not enqueue, retry, or deliver.
```mermaid
flowchart LR
smtpAccept["SMTP 250 accepted"] --> postfixQueue["Postfix queue"]
postfixQueue -->|"deferred: backoff until queue lifetime"| retry["Retry MX"]
retry -->|sent| done["sent"]
retry -->|lifetime elapsed or 5xx| bounce["bounced"]
postfixQueue --> mailQueuePage["/mail-queue policy card"]
postfixQueue --> deliveryPage["/deliveries/id history copy"]
```
[build/postfix-config.sh](../../build/postfix-config.sh) does not set
`queue_run_delay`, `minimal_backoff_time`, `maximal_backoff_time`,
`maximal_queue_lifetime`, `bounce_queue_lifetime`, or `delay_warning_time`.
Debian/Postfix 3.x compiled-in defaults therefore apply unless the operator
overrides them (`postconf -e` inside the container).
### Loading the numbers
Once, when the HTTP role starts ([cmd/panel/httpserver.go](../../cmd/panel/httpserver.go),
after `postfix-config.sh` has run):
```
postconf -h queue_run_delay minimal_backoff_time maximal_backoff_time maximal_queue_lifetime bounce_queue_lifetime delay_warning_time
```
`postconf -h`, not a parse of `/etc/postfix/main.cf`: stock values are not
written to the file. `postconf` is the effective config, including a manual
override.
- Fixed argv, no user input — same pattern as
[postfix.Queue](../../internal/postfix/queue.go) /
[security.md](../security.md).
- Cache on `handlers.Config` (via `web.Config`). The HTMX fragment does not
call `postconf`.
- A live `postconf -e` is visible after the next panel (or container) restart.
While the process is up, the panel shows the start-up snapshot.
- Parse Postfix time units (`300s`, `5d`, `1h`, a bare number is seconds) in
`internal/postfix`. Format human strings (`5 minutes`, `5 days`,
`about 1 hour 7 minutes`) in one place so the Mail queue card and
`deliveryEvents` cannot drift.
- If `postconf` is missing (unit tests on Windows, binary outside the
container): log a warning, fall back to Postfix 3.x compiled-in defaults
(`300s` / `4000s` / `5d` / `0`), and put a muted note on the card. Tests
stub the lookup (as `queueIDs` in the log-tailer) or pass a fixture on
`Config`. Do not fail panel start.
Typical stock values, for orientation only — the UI prints whatever
`postconf` returned:
| Parameter | Stock | Meaning |
|---|---|---|
| `queue_run_delay` / `minimal_backoff_time` | `300s` | First retry and deferred-queue scan |
| `maximal_backoff_time` | `4000s` | Cap on the doubling gap (~1 h 7 min) |
| `maximal_queue_lifetime` | `5d` | Then bounce |
| `delay_warning_time` | `0` | No delay warning to the sender |
## Panel copy
Mail queue card facts: first retry; later retries (doubling, capped);
kept in queue; then bounced. Short prose: there is no fixed attempt count; a
`deferred` message stays in this listing until it is delivered or the queue
lifetime runs out.
`deliveryEvents(row, policy)`:
- `deferred`: retries, first after X, then with increasing gaps up to Y, for
up to Z.
- `bounced`: or Postfix gave up after Z in the queue.
## Tests
- Duration parser: `5d`, `300s`, `4000s`, `1h`, `0`, bare number.
- `/mail-queue` handler: card shows the fixture policy's human strings, not a
live `postconf`.
- Delivery page / `deliveryEvents`: `deferred` and `bounced` contain those
strings ([handlers_monitor_test.go](../../internal/web/handlers/handlers_monitor_test.go)).
- [templates_test.go](../../internal/web/view/templates_test.go): pass the new
fields if rendering `mail_queue` requires them.
`go test` / `go vet` on the touched packages.
## Done when
- `/mail-queue` states this Postfix's first retry, backoff cap, and queue
lifetime.
- A `deferred` / `bounced` delivery page uses the same intervals.
- A manual `postconf -e maximal_queue_lifetime=2d` followed by a panel restart
changes what the panel prints.
- Guide and architecture describe the snapshot; CHANGELOG has an Added entry.
## Risks
- Showing compiled-in fallbacks when `postconf` failed would mislead if the
operator had overridden them — mitigate with the muted note on the card.
- Inventing a max-attempt count would be false; the copy must stay time-based.
-226
View File
@@ -1,226 +0,0 @@
# Plan: visual-style
**Status:** agreed
**Version:** no bearing on semver — presentation only, no schema and no route
changes.
**Order:** independent of the feature roadmap; may be taken up between feature
items.
---
## Goal
Bring the control panel's surface in line with the mark that was approved in
[selfpost-proof.html](../assets/selfpost-proof.html): its palette, its
typography, and the plainness of its components. Today the panel is a default
blue-on-cool-grey admin theme standing next to a warm brick stamp, so the mark
reads as pasted onto someone else's page.
## Scope
**In:**
- `internal/web/view/static/panel.css` — colour tokens, typography, spacing,
every component rule.
- `internal/web/view/static/` — three self-hosted font files.
- Templates, only where a class has to be added or a wrapper introduced for a
rule to have something to attach to.
- `NOTICE` — the OFL attribution the font files oblige.
**Out:**
- Any change to what a page does, which pages exist, or what an operator has to
click. No new features, no copy rewriting.
- The navigation's position and the two-column shell. The proof's panel mock
shows a horizontal bar on a dark header; the panel's left column also carries
the per-page section index (`.sections` plus the scroll-spy in `panel.js`),
which that layout has nowhere to put. Keeping the column is a deliberate
divergence from the mock, not an oversight.
- The mark files themselves (`logo.svg`, `logo-compact.svg`, `favicon.*`) —
already drawn, already converted to outlines.
## Constraint that shapes everything
The panel's Content-Security-Policy is a plain `default-src 'self'` with no
inline-style exemption ([security.md](../security.md)). Every rule lives in
`panel.css`; a `style="..."` attribute in a template is blocked and silently
does nothing. Self-hosted fonts are served from the panel's own origin and are
therefore already covered — no CSP change is needed, and none may be made.
## Typography
IBM Plex, self-hosted. The mark is Plex converted to outlines, so the panel
setting its own name in Segoe UI or Cantarell is the seam this whole item
exists to close.
| File | Covers | Size |
|---|---|---|
| `static/ibm-plex-sans.woff2` | variable, weights 100700, latin | 45.7 KB |
| `static/ibm-plex-mono-400.woff2` | mono regular, latin | 14.8 KB |
| `static/ibm-plex-mono-600.woff2` | mono semibold, latin | 15.7 KB |
76 KB in total, in a 20 MB binary. The variable file replaces what would
otherwise be five static weights and lets the scale below use 300 and 500
without paying per weight.
Monospace is the one the operator actually reads: DKIM records, `mail.log`
lines, application logins, socket paths, generated passwords. `ui-monospace`
resolves to Consolas, SF Mono or DejaVu Sans Mono depending on the operator's
machine, and those differ in advance width — the six-column send log wraps
differently for each. A shipped mono makes those tables one layout.
| Role | Family | Size | Weight |
|---|---|---|---|
| Body | sans | 15px / 1.5 | 400 |
| `h1` | sans | 1.55rem, tracking 0.01em | 300 |
| `h2` | sans | 1.05rem | 600 |
| `label` | sans | 0.9rem | 600 |
| Nav entry / active | sans | 0.95rem | 400 / 600 |
| `th` | **mono**, uppercase, tracking 0.08em | 0.75rem | 500 |
| `.st` status badge | **mono** | 0.78rem | 500 |
| `.code`, `.mono`, `.metric` | **mono** | 0.85rem | 400 |
`font-display: swap`, so a cold load shows the system stack for a frame rather
than blank text.
## Colour tokens
Names stay as they are wherever they already exist: the dark scheme reassigns
the same custom properties, which is why no rule in the file needs
`!important`. Warm neutrals replace the cool greys; brick becomes the accent.
| Token | Light | Dark |
|---|---|---|
| `--bg` | `#F4F2ED` | `#16181B` |
| `--fg` | `#12161C` | `#E9E6E0` |
| `--muted` | `#6B7684` | `#9AA1A9` |
| `--card-bg` | `#FFFFFF` | `#1D2024` |
| `--border` | `#DEDCD7` | `#2C2F34` |
| `--control-border` | `#CBC8C1` | `#3A3E44` |
| `--input-bg` | `#FFFFFF` | `#14161A` |
| `--code-bg` | `#EFEDE9` | `#14161A` |
| `--surface-bg` | `#EAE7E0` | `#23262B` |
| `--accent-fill` / `--on-accent` | `#7A3B2E` / `#FFFFFF` | `#8E4535` / `#FFFFFF` |
| `--accent-text` | `#7A3B2E` | `#CE7B66` |
| `--nav-active-bg` | `#EDE4DE` | `#2A1F1B` |
Brick splits into a fill and a text value because `#7A3B2E` on `#16181B` is
about 2:1 — unreadable as a dark-scheme link. The fill lightens just enough to
keep white on it above 4.5:1; the text value lightens further.
Status families (`--st-ok-*`, `--st-warn-*`, `--st-error-*`, `--st-unknown-*`),
the flash, the credential card and `--danger-*` keep their hues and are only
warmed to sit on paper. The one thing to watch is brick against `st-error` red:
the proof rejected several candidate colours precisely so that the mark would
not read as a status, and the same test now applies to every brick button
standing in a row of `error` badges.
## Components
Everything already in `panel.css`, in the order it appears there: card, form
controls, buttons (filled, outlined, danger), flash, table, status badge,
`.code`, nav (brand, links, sections, session), application list and its
disclosure panels, credential card, status page meters and facts, delivery
timeline, log tables, split layout, encrypt fields, footer.
Two component-level changes rather than pure repaints, both forced by the
accent:
- Row actions (`td.actions a.danger`, `Delete`) become outlined instead of
filled. A filled red button in a table row next to a filled brick button
reads as one block of colour.
- Nav entries carry the active state as brick text on a warm tint rather than
the current blue tint.
## Order of work
1. Fonts into `static/`, `@font-face` and the type scale in `panel.css`,
`NOTICE` attribution. Nothing else changes shape.
2. Token block: light and dark, both schemes in one pass.
3. Chrome: `layout.html`'s nav, footer, shell.
4. Signed-out pages: `login`, `setup` — the mark and one card, where the seam
is worst.
5. `dashboard`, `domain_detail`, `domain_delete`.
6. `deliveries`, `deliveries_rows`, `delivery`.
7. `status` + `status_body`, `mail_queue*`, `system_log*`.
8. `backup`, `account`, `users`, `user_form`, `encrypt_fields`.
9. `CHANGELOG.md` under `[Unreleased]`.
## Verification
- Every page rendered locally and screenshotted in both schemes before and
after (`panel.exe` on Windows, headless Edge), including a 375px-wide pass —
the nav column and the wide tables are where a repaint breaks layout.
- Contrast: body text and every status badge at 4.5:1 or better against its own
background, UI borders at 3:1. Brick on white is 7.3:1 by the proof's own
measurement; the dark-scheme values above are the ones to re-check.
- `go build ./... && go vet ./... && go test ./...`; the template guards in
`internal/web/view/templates_test.go` must stay green, and the static-asset
ETag test grows to cover the three font files.
- No `style=` attribute anywhere in `templates/` — the CSP would drop it.
## What is done
The restyle itself landed in `652f1fe`, with the table-wrapping fixes it
surfaced in `f44f533`. Every page in the order above was rendered in both
schemes from a panel running locally and checked: the signed-out pair, the
domain list empty and with three domains, the whole domain page (credential
card, DKIM/SPF/DMARC, DNS status, applications, rate limit, export, danger
zone), the delete confirmation, the send log with rows, a single delivery,
status, mail queue, system log, backup, settings, users and the user form.
Two of those need standing in for what the container provides: `supervisorctl`,
`saslpasswd2` and `postmap` stubs on `PATH`, `postfix/`, `opendkim/keys/` and
`sasl/` created inside the data dir by hand, and rows seeded into `send_log`
without them the domain, application and send-log pages do not exist locally.
## Outstanding
Nothing here blocks the item; each is written down so it is not rediscovered.
1. **Send-log status is bare text**, while every other status in the panel is a
`.st` badge. Making it one is not a repaint: it needs a mapping from
`sent`/`queued`/`deferred`/`bounced` onto the four badge colours, which is a
judgement about severity (is `deferred` a warning?) rather than a style.
**Needs a decision before it is written.**
2. **The panel overflows horizontally at 375px**`main` and its cards render
wider than the window and the page scrolls sideways. Reproduced with the
stylesheet at `f59befd` too, so it predates this work; tracked separately.
Removing the navigation does not fix it, so it is in `main`/`.card`, not in
the bar the `@media (max-width: 66rem)` block lies down.
3. **Three views were only ever seen empty**: the mail queue with entries, the
system log with lines, and a delivery's own `mail.log` lines (`table.log`).
All three need a running Postfix, so they are a test-server check, not a
local one. `table.log` is the only restyled component with no screenshot
behind it.
4. ~~**CSP and the font ETags**~~**done** on the test server at
`1.1.0-post.669f928`. The policy is unchanged
(`default-src 'self'; object-src 'none'; base-uri 'none'; form-action 'self';
frame-ancestors 'none'`) and admits all three fonts, which come back as
`font/woff2` with `Cache-Control: no-cache` and a content ETag: a matching
`If-None-Match` gets 304, a stale one gets the bytes. The signed-out page
renders in Plex over the network.
5. **`font-display: swap` has never been observed** — every render had the fonts
already on disk. Worth one cold load over the network to see how long the
system stack is on screen.
## Done when
- The panel and the mark read as one design in both schemes, at the reading
measure and on the wide data pages.
- Nothing an operator does changed: same pages, same controls, same copy.
- Fonts are served from the panel's own origin under the unchanged CSP, and the
image works with no network access.
- `NOTICE` credits IBM Plex (OFL-1.1); build, vet and tests are green.
- The three views in **Outstanding** 3 are seen with real data on the test
server, and 4 is confirmed there.
## Risks
- **Visual regression across 21 templates.** The panel has pages that are only
reachable mid-workflow (the credential card, the delete confirmation, the
encrypt fields). Mitigation: the page order above is a checklist, and each
step is screenshotted rather than assumed.
- **Brick against the error red.** If the two fight in a real row of the send
log, the accent gets pulled back to the mark and the buttons stay neutral —
the mark's colour is fixed, the panel's accent is the negotiable one.
- **Thin weights on dark.** `h1` at 300 is the one place a variable font makes
it easy to go too light; check it on the dark scheme before keeping it.
-70
View File
@@ -1,70 +0,0 @@
# Plan: web-split (splitting `internal/web`)
**Status:** done (see [CHANGELOG](../CHANGELOG.md) `[1.1.0]`)
**Version:** `1.x`; an internal refactor, it does not force a break on its own.
---
## What this is
`internal/web` is the project's largest package: ~50 files (templates and
static assets included), ~25 `.go` files and ~4000 lines of Go, with the
handlers for every panel section, sessions, security headers, origin checking,
form validation and template rendering all sitting in one flat namespace.
The candidates to split out are `web/handlers` and `web/auth`, or a cut along
the panel's own domains.
## Why now
At its current size the flat package reads fine: the file names
(`handlers_domains.go`, `handlers_apps.go`, `handlers_monitor.go`) do the work
directories would, and splitting would force exporting what is package-private
today — widening the internal API for cosmetics.
It starts to pay off once the package grows: **domain-admin** and
**inbound-relay** both add code to it — the role brings authorisation into
every handler, the inbound relay brings its own pages and handlers for inbound
domains. The refactor is cheaper before that growth than after it.
## Recommended order
**web-split → domain-admin → inbound-relay** (see the
[roadmap](../roadmap.md)).
1. **web-split** — lay down the package structure (including a place for
`web/auth`) while there are no cross-cutting edits from the role and no new
inbound handlers.
2. **domain-admin** — authorisation in every handler builds on a package layout
already chosen.
3. **inbound-relay** — a new vertical slice; easier to add to an already split
package than to refactor alongside the two features before it.
The order is a recommendation, not a blocker.
## Chosen scheme
**Horizontal split into four packages** (decided at implementation):
```
internal/web/ # Config, Server, New, Handler — composition root; security.go
internal/web/view/ # embed templates/static, render/renderFragment, staticHandler
internal/web/auth/ # session, login/logout/setup, requireAuth, currentUser
internal/web/validate/ # shared form validation (avoids auth ↔ handlers import cycle)
internal/web/handlers/ # all authenticated page handlers (handlers_*.go)
```
`cmd/panel` keeps importing only `internal/web`. Subpackages are not exported
beyond what the composition root needs.
## Done when
The package is split along the scheme above. After the split: `build`/`vet`/`test`
green, the panel's behaviour unchanged.
## Risks
- Splitting too early — a superfluous internal API and churn with nothing to
show for it;
- leaving it until after the growth — a harder refactor, tangled up with the
features.
+12 -9
View File
@@ -47,17 +47,19 @@ Explicitly excluded to prevent scope creep:
- Inbound mail (IMAP/POP3, mailboxes, delivery to user inboxes)
- Webmail
- Multi-user panel / organisations / roles — one administrator; managing
**multiple sending domains** is in scope (see below)
- Organisations / tenancy beyond global + domain-admin roles; managing
**multiple sending domains** under one global administrator is in scope (see
below)
- Inbound antispam/antivirus (rspamd, ClamAV, etc.)
- A custom MTA — Postfix is used as-is
- Dovecot or a full mail stack for SASL — Cyrus SASL (`sasldb2`) only
Agreed **1.x+** extensions (optional inbound relay, domain-admin role) are
tracked in [roadmap.md](roadmap.md) and [plans/](plans/). Inbound relay targets
a 1.x MINOR bump by default; a 2.x major remains possible pending
implementation. Items still marked *candidate* in the roadmap require explicit
approval before coding.
The **domain-admin** role ships in the current line (global administrator plus
domain administrators with assigned domains). The optional **inbound relay** is
the main agreed **1.x+** extension still on the
[roadmap](roadmap.md) — it targets a 1.x MINOR bump by default; a 2.x major
remains possible pending implementation. Items marked *candidate* in the
roadmap require explicit approval before coding.
---
@@ -102,5 +104,6 @@ Adding a domain does **not** create an application automatically.
- **Delete domain** — removes DKIM key and **all** its applications.
- **Delete application** — removes only that app's SASL and map entries.
This is not multi-tenancy (one admin); it is one owner operating several
sending domains with independent application credentials.
This is not multi-tenancy; it is one owner (or a small team with global and
domain-scoped roles) operating several sending domains with independent
application credentials.
+98 -56
View File
@@ -15,7 +15,9 @@ planned but unlisted.
**Versioning:** SemVer MINOR in the **1.x+** line by default (`1.1.0`…), as long
as defaults and migrations stay compatible with `1.0.0`. A major `2.x` only for
an explicit break.
an explicit break. One such break, when 2.x is cut for any reason, is
[schema-squash](#schema-squash) — replacing the 1.x SQLite migration chain
with a baseline. That item does not by itself justify a major.
**Process:** [development.md](development.md). The history of closed phases is
in `git log` and [CHANGELOG.md](../CHANGELOG.md).
@@ -26,22 +28,46 @@ in `git log` and [CHANGELOG.md](../CHANGELOG.md).
| ID | Topic | Status | Plan |
|---|---|---|---|
| domain-admin | Domain administrator role | **agreed** | [plans/domain-admin.md](plans/domain-admin.md) |
| queue-retries | Postfix retry policy in the panel (queue lifetime, backoff) | **agreed** | [plans/queue-retries.md](plans/queue-retries.md) |
| inbound-relay | Inbound relay (backup-MX / forwarding) | **agreed** | [plans/inbound-relay.md](plans/inbound-relay.md) |
| contributing | `CONTRIBUTING.md` | candidate | — |
| visual-style | Panel visual style | **agreed** | [plans/visual-style.md](plans/visual-style.md) |
| dmarc-reports | DMARC aggregate report ingestion and panel UI | candidate | [plans/dmarc-reports.md](plans/dmarc-reports.md) |
| logrotate-mode | `mail.log` stops rotating in some builds | candidate | [plans/logrotate-mode.md](plans/logrotate-mode.md) |
| panel-docs | In-panel operator documentation | candidate | — |
| schema-squash | Squash SQLite migrations into a 2.x baseline | **2.x** | — |
**Recommended order** (not binding): **domain-admin →
inbound-relay** — role-wide authorisation first, then the inbound relay vertical
slice. Deviating is allowed; there are no hard phases here.
**Recommended order** (not binding): **queue-retries** is a small panel item
that can land first or in parallel; the next feature is **inbound-relay**.
The 2026-08-13 full-tree review follow-ups (send-log authorization,
fail-closed paths, docs, GUI, tests, licence) are closed — history in
[CHANGELOG.md](../CHANGELOG.md) `[Unreleased]` and git. Candidates need
explicit agreement before they join the queue.
After a context reset, pick an item marked `agreed` or `in progress`, then work
the checklist in its linked plan.
---
## queue-retries
**Goal:** show on Mail queue and on a delivery's history how this Postfix
retries deferred mail — first delay, backoff cap, queue lifetime — reading
the effective config (`postconf -h`) once at panel start so a manual
override is visible.
**Boundary:** explanation only. Postfix stays as-is; no attempt counter, no
panel knobs for queue lifetime, no schema change. Domain administrators see
the intervals on `/deliveries/{id}` (they cannot open Mail queue).
**Done when:** see the criteria in
[plans/queue-retries.md](plans/queue-retries.md).
**Dependencies / risks:** `postconf` unavailable outside the container
(fallback + muted note). Copy must stay time-based — Postfix has no max
attempt count.
**Version:** patch.
---
## inbound-relay
**Goal:** optional acceptance of mail on port 25 for explicitly configured
@@ -58,33 +84,11 @@ engine stays outside the image, only the attachment point is provided.
**Dependencies / risks:** a finished outbound path; open relay and backscatter;
a wider attack surface (port 25 accepting mail).
**Order:** recommended after [domain-admin](plans/domain-admin.md).
**Version:** target bump `1.x`; `2.x` possible — to be settled once the
implementation lands.
---
## domain-admin
**Goal:** a role with access to one or several assigned domains (the list is
set by the global administrator) — applications, DKIM/DNS, and the send log for
each of them; without global operations (adding domains, full backup, the
queue, `mail.log`).
**Boundary:** an extension of v1.0 — [product.md](product.md) fixes a single
administrator. Not a second all-powerful admin, but limited access to the
assigned domains (one or several).
**Done when:** see [plans/domain-admin.md](plans/domain-admin.md).
**Dependencies / risks:** a users table, the role in the session, authorisation
in every handler, setup and backup. **Order:** recommended **before**
[inbound-relay](plans/inbound-relay.md).
**Version:** `1.x` MINOR, given a compatible migration of the current
administrator into a global one.
---
## contributing
**Goal:** `CONTRIBUTING.md` in the root — the dev loop, the checks to run
@@ -103,30 +107,6 @@ priority.
---
## visual-style
**Goal:** refresh the control panel's visual design — typography, colour tokens,
spacing, and component styling — without changing operator workflows or panel
behaviour.
**Boundary:** presentation only (`panel.css`, templates, static assets); no new
features. Styling must stay compatible with the panel CSP — rules live in
`panel.css`, not inline (see [security.md](security.md) and the stylesheet
header).
**Done when:** see [plans/visual-style.md](plans/visual-style.md). The agreed
direction is the mark's own — brick, warm paper, IBM Plex — taken from
[assets/selfpost-proof.html](assets/selfpost-proof.html); light and dark schemes
remain supported; readability and contrast are preserved.
**Dependencies / risks:** CSP constraints on how styles are applied (rules in
`panel.css`, never inline); visual regression across pages; the accent colour
must not read as a status badge.
**Order:** independent of the feature items; may be taken up between them.
**Version:** no bearing on semver.
---
## dmarc-reports
**Goal:** SelfPost receives DMARC aggregate reports (RFC 7489) on SMTP,
@@ -142,12 +122,74 @@ reports (`ruf=`) out of scope for v1.
**Dependencies / risks:** inbound SMTP in the image (may share infrastructure
with [inbound-relay](plans/inbound-relay.md) but must not require backup-MX);
storage and retention of parsed summaries; the `admin.dmarc_report_email` and
`domains.dmarc_rua` settings added in the DMARC template work must stay the
source of truth for `rua=` in DNS guidance.
storage and retention of parsed summaries; the `dmarc_report_email` setting
(migration `0005` moved it off the old `admin` table into `settings`) and
`domains.dmarc_rua` added in the DMARC template work must stay the source of
truth for `rua=` in DNS guidance.
**Order:** after the DMARC `rua=` settings ship; may follow or overlap with
inbound-relay depending on how port 25 acceptance is structured.
**Version:** `1.x` MINOR.
---
## panel-docs
**Goal:** built-in operator documentation in the panel — short pages (or a
help drawer) that explain what each Status check and other controls mean,
without sending the operator out to `docs/guide.md`.
**Boundary:** in-panel help only; not a second copy of the full operator guide.
Seed content includes the Status blurbs removed from the cards in favour of a
denser layout — Machine (kernel counters / rate window), TLS certificate
(port 465, reverse-proxy mount), Hostname / reverse DNS (forward-confirmed
reverse DNS, PTR at the hosting provider), and similar notes for other panel
surfaces as they lose inline commentary.
**Done when:** an operator can open help from the panel for those topics; the
removed Status blurbs are preserved there (or equivalent); no requirement to
read the git tree for day-to-day meaning of a card.
**Dependencies / risks:** copy ownership and translation; keeping help in sync
when checks change; not bloating every page with a second column of prose.
**Version:** `1.x` MINOR; `candidate` until explicitly agreed.
---
## schema-squash
**Goal:** when 2.x is cut, stop shipping the 1.x migration files
(`0001_init.sql``0005_panel_users.sql`) in the binary and replace them with
one baseline that is the schema as of `user_version = 5`. Fresh 2.x data
directories no longer create-then-drop the historical `admin` table.
**Boundary:** 1.x keeps the full chain so a 1.0.0 data directory still boots.
Do not delete, rename, or reorder those files while MINOR compatibility with
`1.0.0` holds. `migrate()` maps **file order** to `PRAGMA user_version` (`target
= i + 1`); dropping a file in 1.x would skip or mis-apply steps on existing
databases. Git history keeps the old files either way; only the embedded set
in the 2.x image changes.
**Upgrade gate (required with the squash):**
| `user_version` | 2.x behaviour |
|---|---|
| `0` (empty file) | Apply the baseline; set `user_version` to the new chains head |
| `>= 5` (fully migrated 1.x) | Skip; schema is already the baseline |
| `1``4` (mid-chain 1.x) | **Refuse to start** — boot the last 1.x once, then 2.x |
Restore remains a separate lock: the backup manifest version must match the
running binary ([architecture.md](architecture.md) § Persistence). It does not
replace this gate.
**Done when:** 2.x embeds a single baseline (plus any 2.x-only migrations after
it); the gate above is tested; the operator guide says a 2.x image will not
open an unfinished 1.x database.
**Dependencies / risks:** a decided 2.x cut (another breaking change, or an
explicit major). Squashing five short files is not a reason to cut 2.x on its
own. A missed gate leaves a `user_version = 3` database silently stuck.
**Version:** `2.x` major only; not a 1.x item.
+59 -24
View File
@@ -5,7 +5,12 @@ meet; the full v1.0 audit passed. The pre-release review (plan § D, 2026-08-06)
covered the whole diff from the v1.0 audit (Phase 11) to HEAD and the checklist
in full: no exploitable findings; one defence-in-depth change — `--` before the
login in the `saslpasswd2` argv
([internal/app/sasl.go](../internal/app/sasl.go)). (2) **Accepted risks**
([internal/app/sasl.go](../internal/app/sasl.go)). The 2026-08-14 review
(code-review plan § P7, Fable; reviewer ≠ author) covered the P0P1 diff of the
2026-08-13 full-tree review against this document — send-log authorization for
domain administrators, the atomic level-2 admit (`tryAdmit`), fail-closed
session creation, and application-delete ordering: no findings, and nothing
needed adding to the accepted risks. (2) **Accepted risks**
deliberate departures beyond the mandatory, recorded so the decision is not
lost.
@@ -34,8 +39,8 @@ The panel is exposed to the internet — the items below are **not optional**.
- Failed attempts do **not** invalidate the token early (protects setup from
being DoS-ed).
- Once the administrator exists the token is void forever, `/setup/*` → 404.
- The administrator password is bcrypt (or argon2) in SQLite only; no plaintext
and no MD5.
- The administrator password is bcrypt in SQLite only; no plaintext and no
MD5.
- `PANEL_USERNAME` / `PANEL_PASSWORD_HASH` in env are **not used**.
### Application SASL passwords
@@ -112,19 +117,34 @@ deferred item from the roadmap.
- **A `POST` with neither `Sec-Fetch-Site` nor `Origin` is allowed through.**
A client that sends neither — a genuinely old browser, or a webview with a
frozen engine — stays vulnerable to CSRF from any site. Accepted
deliberately: the panel is single-user, the administrator picks the browser,
and a strict mode would not "protect" such a client, it would simply break the
panel in it. Tightening is one line in `originAllowed`
deliberately: every panel user (global or domain-admin) is an operator who
picks their own browser, not an untrusted party the panel needs to defend
against, and a strict mode would not "protect" such a client, it would
simply break the panel in it. Tightening is one line in `originAllowed`
([internal/web/security.go](../internal/web/security.go)): return `false`
instead of `true` in the "neither header present" branch.
- **Session-bound CSRF tokens are not implemented.** The origin check closes the
neighbouring-subdomain case but depends on browser behaviour; a token does
not. The price is a hidden field in roughly two dozen forms. The trigger to
revisit is a requirement for protection that holds regardless of the browser.
A token would not save the panel from XSS inside it either: code executing in
the panel's origin sends the request itself — against that, `html/template`
auto-escaping and CSP do the work, which is why templates must contain no
inline scripts and no inline styles.
revisit is a requirement for protection that holds regardless of the browser,
or a domain-admin population the global administrator does not fully trust
(see the ADR below). A token would not save the panel from XSS inside it
either: code executing in the panel's origin sends the request itself —
against that, `html/template` auto-escaping and CSP do the work, which is why
templates must contain no inline scripts and no inline styles.
- **Destructive-action confirmation (`data-confirm`) is JavaScript-only.**
Delete, regenerate-password, and clear-rate-limit forms carry a
`data-confirm` prompt handled entirely in
[panel.js](../internal/web/view/static/panel.js); with JavaScript disabled
or blocked the form submits immediately, exactly as it did before the
prompts existed. Accepted deliberately: the prompt is a mis-click guard,
not an authorization boundary — the same origin check and session/RBAC
gate every one of these `POST`s whether or not JavaScript ran. Progressive
enhancement means the panel must work with JavaScript off; a
server-rendered confirmation step would need a second page (or a `?confirm=1`
round trip) for every one of these forms, which is what
[`user_delete.html`](../internal/web/view/templates/user_delete.html) and
`domain_delete.html` already do for the two highest-blast-radius deletes.
- **Encrypting backups and exports is an option, not the default.** With the
checkbox cleared the file downloads in the clear, as in 1.0. Otherwise an
operator with nowhere to keep a password would lose the ability to take a
@@ -158,24 +178,35 @@ deferred item from the roadmap.
**Context.** The panel is forms (`POST`) with a cookie session — the classic
CSRF surface. What is needed is a way to tell a request from the panel's own
page apart from one initiated by a third-party site in the logged-in
administrator's browser.
page apart from one initiated by a third-party site in a logged-in user's
browser. The panel is multi-user since 1.2.0 (a global administrator plus
zero or more domain-admin users, each scoped to their assigned domains), but
that is an authorization boundary (who can see or change what), not a change
to the CSRF threat: the attacker in scope here is still an external site
riding a legitimate user's cookie, not one panel user attacking another
through the browser.
**Decision.** `originAllowed` in
[internal/web/security.go](../internal/web/security.go) checks `Sec-Fetch-Site`
(when the browser sends it) or `Origin` (fallback) against the panel's host; a
request carrying neither header is **allowed through** rather than rejected.
There are no session-bound tokens embedded in forms.
There are no session-bound tokens embedded in forms. The check applies the same
way regardless of the requesting user's role.
**Why not tokens.** The panel is single-user (one administrator per instance) —
the threat model does not include cross-user CSRF inside the panel itself, only
an external site making the administrator's browser send a request. The origin
check covers that without touching a single template: a token would need a
hidden field in roughly two dozen forms and synchronisation with every new form,
and it would still not protect against XSS inside the panel — code executing in
the panel's origin reads the token and sends the request itself. XSS is handled
by `html/template` auto-escaping and CSP, so that is a separate line of defence,
not a CSRF token.
**Why not tokens.** Cross-user CSRF is not the threat model here: a
domain-admin's browser sending a request still needs that domain-admin's own
cookie, so a token would not add a boundary between roles that the
authorization checks (`Principal.CanAccessDomain`,
[internal/web/auth/principal.go](../internal/web/auth/principal.go); route
gating in [internal/web/handlers/authz.go](../internal/web/handlers/authz.go))
don't already enforce. The remaining case is an external site making a
logged-in user's browser send a request, which the origin check covers without
touching a single template.
A token would need a hidden field in roughly two dozen forms and
synchronisation with every new form, and it would still not protect against
XSS inside the panel — code executing in the panel's origin reads the token
and sends the request itself. XSS is handled by `html/template` auto-escaping
and CSP, so that is a separate line of defence, not a CSRF token.
**Trade-off.** A client that sends neither `Sec-Fetch-Site` nor `Origin` (a
genuinely old browser, or a webview with a frozen engine) stays vulnerable — see
@@ -183,7 +214,11 @@ genuinely old browser, or a webview with a frozen engine) stays vulnerable — s
such a client, at the price of a narrow residual surface.
**Revisit if:** a requirement appears for protection that does not depend on
browser behaviour, or the panel becomes multi-user.
browser behaviour, or domain-admin accounts stop being trusted operators (for
example, if a future release lets a global administrator invite domain-admins
whose browsers/devices are not vetted) — at that point cross-role request
forgery inside the panel would need its own analysis, separate from the
external-site case this ADR covers.
## How this list grows
+14 -4
View File
@@ -166,14 +166,24 @@ func (s *Service) RegeneratePassword(id int64) (string, error) {
// Delete removes an application: its SASL account, its registry row (and address
// rows via cascade) and its sender-map bindings, then reloads Postfix (spec
// 7.2.8). The domain and other applications are untouched.
//
// The order matches domain deletion: the SASL account goes first, while the
// login is still in the registry. Dropping the row first would, on a
// saslpasswd2 failure, leave an account that can still authenticate to Postfix
// but that the panel no longer knows about — an orphan no operator can see or
// remove. Failing before the row is deleted is recoverable: the application is
// still listed and the delete can be retried.
func (s *Service) Delete(id int64) error {
a, err := s.store.DeleteApplication(id)
a, err := s.store.GetApplication(id)
if err != nil {
return err
}
if err := s.sasl.Delete(a.Login); err != nil {
return err
}
if _, err := s.store.DeleteApplication(id); err != nil {
return err
}
// Drop the application's level-2 limit, if any (guide § Rate limiting);
// rate_limits has no cascade of its own.
if err := s.store.DeleteRateLimit(store.RateLimitScopeApp, id); err != nil {
@@ -189,9 +199,9 @@ func (s *Service) RateLimit(appID int64) (store.RateLimit, bool, error) {
return s.store.GetRateLimit(store.RateLimitScopeApp, appID)
}
// SaveRateLimit stores the application-level rate limit. The caller has validated
// the IPs and numbers (security.md); the milter reads the row live, so no reload
// is needed.
// SaveRateLimit stores the application-level trusted-IP override (guide § Rate
// limiting). The caller has validated the IPs and numbers (security.md); the
// milter reads the row live, so no reload is needed.
func (s *Service) SaveRateLimit(appID int64, ips []string, maxMessages, windowSeconds int) error {
return s.store.SetRateLimit(store.RateLimit{
Scope: store.RateLimitScopeApp,
+31
View File
@@ -181,6 +181,37 @@ func TestServiceDelete(t *testing.T) {
}
}
// If sasldb2 cannot be updated the application must stay in the registry: an
// account that still authenticates but has no panel row is invisible to the
// operator and cannot be deleted again.
func TestServiceDeleteKeepsRowWhenSASLFails(t *testing.T) {
svc, st, rec, _ := newServiceHarness(t)
d := addDomain(t, st, "example.com")
a, _, err := svc.Create(d.ID, "app1", store.AddressModeWildcard, nil)
if err != nil {
t.Fatal(err)
}
rec.failNext = true // saslpasswd2 -d fails
if err := svc.Delete(a.ID); err == nil {
t.Fatal("Delete reported success although the SASL account was not removed")
}
apps, _ := st.ListApplicationsByDomain(d.ID)
if len(apps) != 1 {
t.Fatalf("registry row dropped while the SASL account can still authenticate: %+v", apps)
}
if _, ok := rec.set["app1"]; !ok {
t.Fatal("SASL account gone despite the failure — the harness no longer proves the ordering")
}
// The delete is retryable now that the row is still there.
if err := svc.Delete(a.ID); err != nil {
t.Fatalf("retried Delete: %v", err)
}
if _, ok := rec.set["app1"]; ok {
t.Error("SASL account not deleted on retry")
}
}
func TestServiceUpdateMode(t *testing.T) {
svc, st, _, maps := newServiceHarness(t)
d := addDomain(t, st, "example.com")
+23 -16
View File
@@ -6,13 +6,19 @@
// it. TLS certificates (the reverse proxy's responsibility) and the Postfix
// queue are deliberately excluded (architecture.md § Persistence).
//
// Restore is intentionally not a separate code path: a backup is extracted
// into the /data bind mount before first start, and the panel regenerates
// Postfix and OpenDKIM from the restored SQLite state exactly as on any normal
// start. The only restore-specific step is CheckRestore, which refuses to boot
// if the manifest's version does not match the running binary, so
// Restore is not a separate code path in the panel: a backup is extracted into
// the /data bind mount before first start, and the archive already carries
// everything the mail path needs — DKIM keys, sasldb2, and Postfix's sender
// map — so the daemons can start on the extracted files alone. The
// restore-specific steps the panel runs are CheckRestore, which refuses to
// boot if the manifest's version does not match the running binary so
// schema/format skew between versions cannot silently corrupt state
// (architecture.md § Persistence).
// (architecture.md § Persistence), and a one-time Resync of OpenDKIM's tables
// and the Postfix sender map from SQLite on that first boot, so any drift
// between the archive and the database is healed before mail flows. If
// on-disk state drifts again later — for example after a manual edit under
// /data — the Status page's "Reload configuration" button runs the same
// Resync on demand.
package backup
import (
@@ -264,33 +270,34 @@ func snapshotDB(dbPath string) (path string, cleanup func(), err error) {
// directory), its version must match binaryVersion or the panel refuses to
// start, telling the operator which image tag to use. On a match the manifest
// is consumed (deleted) so it guards only the first boot after a restore and
// never blocks a later in-place image upgrade. Absence of the manifest is the
// normal case and returns nil.
func CheckRestore(manifestPath, binaryVersion string) error {
// never blocks a later in-place image upgrade, and restored is true so the
// caller can heal drifted daemon maps once. Absence of the manifest is the
// normal case and returns restored == false with a nil error.
func CheckRestore(manifestPath, binaryVersion string) (restored bool, err error) {
data, err := os.ReadFile(manifestPath)
if os.IsNotExist(err) {
return nil // ordinary start, not a restore
return false, nil // ordinary start, not a restore
}
if err != nil {
return fmt.Errorf("backup: read restore manifest: %w", err)
return false, fmt.Errorf("backup: read restore manifest: %w", err)
}
var m Manifest
if err := json.Unmarshal(data, &m); err != nil {
return fmt.Errorf("backup: restore manifest %s is not valid JSON: %w", manifestPath, err)
return false, fmt.Errorf("backup: restore manifest %s is not valid JSON: %w", manifestPath, err)
}
if m.Format != FormatFull {
return fmt.Errorf("backup: %s is not a SelfPost full backup manifest (format %q)", manifestPath, m.Format)
return false, fmt.Errorf("backup: %s is not a SelfPost full backup manifest (format %q)", manifestPath, m.Format)
}
if m.Version != binaryVersion {
return fmt.Errorf(
return false, fmt.Errorf(
"backup: this backup was created by SelfPost %s but this image is %s — restore into the matching image (selfpost:%s)",
m.Version, binaryVersion, m.Version)
}
// Version matches: consume the manifest so subsequent normal starts (and
// in-place upgrades) are not gated by it.
if err := os.Remove(manifestPath); err != nil {
return fmt.Errorf("backup: consume restore manifest: %w", err)
return false, fmt.Errorf("backup: consume restore manifest: %w", err)
}
return nil
return true, nil
}
+19 -4
View File
@@ -156,17 +156,25 @@ func writeManifest(t *testing.T, dir, format, version string) string {
}
func TestCheckRestoreNoManifestIsNormalStart(t *testing.T) {
if err := CheckRestore(filepath.Join(t.TempDir(), "manifest.json"), "1.0.0"); err != nil {
restored, err := CheckRestore(filepath.Join(t.TempDir(), "manifest.json"), "1.0.0")
if err != nil {
t.Errorf("CheckRestore with no manifest = %v, want nil", err)
}
if restored {
t.Error("CheckRestore with no manifest reported a restore")
}
}
func TestCheckRestoreMatchConsumesManifest(t *testing.T) {
dir := t.TempDir()
path := writeManifest(t, dir, FormatFull, "1.0.0")
if err := CheckRestore(path, "1.0.0"); err != nil {
restored, err := CheckRestore(path, "1.0.0")
if err != nil {
t.Fatalf("CheckRestore matching = %v, want nil", err)
}
if !restored {
t.Fatal("CheckRestore matching did not report a restore")
}
if _, err := os.Stat(path); !os.IsNotExist(err) {
t.Errorf("manifest should be consumed after a matching restore, stat err = %v", err)
}
@@ -175,10 +183,13 @@ func TestCheckRestoreMatchConsumesManifest(t *testing.T) {
func TestCheckRestoreVersionMismatchRefusesAndKeeps(t *testing.T) {
dir := t.TempDir()
path := writeManifest(t, dir, FormatFull, "1.0.0")
err := CheckRestore(path, "2.0.0")
restored, err := CheckRestore(path, "2.0.0")
if err == nil {
t.Fatal("CheckRestore mismatch = nil, want error")
}
if restored {
t.Error("CheckRestore mismatch reported a restore")
}
if !strings.Contains(err.Error(), "1.0.0") || !strings.Contains(err.Error(), "2.0.0") {
t.Errorf("error should name both versions: %v", err)
}
@@ -190,7 +201,11 @@ func TestCheckRestoreVersionMismatchRefusesAndKeeps(t *testing.T) {
func TestCheckRestoreWrongFormatRejected(t *testing.T) {
dir := t.TempDir()
path := writeManifest(t, dir, "something-else", "1.0.0")
if err := CheckRestore(path, "1.0.0"); err == nil {
restored, err := CheckRestore(path, "1.0.0")
if err == nil {
t.Error("CheckRestore accepted a non-backup manifest")
}
if restored {
t.Error("CheckRestore wrong format reported a restore")
}
}
+8
View File
@@ -167,6 +167,14 @@ func (o *OpenDKIM) Reload() error {
return o.reload()
}
// SetReloadHook replaces how configuration is applied after a rebuild. Tests
// that cannot reach supervisord use this to verify file regeneration alone.
func (o *OpenDKIM) SetReloadHook(fn func() error) {
if fn != nil {
o.reload = fn
}
}
// renderTables builds the KeyTable and SigningTable byte contents for a domain
// set, sorted by name so the output is deterministic. Every domain is
// re-checked for shell/config-injection safety before being written
+3 -3
View File
@@ -145,13 +145,13 @@ func (s *Service) RateLimit(domainID int64) (store.RateLimit, bool, error) {
}
// SaveRateLimit stores the domain-level rate limit. The caller has validated the
// IPs and numbers (security.md); the milter reads the row live, so no reload is
// needed.
// numbers (security.md); the milter reads the row live, so no reload is needed.
// Domain limits do not use an IP allowlist.
func (s *Service) SaveRateLimit(domainID int64, ips []string, maxMessages, windowSeconds int) error {
return s.store.SetRateLimit(store.RateLimit{
Scope: store.RateLimitScopeDomain,
RefID: domainID,
AllowedIPs: ips,
AllowedIPs: ips, // unused for domain enforcement; kept empty by the panel
MaxMessages: maxMessages,
WindowSeconds: windowSeconds,
})
+100 -35
View File
@@ -62,10 +62,14 @@ type CPU struct {
// nothing until it is true.
Measured bool
BusyPct float64
// Cores and Threads are physical cores and logical CPUs (SMT). Both come
// from /proc/cpuinfo when it is readable; otherwise Threads falls back to
// the cpuN lines in /proc/stat and Cores matches Threads.
Cores int
Threads int
// Load is the 1/5/15-minute load average, present when /proc/loadavg
// could be read. Unlike BusyPct it needs no previous sample, so it is
// there on the very first page load.
// could be read. Unlike BusyPct it needs no previous sample. It is kept
// for the sampler but is no longer shown on the Status card.
Load [3]float64
HasLoad bool
Status Status
@@ -176,7 +180,14 @@ type netCounters struct {
func (m *MachineSampler) Sample() Machine {
root := m.root()
now := time.Now()
cpuNow, cores, cpuErr := readCPUTimes(root)
cpuNow, logical, cpuErr := readCPUTimes(root)
cores, threads := readCPUTopology(root)
if threads == 0 {
threads = logical
}
if cores == 0 {
cores = threads
}
netNow, netErr := readNetDev(root)
m.mu.Lock()
@@ -198,7 +209,7 @@ func (m *MachineSampler) Sample() Machine {
fresh := !prevAt.IsZero() && window > 0 && window <= machineSampleWindow
mach := Machine{
CPU: cpuUsage(prevCPU, cpuNow, cores, readLoadAvg(root), fresh, cpuErr),
CPU: cpuUsage(prevCPU, cpuNow, cores, threads, readLoadAvg(root), fresh, cpuErr),
Memory: readMemory(root),
Network: networkUsage(prevNet, netNow, window, fresh, netErr),
}
@@ -209,12 +220,11 @@ func (m *MachineSampler) Sample() Machine {
return mach
}
// cpuUsage grades the processor over the window. The load average is reported
// alongside it because the two answer different questions — how busy the
// processor was, and how many tasks were waiting for it — and a machine can
// look idle while work queues up behind a slow disk.
func cpuUsage(prev, cur cpuTimes, cores int, load [3]float64, fresh bool, err error) CPU {
c := CPU{Cores: cores}
// cpuUsage grades the processor over the window. Detail carries only the
// core and thread counts — load average and busy prose stay out of the Status
// card (see roadmap panel-docs for operator-facing explanation later).
func cpuUsage(prev, cur cpuTimes, cores, threads int, load [3]float64, fresh bool, err error) CPU {
c := CPU{Cores: cores, Threads: threads}
if !isZeroLoad(load) {
c.Load, c.HasLoad = load, true
}
@@ -223,9 +233,9 @@ func cpuUsage(prev, cur cpuTimes, cores int, load [3]float64, fresh bool, err er
c.Detail = "The kernel's processor counters (/proc/stat) could not be read here."
return c
}
c.Detail = c.contextText()
if !fresh || cur.total <= prev.total {
c.Status = StatusUnknown
c.Detail = joinDetail("Measuring — this reading sets the baseline; the next refresh has the figure.", c.contextText())
return c
}
@@ -242,25 +252,24 @@ func cpuUsage(prev, cur cpuTimes, cores int, load [3]float64, fresh bool, err er
if c.BusyPct >= cpuWarnPct {
c.Status = StatusWarn
c.Detail = joinDetail(c.contextText(), "The processor is close to fully busy, which slows queue processing and every panel page.")
} else {
c.Status = StatusOK
c.Detail = c.contextText()
}
return c
}
// contextText is the CPU's supporting figures: what the percentage is a
// percentage of, and how deep the run queue is.
// contextText is the CPU detail column: physical cores and logical threads.
func (c CPU) contextText() string {
var parts []string
if c.Cores > 0 {
parts = append(parts, fmt.Sprintf("%d core(s)", c.Cores))
switch {
case c.Cores > 0 && c.Threads > 0:
return fmt.Sprintf("%d cores · %d threads", c.Cores, c.Threads)
case c.Threads > 0:
return fmt.Sprintf("%d threads", c.Threads)
case c.Cores > 0:
return fmt.Sprintf("%d cores", c.Cores)
default:
return ""
}
if c.HasLoad {
parts = append(parts, fmt.Sprintf("load average %.2f, %.2f, %.2f", c.Load[0], c.Load[1], c.Load[2]))
}
return strings.Join(parts, " · ")
}
// readMemory reports main memory from /proc/meminfo. Used is derived from
@@ -296,8 +305,7 @@ func readMemory(root string) Memory {
m.SwapUsedBytes = m.SwapTotalBytes - swapFree
}
detail := fmt.Sprintf("%s used of %s; %s available to new work.",
humanBytes(m.UsedBytes), humanBytes(total), humanBytes(available))
detail := fmt.Sprintf("%s used of %s.", humanBytes(m.UsedBytes), humanBytes(total))
if m.SwapTotalBytes > 0 {
detail += fmt.Sprintf(" Swap: %s of %s.", humanBytes(m.SwapUsedBytes), humanBytes(m.SwapTotalBytes))
}
@@ -418,6 +426,74 @@ func readCPUTimes(root string) (cpuTimes, int, error) {
return times, cores, nil
}
// readCPUTopology returns physical core and logical thread counts from
// /proc/cpuinfo. Zeroes mean the file was missing or empty; the caller falls
// back to the cpuN count from /proc/stat.
func readCPUTopology(root string) (cores, threads int) {
data, err := os.ReadFile(filepath.Join(root, "cpuinfo"))
if err != nil {
return 0, 0
}
type coreKey struct{ phys, core int }
seen := map[coreKey]bool{}
var (
inCPU bool
phys = -1
core = -1
idx int
)
flush := func() {
if !inCPU {
return
}
p, c := phys, core
if p < 0 {
p = idx
}
if c < 0 {
c = idx
}
seen[coreKey{p, c}] = true
inCPU, phys, core = false, -1, -1
}
for _, line := range strings.Split(string(data), "\n") {
line = strings.TrimSpace(line)
if line == "" {
flush()
continue
}
key, val, ok := strings.Cut(line, ":")
if !ok {
continue
}
key = strings.TrimSpace(key)
val = strings.TrimSpace(val)
switch key {
case "processor":
flush()
inCPU = true
threads++
idx = threads - 1
if n, err := strconv.Atoi(val); err == nil {
idx = n
}
case "physical id":
if n, err := strconv.Atoi(val); err == nil {
phys = n
}
case "core id":
if n, err := strconv.Atoi(val); err == nil {
core = n
}
}
}
flush()
if threads == 0 {
return 0, 0
}
return len(seen), threads
}
// readLoadAvg reads the 1/5/15-minute load averages. A machine without
// /proc/loadavg simply has none reported, so the failure is a zero value rather
// than an error.
@@ -544,14 +620,3 @@ func humanRate(perSec float64) string {
}
return humanBytes(uint64(perSec+0.5)) + "/s"
}
// joinDetail joins the non-empty parts of a detail line.
func joinDetail(parts ...string) string {
var kept []string
for _, p := range parts {
if p != "" {
kept = append(kept, p)
}
}
return strings.Join(kept, " ")
}
+45 -10
View File
@@ -47,7 +47,7 @@ func TestCPUUsageOverTwoReadings(t *testing.T) {
// 1000 ticks pass, 250 of them idle: 75% busy.
prev := cpuTimes{total: 10000, idle: 8000}
cur := cpuTimes{total: 11000, idle: 8250}
got := cpuUsage(prev, cur, 4, [3]float64{0.5, 0.4, 0.3}, true, nil)
got := cpuUsage(prev, cur, 4, 8, [3]float64{0.5, 0.4, 0.3}, true, nil)
if !got.Measured {
t.Fatalf("reading not marked measured: %+v", got)
}
@@ -57,16 +57,22 @@ func TestCPUUsageOverTwoReadings(t *testing.T) {
if got.Status != StatusOK {
t.Errorf("status = %q, want ok", got.Status)
}
if got.Cores != 4 || !got.HasLoad {
t.Errorf("cores/load not reported: %+v", got)
if got.Cores != 4 || got.Threads != 8 || !got.HasLoad {
t.Errorf("cores/threads/load not reported: %+v", got)
}
if got.Detail != "4 cores · 8 threads" {
t.Errorf("detail = %q, want cores and threads only", got.Detail)
}
}
func TestCPUUsageWarnsWhenFullyBusy(t *testing.T) {
got := cpuUsage(cpuTimes{total: 10000, idle: 5000}, cpuTimes{total: 11000, idle: 5010}, 1, [3]float64{}, true, nil)
got := cpuUsage(cpuTimes{total: 10000, idle: 5000}, cpuTimes{total: 11000, idle: 5010}, 1, 1, [3]float64{}, true, nil)
if got.Status != StatusWarn {
t.Errorf("99%% busy graded %q, want warn (%s)", got.Status, got.Detail)
}
if got.Detail != "1 cores · 1 threads" {
t.Errorf("detail = %q, want topology only (no warn prose)", got.Detail)
}
if got.HasLoad {
t.Error("a missing load average should not be reported as zeros")
}
@@ -84,17 +90,44 @@ func TestCPUUsageWithoutAUsableWindow(t *testing.T) {
{"no previous reading", cpuTimes{}, cpuTimes{total: 11000, idle: 8250}, false},
{"counters did not advance", cpuTimes{total: 11000, idle: 8250}, cpuTimes{total: 11000, idle: 8250}, true},
} {
got := cpuUsage(c.prev, c.cur, 2, [3]float64{}, c.fresh, nil)
got := cpuUsage(c.prev, c.cur, 2, 2, [3]float64{}, c.fresh, nil)
if got.Measured || got.Status != StatusUnknown {
t.Errorf("%s: measured=%v status=%q, want unmeasured/unknown", c.name, got.Measured, got.Status)
}
}
if got := cpuUsage(cpuTimes{}, cpuTimes{}, 0, [3]float64{}, false, os.ErrNotExist); got.Status != StatusUnknown {
if got := cpuUsage(cpuTimes{}, cpuTimes{}, 0, 0, [3]float64{}, false, os.ErrNotExist); got.Status != StatusUnknown {
t.Errorf("unreadable /proc/stat: status %q, want unknown", got.Status)
}
}
func TestReadCPUTopology(t *testing.T) {
dir := t.TempDir()
const cpuinfo = `processor : 0
physical id : 0
core id : 0
processor : 1
physical id : 0
core id : 0
processor : 2
physical id : 0
core id : 1
processor : 3
physical id : 0
core id : 1
`
if err := os.WriteFile(filepath.Join(dir, "cpuinfo"), []byte(cpuinfo), 0o600); err != nil {
t.Fatal(err)
}
cores, threads := readCPUTopology(dir)
if cores != 2 || threads != 4 {
t.Errorf("topology = %d cores / %d threads, want 2/4", cores, threads)
}
}
func TestReadMemory(t *testing.T) {
dir := fakeProc(t, "cpu 1 1 1 1 1 1 1 1\n", meminfoSample, "0.1 0.2 0.3 1/2 3\n", netdevSample)
got := readMemory(dir)
@@ -104,8 +137,7 @@ func TestReadMemory(t *testing.T) {
if got.TotalBytes != 4*1024*1024*1024 {
t.Errorf("total = %d bytes (%s), want 4 GiB", got.TotalBytes, got.TotalText())
}
// 4 GiB total, 2 GiB available to new work: half used, cache included in
// what is available.
// 4 GiB total, 2 GiB MemAvailable: half used (cache counted as available).
if got.Percent() != 50 {
t.Errorf("used = %.1f%% (%s), want 50%%", got.UsedPct, got.PctText())
}
@@ -221,8 +253,11 @@ func TestMachineSamplerNeedsTwoReadings(t *testing.T) {
if !first.Memory.Measured {
t.Error("memory is a level, not a rate: it must be reported on the first sample")
}
if first.CPU.Cores != 2 {
t.Errorf("cores = %d, want 2", first.CPU.Cores)
if first.CPU.Cores != 2 || first.CPU.Threads != 2 {
t.Errorf("cores/threads = %d/%d, want 2/2", first.CPU.Cores, first.CPU.Threads)
}
if first.CPU.Detail != "2 cores · 2 threads" {
t.Errorf("detail = %q, want topology only", first.CPU.Detail)
}
if !first.CPU.HasLoad || first.CPU.Load[0] != 0.42 {
t.Errorf("load average not read: %+v", first.CPU.Load)
+1 -1
View File
@@ -36,7 +36,7 @@ func CheckSocket(name, path string, required bool) Socket {
default:
s.Present = true
s.Status = StatusOK
s.Detail = "Listening."
s.Detail = "Listening"
}
return s
}
+3 -2
View File
@@ -14,8 +14,9 @@ import (
var License []byte
// SourceURL is where Corresponding Source for the published upstream
// releases lives. Operators who ship a modified version must point their
// users at their own sources instead (NOTICE; AGPL-3.0 §13).
// releases lives. It is injected into the panel footer by view.go.
// Operators who ship a modified version must change this constant so the
// footer points at their own sources (NOTICE; AGPL-3.0 §13).
const SourceURL = "https://github.com/mixeme/selfpost"
// CopyrightLine is the short copyright notice shown in the panel footer.
+19
View File
@@ -19,3 +19,22 @@ func TestEmbeddedLicenseMatchesRoot(t *testing.T) {
t.Fatal("embedded LICENSE is empty")
}
}
// NOTICE used to tell modifiers to edit layout.html for the Source URL. The
// footer reads legal.SourceURL; a fork that only changed the template would
// still advertise the upstream repo.
func TestNoticePointsAtSourceURLConstant(t *testing.T) {
notice, err := os.ReadFile(filepath.Join("..", "..", "NOTICE"))
if err != nil {
t.Fatalf("read NOTICE: %v", err)
}
if !bytes.Contains(notice, []byte("internal/legal/legal.go")) {
t.Error("NOTICE must tell modifiers to update SourceURL in internal/legal/legal.go")
}
if bytes.Contains(notice, []byte("layout.html")) {
t.Error("NOTICE still tells modifiers to edit layout.html for the Source URL")
}
if !bytes.Contains(notice, []byte("internal/web/view/static/OFL.txt")) {
t.Error("NOTICE must point at the OFL text that travels with the Plex fonts")
}
}
+30 -8
View File
@@ -36,6 +36,29 @@ type inflight struct {
m map[string]map[*reservation]struct{}
}
// tryAdmit decides whether one more message may be sent for key and, if so,
// claims its slot — both under a single lock. Counting and reserving in two
// separate critical sections would let two SMTP sessions that reach MAIL FROM
// at the same moment observe the same total and each take the last free slot,
// which is exactly the overshoot the in-flight registry exists to prevent.
//
// stored is the count the send log already holds for the limit's window and max
// is the ceiling; the caller supplies both because only it can query the store.
// The returned total is what was measured, for the refusal log line.
func (f *inflight) tryAdmit(key string, since time.Time, stored, max int64) (*reservation, int64, bool) {
if f == nil {
return nil, stored, stored < max // no in-flight accounting (tests)
}
f.mu.Lock()
defer f.mu.Unlock()
total := stored + f.countLocked(key, since)
if total >= max {
return nil, total, false
}
return f.reserveLocked(key), total, true
}
// count returns how many reservations for key were taken within the limit's
// window (at or after since), pruning any that outlived reservationTTL.
func (f *inflight) count(key string, since time.Time) int64 {
@@ -45,6 +68,11 @@ func (f *inflight) count(key string, since time.Time) int64 {
f.mu.Lock()
defer f.mu.Unlock()
return f.countLocked(key, since)
}
// countLocked is count's body; the caller holds mu.
func (f *inflight) countLocked(key string, since time.Time) int64 {
set := f.m[key]
cutoff := time.Now().Add(-reservationTTL)
var n int64
@@ -63,14 +91,8 @@ func (f *inflight) count(key string, since time.Time) int64 {
return n
}
// reserve claims a slot for key until the message is recorded or released.
func (f *inflight) reserve(key string) *reservation {
if f == nil {
return nil
}
f.mu.Lock()
defer f.mu.Unlock()
// reserveLocked is reserve's body; the caller holds mu.
func (f *inflight) reserveLocked(key string) *reservation {
if f.m == nil {
f.m = make(map[string]map[*reservation]struct{})
}
+194 -24
View File
@@ -3,6 +3,7 @@ package milter
import (
"errors"
"net"
"sync"
"testing"
"time"
@@ -15,7 +16,10 @@ import (
// swallows recorder errors and still accepts the message. By default it reports
// no configured rate limit, so the level-2 check is inert unless a test sets
// limits (see fakeRecorder fields).
// mu guards the recorded slices so several sessions may drive one recorder
// concurrently, as they do in the real server.
type fakeRecorder struct {
mu sync.Mutex
entries []store.SendLogEntry
rejected []store.SendLogEntry
fail bool
@@ -27,17 +31,25 @@ type fakeRecorder struct {
counts map[string]int64
lookupErr error
countErr error
// onCount, if set, runs inside CountMessages. It lets a test hold every
// racing session at the store lookup until they can all proceed together.
onCount func()
}
func (f *fakeRecorder) InsertQueued(e store.SendLogEntry) error {
if f.fail {
return errors.New("boom")
}
f.mu.Lock()
defer f.mu.Unlock()
f.entries = append(f.entries, e)
return nil
}
func (f *fakeRecorder) InsertRejected(e store.SendLogEntry) error {
f.mu.Lock()
defer f.mu.Unlock()
f.rejected = append(f.rejected, e)
return nil
}
@@ -47,6 +59,9 @@ func (f *fakeRecorder) RateLimit(scope, ref string) (store.RateLimit, bool, erro
return store.RateLimit{}, false, f.lookupErr
}
rl, ok := f.limits[scope+"|"+ref]
if ok {
rl.Scope = scope
}
return rl, ok, nil
}
@@ -54,6 +69,9 @@ func (f *fakeRecorder) CountMessages(scope, ref string, _ time.Time) (int64, err
if f.countErr != nil {
return 0, f.countErr
}
if f.onCount != nil {
f.onCount()
}
return f.counts[scope+"|"+ref], nil
}
@@ -186,11 +204,14 @@ func TestBracedMacros(t *testing.T) {
}
}
// limitAt is the client IP the rate-limit tests connect from; the limits below
// register it so the differentiated check applies.
// limitIP is the client IP rate-limit tests connect from.
const limitIP = "203.0.113.7"
func activeLimit(ips ...string) store.RateLimit {
func domainLimit() store.RateLimit {
return store.RateLimit{MaxMessages: 5, WindowSeconds: 3600}
}
func appLimit(ips ...string) store.RateLimit {
return store.RateLimit{AllowedIPs: ips, MaxMessages: 5, WindowSeconds: 3600}
}
@@ -212,7 +233,7 @@ func mailFrom(t *testing.T, rec Store, ip, from, login string) milter.Response {
func TestRateLimitRefusesWhenDomainOverLimit(t *testing.T) {
rec := &fakeRecorder{
limits: map[string]store.RateLimit{
store.RateLimitScopeDomain + "|example.com": activeLimit(limitIP),
store.RateLimitScopeDomain + "|example.com": domainLimit(),
},
counts: map[string]int64{store.RateLimitScopeDomain + "|example.com": 5}, // == max
}
@@ -227,7 +248,7 @@ func TestRateLimitRefusesWhenDomainOverLimit(t *testing.T) {
func TestRateLimitRefusesWhenAppOverLimit(t *testing.T) {
rec := &fakeRecorder{
limits: map[string]store.RateLimit{
store.RateLimitScopeApp + "|app1": activeLimit(limitIP),
store.RateLimitScopeApp + "|app1": appLimit(limitIP),
},
counts: map[string]int64{store.RateLimitScopeApp + "|app1": 9}, // over max
}
@@ -239,7 +260,7 @@ func TestRateLimitRefusesWhenAppOverLimit(t *testing.T) {
func TestRateLimitAllowsUnderLimit(t *testing.T) {
rec := &fakeRecorder{
limits: map[string]store.RateLimit{
store.RateLimitScopeDomain + "|example.com": activeLimit(limitIP),
store.RateLimitScopeDomain + "|example.com": domainLimit(),
},
counts: map[string]int64{store.RateLimitScopeDomain + "|example.com": 4}, // < max
}
@@ -251,25 +272,76 @@ func TestRateLimitAllowsUnderLimit(t *testing.T) {
}
}
func TestRateLimitIgnoresUnregisteredIP(t *testing.T) {
func TestRateLimitDomainAppliesToAnyIP(t *testing.T) {
rec := &fakeRecorder{
limits: map[string]store.RateLimit{
store.RateLimitScopeDomain + "|example.com": activeLimit("198.51.100.1"), // not limitIP
store.RateLimitScopeDomain + "|example.com": domainLimit(),
},
counts: map[string]int64{store.RateLimitScopeDomain + "|example.com": 999},
}
// The sender's IP is not in the domain's registered set, so level-2 does not
// apply even though the count is huge (level-1 anvil would still cover it).
// Domain ceilings apply to every client IP; leftover AllowedIPs on the row
// are ignored.
if resp := mailFrom(t, rec, limitIP, "a@example.com", "app1"); resp != milter.RespTempFail {
t.Fatalf("domain over limit from any IP = %v, want TempFail", resp)
}
}
func TestRateLimitTrustedAppSkipsDomain(t *testing.T) {
rec := &fakeRecorder{
limits: map[string]store.RateLimit{
store.RateLimitScopeDomain + "|example.com": {MaxMessages: 1, WindowSeconds: 3600},
store.RateLimitScopeApp + "|app1": {
AllowedIPs: []string{limitIP}, MaxMessages: 10, WindowSeconds: 3600,
},
},
counts: map[string]int64{
store.RateLimitScopeDomain + "|example.com": 5, // over domain
store.RateLimitScopeApp + "|app1": 2, // under app
},
}
if resp := mailFrom(t, rec, limitIP, "a@example.com", "app1"); resp != milter.RespContinue {
t.Fatalf("unregistered IP = %v, want Continue (level-2 n/a)", resp)
t.Fatalf("trusted app under its ceiling = %v, want Continue (domain skipped)", resp)
}
}
func TestRateLimitUnlistedIPHitsDomain(t *testing.T) {
rec := &fakeRecorder{
limits: map[string]store.RateLimit{
store.RateLimitScopeDomain + "|example.com": {MaxMessages: 1, WindowSeconds: 3600},
store.RateLimitScopeApp + "|app1": {
AllowedIPs: []string{"198.51.100.1"}, MaxMessages: 100, WindowSeconds: 3600,
},
},
counts: map[string]int64{
store.RateLimitScopeDomain + "|example.com": 1,
store.RateLimitScopeApp + "|app1": 0,
},
}
if resp := mailFrom(t, rec, limitIP, "a@example.com", "app1"); resp != milter.RespTempFail {
t.Fatalf("unlisted IP under domain = %v, want TempFail", resp)
}
}
func TestRateLimitAppWithoutIPsDoesNotPrivilege(t *testing.T) {
rec := &fakeRecorder{
limits: map[string]store.RateLimit{
store.RateLimitScopeDomain + "|example.com": {MaxMessages: 1, WindowSeconds: 3600},
store.RateLimitScopeApp + "|app1": {MaxMessages: 100, WindowSeconds: 3600}, // no IPs
},
counts: map[string]int64{
store.RateLimitScopeDomain + "|example.com": 1,
store.RateLimitScopeApp + "|app1": 0,
},
}
if resp := mailFrom(t, rec, limitIP, "a@example.com", "app1"); resp != milter.RespTempFail {
t.Fatalf("app without IPs must not skip domain = %v, want TempFail", resp)
}
}
func TestRateLimitInactiveWithoutCeiling(t *testing.T) {
rec := &fakeRecorder{
// IP registered but no ceiling/window: an inert draft, must not enforce.
limits: map[string]store.RateLimit{
store.RateLimitScopeDomain + "|example.com": {AllowedIPs: []string{limitIP}},
store.RateLimitScopeDomain + "|example.com": {AllowedIPs: []string{limitIP}}, // no max/window
},
counts: map[string]int64{store.RateLimitScopeDomain + "|example.com": 999},
}
@@ -288,7 +360,7 @@ func TestRateLimitFailsOpenOnLookupError(t *testing.T) {
func TestRateLimitFailsOpenOnCountError(t *testing.T) {
rec := &fakeRecorder{
limits: map[string]store.RateLimit{
store.RateLimitScopeDomain + "|example.com": activeLimit(limitIP),
store.RateLimitScopeDomain + "|example.com": domainLimit(),
},
countErr: errors.New("db down"),
}
@@ -300,7 +372,7 @@ func TestRateLimitFailsOpenOnCountError(t *testing.T) {
func TestRateLimitNoIPKeyDoesNotApply(t *testing.T) {
rec := &fakeRecorder{
limits: map[string]store.RateLimit{
store.RateLimitScopeDomain + "|example.com": activeLimit(limitIP),
store.RateLimitScopeDomain + "|example.com": domainLimit(),
},
counts: map[string]int64{store.RateLimitScopeDomain + "|example.com": 999},
}
@@ -333,7 +405,7 @@ func mailFromIn(t *testing.T, rec Store, fl *inflight, ip, from, login string) (
func limitedRecorder(count int64) *fakeRecorder {
return &fakeRecorder{
limits: map[string]store.RateLimit{
store.RateLimitScopeDomain + "|example.com": activeLimit(limitIP),
store.RateLimitScopeDomain + "|example.com": domainLimit(),
},
counts: map[string]int64{store.RateLimitScopeDomain + "|example.com": count},
}
@@ -358,6 +430,58 @@ func TestRateLimitCountsInFlightMessages(t *testing.T) {
}
}
// The case above is sequential: the second session reads the stored count after
// the first has already reserved. Here every session reads it first — the gate
// holds them all inside the lookup — which is what concurrent SMTP connections
// actually do. However many then race for the single free slot, exactly one may
// pass. (TestTryAdmitHandsOutEachSlotOnce is the test that fails when counting
// and reserving are not one step; this one pins the session-level behaviour.)
func TestRateLimitAdmitsOnlyOneRacingSession(t *testing.T) {
rec := limitedRecorder(4) // one below the ceiling of 5
fl := &inflight{}
gate := make(chan struct{})
// Every session is held inside the stored-count lookup until all of them
// have read it, which is the state the race needs: none of them can see
// another's reservation, because none has been taken yet.
const racers = 32
var atCount, done sync.WaitGroup
atCount.Add(racers)
go func() { atCount.Wait(); close(gate) }()
rec.onCount = func() { atCount.Done(); <-gate }
responses := make([]milter.Response, racers)
for i := range racers {
done.Add(1)
go func() {
defer done.Done()
// Connect is skipped so every goroutine starts from the same point;
// the client IP is what Connect would have captured.
s := &session{rec: rec, flight: fl, clientIP: limitIP}
resp, err := s.MailFrom("a@example.com", mods(map[string]string{"auth_authen": "app1"}))
if err != nil {
resp = nil // reported as a missing Continue below
}
responses[i] = resp
}()
}
done.Wait()
admitted := 0
for _, resp := range responses {
if resp == milter.RespContinue {
admitted++
}
}
if admitted != 1 {
t.Fatalf("%d of %d racing sessions admitted, want exactly 1 — the last slot was handed out twice",
admitted, racers)
}
if n := fl.count(store.RateLimitScopeDomain+"|example.com", time.Now().Add(-time.Hour)); n != 1 {
t.Fatalf("in-flight reservations = %d, want 1", n)
}
}
// Once the message is recorded the stored count sees it, so its reservation
// must be given back — otherwise it would be counted twice and the ceiling
// would drift closed.
@@ -395,17 +519,17 @@ func TestReservationReleasedOnAbort(t *testing.T) {
}
}
// A refused message must not leave the slots it claimed for the limits checked
// before the one that tripped, or every refusal would tighten the ceiling.
func TestRefusalReleasesEarlierReservation(t *testing.T) {
// A trusted app at its ceiling refuses without touching the domain counter;
// no domain reservation should linger after the refusal.
func TestRefusalDoesNotLeaveDomainReservation(t *testing.T) {
rec := &fakeRecorder{
limits: map[string]store.RateLimit{
store.RateLimitScopeDomain + "|example.com": activeLimit(limitIP),
store.RateLimitScopeApp + "|app1": activeLimit(limitIP),
store.RateLimitScopeDomain + "|example.com": domainLimit(),
store.RateLimitScopeApp + "|app1": appLimit(limitIP),
},
counts: map[string]int64{
store.RateLimitScopeDomain + "|example.com": 0, // domain: plenty of room
store.RateLimitScopeApp + "|app1": 5, // app: at the ceiling
store.RateLimitScopeDomain + "|example.com": 0,
store.RateLimitScopeApp + "|app1": 5, // app at ceiling
},
}
fl := &inflight{}
@@ -415,6 +539,49 @@ func TestRefusalReleasesEarlierReservation(t *testing.T) {
if n := fl.count(store.RateLimitScopeDomain+"|example.com", time.Now().Add(-time.Hour)); n != 0 {
t.Fatalf("domain reservation left behind after refusal: %d", n)
}
if n := fl.count(store.RateLimitScopeApp+"|app1", time.Now().Add(-time.Hour)); n != 0 {
t.Fatalf("app reservation left behind after refusal: %d", n)
}
}
// The ceiling is handed out exactly max times however the sessions interleave.
// Counting and reserving in two critical sections passes the sequential tests
// above and still overshoots here, because between one session's count and its
// reservation any number of others can pass the same check.
func TestTryAdmitHandsOutEachSlotOnce(t *testing.T) {
const (
max = 500
workers = 8
)
fl := &inflight{}
since := time.Now().Add(-time.Hour)
start := make(chan struct{})
admitted := make([]int, workers)
var wg sync.WaitGroup
for i := range workers {
wg.Add(1)
go func() {
defer wg.Done()
<-start
for {
_, _, ok := fl.tryAdmit("domain|example.com", since, 0, max)
if !ok {
return
}
admitted[i]++
}
}()
}
close(start)
wg.Wait()
total := 0
for _, n := range admitted {
total += n
}
if total != max {
t.Fatalf("admitted %d messages under a ceiling of %d", total, max)
}
}
// The in-flight count only covers the limit's own window: a reservation older
@@ -422,7 +589,10 @@ func TestRefusalReleasesEarlierReservation(t *testing.T) {
// counted against a window it no longer belongs to.
func TestInflightIgnoresReservationsOutsideWindow(t *testing.T) {
fl := &inflight{}
r := fl.reserve("domain|example.com")
r, _, ok := fl.tryAdmit("domain|example.com", time.Now().Add(-time.Hour), 0, 1)
if !ok {
t.Fatal("tryAdmit refused the first message under a ceiling of 1")
}
r.at = time.Now().Add(-time.Minute)
if n := fl.count("domain|example.com", time.Now().Add(-time.Hour)); n != 1 {
+39 -34
View File
@@ -8,9 +8,11 @@ import (
)
// overLimit reports whether the message currently being received should be
// refused under a level-2 differentiated limit (guide § Rate limiting). It
// checks the domain-level and application-level limits in turn; either being
// exceeded is enough to refuse.
// refused under a level-2 differentiated limit (guide § Rate limiting).
//
// Trusted application IPs (app limit active and client IP listed) use only the
// app ceiling and skip the domain check. Everyone else is under the domain
// ceiling when one is configured; otherwise only level 1 applies.
//
// It is deliberately fail-open: any store error, or the absence of a usable
// limit, is treated as "not over limit" so a malfunction of the level-2
@@ -26,46 +28,49 @@ func (s *session) overLimit() bool {
if s.clientIP == "" {
return false // no client IP to key on; level-2 does not apply
}
checks := []struct{ scope, ref string }{
{store.RateLimitScopeDomain, domainOf(s.from)},
{store.RateLimitScopeApp, s.login},
}
var taken []*reservation
for _, c := range checks {
if c.ref == "" {
continue
}
rl, ok, err := s.rec.RateLimit(c.scope, c.ref)
if s.login != "" {
rl, ok, err := s.rec.RateLimit(store.RateLimitScopeApp, s.login)
if err != nil {
log.Printf("journal-milter: rate-limit lookup %s %q: %v (fail-open)", c.scope, c.ref, err)
continue
log.Printf("journal-milter: rate-limit lookup application %q: %v (fail-open)", s.login, err)
} else if ok && rl.Active() && rl.AllowsIP(s.clientIP) {
return s.enforceLimit(store.RateLimitScopeApp, s.login, rl)
}
// No limit configured, an inert draft, or a client IP outside the
// registered set: the differentiated limit does not apply here.
if !ok || !rl.Active() || !rl.AllowsIP(s.clientIP) {
continue
}
domain := domainOf(s.from)
if domain == "" {
return false
}
rl, ok, err := s.rec.RateLimit(store.RateLimitScopeDomain, domain)
if err != nil {
log.Printf("journal-milter: rate-limit lookup domain %q: %v (fail-open)", domain, err)
return false
}
if !ok || !rl.Active() {
return false
}
return s.enforceLimit(store.RateLimitScopeDomain, domain, rl)
}
// enforceLimit counts recent messages for scope/ref and refuses when at or
// above the ceiling. The stored count and the in-flight slots are weighed and
// the admitted message's own slot is taken in one atomic step (tryAdmit), so
// two sessions racing at MAIL FROM cannot both claim the last free slot.
func (s *session) enforceLimit(scope, ref string, rl store.RateLimit) bool {
since := time.Now().Add(-time.Duration(rl.WindowSeconds) * time.Second)
n, err := s.rec.CountMessages(c.scope, c.ref, since)
stored, err := s.rec.CountMessages(scope, ref, since)
if err != nil {
log.Printf("journal-milter: rate-limit count %s %q: %v (fail-open)", c.scope, c.ref, err)
continue
log.Printf("journal-milter: rate-limit count %s %q: %v (fail-open)", scope, ref, err)
return false
}
key := c.scope + "|" + c.ref
n += s.flight.count(key, since)
if n >= int64(rl.MaxMessages) {
r, n, ok := s.flight.tryAdmit(scope+"|"+ref, since, stored, int64(rl.MaxMessages))
if !ok {
log.Printf("journal-milter: %s %q over limit: %d/%d in %ds from %s — refusing 4xx",
c.scope, c.ref, n, rl.MaxMessages, rl.WindowSeconds, s.clientIP)
// The message is refused, so the slots claimed for the limits
// checked before this one must not stay claimed.
for _, r := range taken {
s.flight.release(r)
}
scope, ref, n, rl.MaxMessages, rl.WindowSeconds, s.clientIP)
return true
}
taken = append(taken, s.flight.reserve(key))
}
s.reserved = append(s.reserved, taken...)
s.reserved = append(s.reserved, r)
return false
}
+8
View File
@@ -75,6 +75,14 @@ func (p *Postfix) Reload() error {
return p.reload()
}
// SetReloadHook replaces how configuration is applied after a rebuild. Tests
// that cannot reach supervisord use this to verify file regeneration alone.
func (p *Postfix) SetReloadHook(fn func() error) {
if fn != nil {
p.reload = fn
}
}
// renderSenderLoginMaps builds the sender_login_maps file contents. Keys are
// sorted for deterministic output and the logins under each key are sorted and
// de-duplicated. Every address and login is re-checked for injection safety
+26
View File
@@ -75,6 +75,32 @@ func (s *Store) ListDomains() ([]Domain, error) {
return out, rows.Err()
}
// ListDomainsForUser returns domains assigned to userID with application counts,
// ordered by name.
func (s *Store) ListDomainsForUser(userID int64) ([]Domain, error) {
rows, err := s.db.Query(`
SELECT d.id, d.name, d.dkim_selector, d.dmarc_rua, d.created_at,
(SELECT COUNT(*) FROM applications a WHERE a.domain_id = d.id)
FROM domains d
INNER JOIN user_domains ud ON ud.domain_id = d.id
WHERE ud.user_id = ?
ORDER BY d.name`, userID)
if err != nil {
return nil, fmt.Errorf("list domains for user: %w", err)
}
defer rows.Close()
var out []Domain
for rows.Next() {
d, err := scanDomain(rows)
if err != nil {
return nil, err
}
out = append(out, d)
}
return out, rows.Err()
}
// GetDomain returns a single domain (with its application count) by id, or
// ErrDomainNotFound.
func (s *Store) GetDomain(id int64) (Domain, error) {
+24 -20
View File
@@ -18,36 +18,40 @@ const (
)
// RateLimit is a differentiated level-2 rate limit (guide § Rate limiting):
// an optional set of expected client IPs plus a message ceiling over a sliding
// window, attached to a domain or an application. It is enforced in the
// journal-milter; level 1 (Postfix anvil, architecture.md § Mail path) is the
// IP backstop that always applies even when this is absent or the milter is
// down.
// a message ceiling over a sliding window, attached to a domain or an
// application. It is enforced in the journal-milter; level 1 (Postfix anvil,
// architecture.md § Mail path) is the IP backstop that always applies even
// when this is absent or the milter is down.
//
// Both the IP binding and the ceiling are optional in the schema, but a limit
// is only enforced when it is Active(): the design deliberately allows an
// admin to leave the IP binding empty for apps that send from changing IPs, in
// which case only level 1 protects them (guide § Rate limiting).
// Domain limits apply to every client IP once max and window are set. Application
// limits additionally require AllowedIPs: those trusted addresses get the app
// ceiling (above the domain) and skip the domain check; other IPs stay under
// the domain limit or level 1 alone (guide § Rate limiting).
type RateLimit struct {
Scope string
RefID int64
AllowedIPs []string // canonical client IPs this limit applies to
AllowedIPs []string // trusted client IPs for an application override
MaxMessages int
WindowSeconds int
}
// Active reports whether the limit is fully configured and should be enforced.
// A missing IP binding, ceiling or window leaves the differentiated limit
// inert (guide § Rate limiting): the IP binding is what scopes the limit to a
// known sender.
// Domain: max and window only. Application: also needs at least one trusted IP
// (the privilege that raises the ceiling above the domain).
func (r RateLimit) Active() bool {
return len(r.AllowedIPs) > 0 && r.MaxMessages > 0 && r.WindowSeconds > 0
if r.MaxMessages <= 0 || r.WindowSeconds <= 0 {
return false
}
if r.Scope == RateLimitScopeApp {
return len(r.AllowedIPs) > 0
}
// Domain (and any unset/legacy scope treated as domain-style): no IP list.
return true
}
// AllowsIP reports whether ip is one of the limit's registered client IPs. The
// comparison parses both sides so equivalent textual forms of the same address
// match; a client IP outside the list means the differentiated limit does not
// apply to it (level 1 still does).
// AllowsIP reports whether ip is one of the application's trusted client IPs.
// Used only for application overrides; domain limits do not consult this list.
// Equivalent textual forms of the same address match.
func (r RateLimit) AllowsIP(ip string) bool {
c := net.ParseIP(ip)
if c == nil {
@@ -183,8 +187,8 @@ func (s *Store) CountMessages(scope, ref string, since time.Time) (int64, error)
}
// scanRateLimit reads the three stored columns, tolerating NULL numeric columns
// (an IP-only draft) by leaving the corresponding field zero, which makes the
// limit inert via Active().
// by leaving the corresponding field zero, which makes the limit inert via
// Active() until max and window are both set.
func scanRateLimit(r scanRow) (RateLimit, error) {
var (
ips sql.NullString
+17 -9
View File
@@ -152,27 +152,35 @@ func TestDeleteRateLimitsForDomain(t *testing.T) {
func TestRateLimitActiveAndAllowsIP(t *testing.T) {
inactive := []RateLimit{
{},
{AllowedIPs: []string{"203.0.113.1"}}, // no ceiling
{AllowedIPs: []string{"203.0.113.1"}, MaxMessages: 5}, // no window
{MaxMessages: 5, WindowSeconds: 60}, // no IPs
{Scope: RateLimitScopeDomain, AllowedIPs: []string{"203.0.113.1"}}, // no ceiling
{Scope: RateLimitScopeDomain, MaxMessages: 5}, // no window
{Scope: RateLimitScopeApp, MaxMessages: 5, WindowSeconds: 60}, // app needs IPs
{Scope: RateLimitScopeApp, AllowedIPs: []string{"203.0.113.1"}, MaxMessages: 5}, // no window
}
for i, rl := range inactive {
if rl.Active() {
t.Fatalf("case %d: %+v should be inactive", i, rl)
}
}
active := RateLimit{AllowedIPs: []string{"203.0.113.1", "2001:db8::1"}, MaxMessages: 5, WindowSeconds: 60}
if !active.Active() {
t.Fatalf("should be active: %+v", active)
domainActive := RateLimit{Scope: RateLimitScopeDomain, MaxMessages: 5, WindowSeconds: 60}
if !domainActive.Active() {
t.Fatalf("domain without IPs should be active: %+v", domainActive)
}
if !active.AllowsIP("203.0.113.1") || !active.AllowsIP("2001:db8::1") {
appActive := RateLimit{
Scope: RateLimitScopeApp, AllowedIPs: []string{"203.0.113.1", "2001:db8::1"},
MaxMessages: 5, WindowSeconds: 60,
}
if !appActive.Active() {
t.Fatalf("should be active: %+v", appActive)
}
if !appActive.AllowsIP("203.0.113.1") || !appActive.AllowsIP("2001:db8::1") {
t.Fatalf("registered IPs should match")
}
// Equivalent textual form of the IPv6 address must still match.
if !active.AllowsIP("2001:0db8:0000:0000:0000:0000:0000:0001") {
if !appActive.AllowsIP("2001:0db8:0000:0000:0000:0000:0000:0001") {
t.Fatalf("expanded IPv6 form should match")
}
if active.AllowsIP("198.51.100.7") || active.AllowsIP("not-an-ip") || active.AllowsIP("") {
if appActive.AllowsIP("198.51.100.7") || appActive.AllowsIP("not-an-ip") || appActive.AllowsIP("") {
t.Fatalf("unregistered/invalid IPs must not match")
}
}
+30 -2
View File
@@ -178,15 +178,29 @@ func (s *Store) GetSendLog(id int64) (SendLogRow, error) {
return row, nil
}
// SendLogFilter narrows QuerySendLog/CountSendLog by domain and/or
// application login. An empty field matches everything.
// SendLogFilter narrows QuerySendLog/CountSendLog. It carries two kinds of
// narrowing, and they behave in opposite ways on purpose.
//
// Domain and AppLogin are the operator's own filters, chosen in the UI: an
// empty field matches everything.
//
// Domains and AllDomains are the authorization scope, which no query parameter
// may widen. Domains is the exhaustive set of domain names the caller is
// entitled to read, applied as an IN constraint; AllDomains lifts that
// restriction and is the only way to read the whole journal. A zero-valued
// filter therefore matches *no* rows: a caller that forgets to state a scope
// gets an empty log rather than every tenant's mail, which is the failure mode
// this struct exists to make impossible.
type SendLogFilter struct {
Domain string
AppLogin string
Domains []string
AllDomains bool
}
// QuerySendLog returns send-log rows matching filter, newest first, for the
// monitoring screen's server-side pagination (product.md's send-log view).
// The filter's authorization scope is mandatory: see SendLogFilter.
func (s *Store) QuerySendLog(filter SendLogFilter, limit, offset int) ([]SendLogRow, error) {
where, args := sendLogWhere(filter)
args = append(args, limit, offset)
@@ -233,6 +247,20 @@ func (s *Store) CountSendLog(filter SendLogFilter) (int64, error) {
func sendLogWhere(f SendLogFilter) (string, []any) {
var clauses []string
var args []any
if !f.AllDomains {
// No scope is not "no restriction": a domain administrator whose last
// assignment was deleted owns nothing in the journal and must see
// nothing, and the same clause catches a caller that never set a scope.
if len(f.Domains) == 0 {
return " WHERE 1 = 0", nil
}
marks := make([]string, len(f.Domains))
for i, name := range f.Domains {
marks[i] = "?"
args = append(args, name)
}
clauses = append(clauses, "domain IN ("+strings.Join(marks, ", ")+")")
}
if f.Domain != "" {
clauses = append(clauses, "domain = ?")
args = append(args, f.Domain)
+57
View File
@@ -1,6 +1,8 @@
package store
import (
"sort"
"strings"
"testing"
"time"
)
@@ -162,6 +164,61 @@ func TestUpdateStatusNoMatch(t *testing.T) {
}
}
// The journal is read by principals who are only entitled to part of it, so
// the scope is part of the query rather than something the caller remembers to
// apply afterwards. A filter that states no scope is a caller that has not
// decided who is asking, and the safe answer to that is nothing.
func TestSendLogScopeIsMandatory(t *testing.T) {
st := openTestStore(t)
for _, domain := range []string{"first.example.ru", "second.example.ru"} {
if err := st.InsertQueued(SendLogEntry{
QueueID: "Q-" + domain, Domain: domain, AppLogin: "app-" + domain,
From: "noreply@" + domain, To: "public@example.net", Subject: domain,
}); err != nil {
t.Fatalf("InsertQueued: %v", err)
}
}
for name, tc := range map[string]struct {
filter SendLogFilter
want []string
}{
"no scope": {SendLogFilter{}, nil},
"empty scope": {SendLogFilter{Domains: []string{}}, nil},
"all domains": {SendLogFilter{AllDomains: true}, []string{"first.example.ru", "second.example.ru"}},
"one domain": {SendLogFilter{Domains: []string{"first.example.ru"}}, []string{"first.example.ru"}},
"two domains": {SendLogFilter{Domains: []string{"first.example.ru", "second.example.ru"}}, []string{"first.example.ru", "second.example.ru"}},
"unknown domain": {SendLogFilter{Domains: []string{"third.example.ru"}}, nil},
"filter within": {SendLogFilter{Domain: "first.example.ru", Domains: []string{"first.example.ru", "second.example.ru"}}, []string{"first.example.ru"}},
"filter outside": {SendLogFilter{Domain: "second.example.ru", Domains: []string{"first.example.ru"}}, nil},
"app filter outside": {SendLogFilter{AppLogin: "app-second.example.ru", Domains: []string{"first.example.ru"}}, nil},
} {
rows, err := st.QuerySendLog(tc.filter, 50, 0)
if err != nil {
t.Fatalf("%s: QuerySendLog: %v", name, err)
}
var got []string
for _, r := range rows {
got = append(got, r.Domain)
}
// Which rows came back is the question here; the page's own order is
// newest-first and is tested where it matters.
sort.Strings(got)
if strings.Join(got, ",") != strings.Join(tc.want, ",") {
t.Errorf("%s: rows for %v, want %v", name, got, tc.want)
}
// The count drives pagination, so it has to agree with the page or the
// UI advertises pages of rows the reader is not allowed to see.
n, err := st.CountSendLog(tc.filter)
if err != nil {
t.Fatalf("%s: CountSendLog: %v", name, err)
}
if int(n) != len(tc.want) {
t.Errorf("%s: count %d, want %d", name, n, len(tc.want))
}
}
}
func TestDeleteSendLogBefore(t *testing.T) {
st := openTestStore(t)
+4 -3
View File
@@ -1,7 +1,8 @@
// Package store owns the SelfPost SQLite database: the single file under /data
// that persists the administrator account, sending domains and applications,
// the send log and rate-limit settings (architecture.md § Persistence). It
// exposes typed queries so the rest of the panel never builds SQL by hand.
// that persists panel users (global administrators and domain-admins), sending
// domains and applications, the send log and rate-limit settings
// (architecture.md § Persistence). It exposes typed queries so the rest of the
// panel never builds SQL by hand.
package store
import (
+2
View File
@@ -51,6 +51,8 @@ func New(st *store.Store, cfg Config, v *view.Engine, setupTokenPath string) *Mo
loginLimiter: newRateLimiter(10, 15*time.Minute),
trustedProxies: cfg.TrustedProxyCIDRs,
}
m.setupLimiter.startSweeper()
m.loginLimiter.startSweeper()
m.setup = newSetupManager(st, cfg.Hostname, setupTokenPath)
return m
}
+74 -7
View File
@@ -1,13 +1,17 @@
package auth
import (
"database/sql"
"net/http"
"net/http/httptest"
"net/url"
"path/filepath"
"strings"
"testing"
"time"
"golang.org/x/crypto/bcrypt"
"github.com/mixeme/selfpost/internal/store"
"github.com/mixeme/selfpost/internal/web/view"
)
@@ -22,6 +26,15 @@ func newTestSessionStore(t *testing.T) *sessionStore {
return newSessionStore(st, 7*24*time.Hour)
}
func mustCreate(t *testing.T, s *sessionStore, username string) string {
t.Helper()
token, err := s.Create(username)
if err != nil {
t.Fatalf("create session: %v", err)
}
return token
}
func mustView(t *testing.T) *view.Engine {
t.Helper()
v, err := view.New("test")
@@ -86,7 +99,7 @@ func TestSessionTokenIgnoresTheOtherName(t *testing.T) {
func TestRequireAuthRejectsDuplicateCookies(t *testing.T) {
m := testModule(t, false)
token := m.sessions.Create("admin")
token := mustCreate(t, m.sessions, "admin")
reached := false
h := m.RequireAuth(http.HandlerFunc(func(http.ResponseWriter, *http.Request) { reached = true }))
@@ -107,7 +120,7 @@ func TestRequireAuthRejectsDuplicateCookies(t *testing.T) {
func TestLogoutClearsBothCookieNames(t *testing.T) {
m := testModule(t, true)
token := m.sessions.Create("admin")
token := mustCreate(t, m.sessions, "admin")
r := httptest.NewRequest(http.MethodPost, "http://panel.example.com/logout", nil)
r.Host = "panel.example.com"
@@ -132,9 +145,63 @@ func TestLogoutClearsBothCookieNames(t *testing.T) {
}
}
// A session that could not be stored must not turn into a cookie: the browser
// would look signed in, and every request it made would be bounced to /login
// with no explanation. Only the sessions table is broken here, so the request
// gets past the user lookup and password check and fails exactly where the
// session is written.
func TestLoginSetsNoCookieWhenTheSessionCannotBeStored(t *testing.T) {
path := filepath.Join(t.TempDir(), "test.db")
st, err := store.Open(path)
if err != nil {
t.Fatalf("open store: %v", err)
}
t.Cleanup(func() { st.Close() })
hash, err := bcrypt.GenerateFromPassword([]byte("correct-horse-battery"), bcrypt.MinCost)
if err != nil {
t.Fatalf("hash password: %v", err)
}
if err := st.CreateGlobalUser("admin", string(hash)); err != nil {
t.Fatalf("create user: %v", err)
}
dropSessionsTable(t, path)
m := New(st, Config{}, mustView(t), "")
form := url.Values{"username": {"admin"}, "password": {"correct-horse-battery"}}
r := httptest.NewRequest(http.MethodPost, "http://panel.example.com/login",
strings.NewReader(form.Encode()))
r.Header.Set("Content-Type", "application/x-www-form-urlencoded")
rec := httptest.NewRecorder()
m.HandleLogin(rec, r)
if got := rec.Header().Values("Set-Cookie"); len(got) != 0 {
t.Errorf("a session cookie was issued for a session that was never stored: %v", got)
}
if rec.Code != http.StatusInternalServerError {
t.Errorf("status = %d, want 500 (the login failed)", rec.Code)
}
if loc := rec.Header().Get("Location"); loc != "" {
t.Errorf("the browser was sent to %q as if it were signed in", loc)
}
}
// dropSessionsTable breaks session persistence while leaving the rest of the
// schema usable. The SQLite driver is registered by internal/store.
func dropSessionsTable(t *testing.T, path string) {
t.Helper()
db, err := sql.Open("sqlite", "file:"+path+"?_pragma=busy_timeout(5000)")
if err != nil {
t.Fatalf("open database directly: %v", err)
}
defer db.Close()
if _, err := db.Exec("DROP TABLE sessions"); err != nil {
t.Fatalf("drop sessions table: %v", err)
}
}
func TestSessionRename(t *testing.T) {
s := newTestSessionStore(t)
token := s.Create("admin")
token := mustCreate(t, s, "admin")
s.Rename(token, "operator")
@@ -149,8 +216,8 @@ func TestSessionRename(t *testing.T) {
func TestSessionDestroyOthers(t *testing.T) {
s := newTestSessionStore(t)
keep := s.Create("admin")
other := s.Create("admin")
keep := mustCreate(t, s, "admin")
other := mustCreate(t, s, "admin")
s.DestroyOthers(keep)
@@ -165,7 +232,7 @@ func TestSessionDestroyOthers(t *testing.T) {
func TestSessionLookupRejectsExpired(t *testing.T) {
s := newTestSessionStore(t)
s.idle = -time.Minute
token := s.Create("admin")
token := mustCreate(t, s, "admin")
if _, ok := s.Lookup(token); ok {
t.Fatal("expired session was accepted")
@@ -174,7 +241,7 @@ func TestSessionLookupRejectsExpired(t *testing.T) {
func TestSessionTouchThrottled(t *testing.T) {
s := newTestSessionStore(t)
token := s.Create("admin")
token := mustCreate(t, s, "admin")
if s.Touch(token) {
t.Fatal("touch renewed a session created moments ago")
+6 -1
View File
@@ -120,7 +120,12 @@ func (m *Module) submitLogin(w http.ResponseWriter, r *http.Request) {
return
}
token := m.sessions.Create(user.Username)
token, err := m.sessions.Create(user.Username)
if err != nil {
logf("panel: login: create session failed: %v", err)
m.renderLogin(w, http.StatusInternalServerError, "Internal error. Please try again.")
return
}
m.setSessionCookie(w, token)
http.Redirect(w, r, "/", http.StatusSeeOther)
}
+284
View File
@@ -0,0 +1,284 @@
package auth
import (
"net/http"
"net/http/httptest"
"net/url"
"strings"
"testing"
"golang.org/x/crypto/bcrypt"
)
const testPassword = "correct-horse-battery"
// moduleWithAdmin returns a panel that has already been through setup, with one
// global administrator whose password is testPassword.
func moduleWithAdmin(t *testing.T) *Module {
t.Helper()
m := testModule(t, false)
hash, err := bcrypt.GenerateFromPassword([]byte(testPassword), bcrypt.MinCost)
if err != nil {
t.Fatalf("hash password: %v", err)
}
if err := m.store.CreateGlobalUser("admin", string(hash)); err != nil {
t.Fatalf("create user: %v", err)
}
return m
}
// postLogin submits the sign-in form from remoteAddr (the limiter's key) and
// returns what the handler wrote.
func postLogin(m *Module, remoteAddr, username, password string) *httptest.ResponseRecorder {
form := url.Values{"username": {username}, "password": {password}}
r := httptest.NewRequest(http.MethodPost, "http://panel.example.com/login",
strings.NewReader(form.Encode()))
r.Header.Set("Content-Type", "application/x-www-form-urlencoded")
r.RemoteAddr = remoteAddr
rec := httptest.NewRecorder()
m.HandleLogin(rec, r)
return rec
}
// sessionCookieValue returns the session token the response issued, or "" if it
// issued none.
func sessionCookieValue(t *testing.T, m *Module, rec *httptest.ResponseRecorder) string {
t.Helper()
for _, c := range rec.Result().Cookies() {
if c.Name == m.sessionCookie() {
return c.Value
}
}
return ""
}
func TestLoginSignsInWithTheRightPassword(t *testing.T) {
m := moduleWithAdmin(t)
rec := postLogin(m, "203.0.113.7:5000", "admin", testPassword)
if rec.Code != http.StatusSeeOther || rec.Header().Get("Location") != "/" {
t.Fatalf("status = %d, Location = %q; want a redirect to /", rec.Code, rec.Header().Get("Location"))
}
token := sessionCookieValue(t, m, rec)
if token == "" {
t.Fatal("no session cookie was issued")
}
name, ok := m.sessions.Lookup(token)
if !ok || name != "admin" {
t.Fatalf("the cookie's session resolves to %q, %t; want admin", name, ok)
}
}
// A refused sign-in must not say which half was wrong: the panel is public, and
// distinguishable answers would turn the form into a list of usernames.
func TestLoginRefusesBadCredentialsWithoutSayingWhy(t *testing.T) {
m := moduleWithAdmin(t)
bodies := make(map[string]string, 2)
for name, creds := range map[string][2]string{
"wrong password": {"admin", "not-the-password"},
"unknown user": {"nobody", testPassword},
} {
rec := postLogin(m, "203.0.113.7:5000", creds[0], creds[1])
if rec.Code != http.StatusUnauthorized {
t.Errorf("%s: status = %d, want 401", name, rec.Code)
}
if got := sessionCookieValue(t, m, rec); got != "" {
t.Errorf("%s: a session cookie was issued: %q", name, got)
}
bodies[name] = rec.Body.String()
}
if bodies["wrong password"] != bodies["unknown user"] {
t.Error("the two refusals differ, so the form tells an attacker which usernames exist")
}
}
// The lockout is what makes online guessing pointless, so it has to hold even
// for the request that finally carries the right password — and it has to be
// scoped to the address that spent the attempts.
func TestLoginLocksOutAfterTooManyAttempts(t *testing.T) {
m := moduleWithAdmin(t)
const attacker = "203.0.113.7:5000"
for i := 0; i < 10; i++ {
if rec := postLogin(m, attacker, "admin", "guess"); rec.Code != http.StatusUnauthorized {
t.Fatalf("attempt %d: status = %d, want 401 (still under the limit)", i+1, rec.Code)
}
}
rec := postLogin(m, attacker, "admin", testPassword)
if rec.Code != http.StatusTooManyRequests {
t.Errorf("status = %d, want 429; the lockout was bypassed by guessing right", rec.Code)
}
if got := sessionCookieValue(t, m, rec); got != "" {
t.Errorf("a locked-out request was signed in: %q", got)
}
if rec := postLogin(m, "198.51.100.9:5000", "admin", testPassword); rec.Code != http.StatusSeeOther {
t.Errorf("another address got %d; one guesser locked out the whole internet", rec.Code)
}
}
// Before the first administrator exists there is nothing to sign in as, so the
// form is replaced by a pointer to the setup link rather than a password box
// that can never succeed.
func TestLoginPointsAtSetupBeforeTheFirstAdministrator(t *testing.T) {
m := testModule(t, false)
rec := httptest.NewRecorder()
m.HandleLogin(rec, httptest.NewRequest(http.MethodGet, "http://panel.example.com/login", nil))
if rec.Code != http.StatusOK {
t.Fatalf("status = %d, want 200", rec.Code)
}
body := rec.Body.String()
if !strings.Contains(body, "No administrator has been created yet") {
t.Errorf("the login page does not point at the setup link:\n%s", body)
}
if strings.Contains(body, `name="password"`) {
t.Errorf("the login page offers a password field with no account to use it:\n%s", body)
}
}
// getSetup performs the GET the operator's browser makes when it follows the
// one-time link.
func getSetup(m *Module, token string) *httptest.ResponseRecorder {
rec := httptest.NewRecorder()
m.HandleSetup(rec, httptest.NewRequest(http.MethodGet, "http://panel.example.com/setup/"+token, nil))
return rec
}
func postSetup(m *Module, token string, form url.Values) *httptest.ResponseRecorder {
r := httptest.NewRequest(http.MethodPost, "http://panel.example.com/setup/"+token,
strings.NewReader(form.Encode()))
r.Header.Set("Content-Type", "application/x-www-form-urlencoded")
rec := httptest.NewRecorder()
m.HandleSetup(rec, r)
return rec
}
func setupForm(username, password, confirm string) url.Values {
return url.Values{
"username": {username},
"password": {password},
"password_confirm": {confirm},
}
}
// The setup link creates the first global administrator and then stops
// existing: the persistent fact is the user row, so the link is dead after a
// restart too, not only for the process that served it.
func TestSetupCreatesTheFirstAdministratorAndThenCloses(t *testing.T) {
m := testModule(t, false)
token, ok := m.setup.activeToken()
if !ok {
t.Fatal("no setup token on a panel with no users")
}
if rec := getSetup(m, token); rec.Code != http.StatusOK {
t.Fatalf("GET the setup link = %d, want the form", rec.Code)
}
rec := postSetup(m, token, setupForm("operator", "a-long-enough-password", "a-long-enough-password"))
if rec.Code != http.StatusSeeOther || rec.Header().Get("Location") != "/login" {
t.Fatalf("status = %d, Location = %q; want a redirect to /login", rec.Code, rec.Header().Get("Location"))
}
u, err := m.store.GetUserByUsername("operator")
if err != nil {
t.Fatalf("the administrator was not created: %v", err)
}
if u.Role != RoleGlobal {
t.Errorf("the first administrator has role %q, want global", u.Role)
}
if err := bcrypt.CompareHashAndPassword([]byte(u.PasswordHash), []byte("a-long-enough-password")); err != nil {
t.Errorf("the stored hash does not match the password that was set: %v", err)
}
if rec := getSetup(m, token); rec.Code != http.StatusNotFound {
t.Errorf("the setup link still answers %d after setup completed, want 404", rec.Code)
}
if rec := postSetup(m, token, setupForm("second", "a-long-enough-password", "a-long-enough-password")); rec.Code != http.StatusNotFound {
t.Errorf("a second administrator could be created through the setup link (%d)", rec.Code)
}
}
// A token that is wrong, or one that has aged out and been replaced, is not a
// hint that setup exists: both answer 404, the same as any unknown path.
func TestSetupRejectsAWrongOrExpiredToken(t *testing.T) {
m := testModule(t, false)
token, ok := m.setup.activeToken()
if !ok {
t.Fatal("no setup token on a panel with no users")
}
if rec := getSetup(m, token+"x"); rec.Code != http.StatusNotFound {
t.Errorf("a wrong token answered %d, want 404", rec.Code)
}
expireSetupToken(m)
if rec := getSetup(m, token); rec.Code != http.StatusNotFound {
t.Errorf("the expired token still opens setup (%d)", rec.Code)
}
fresh, _ := m.setup.activeToken()
if fresh == token {
t.Fatal("the expired token was not replaced")
}
if rec := getSetup(m, fresh); rec.Code != http.StatusOK {
t.Errorf("the reissued token does not open setup (%d)", rec.Code)
}
}
// The first account is the one that can never be locked out of the panel from
// outside, so the rules that apply to every other user apply here too — before
// anything is written.
func TestSetupRejectsCredentialsItWouldNotAcceptLater(t *testing.T) {
for name, form := range map[string]url.Values{
"username too short": setupForm("op", "a-long-enough-password", "a-long-enough-password"),
"username not ASCII": setupForm("оператор", "a-long-enough-password", "a-long-enough-password"),
"passwords differ": setupForm("operator", "a-long-enough-password", "a-long-enough-passwerd"),
"password too short": setupForm("operator", "short", "short"),
"no password at all": setupForm("operator", "", ""),
"no username at all": setupForm("", "a-long-enough-password", "a-long-enough-password"),
} {
m := testModule(t, false)
token, _ := m.setup.activeToken()
rec := postSetup(m, token, form)
if rec.Code != http.StatusBadRequest {
t.Errorf("%s: status = %d, want 400", name, rec.Code)
}
if exists, err := m.store.UserExists(); err != nil || exists {
t.Errorf("%s: an administrator was created anyway (err=%v)", name, err)
}
if rec := getSetup(m, token); rec.Code != http.StatusOK {
t.Errorf("%s: the setup link was burned by a rejected form (%d)", name, rec.Code)
}
}
}
// Setup is unauthenticated by definition, so the only thing between the token
// and an offline guesser is the limiter in front of it.
func TestSetupIsRateLimited(t *testing.T) {
m := testModule(t, false)
for i := 0; i < 10; i++ {
if rec := getSetup(m, "wrong-token"); rec.Code != http.StatusNotFound {
t.Fatalf("attempt %d: status = %d, want 404 (still under the limit)", i+1, rec.Code)
}
}
if rec := getSetup(m, "wrong-token"); rec.Code != http.StatusTooManyRequests {
t.Errorf("status = %d, want 429 after the eleventh attempt", rec.Code)
}
}
// expireSetupToken ages the current token out, the state the panel reaches when
// nobody follows the link within setupTokenTTL.
func expireSetupToken(m *Module) {
m.setup.mu.Lock()
defer m.setup.mu.Unlock()
m.setup.expiresAt = m.setup.expiresAt.Add(-2 * setupTokenTTL)
}
-12
View File
@@ -44,15 +44,3 @@ func CurrentUser(r *http.Request) string {
}
return ""
}
// RequireGlobal wraps a handler that only global administrators may reach.
func RequireGlobal(next http.Handler) http.Handler {
return http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
p, ok := CurrentPrincipal(r.Context())
if !ok || !p.IsGlobal() {
http.NotFound(w, r)
return
}
next.ServeHTTP(w, r)
})
}
+43
View File
@@ -5,11 +5,14 @@ import (
"time"
)
const defaultMaxBuckets = 4096
// rateLimiter is a simple fixed-window per-key counter used to throttle the
// setup and login routes (security.md). Keys are client IPs.
type rateLimiter struct {
max int
window time.Duration
maxBuckets int
mu sync.Mutex
buckets map[string]*rlBucket
@@ -24,10 +27,23 @@ func newRateLimiter(max int, window time.Duration) *rateLimiter {
return &rateLimiter{
max: max,
window: window,
maxBuckets: defaultMaxBuckets,
buckets: make(map[string]*rlBucket),
}
}
func (r *rateLimiter) startSweeper() {
go func() {
ticker := time.NewTicker(r.window)
defer ticker.Stop()
for range ticker.C {
r.mu.Lock()
r.sweep(time.Now())
r.mu.Unlock()
}
}()
}
func (r *rateLimiter) Allow(key string) bool {
now := time.Now()
r.mu.Lock()
@@ -35,6 +51,7 @@ func (r *rateLimiter) Allow(key string) bool {
b := r.buckets[key]
if b == nil || now.After(b.windowEnds) {
r.makeRoom(now)
r.buckets[key] = &rlBucket{count: 1, windowEnds: now.Add(r.window)}
r.sweep(now)
return true
@@ -46,6 +63,32 @@ func (r *rateLimiter) Allow(key string) bool {
return true
}
func (r *rateLimiter) makeRoom(now time.Time) {
if r.maxBuckets <= 0 || len(r.buckets) < r.maxBuckets {
return
}
r.sweep(now)
for len(r.buckets) >= r.maxBuckets {
r.evictOldest()
}
}
func (r *rateLimiter) evictOldest() {
var oldestKey string
var oldestEnds time.Time
first := true
for k, b := range r.buckets {
if first || b.windowEnds.Before(oldestEnds) {
oldestKey = k
oldestEnds = b.windowEnds
first = false
}
}
if oldestKey != "" {
delete(r.buckets, oldestKey)
}
}
func (r *rateLimiter) sweep(now time.Time) {
for k, b := range r.buckets {
if now.After(b.windowEnds) {
+125
View File
@@ -0,0 +1,125 @@
package auth
import (
"testing"
"time"
)
// The limiter is what stands between the public login form and an unlimited
// guessing rate (security.md), so the ceiling has to be exact: the configured
// number of attempts go through and the next one does not, however often it is
// repeated.
func TestRateLimiterStopsAtTheCeiling(t *testing.T) {
r := newRateLimiter(3, time.Minute)
for i := 1; i <= 3; i++ {
if !r.Allow("203.0.113.7") {
t.Fatalf("attempt %d of 3 was refused before the ceiling", i)
}
}
for i := 4; i <= 6; i++ {
if r.Allow("203.0.113.7") {
t.Fatalf("attempt %d passed after the ceiling of 3", i)
}
}
}
// Buckets are per key, so one locked-out address must not lock out the rest of
// the internet — a shared counter would turn a single guesser into a denial of
// service against every operator.
func TestRateLimiterKeepsKeysApart(t *testing.T) {
r := newRateLimiter(1, time.Minute)
if !r.Allow("203.0.113.7") || r.Allow("203.0.113.7") {
t.Fatal("the first key did not use up its single attempt")
}
if !r.Allow("198.51.100.9") {
t.Fatal("a second address was refused because another one was locked out")
}
}
// The window is fixed, not sliding: once it has elapsed the count starts again
// from zero rather than being carried over. Time is moved by ageing the bucket
// instead of sleeping, so the test states the boundary rather than approaching
// it.
func TestRateLimiterReopensAfterTheWindow(t *testing.T) {
r := newRateLimiter(2, time.Minute)
r.Allow("203.0.113.7")
r.Allow("203.0.113.7")
if r.Allow("203.0.113.7") {
t.Fatal("the ceiling was not reached")
}
expire(r, "203.0.113.7")
if !r.Allow("203.0.113.7") {
t.Fatal("the key is still locked out after its window ended")
}
if !r.Allow("203.0.113.7") {
t.Fatal("the new window did not start from an empty count")
}
if r.Allow("203.0.113.7") {
t.Fatal("the new window allowed more than the ceiling")
}
}
// Every address that ever tried to sign in gets a bucket, and the only thing
// that removes the finished ones is the sweep on a new window. It runs on the
// key that triggered it as well as on the others, so a long-running panel does
// not accumulate a bucket per source address for ever.
func TestRateLimiterSweepsFinishedBuckets(t *testing.T) {
r := newRateLimiter(2, time.Minute)
for _, key := range []string{"203.0.113.7", "198.51.100.9"} {
r.Allow(key)
expire(r, key)
}
r.Allow("192.0.2.5") // still inside its window
r.Allow("203.0.113.7") // new window for this key: sweeps the rest
r.mu.Lock()
defer r.mu.Unlock()
if _, ok := r.buckets["198.51.100.9"]; ok {
t.Error("a bucket whose window ended is still held")
}
for _, key := range []string{"203.0.113.7", "192.0.2.5"} {
if _, ok := r.buckets[key]; !ok {
t.Errorf("the sweep dropped %s, whose window is still open", key)
}
}
}
// expire moves a key's window into the past, the same state it would reach by
// waiting for the window to elapse.
func expire(r *rateLimiter, key string) {
r.mu.Lock()
defer r.mu.Unlock()
if b := r.buckets[key]; b != nil {
b.windowEnds = time.Now().Add(-time.Second)
}
}
// A long-running panel can see many unique client addresses. Finished buckets
// are swept on every new window, and a hard cap evicts the oldest when the map
// would otherwise grow without bound.
func TestRateLimiterCapsBucketCount(t *testing.T) {
r := newRateLimiter(1, time.Minute)
r.maxBuckets = 3
for i, key := range []string{"203.0.113.7", "198.51.100.9", "192.0.2.5"} {
if !r.Allow(key) {
t.Fatalf("attempt %d for %s was refused under the cap", i+1, key)
}
expire(r, key)
}
if !r.Allow("203.0.113.8") {
t.Fatal("a fourth address was refused even though room was made")
}
r.mu.Lock()
defer r.mu.Unlock()
if len(r.buckets) > 3 {
t.Fatalf("bucket count = %d, want at most 3", len(r.buckets))
}
}
+9 -4
View File
@@ -41,17 +41,22 @@ func hashToken(token string) string {
return hex.EncodeToString(sum[:])
}
// Create issues a new session for username and returns its token.
func (s *sessionStore) Create(username string) string {
// Create issues a new session for username and returns its token. It fails
// closed: if the row cannot be written the caller gets an error and must not
// hand out a cookie, because a token that is not in the database looks like a
// signed-in browser while every request it makes bounces back to /login.
func (s *sessionStore) Create(username string) (string, error) {
token := randomToken(32)
now := time.Now()
if err := s.store.CreateSession(hashToken(token), username, now.Add(s.idle)); err != nil {
logf("panel: session: create failed: %v", err)
return "", err
}
// Pruning is housekeeping: the new session is already valid, so a failure
// here is logged and does not fail the login.
if _, err := s.store.DeleteExpiredSessions(now); err != nil {
logf("panel: session: prune expired failed: %v", err)
}
return token
return token, nil
}
// Lookup returns the session username for a token if it exists and is
+3 -2
View File
@@ -16,8 +16,9 @@ const setupTokenTTL = 10 * time.Minute
// setupManager owns the one-time administrator setup token. The token itself is
// ephemeral (regenerated on restart or expiry) and lives only in memory; the
// persistent "setup complete" fact is the presence of the admin row in the
// store, so once that exists the token is gone for good (security.md).
// persistent "setup complete" fact is the presence of any row in the store's
// users table (`store.UserExists`), so once the first global administrator is
// created the token is gone for good (security.md).
type setupManager struct {
store *store.Store
hostname string
+1 -11
View File
@@ -32,17 +32,7 @@ func (h *Handlers) assignedDomains(p auth.Principal) ([]store.Domain, error) {
if p.IsGlobal() {
return h.store.ListDomains()
}
all, err := h.store.ListDomains()
if err != nil {
return nil, err
}
var out []store.Domain
for _, d := range all {
if p.CanAccessDomain(d.ID) {
out = append(out, d)
}
}
return out, nil
return h.store.ListDomainsForUser(p.ID)
}
func domainNameSet(domains []store.Domain) map[string]bool {
+125
View File
@@ -0,0 +1,125 @@
package handlers
import (
"net/http"
"net/http/httptest"
"testing"
"github.com/mixeme/selfpost/internal/web/auth"
)
// route is one entry of the authenticated mux, named the way web.go registers
// it so a route added there without a guard is visible as a missing case here.
type route struct {
method string
target string
handler func(*Handlers) http.HandlerFunc
// pathValues are the {id}-style segments the router would have bound.
pathValues map[string]string
}
// globalOnlyRoutes is every page and action that only a global administrator
// may reach: the panel's users, the whole-server backup and domain import, the
// machine-wide status and log views, and the domain lifecycle. A domain
// administrator is answered 404 rather than 403 so the panel does not confirm
// that the page exists (security.md).
var globalOnlyRoutes = []route{
{"GET", "/users", func(h *Handlers) http.HandlerFunc { return h.HandleUsers }, nil},
{"GET", "/users/new", func(h *Handlers) http.HandlerFunc { return h.HandleUserNew }, nil},
{"POST", "/users/new", func(h *Handlers) http.HandlerFunc { return h.HandleUserNew }, nil},
{"GET", "/users/1", func(h *Handlers) http.HandlerFunc { return h.HandleUserEdit }, map[string]string{"uid": "1"}},
{"POST", "/users/1", func(h *Handlers) http.HandlerFunc { return h.HandleUserEdit }, map[string]string{"uid": "1"}},
{"GET", "/users/1/delete", func(h *Handlers) http.HandlerFunc { return h.HandleUserDeleteConfirm }, map[string]string{"uid": "1"}},
{"POST", "/users/1/delete", func(h *Handlers) http.HandlerFunc { return h.HandleUserDelete }, map[string]string{"uid": "1"}},
{"GET", "/backup", func(h *Handlers) http.HandlerFunc { return h.HandleBackupPage }, nil},
{"POST", "/backup", func(h *Handlers) http.HandlerFunc { return h.HandleBackup }, nil},
{"POST", "/domains/import", func(h *Handlers) http.HandlerFunc { return h.HandleImportDomain }, nil},
{"GET", "/status", func(h *Handlers) http.HandlerFunc { return h.HandleStatus }, nil},
{"GET", "/status/fragment", func(h *Handlers) http.HandlerFunc { return h.HandleStatusFragment }, nil},
{"POST", "/status/recheck", func(h *Handlers) http.HandlerFunc { return h.HandleStatusRecheck }, nil},
{"GET", "/mail-queue", func(h *Handlers) http.HandlerFunc { return h.HandleMailQueue }, nil},
{"GET", "/mail-queue/body", func(h *Handlers) http.HandlerFunc { return h.HandleMailQueueBody }, nil},
{"GET", "/system-log", func(h *Handlers) http.HandlerFunc { return h.HandleSystemLog }, nil},
{"GET", "/system-log/body", func(h *Handlers) http.HandlerFunc { return h.HandleSystemLogBody }, nil},
{"POST", "/domains", func(h *Handlers) http.HandlerFunc { return h.HandleAddDomain }, nil},
{"GET", "/domains/1/delete", func(h *Handlers) http.HandlerFunc { return h.HandleDeleteConfirm }, map[string]string{"id": "1"}},
{"POST", "/domains/1/delete", func(h *Handlers) http.HandlerFunc { return h.HandleDeleteDomain }, map[string]string{"id": "1"}},
{"POST", "/reload", func(h *Handlers) http.HandlerFunc { return h.HandleReload }, nil},
}
// A domain administrator has an account on the panel, so authentication is not
// what keeps them off these pages — the per-handler role check is. Each of them
// is reached here with a valid session for a principal that owns a domain, the
// case the send-log leak (P0, code-review.md) showed is easy to get wrong.
func TestGlobalOnlyRoutesAnswerADomainAdmin404(t *testing.T) {
h, domains := serverWithTwoDomains(t)
p := domainAdmin(t, h.store, "global-only", domains["first.example.ru"].ID)
for _, rt := range globalOnlyRoutes {
rec := call(h, rt, p)
if rec.Code != http.StatusNotFound {
t.Errorf("%s %s as a domain administrator = %d, want 404:\n%s",
rt.method, rt.target, rec.Code, rec.Body.String())
}
}
}
// The same 404 covers a request that carries no principal at all: the auth
// middleware normally redirects those, but a handler must not depend on
// middleware it cannot see for the role it enforces itself.
func TestGlobalOnlyRoutesAnswerAnUnknownPrincipal404(t *testing.T) {
h, _ := serverWithTwoDomains(t)
for _, rt := range globalOnlyRoutes {
rec := httptest.NewRecorder()
req := httptest.NewRequest(rt.method, rt.target, nil)
for k, v := range rt.pathValues {
req.SetPathValue(k, v)
}
rt.handler(h)(rec, req)
if rec.Code != http.StatusNotFound {
t.Errorf("%s %s with no principal = %d, want 404", rt.method, rt.target, rec.Code)
}
}
}
// The 404s above would also pass if a handler were broken into always returning
// one, so at least the two pages that need nothing but the store and the view
// have to be shown opening for a global administrator.
func TestGlobalOnlyRoutesOpenForAGlobalAdministrator(t *testing.T) {
h, _ := serverWithTwoDomains(t)
for _, target := range []string{"/users", "/backup"} {
rt := getRoute(t, target)
if rec := call(h, rt, globalPrincipal); rec.Code != http.StatusOK {
t.Errorf("GET %s as a global administrator = %d, want 200:\n%s",
target, rec.Code, rec.Body.String())
}
}
}
func getRoute(t *testing.T, target string) route {
t.Helper()
for _, rt := range globalOnlyRoutes {
if rt.method == http.MethodGet && rt.target == target {
return rt
}
}
t.Fatalf("no GET %s among the global-only routes", target)
return route{}
}
func call(h *Handlers, rt route, p auth.Principal) *httptest.ResponseRecorder {
rec := httptest.NewRecorder()
req := httptest.NewRequest(rt.method, rt.target, nil)
req = auth.RequestWithPrincipal(req, p)
for k, v := range rt.pathValues {
req.SetPathValue(k, v)
}
rt.handler(h)(rec, req)
return rec
}
@@ -0,0 +1,46 @@
package handlers
import (
"regexp"
"strings"
"testing"
)
// TestE2ESendLogStatusMarkupDrift renders the real deliveries_rows fragment and
// checks whether the e2e gate's HTML scrapers still match it.
func TestE2ESendLogStatusMarkupDrift(t *testing.T) {
h, _ := serverWithDelivery(t)
rendered := getBody(t, h.HandleDeliveriesRows, "/deliveries/rows")
// These mirror test/e2e/main_test.go — keep in sync when fixing the e2e gate.
statusCellPattern := regexp.MustCompile(`class="st st-[^"]+">(queued|sent|deferred|bounced|rejected)</span>`)
containsCell := func(html, needle string) bool {
return strings.Contains(html, `<span class="st st-`) && strings.Contains(html, `">`+needle+`</span>`)
}
if statusCellPattern.FindStringSubmatch(rendered) == nil {
t.Fatalf("e2e statusCellPattern does not match rendered send-log rows:\n%s", snippet(rendered, `class="status"`))
}
if !containsCell(rendered, "sent") {
t.Fatalf("e2e containsCell does not match rendered send-log rows:\n%s", snippet(rendered, `class="status"`))
}
}
func snippet(s, needle string) string {
i := strings.Index(s, needle)
if i < 0 {
if len(s) > 200 {
return s[:200] + "..."
}
return s
}
start := i - 20
if start < 0 {
start = 0
}
end := i + 120
if end > len(s) {
end = len(s)
}
return s[start:end]
}
+4
View File
@@ -24,6 +24,10 @@ type Config struct {
TLSCertFile string
OpenDKIMSocket string
JournalSocket string
// Level-1 Postfix anvil backstop (env RATE_LIMIT_*), shown in the panel
// and used to cap domain/app level-2 ceilings (guide § Rate limiting).
RateLimitMessagesPerIP int
RateLimitWindowSeconds int
}
// Handlers holds dependencies for authenticated panel routes.
+3 -1
View File
@@ -165,9 +165,11 @@ func (h *Handlers) renderDomainDetail(w http.ResponseWriter, r *http.Request, st
data["ExportErr"] = view.ExportErr
data["MinPwLen"] = validate.MinSecretFilePasswordLen
data["DomainHasRL"] = domainRLok && domainRL.Active()
data["DomainRLIPs"] = strings.Join(domainRL.AllowedIPs, "\n")
data["DomainRLMax"] = intOrBlank(domainRL.MaxMessages)
data["DomainRLWin"] = windowOrDefault(domainRL.WindowSeconds)
data["DomainRLMaxNum"] = domainRL.MaxMessages
data["L1Messages"] = h.l1Messages()
data["L1Window"] = h.l1Window()
h.view.Render(w, status, "domain_detail", data)
}
+1 -1
View File
@@ -45,7 +45,7 @@ func (h *Handlers) renderBackupPage(w http.ResponseWriter, r *http.Request, stat
// wrong form.
func (h *Handlers) renderBackupPageWith(w http.ResponseWriter, r *http.Request, status int, importErr, backupErr string) {
data := h.pageBase(r)
data["Title"] = "SelfPost — backup"
data["Title"] = "SelfPost — backup & migration"
data["Active"] = "backup"
data["ImportErr"] = importErr
data["BackupErr"] = backupErr
+1 -1
View File
@@ -41,7 +41,7 @@ func (h *Handlers) renderDashboard(w http.ResponseWriter, r *http.Request, statu
return
}
data := h.pageBase(r)
data["Title"] = "SelfPost"
data["Title"] = "SelfPost — domains"
data["Active"] = "domains"
data["Domains"] = h.domainRows(domains)
data["Error"] = formErr
+40 -30
View File
@@ -299,16 +299,22 @@ func deliveriesBackURL(r *http.Request) string {
// sendLogData reads the domain/app filters and page number off the query
// string, queries the store, and assembles everything the template needs
// (filter dropdown options plus the current selection, rows, and pagination).
//
// The invariant this function owes the journal: a principal who is not global
// only ever reads rows for the domains assigned to them. That scope is stated
// to the store as SendLogFilter.Domains and holds for every number of
// assignments, including none — a domain administrator whose last domain was
// deleted gets an empty log, not the whole one. The query parameters are
// filters *within* that scope and can only narrow it: both are checked against
// the assigned domains and their applications before the query runs, because a
// dropdown that offers only permitted values is a courtesy to the browser, not
// a check on the request.
func (h *Handlers) sendLogData(r *http.Request) (map[string]any, error) {
p, ok := h.principal(r)
if !ok {
return nil, errors.New("no principal")
}
q := r.URL.Query()
filter := store.SendLogFilter{
Domain: q.Get("domain"),
AppLogin: q.Get("app"),
}
assigned, err := h.assignedDomains(p)
if err != nil {
@@ -316,31 +322,6 @@ func (h *Handlers) sendLogData(r *http.Request) (map[string]any, error) {
}
allowedNames := domainNameSet(assigned)
if !p.IsGlobal() {
if filter.Domain != "" && !allowedNames[filter.Domain] {
filter.Domain = ""
}
if filter.Domain == "" && len(assigned) == 1 {
filter.Domain = assigned[0].Name
}
}
page := parsePage(q.Get("p"))
total, err := h.store.CountSendLog(filter)
if err != nil {
return nil, err
}
rows, err := h.store.QuerySendLog(filter, sendLogPageSize, (page-1)*sendLogPageSize)
if err != nil {
return nil, err
}
view := make([]sendLogRow, len(rows))
for i := range rows {
rows[i].Subject = mailhdr.DecodeSubject(rows[i].Subject)
view[i] = sendLogRow{SendLogRow: rows[i], Level: deliveryLevel(rows[i].Status)}
}
domainNames := make([]string, 0, len(assigned))
for _, d := range assigned {
domainNames = append(domainNames, d.Name)
@@ -362,9 +343,38 @@ func (h *Handlers) sendLogData(r *http.Request) (map[string]any, error) {
}
sort.Strings(logins)
if !p.IsGlobal() && filter.AppLogin != "" && !loginSet[filter.AppLogin] {
filter := store.SendLogFilter{
Domain: q.Get("domain"),
AppLogin: q.Get("app"),
// A global administrator reads the whole journal, including rows left
// behind by a domain that has since been deleted.
Domains: domainNames,
AllDomains: p.IsGlobal(),
}
if !p.IsGlobal() {
if filter.Domain != "" && !allowedNames[filter.Domain] {
filter.Domain = ""
}
if filter.AppLogin != "" && !loginSet[filter.AppLogin] {
filter.AppLogin = ""
}
}
page := parsePage(q.Get("p"))
total, err := h.store.CountSendLog(filter)
if err != nil {
return nil, err
}
rows, err := h.store.QuerySendLog(filter, sendLogPageSize, (page-1)*sendLogPageSize)
if err != nil {
return nil, err
}
view := make([]sendLogRow, len(rows))
for i := range rows {
rows[i].Subject = mailhdr.DecodeSubject(rows[i].Subject)
view[i] = sendLogRow{SendLogRow: rows[i], Level: deliveryLevel(rows[i].Status)}
}
lastPage := 1
if total > 0 {
+180 -10
View File
@@ -129,7 +129,7 @@ func TestDeliveryPageMarksAQueuedMessageAsStillWaiting(t *testing.T) {
}); err != nil {
t.Fatalf("insert: %v", err)
}
rows, err := h.store.QuerySendLog(store.SendLogFilter{}, 1, 0)
rows, err := h.store.QuerySendLog(store.SendLogFilter{AllDomains: true}, 1, 0)
if err != nil || len(rows) != 1 {
t.Fatalf("query: %v (%d rows)", err, len(rows))
}
@@ -215,6 +215,174 @@ func TestDeliveryPageNotFound(t *testing.T) {
}
}
// A domain administrator reads the journal of the domains assigned to them and
// nothing else. The list used to be scoped only when exactly one domain was
// assigned, which meant two assignments read as none at all.
func TestSendLogScopedToAssignedDomains(t *testing.T) {
h, domains := serverWithTwoDomains(t)
for name, tc := range map[string]struct {
username string
domainIDs []int64
want []string
unwanted []string
}{
"global sees both": {
"", nil, []string{"First message", "Second message"}, nil,
},
"one assigned domain": {
"one-domain", []int64{domains["first.example.ru"].ID},
[]string{"First message"}, []string{"Second message", "second-app"},
},
"two assigned domains": {
"two-domains", []int64{domains["first.example.ru"].ID, domains["second.example.ru"].ID},
[]string{"First message", "Second message"}, nil,
},
// Every assigned domain deleted cascades the assignments away. That
// leaves a principal entitled to nothing, which is an empty log — the
// case that used to hand over the whole journal.
"no assigned domains": {
"no-domains", nil, []string{"No messages logged yet."},
[]string{"First message", "Second message"},
},
} {
var p auth.Principal
if tc.username == "" {
p = globalPrincipal
} else {
p = domainAdmin(t, h.store, tc.username, tc.domainIDs...)
}
for view, handler := range map[string]http.HandlerFunc{
"page": h.HandleDeliveries,
"fragment": h.HandleDeliveriesRows,
} {
out := getBodyAs(t, handler, "/deliveries", p)
for _, want := range tc.want {
if !strings.Contains(out, want) {
t.Errorf("%s (%s): missing %q:\n%s", name, view, want, out)
}
}
for _, unwanted := range tc.unwanted {
if strings.Contains(out, unwanted) {
t.Errorf("%s (%s): leaks %q:\n%s", name, view, unwanted, out)
}
}
}
}
}
// The filter dropdowns offer only permitted values, so a leak through them can
// only come from a hand-written URL — which is exactly why the values are
// checked against the principal's own domains and applications rather than
// trusted for having been rendered by us.
func TestSendLogIgnoresForgedFilters(t *testing.T) {
h, domains := serverWithTwoDomains(t)
p := domainAdmin(t, h.store, "forged-filter", domains["first.example.ru"].ID)
for _, target := range []string{
"/deliveries?domain=second.example.ru",
"/deliveries?app=second-app",
"/deliveries?domain=second.example.ru&app=second-app",
} {
out := getBodyAs(t, h.HandleDeliveries, target, p)
if strings.Contains(out, "Second message") {
t.Errorf("GET %s leaks another domain's journal:\n%s", target, out)
}
if !strings.Contains(out, "First message") {
t.Errorf("GET %s hid the principal's own journal:\n%s", target, out)
}
}
}
// The detail page has always checked membership; keep it checked, because the
// list and the page are two ways to the same row.
func TestDeliveryPageForeignDomainNotFound(t *testing.T) {
h, domains := serverWithTwoDomains(t)
rows, err := h.store.QuerySendLog(store.SendLogFilter{Domain: "second.example.ru", AllDomains: true}, 1, 0)
if err != nil || len(rows) != 1 {
t.Fatalf("query: %v (%d rows)", err, len(rows))
}
rec := httptest.NewRecorder()
req := httptest.NewRequest(http.MethodGet, "/deliveries/"+itoa(rows[0].ID), nil)
req.SetPathValue("id", itoa(rows[0].ID))
req = auth.RequestWithPrincipal(req, domainAdmin(t, h.store, "foreign-detail", domains["first.example.ru"].ID))
h.HandleDelivery(rec, req)
if rec.Code != http.StatusNotFound {
t.Errorf("delivery page for a foreign domain = %d, want 404", rec.Code)
}
}
// serverWithTwoDomains builds a panel over a store holding two domains, one
// application and one delivered message each, so a scoping test can tell "my
// rows" from "every row" by reading the page.
func serverWithTwoDomains(t *testing.T) (*Handlers, map[string]store.Domain) {
t.Helper()
st, err := store.Open(filepath.Join(t.TempDir(), "test.db"))
if err != nil {
t.Fatalf("open store: %v", err)
}
t.Cleanup(func() { _ = st.Close() })
domains := make(map[string]store.Domain, 2)
for _, d := range []struct{ name, app, subject string }{
{"first.example.ru", "first-app", "First message"},
{"second.example.ru", "second-app", "Second message"},
} {
dom, err := st.AddDomain(d.name, "mail")
if err != nil {
t.Fatalf("add domain %s: %v", d.name, err)
}
if _, err := st.AddApplication(dom.ID, d.app, store.AddressModeWildcard, nil); err != nil {
t.Fatalf("add application %s: %v", d.app, err)
}
if err := st.InsertQueued(store.SendLogEntry{
QueueID: "Q" + d.app, Domain: d.name, AppLogin: d.app,
From: "noreply@" + d.name, To: "public@example.net", Subject: d.subject,
}); err != nil {
t.Fatalf("insert %s: %v", d.subject, err)
}
domains[d.name] = dom
}
return &Handlers{store: st, view: mustView(t), cfg: Config{Version: "test"}}, domains
}
var globalPrincipal = auth.Principal{ID: 1, Username: "admin", Role: auth.RoleGlobal}
func domainAdmin(t *testing.T, st *store.Store, username string, domainIDs ...int64) auth.Principal {
t.Helper()
const hash = "test-hash"
if len(domainIDs) == 0 {
placeholder, err := st.AddDomain(username+".placeholder.invalid", "mail")
if err != nil {
t.Fatalf("add placeholder domain: %v", err)
}
id, err := st.CreateUser(username, hash, store.RoleDomainAdmin, []int64{placeholder.ID})
if err != nil {
t.Fatalf("create domain admin %s: %v", username, err)
}
if err := st.DeleteDomain(placeholder.ID); err != nil {
t.Fatalf("delete placeholder domain: %v", err)
}
domainIDs = nil
u, err := st.GetUser(id)
if err != nil {
t.Fatalf("get domain admin %s: %v", username, err)
}
return auth.Principal{ID: u.ID, Username: u.Username, Role: u.Role, Domains: u.DomainIDs}
}
id, err := st.CreateUser(username, hash, store.RoleDomainAdmin, domainIDs)
if err != nil {
t.Fatalf("create domain admin %s: %v", username, err)
}
u, err := st.GetUser(id)
if err != nil {
t.Fatalf("get domain admin %s: %v", username, err)
}
return auth.Principal{ID: u.ID, Username: u.Username, Role: u.Role, Domains: u.DomainIDs}
}
// serverWithDelivery builds a panel over a store holding one delivery, written
// the way the journal-milter wrote them before it decoded subjects itself.
func serverWithDelivery(t *testing.T) (*Handlers, store.SendLogRow) {
@@ -238,7 +406,7 @@ func serverWithDelivery(t *testing.T) (*Handlers, store.SendLogRow) {
if _, err := st.UpdateStatus("4A1B2C3D", "public@example.ru", store.StatusSent); err != nil {
t.Fatalf("update status: %v", err)
}
rows, err := st.QuerySendLog(store.SendLogFilter{}, 1, 0)
rows, err := st.QuerySendLog(store.SendLogFilter{AllDomains: true}, 1, 0)
if err != nil || len(rows) != 1 {
t.Fatalf("query: %v (%d rows)", err, len(rows))
}
@@ -246,18 +414,20 @@ func serverWithDelivery(t *testing.T) (*Handlers, store.SendLogRow) {
return &Handlers{store: st, view: mustView(t), cfg: Config{Version: "test"}}, rows[0]
}
// getBody runs one handler over a GET and returns the page it wrote, failing
// the test on any non-200. The path's {id} is bound by hand because these calls
// bypass the router that would otherwise fill it in.
// getBody runs one handler over a GET as the global administrator.
func getBody(t *testing.T, h http.HandlerFunc, target string) string {
t.Helper()
return getBodyAs(t, h, target, globalPrincipal)
}
// getBodyAs runs one handler over a GET as the given principal and returns the
// page it wrote, failing the test on any non-200. The path's {id} is bound by
// hand because these calls bypass the router that would otherwise fill it in.
func getBodyAs(t *testing.T, h http.HandlerFunc, target string, p auth.Principal) string {
t.Helper()
rec := httptest.NewRecorder()
req := httptest.NewRequest(http.MethodGet, target, nil)
req = auth.RequestWithPrincipal(req, auth.Principal{
ID: 1,
Username: "admin",
Role: auth.RoleGlobal,
})
req = auth.RequestWithPrincipal(req, p)
if rest, ok := strings.CutPrefix(req.URL.Path, "/deliveries/"); ok && rest != "rows" {
req.SetPathValue("id", rest)
}
+61 -5
View File
@@ -19,24 +19,73 @@ type rateLimitInput struct {
windowSeconds int
}
func parseRateLimitForm(r *http.Request) (rateLimitInput, error) {
func (h *Handlers) l1Messages() int {
if h.cfg.RateLimitMessagesPerIP > 0 {
return h.cfg.RateLimitMessagesPerIP
}
return 100
}
func (h *Handlers) l1Window() int {
if h.cfg.RateLimitWindowSeconds > 0 {
return h.cfg.RateLimitWindowSeconds
}
return defaultRateLimitWindowSeconds
}
func parseDomainRateLimitForm(r *http.Request, l1Max int) (rateLimitInput, error) {
if err := r.ParseForm(); err != nil {
return rateLimitInput{}, fmt.Errorf("invalid form submission")
}
if r.PostFormValue("clear") != "" {
return rateLimitInput{clear: true}, nil
}
rawMax := strings.TrimSpace(r.PostFormValue("max_messages"))
if rawMax == "" {
return rateLimitInput{clear: true}, nil
}
maxMessages, err := parsePositiveInt(rawMax, 0)
if err != nil || maxMessages <= 0 {
return rateLimitInput{}, fmt.Errorf("enter a message limit greater than zero")
}
if maxMessages > l1Max {
return rateLimitInput{}, fmt.Errorf("message limit cannot exceed the level-1 backstop (%d)", l1Max)
}
windowSeconds, err := parsePositiveInt(r.PostFormValue("window_seconds"), defaultRateLimitWindowSeconds)
if err != nil || windowSeconds <= 0 {
return rateLimitInput{}, fmt.Errorf("enter a time window greater than zero seconds")
}
return rateLimitInput{maxMessages: maxMessages, windowSeconds: windowSeconds}, nil
}
func parseAppRateLimitForm(r *http.Request, l1Max, domainMax int, domainActive bool) (rateLimitInput, error) {
if err := r.ParseForm(); err != nil {
return rateLimitInput{}, fmt.Errorf("invalid form submission")
}
if r.PostFormValue("clear") != "" {
return rateLimitInput{clear: true}, nil
}
rawMax := strings.TrimSpace(r.PostFormValue("max_messages"))
if rawMax == "" {
return rateLimitInput{clear: true}, nil
}
ips, err := parseIPList(r.PostFormValue("allowed_ips"))
if err != nil {
return rateLimitInput{}, err
}
if len(ips) == 0 {
return rateLimitInput{clear: true}, nil
return rateLimitInput{}, fmt.Errorf("enter at least one trusted client IP for an application override")
}
maxMessages, err := parsePositiveInt(r.PostFormValue("max_messages"), 0)
maxMessages, err := parsePositiveInt(rawMax, 0)
if err != nil || maxMessages <= 0 {
return rateLimitInput{}, fmt.Errorf("enter a message limit greater than zero")
}
if maxMessages > l1Max {
return rateLimitInput{}, fmt.Errorf("message limit cannot exceed the level-1 backstop (%d)", l1Max)
}
if domainActive && maxMessages <= domainMax {
return rateLimitInput{}, fmt.Errorf("application override must be greater than the domain limit (%d)", domainMax)
}
windowSeconds, err := parsePositiveInt(r.PostFormValue("window_seconds"), defaultRateLimitWindowSeconds)
if err != nil || windowSeconds <= 0 {
return rateLimitInput{}, fmt.Errorf("enter a time window greater than zero seconds")
@@ -77,7 +126,7 @@ func (h *Handlers) HandleDomainRateLimit(w http.ResponseWriter, r *http.Request)
if !ok {
return
}
in, err := parseRateLimitForm(r)
in, err := parseDomainRateLimitForm(r, h.l1Messages())
if err != nil {
h.renderDomainDetail(w, r, http.StatusBadRequest, d, detailView{
FormMode: store.AddressModeWildcard,
@@ -103,7 +152,14 @@ func (h *Handlers) HandleAppRateLimit(w http.ResponseWriter, r *http.Request) {
http.Error(w, "internal error", http.StatusInternalServerError)
return
}
in, err := parseRateLimitForm(r)
domainRL, domainOK, err := h.domains.RateLimit(d.ID)
if err != nil {
logf("panel: domain %d: rate limit: %v", d.ID, err)
http.Error(w, "internal error", http.StatusInternalServerError)
return
}
domainActive := domainOK && domainRL.Active()
in, err := parseAppRateLimitForm(r, h.l1Messages(), domainRL.MaxMessages, domainActive)
if err != nil {
h.renderDomainDetail(w, r, http.StatusBadRequest, d, detailView{
FormMode: store.AddressModeWildcard,
@@ -0,0 +1,93 @@
package handlers
import (
"net/http"
"net/http/httptest"
"net/url"
"strings"
"testing"
)
func TestParseDomainRateLimitForm(t *testing.T) {
t.Parallel()
form := func(vals url.Values) *http.Request {
r := httptest.NewRequest(http.MethodPost, "/", strings.NewReader(vals.Encode()))
r.Header.Set("Content-Type", "application/x-www-form-urlencoded")
return r
}
in, err := parseDomainRateLimitForm(form(url.Values{
"max_messages": {"50"},
"window_seconds": {"3600"},
}), 100)
if err != nil || in.clear || in.maxMessages != 50 || in.windowSeconds != 3600 || len(in.ips) != 0 {
t.Fatalf("valid domain = %+v err=%v", in, err)
}
in, err = parseDomainRateLimitForm(form(url.Values{"max_messages": {""}}), 100)
if err != nil || !in.clear {
t.Fatalf("empty max should clear: %+v err=%v", in, err)
}
_, err = parseDomainRateLimitForm(form(url.Values{
"max_messages": {"150"},
"window_seconds": {"3600"},
}), 100)
if err == nil || !strings.Contains(err.Error(), "level-1") {
t.Fatalf("over L1 want error, got %v", err)
}
}
func TestParseAppRateLimitForm(t *testing.T) {
t.Parallel()
form := func(vals url.Values) *http.Request {
r := httptest.NewRequest(http.MethodPost, "/", strings.NewReader(vals.Encode()))
r.Header.Set("Content-Type", "application/x-www-form-urlencoded")
return r
}
in, err := parseAppRateLimitForm(form(url.Values{
"allowed_ips": {"203.0.113.10"},
"max_messages": {"80"},
"window_seconds": {"3600"},
}), 100, 40, true)
if err != nil || in.maxMessages != 80 || len(in.ips) != 1 {
t.Fatalf("valid app override = %+v err=%v", in, err)
}
_, err = parseAppRateLimitForm(form(url.Values{
"max_messages": {"80"},
"window_seconds": {"3600"},
}), 100, 40, true)
if err == nil || !strings.Contains(err.Error(), "trusted client IP") {
t.Fatalf("missing IPs want error, got %v", err)
}
_, err = parseAppRateLimitForm(form(url.Values{
"allowed_ips": {"203.0.113.10"},
"max_messages": {"40"},
"window_seconds": {"3600"},
}), 100, 40, true)
if err == nil || !strings.Contains(err.Error(), "greater than the domain") {
t.Fatalf("app <= domain want error, got %v", err)
}
_, err = parseAppRateLimitForm(form(url.Values{
"allowed_ips": {"203.0.113.10"},
"max_messages": {"150"},
"window_seconds": {"3600"},
}), 100, 0, false)
if err == nil || !strings.Contains(err.Error(), "level-1") {
t.Fatalf("over L1 want error, got %v", err)
}
// No domain limit: any app ceiling ≤ L1 is fine.
in, err = parseAppRateLimitForm(form(url.Values{
"allowed_ips": {"203.0.113.10"},
"max_messages": {"50"},
"window_seconds": {"3600"},
}), 100, 0, false)
if err != nil || in.maxMessages != 50 {
t.Fatalf("app without domain = %+v err=%v", in, err)
}
}
@@ -14,8 +14,8 @@ import (
"golang.org/x/crypto/bcrypt"
)
// HandleAccount serves the signed-in user's account settings.
func (h *Handlers) HandleAccount(w http.ResponseWriter, r *http.Request) {
// HandleSettings serves the signed-in user's panel settings.
func (h *Handlers) HandleSettings(w http.ResponseWriter, r *http.Request) {
switch r.Method {
case http.MethodGet:
p, ok := h.principal(r)
@@ -25,20 +25,20 @@ func (h *Handlers) HandleAccount(w http.ResponseWriter, r *http.Request) {
}
u, err := h.store.GetUser(p.ID)
if err != nil {
logf("panel: account: get user failed: %v", err)
logf("panel: settings: get user failed: %v", err)
http.Error(w, "internal error", http.StatusInternalServerError)
return
}
h.renderAccount(w, r, http.StatusOK, "", u.Username, u.DMARCReportEmail, p.IsGlobal())
h.renderSettings(w, r, http.StatusOK, "", u.Username, u.DMARCReportEmail, p.IsGlobal())
case http.MethodPost:
h.submitAccount(w, r)
h.submitSettings(w, r)
default:
w.Header().Set("Allow", "GET, POST")
http.Error(w, "method not allowed", http.StatusMethodNotAllowed)
}
}
func (h *Handlers) renderAccount(w http.ResponseWriter, r *http.Request, status int, formErr, formUsername, formDMARCEmail string, showDMARC bool) {
func (h *Handlers) renderSettings(w http.ResponseWriter, r *http.Request, status int, formErr, formUsername, formDMARCEmail string, showDMARC bool) {
var reportAuth dnscheck.Result
if showDMARC && formDMARCEmail != "" {
if hub := dnscheck.EmailDomain(formDMARCEmail); hub != "" {
@@ -49,7 +49,7 @@ func (h *Handlers) renderAccount(w http.ResponseWriter, r *http.Request, status
}
data := h.pageBase(r)
data["Title"] = "SelfPost — settings"
data["Active"] = "account"
data["Active"] = "settings"
data["FormUsername"] = formUsername
data["FormDMARCEmail"] = formDMARCEmail
data["ShowDMARC"] = showDMARC
@@ -58,11 +58,13 @@ func (h *Handlers) renderAccount(w http.ResponseWriter, r *http.Request, status
data["ReportAuthDNS"] = reportAuth
data["ReportAuthHub"] = dnscheck.EmailDomain(formDMARCEmail)
data["Error"] = formErr
data["Flash"] = accountFlash(r)
h.view.Render(w, status, "account", data)
data["Flash"] = settingsFlash(r)
data["L1Messages"] = h.l1Messages()
data["L1Window"] = h.l1Window()
h.view.Render(w, status, "settings", data)
}
func accountFlash(r *http.Request) string {
func settingsFlash(r *http.Request) string {
switch r.URL.Query().Get("updated") {
case "username":
return "Username changed."
@@ -83,16 +85,16 @@ func accountFlash(r *http.Request) string {
}
}
func (h *Handlers) submitAccount(w http.ResponseWriter, r *http.Request) {
func (h *Handlers) submitSettings(w http.ResponseWriter, r *http.Request) {
if !h.auth.AllowLoginAttempt(r) {
p, _ := h.principal(r)
h.renderAccount(w, r, http.StatusTooManyRequests,
h.renderSettings(w, r, http.StatusTooManyRequests,
"Too many attempts. Please wait and try again.", auth.CurrentUser(r), "", p.IsGlobal())
return
}
if err := r.ParseForm(); err != nil {
p, _ := h.principal(r)
h.renderAccount(w, r, http.StatusBadRequest, "Invalid form submission.", auth.CurrentUser(r), "", p.IsGlobal())
h.renderSettings(w, r, http.StatusBadRequest, "Invalid form submission.", auth.CurrentUser(r), "", p.IsGlobal())
return
}
@@ -103,7 +105,7 @@ func (h *Handlers) submitAccount(w http.ResponseWriter, r *http.Request) {
}
user, err := h.store.GetUser(p.ID)
if err != nil {
logf("panel: account: get user failed: %v", err)
logf("panel: settings: get user failed: %v", err)
http.Error(w, "internal error", http.StatusInternalServerError)
return
}
@@ -121,21 +123,21 @@ func (h *Handlers) submitAccount(w http.ResponseWriter, r *http.Request) {
}
if err := bcrypt.CompareHashAndPassword([]byte(user.PasswordHash), []byte(current)); err != nil {
h.renderAccount(w, r, http.StatusUnauthorized, "Current password is incorrect.", username, dmarcEmail, p.IsGlobal())
h.renderSettings(w, r, http.StatusUnauthorized, "Current password is incorrect.", username, dmarcEmail, p.IsGlobal())
return
}
renaming := username != user.Username
if renaming {
if err := validate.Username(username); err != nil {
h.renderAccount(w, r, http.StatusBadRequest, err.Error(), username, dmarcEmail, p.IsGlobal())
h.renderSettings(w, r, http.StatusBadRequest, err.Error(), username, dmarcEmail, p.IsGlobal())
return
}
}
if p.IsGlobal() {
if err := validate.Email(dmarcEmail); err != nil {
h.renderAccount(w, r, http.StatusBadRequest, err.Error(), username, dmarcEmail, true)
h.renderSettings(w, r, http.StatusBadRequest, err.Error(), username, dmarcEmail, true)
return
}
}
@@ -145,16 +147,16 @@ func (h *Handlers) submitAccount(w http.ResponseWriter, r *http.Request) {
repassword := password != "" || confirm != ""
if repassword {
if password != confirm {
h.renderAccount(w, r, http.StatusBadRequest, "New passwords do not match.", username, dmarcEmail, p.IsGlobal())
h.renderSettings(w, r, http.StatusBadRequest, "New passwords do not match.", username, dmarcEmail, p.IsGlobal())
return
}
if err := validate.AdminPassword(password); err != nil {
h.renderAccount(w, r, http.StatusBadRequest, err.Error(), username, dmarcEmail, p.IsGlobal())
h.renderSettings(w, r, http.StatusBadRequest, err.Error(), username, dmarcEmail, p.IsGlobal())
return
}
}
if !renaming && !repassword && !emailChanging {
h.renderAccount(w, r, http.StatusBadRequest,
h.renderSettings(w, r, http.StatusBadRequest,
"Nothing to change: enter a new username, password, or DMARC report address.", username, dmarcEmail, p.IsGlobal())
return
}
@@ -163,8 +165,8 @@ func (h *Handlers) submitAccount(w http.ResponseWriter, r *http.Request) {
if repassword {
newHash, err := bcrypt.GenerateFromPassword([]byte(password), bcrypt.DefaultCost)
if err != nil {
logf("panel: account: hashing password failed: %v", err)
h.renderAccount(w, r, http.StatusInternalServerError,
logf("panel: settings: hashing password failed: %v", err)
h.renderSettings(w, r, http.StatusInternalServerError,
"Internal error. Please try again.", username, dmarcEmail, p.IsGlobal())
return
}
@@ -172,17 +174,17 @@ func (h *Handlers) submitAccount(w http.ResponseWriter, r *http.Request) {
}
if err := h.store.UpdateUser(user.ID, username, hash, dmarcEmail); err != nil {
logf("panel: account: update user failed: %v", err)
logf("panel: settings: update user failed: %v", err)
msg := "Could not save the changes. Please check the logs and try again."
if errors.Is(err, store.ErrUserNotFound) {
msg = "There is no user account to update."
}
if errors.Is(err, store.ErrUserExists) {
msg = "That username is already in use."
h.renderAccount(w, r, http.StatusConflict, msg, username, dmarcEmail, p.IsGlobal())
h.renderSettings(w, r, http.StatusConflict, msg, username, dmarcEmail, p.IsGlobal())
return
}
h.renderAccount(w, r, http.StatusInternalServerError, msg, username, dmarcEmail, p.IsGlobal())
h.renderSettings(w, r, http.StatusInternalServerError, msg, username, dmarcEmail, p.IsGlobal())
return
}
@@ -195,8 +197,8 @@ func (h *Handlers) submitAccount(w http.ResponseWriter, r *http.Request) {
}
}
logf("panel: user %d account updated (username: %t, password: %t, dmarc email: %t)", user.ID, renaming, repassword, emailChanging)
http.Redirect(w, r, "/account?updated="+updatedFlag(renaming, repassword, emailChanging), http.StatusSeeOther)
logf("panel: user %d settings updated (username: %t, password: %t, dmarc email: %t)", user.ID, renaming, repassword, emailChanging)
http.Redirect(w, r, "/settings?updated="+updatedFlag(renaming, repassword, emailChanging), http.StatusSeeOther)
}
func updatedFlag(renamed, repassword, emailChanged bool) string {
+3 -4
View File
@@ -13,15 +13,11 @@ func (h *Handlers) HandleStatus(w http.ResponseWriter, r *http.Request) {
return
}
data := h.statusBody()
srv := h.dns.Server(h.cfg.Hostname, false)
data["Title"] = "SelfPost — status"
data["User"] = auth.CurrentUser(r)
data["Active"] = "status"
data["IsGlobal"] = true
data["Flash"] = statusFlash(r)
data["Hostname"] = h.cfg.Hostname
data["PTR"] = srv.PTR
h.view.Render(w, http.StatusOK, "status", data)
}
@@ -68,6 +64,7 @@ func (h *Handlers) statusBody() map[string]any {
}
machine := h.machine.Sample()
srv := h.dns.Server(h.cfg.Hostname, false)
overall := health.Worst(procStatus, queueStatus, cert.Status, socketStatus, machine.Status)
return map[string]any{
@@ -81,6 +78,8 @@ func (h *Handlers) statusBody() map[string]any {
"Cert": cert,
"Sockets": sockets,
"SocketStatus": socketStatus,
"Hostname": h.cfg.Hostname,
"PTR": srv.PTR,
"OverallStatus": overall,
"OverallHeading": overallHeading(overall),
}
+67 -5
View File
@@ -102,10 +102,6 @@ func (h *Handlers) HandleUserEdit(w http.ResponseWriter, r *http.Request) {
h.renderUserForm(w, r, http.StatusBadRequest, u.ID, userFormView{FormErr: "Invalid form submission.", FormUsername: u.Username, FormRole: string(u.Role)})
return
}
if r.PostFormValue("action") == "delete" {
h.submitUserDelete(w, r, u)
return
}
h.submitUserUpdate(w, r, u)
default:
w.Header().Set("Allow", "GET, POST")
@@ -121,19 +117,33 @@ func (h *Handlers) renderUserForm(w http.ResponseWriter, r *http.Request, status
return
}
data := h.pageBase(r)
data["Title"] = "SelfPost — user"
if userID != 0 {
data["Title"] = "SelfPost — edit user"
} else {
data["Title"] = "SelfPost — create user"
}
data["Active"] = "users"
data["UserID"] = userID
data["Domains"] = domains
data["Error"] = view.FormErr
data["FormUsername"] = view.FormUsername
data["FormRole"] = view.FormRole
data["GlobalRole"] = store.RoleGlobal
data["FormDomains"] = view.FormDomains
data["FormPassword"] = view.FormPassword
data["IsEdit"] = userID != 0
data["LastGlobalLocked"] = lastGlobalLocked(h, userID, view.FormRole)
h.view.Render(w, status, "user_form", data)
}
func lastGlobalLocked(h *Handlers, userID int64, formRole string) bool {
if userID == 0 || formRole != string(store.RoleGlobal) {
return false
}
n, err := h.store.CountGlobalUsers()
return err == nil && n <= 1
}
func (h *Handlers) submitUserCreate(w http.ResponseWriter, r *http.Request) {
if err := r.ParseForm(); err != nil {
h.renderUserForm(w, r, http.StatusBadRequest, 0, userFormView{FormErr: "Invalid form submission."})
@@ -277,6 +287,58 @@ func (h *Handlers) submitUserUpdate(w http.ResponseWriter, r *http.Request, u st
http.Redirect(w, r, "/users?done=updated", http.StatusSeeOther)
}
// HandleUserDeleteConfirm shows the cascade warning before a panel user is
// removed — the same pattern as HandleDeleteConfirm for domains, so a single
// mis-click on Delete cannot remove a user (P3, code-review.md).
func (h *Handlers) HandleUserDeleteConfirm(w http.ResponseWriter, r *http.Request) {
if _, ok := h.requireGlobal(w, r); !ok {
return
}
uid, ok := parseUserID(w, r)
if !ok {
return
}
u, err := h.store.GetUser(uid)
if err != nil {
if errors.Is(err, store.ErrUserNotFound) {
http.NotFound(w, r)
return
}
logf("panel: get user %d: %v", uid, err)
http.Error(w, "internal error", http.StatusInternalServerError)
return
}
data := h.pageBase(r)
data["Title"] = "SelfPost — delete " + u.Username
data["Active"] = "users"
data["TargetID"] = u.ID
data["TargetUsername"] = u.Username
h.view.Render(w, http.StatusOK, "user_delete", data)
}
// HandleUserDelete performs the deletion confirmed on HandleUserDeleteConfirm
// and returns to the user list.
func (h *Handlers) HandleUserDelete(w http.ResponseWriter, r *http.Request) {
if _, ok := h.requireGlobal(w, r); !ok {
return
}
uid, ok := parseUserID(w, r)
if !ok {
return
}
u, err := h.store.GetUser(uid)
if err != nil {
if errors.Is(err, store.ErrUserNotFound) {
http.NotFound(w, r)
return
}
logf("panel: get user %d: %v", uid, err)
http.Error(w, "internal error", http.StatusInternalServerError)
return
}
h.submitUserDelete(w, r, u)
}
func (h *Handlers) submitUserDelete(w http.ResponseWriter, r *http.Request, u store.User) {
p, ok := h.principal(r)
if !ok {
+93
View File
@@ -0,0 +1,93 @@
Copyright © 2017 IBM Corp. with Reserved Font Name "Plex"
This Font Software is licensed under the SIL Open Font License, Version 1.1.
This license is copied below, and is also available with a FAQ at: http://scripts.sil.org/OFL
-----------------------------------------------------------
SIL OPEN FONT LICENSE Version 1.1 - 26 February 2007
-----------------------------------------------------------
PREAMBLE
The goals of the Open Font License (OFL) are to stimulate worldwide
development of collaborative font projects, to support the font creation
efforts of academic and linguistic communities, and to provide a free and
open framework in which fonts may be shared and improved in partnership
with others.
The OFL allows the licensed fonts to be used, studied, modified and
redistributed freely as long as they are not sold by themselves. The
fonts, including any derivative works, can be bundled, embedded,
redistributed and/or sold with any software provided that any reserved
names are not used by derivative works. The fonts and derivatives,
however, cannot be released under any other type of license. The
requirement for fonts to remain under this license does not apply
to any document created using the fonts or their derivatives.
DEFINITIONS
"Font Software" refers to the set of files released by the Copyright
Holder(s) under this license and clearly marked as such. This may
include source files, build scripts and documentation.
"Reserved Font Name" refers to any names specified as such after the
copyright statement(s).
"Original Version" refers to the collection of Font Software components as
distributed by the Copyright Holder(s).
"Modified Version" refers to any derivative made by adding to, deleting,
or substituting -- in part or in whole -- any of the components of the
Original Version, by changing formats or by porting the Font Software to a
new environment.
"Author" refers to any designer, engineer, programmer, technical
writer or other person who contributed to the Font Software.
PERMISSION & CONDITIONS
Permission is hereby granted, free of charge, to any person obtaining
a copy of the Font Software, to use, study, copy, merge, embed, modify,
redistribute, and sell modified and unmodified copies of the Font
Software, subject to the following conditions:
1) Neither the Font Software nor any of its individual components,
in Original or Modified Versions, may be sold by itself.
2) Original or Modified Versions of the Font Software may be bundled,
redistributed and/or sold with any software, provided that each copy
contains the above copyright notice and this license. These can be
included either as stand-alone text files, human-readable headers or
in the appropriate machine-readable metadata fields within text or
binary files as long as those fields can be easily viewed by the user.
3) No Modified Version of the Font Software may use the Reserved Font
Name(s) unless explicit written permission is granted by the corresponding
Copyright Holder. This restriction only applies to the primary font name as
presented to the users.
4) The name(s) of the Copyright Holder(s) or the Author(s) of the Font
Software shall not be used to promote, endorse or advertise any
Modified Version, except to acknowledge the contribution(s) of the
Copyright Holder(s) and the Author(s) or with their explicit written
permission.
5) The Font Software, modified or unmodified, in part or in whole,
must be distributed entirely under this license, and must not be
distributed under any other license. The requirement for fonts to
remain under this license does not apply to any document created
using the Font Software.
TERMINATION
This license becomes null and void if any of the above conditions are
not met.
DISCLAIMER
THE FONT SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO ANY WARRANTIES OF
MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT
OF COPYRIGHT, PATENT, TRADEMARK, OR OTHER RIGHT. IN NO EVENT SHALL THE
COPYRIGHT HOLDER BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY,
INCLUDING ANY GENERAL, SPECIAL, INDIRECT, INCIDENTAL, OR CONSEQUENTIAL
DAMAGES, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
FROM, OUT OF THE USE OR INABILITY TO USE THE FONT SOFTWARE OR FROM
OTHER DEALINGS IN THE FONT SOFTWARE.
+184 -78
View File
@@ -161,11 +161,41 @@ h1 { font-size: 1.5rem; font-weight: 300; letter-spacing: -0.01em; margin-top: 0
padding: 1.5rem; margin: 0 auto;
}
.card.narrow { max-width: 24rem; }
/* A narrow card as a direct child of main means the page is a short form:
every sibling in the column (heading, flash, footer) must share that 24rem
measure, or they centre on 48rem while the card centres on 24rem and the
left edges drift by half the difference (12rem / 192px at a 16px root
measured). The column itself stays 64rem so the navigation does not move;
only the stack inside the column narrows. :has tracks the card rather than
a per-page class, so a new signed-in page that adds .card.narrow inherits
the alignment without a second stylesheet rule.
login/setup already shrink main to 24rem (below); this rule is then a no-op
relative to the parent. Do not replace that with :has alone: narrowing main
is what centres the signed-out block on the page when there is no nav. */
main:has(> .card.narrow) > * { max-width: 24rem; }
label { display: block; font-weight: 600; margin: 0.9rem 0 0.3rem; }
/* Checkbox rows (domain assignment, encrypt toggle) opt out of the block
label and full-width input rules above: the box and its caption sit on one
line. Without this the caption drops under a stretched checkbox. */
label.check {
display: flex; align-items: center; gap: 0.5rem; margin: 0.45rem 0 0; font-weight: 600;
cursor: pointer;
}
label.check input { width: auto; margin: 0; padding: 0; border: 0; border-radius: 0; background: none; }
input {
width: 100%; padding: 0.55rem 0.7rem; font-family: inherit; font-size: 1rem;
border: 1px solid var(--control-border); border-radius: 5px; background: var(--input-bg); color: inherit;
}
/* Domain assignment on the user form: a bordered group under Role, not another
stacked field label. Legend replaces the usual label; the muted line is the
hint that used to sit where a field caption would. */
fieldset {
margin: 0.9rem 0 0; padding: 0.55rem 0.85rem 0.85rem;
border: 1px solid var(--control-border); border-radius: 5px;
}
fieldset legend { padding: 0 0.25rem; font-weight: 600; }
fieldset > .muted { margin: 0.15rem 0 0.35rem; font-size: 0.85rem; font-weight: 400; }
fieldset > label.check:first-of-type { margin-top: 0.15rem; }
/* One vocabulary for actions. Anything that performs an action looks like a
button: a <button>, or an <a> carrying .btn/.danger where the action is a
plain navigation (the delete confirmation page, the full queue view).
@@ -181,6 +211,14 @@ button, a.btn, a.danger {
color: var(--on-accent); background: var(--accent-fill); border: 0; border-radius: 5px; cursor: pointer;
}
button:hover, a.btn:hover { background: var(--accent-fill-hover); }
button:disabled, button.danger:disabled {
opacity: 0.45; cursor: not-allowed;
}
button:disabled:hover { background: var(--accent-fill); }
button.danger:disabled:hover { background: var(--danger-fill); }
select:disabled {
opacity: 0.65; cursor: not-allowed; background: var(--code-bg);
}
.error { color: var(--danger-fg); margin: 0.6rem 0 0; font-weight: 600; }
.muted { color: var(--muted); }
/* Links take the accent. The panel never set a colour here and ran on the
@@ -200,15 +238,18 @@ form.inline { display: inline; margin: 0; }
left, three alignments on a page with four elements. Narrowing the column to
the card's own width makes the three line up and puts the block as a whole in
the middle of the page.
By page name and not by "the page holds a narrow card", which is what this
briefly was. Settings and the user form hold one too, and narrowing their
column moves the navigation sideways with it the shell centres the column
and the page as a pair, so 296px of it, measured. Their heading and card are
still not aligned with each other; that is worth fixing on its own terms and
not by making every page's chrome move. */
By page name and not by "the page holds a narrow card": signed-in pages that
hold a narrow card keep the 64rem column (see main:has(> .card.narrow)
above) so the navigation stays put; only these two pages have no nav and
need the column itself narrowed. */
main.page-login, main.page-setup { max-width: 24rem; }
.card + .card { margin-top: 1.2rem; }
.flash { background: var(--flash-bg); border: 1px solid var(--flash-border); color: var(--flash-fg); padding: 0.7rem 1rem; border-radius: 6px; margin-bottom: 1.2rem; }
/* RateLimitErr renders as .flash.error: a validation failure, not a success
notice. Without this rule it inherited the success surface (green) and only
.error's text colour, so a rejected rate limit read as good news in red
text. Same box, danger palette. */
.flash.error { background: var(--danger-bg); border-color: var(--danger-border); color: var(--danger-fg); }
table { width: 100%; border-collapse: collapse; }
/* A table column is at least as wide as the longest unbreakable run inside it,
and the panel's tables are full of runs with nothing to break on: email
@@ -223,9 +264,9 @@ th, td { text-align: left; padding: 0.5rem 0.4rem; border-bottom: 1px solid var(
td.time, td.status { white-space: nowrap; }
/* Column headings are set in the mono face: they are labels for machine data
rather than prose, and the wider tracking a small monospaced capital wants
also holds them apart from the first row of values. Same treatment on the
other two small all-caps labels in the panel (.sections-title, .fact-label),
so a heading of that size reads as one thing wherever it appears. */
also holds them apart from the first row of values. Same treatment on
.fact-label, so a heading of that size reads as one thing wherever it
appears. */
th {
font-family: var(--font-mono); font-size: 0.75rem; font-weight: 500;
text-transform: uppercase; letter-spacing: 0.08em; color: var(--muted);
@@ -293,23 +334,19 @@ button.danger:hover, a.danger:hover { background: var(--danger-fill-hover); }
.app .actions { display: flex; flex-wrap: wrap; gap: 0.4rem; margin-top: 0.7rem; }
/* Mode and rate limit open under the whole row rather than inside it. As
<details> the panel sat where its summary was, so opening one cut the row of
four controls in half and pushed the rest below a block of fields the
buttons moved every time a panel opened or closed. The panels are therefore
the last children of the row, and the checkbox that opens each one is
visually hidden earlier in the row with its label drawn as the button. The
checkbox stays in the tab order and keeps its focus ring on the label, so it
works from the keyboard, and being pure CSS it also works with JavaScript
blocked, as <details> did. */
controls in half and pushed the rest below a block of fields the buttons
moved every time a panel opened or closed. The panel is therefore the last
child of the row, and the checkbox that opens it is visually hidden earlier
in the row with its label drawn as the button. The checkbox stays in the tab
order and keeps its focus ring on the label, so it works from the keyboard,
and being pure CSS it also works with JavaScript blocked, as <details> did. */
.app .actions > .panel-toggle {
position: absolute; width: 1px; height: 1px; margin: 0; opacity: 0; pointer-events: none;
}
.app .actions > .panel { display: none; flex: 1 0 100%; }
.app .actions > .t-mode:checked ~ .panel-mode,
.app .actions > .t-limit:checked ~ .panel-limit { display: block; }
.app .actions > .t-mode:checked ~ .for-mode,
.app .actions > .t-limit:checked ~ .for-limit { background: var(--surface-open-bg); }
.app .actions > .t-mode:focus-visible ~ .for-mode,
.app .actions > .t-limit:focus-visible ~ .for-limit { outline: 2px solid var(--accent-fill); outline-offset: 2px; }
.app .actions > .t-edit:checked ~ .panel-edit { display: block; }
.app .actions > .t-edit:checked ~ .for-edit { background: var(--surface-open-bg); }
.app .actions > .t-edit:focus-visible ~ .for-edit { outline: 2px solid var(--accent-fill); outline-offset: 2px; }
.panel form { margin-top: 0.6rem; }
/* A panel's own submit is a form button, not one of the controls in the row
above, so it takes back the spacing the compact .actions rule zeroes out:
@@ -326,6 +363,21 @@ button.danger:hover, a.danger:hover { background: var(--danger-fill-hover); }
}
.app .actions > .panel .panel-buttons button,
.app .actions > .panel .panel-buttons form { margin-top: 0; }
.app .actions > .panel .check-cols { margin-top: 0.6rem; }
.app .actions > .panel .check-col > form { margin-top: 0; }
.app .actions > .panel .check-col > .check-col-title { margin-top: 0; }
.app .actions > .panel .check-col > .muted { margin: 0; }
/* Address mode select ‖ trusted-IP field: same top offset and control height. */
.app .actions > .panel .check-col > .muted + form > select:first-of-type,
.app .actions > .panel .check-col > .muted + form > textarea[name="allowed_ips"] {
margin-top: 0.45rem;
min-height: 2.55rem;
box-sizing: border-box;
}
.app .actions > .panel .check-col > .muted + form > textarea[name="allowed_ips"] {
field-sizing: fixed;
overflow-y: auto;
}
.credential { border-color: var(--credential-border); background: var(--credential-bg); }
/* Panel navigation: rendered once from the layout, so it is present on every
authenticated page without each content template having to include it. */
@@ -334,11 +386,9 @@ button.danger:hover, a.danger:hover { background: var(--danger-fill-hover); }
738px the panel was wide, so as a bar it had to be split over two rows and
even then it cost the top of every page. Standing it up removes that: the
entries share one left edge to scan down, the session sits at the foot where
it is out of the way, and there is room between them for the current page's
own sections (.sections below), which is what makes the long pages navigable.
Sticky, so both lists stay in view while the page scrolls past them. The
layout template lists the blocks in the order they are drawn, so reading and
tab order follow the eye without a CSS `order`. */
it is out of the way. Sticky, so the list stays in view while the page
scrolls past it. The layout template lists the blocks in the order they are
drawn, so reading and tab order follow the eye without a CSS `order`. */
.nav {
position: sticky; top: 2rem; align-self: flex-start;
flex: none; width: 13.5rem;
@@ -356,6 +406,9 @@ button.danger:hover, a.danger:hover { background: var(--danger-fill-hover); }
of them. It may be long and there is nothing to break it on, so it is allowed
to break mid word rather than widen the column. */
.nav .session .muted { padding: 0 0.6rem; font-size: 0.85rem; overflow-wrap: anywhere; }
.nav .session-user {
display: flex; align-items: center; gap: 0.5rem;
}
/* The mark is a link, but not one of the column's entries: it takes none of the
padding and rounding the entry rule below applies, so its own edge lines up
with the entries' icons rather than sitting half a step inside them. */
@@ -400,32 +453,7 @@ button.danger:hover, a.danger:hover { background: var(--danger-fill-hover); }
/* The icons draw in the entry's own colour, so the active entry's darker text,
a link's blue and Sign out's red all carry through without a rule apiece. */
.nav .icon { width: 1rem; height: 1rem; flex: none; }
/* The current page's own sections, listed under the page entries. The domain
page and the status page are eight and nine cards tall, and the only way to
the card you came for was to scroll past all the ones you did not. Only pages
that long carry an index it comes from the page's own "sections" template
(see layout.html), so a page with two cards renders nothing here. */
.sections {
display: flex; flex-direction: column; gap: 0.1rem;
padding-top: 0.75rem; border-top: 1px solid var(--border);
}
.sections-title {
margin: 0 0 0.25rem 0.6rem;
font-family: var(--font-mono);
font-size: 0.7rem; font-weight: 500; text-transform: uppercase; letter-spacing: 0.1em;
color: var(--muted);
}
/* Quieter and a step in from the page entries: this is an index of one page,
subordinate to the list of pages above it. */
.nav .sections a {
padding: 0.2rem 0.6rem 0.2rem 0.9rem;
font-size: 0.85rem; color: var(--muted); text-decoration: none;
}
.nav .sections a:hover { color: var(--fg); }
/* panel.js marks the section the page is scrolled to. Without JavaScript
nothing is marked and the list is still a working index. */
.nav .sections a.current { color: var(--fg); font-weight: 600; background: var(--nav-active-bg); }
/* Jumping to a card should not leave it touching the top edge of the window. */
/* Jumping to a card (in-page links) should not leave it touching the top edge. */
.card[id] { scroll-margin-top: 1rem; }
/* Below the width the two columns need (13.5rem of navigation, 1.75rem of gap
and the 48rem measure, plus the body's padding the measure, not the column
@@ -449,11 +477,10 @@ button.danger:hover, a.danger:hover { background: var(--danger-fill-hover); }
.nav .brand img { width: 110px; }
/* Each block keeps its own group of entries together and wraps as one; the
rules that separated the blocks vertically become the space between them. */
.nav .links, .nav .session, .nav .sections {
.nav .links, .nav .session {
flex-direction: row; flex-wrap: wrap; align-items: center;
gap: 0.2rem 0.6rem; padding-top: 0; border-top: 0;
}
.sections-title { margin: 0; }
}
/* Status badges: one vocabulary (ok/warn/error/unknown) shared by the server
status page and the per-domain DNS checks, so a colour means the same thing
@@ -464,16 +491,14 @@ button.danger:hover, a.danger:hover { background: var(--danger-fill-hover); }
the pill's own emphasis dropped with the pill, since the colour already
says how loud the badge is. */
/* line-height 1 and the padding doing the centring, rather than the 1.5 the
badge would inherit from the body. In a tall line box a word of lowercase
with no descender ok, warn, unknown sits above the optical centre of the
space around it, and inside a bordered box that reads as text stuck to the
top. With the box hugging the line, the padding is what centres the word, and
it is uneven on purpose: the extra tenth at the top pays for the descender
space every line box reserves below the baseline and these words rarely use.
The one that does use it, queued, keeps the same box the tail hangs into
the bottom padding rather than growing the badge. */
badge would inherit from the body. IBM Plex Mono sits a shade low in its em
square, so equal padding leaves "ok" / "warn" sunk in the box (and visibly
below the heading or label it sits next to). The bottom padding is therefore
heavier than the top enough to lift the word to the optical centre. A
top-heavy pad was tried earlier and made the sink worse. Words with a
descender (queued) hang into that bottom pad rather than growing the badge. */
.st {
display: inline-block; padding: 0.28rem 0.45rem 0.2rem; border-radius: 4px;
display: inline-block; padding: 0.14rem 0.45rem 0.28rem; border-radius: 4px;
font-family: var(--font-mono); font-size: 0.78rem; font-weight: 500; line-height: 1;
letter-spacing: 0.02em;
vertical-align: middle; border: 1px solid transparent;
@@ -501,8 +526,20 @@ meter { width: 5rem; height: 0.7rem; vertical-align: middle; margin-right: 0.4re
beside them would otherwise win the width and leave "Memory" broken across
two lines. */
.metric { white-space: nowrap; }
/* Copy sits at the top of the value right for a long DKIM key. The row's
.code uses the compact button's vertical padding so a one-line value is the
same height as Copy and does not leave the button hanging under an empty
band of padding. */
.code-row { display: flex; align-items: flex-start; gap: 0.5rem; }
.code-row .code { flex: 1; min-width: 0; }
.code-row .code {
flex: 1; min-width: 0; margin-top: 0;
padding-top: 0.45rem; padding-bottom: 0.45rem;
}
/* A short field with its submit on the same row (Add domain). The button
keeps the card's usual filled look; its top margin is for stacked forms. */
.input-row { display: flex; align-items: stretch; gap: 0.5rem; }
.input-row input { flex: 1; min-width: 0; }
.input-row button { margin-top: 0; flex: none; white-space: nowrap; }
/* Two cards abreast, on the one page that has a pair of them worth reading
together (a delivery's facts beside its history). auto-fit with a minimum
@@ -513,12 +550,85 @@ meter { width: 5rem; height: 0.7rem; vertical-align: middle; margin-right: 0.4re
display: grid; grid-template-columns: repeat(auto-fit, minmax(22rem, 1fr));
gap: 1.2rem; margin-top: 1.2rem;
}
/* Cards use margin: 0 auto so a lone card centres in the column. Inside the
grid those auto side margins absorb free space and shrink each card to its
content instead of filling the 1fr track so cancel them here, and allow
the cell to shrink below the card's intrinsic minimum when the column is
tight. */
.split > .card { margin-left: 0; margin-right: 0; min-width: 0; }
/* The gap already spaces the cards inside the grid; the general .card + .card
rule would add a second gap's worth on top of it, and only to the second
card, which in one column reads as an uneven stack. The blocks around the
grid keep their spacing from it instead. */
.split > .card + .card { margin-top: 0; }
.split + .card { margin-top: 1.2rem; }
.card + .split { margin-top: 1.2rem; }
/* Configuration sits outside the polled fragment; without this it would
touch the last .split inside #status-body (a nested .split is not a sibling). */
#status-body + .card { margin-top: 1.2rem; }
/* Peer checks or paired forms inside one card (domain DNS status: two rows of
two; domain settings; application edit). Same auto-fit idea as .split, so a
wide card gets two columns not three and falls to one when narrow. */
.check-cols {
display: grid; grid-template-columns: repeat(auto-fit, minmax(22rem, 1fr));
gap: 1rem 1.2rem; margin-top: 1rem;
}
.check-col { min-width: 0; }
.check-col > label:first-child { margin-top: 0; }
.check-col-title {
margin: 1rem 0 0.35rem; font-size: 1.05rem; font-weight: 600;
}
.card > h2 + .check-col-title { margin-top: 0.55rem; }
/* Domain settings pairs DMARC rate limit: share four row tracks so titles,
leads, fields and action buttons line up across columns. */
@supports (grid-template-rows: subgrid) {
.check-cols-rows {
grid-template-rows: auto auto 1fr auto;
}
.check-cols-rows > .check-col {
display: grid;
grid-template-rows: subgrid;
grid-row: span 4;
}
}
.check-cols-rows > .check-col > .check-col-title { margin-top: 0; }
.check-cols-rows > .check-col > .muted { margin: 0; }
.check-col-fields { min-width: 0; }
.check-col-fields > form { margin: 0; }
.check-col-fields .field-pair { margin-top: 0; }
.check-col-fields .field-pair > div > label { margin-top: 0.9rem; }
.check-col-actions {
display: flex; flex-wrap: wrap; align-items: center; gap: 0.6rem 0.75rem;
margin-top: 0;
}
.check-col-actions > button { margin-top: 1.2rem; }
.check-col-actions > form.inline { margin: 0; }
.check-col-actions > form.inline > button { margin-top: 1.2rem; }
/* Two short fields on one row (rate-limit message count window). Falls to
one column when the parent is too narrow for both. */
.field-pair {
display: grid; grid-template-columns: repeat(auto-fit, minmax(8rem, 1fr));
gap: 0 1rem; margin-top: 0.45rem;
}
.field-pair > div { min-width: 0; }
.field-pair label { margin-top: 0.45rem; }
/* Hint under a field-pair sits tight above the form's submit button (the
global button margin-top would leave a blank line). */
.field-pair + p.muted { margin: 0.45rem 0 0; }
.field-pair + p.muted + button { margin-top: 0.45rem; }
/* Host / name beside a DNS Type that is always TXT size the type column to
that token rather than giving it half the row. Vertical padding matches
.code-row .code so TXT is the same height as Host when Copy sits beside it. */
.field-pair.host-type {
grid-template-columns: minmax(0, 1fr) auto;
}
.field-pair.host-type .code {
padding-top: 0.45rem; padding-bottom: 0.45rem;
}
.field-type { width: max-content; }
.field-type .code {
width: fit-content; min-width: 2.75rem; text-align: center; box-sizing: border-box;
}
/* The subject heads a delivery's page, and it is the one heading in the panel
whose text we do not control: it may be a hundred characters with nothing to
break on. It wraps to as many lines as it needs (this is the page's name, not
@@ -608,7 +718,7 @@ button.copy, .actions button, .actions > label.toggle, .actions a.danger, .nav b
}
button.copy:hover, .actions button:hover, .actions > label.toggle:hover,
.actions a.danger:hover, .nav button:hover { background: var(--surface-bg-hover); }
button.copy { flex: none; margin-top: 0.3rem; }
button.copy { flex: none; }
.actions button.danger, .actions a.danger, .nav button.danger {
color: var(--danger-fg); background: var(--danger-bg); border-color: var(--danger-border);
}
@@ -626,19 +736,15 @@ button.copy { flex: none; margin-top: 0.3rem; }
}
/* The optional "encrypt this download" block on the backup and export forms.
Its label is the one checkbox in the panel, so it opts out of the
block-level label rule above and sits on one line with its box; the fields it
reveals are indented under it to read as its consequence rather than as three
more fields of the form. panel.js hides the inner block until the box is
ticked (and empties it when unticked); without JavaScript everything stays
visible, which the server handles identically. The import form reuses the
same indented .encrypt-fields look for its password field, but reveals it
by file extension instead of a checkbox (see panel.js). */
The checkbox uses the shared label.check rule above; the fields it reveals
are indented under it to read as its consequence rather than as three more
fields of the form. panel.js hides the inner block until the box is ticked
(and empties it when unticked); without JavaScript everything stays visible,
which the server handles identically. The import form reuses the same
indented .encrypt-fields look for its password field, but reveals it by
file extension instead of a checkbox (see panel.js). */
.encrypt { margin-top: 1.2rem; }
.encrypt label.check {
display: flex; align-items: center; gap: 0.5rem; margin: 0; font-weight: 600;
}
.encrypt label.check input { width: auto; margin: 0; }
.encrypt label.check { margin: 0; }
.encrypt-fields {
margin-top: 1.2rem; margin-left: 1.6rem; padding-left: 0.9rem;
border-left: 2px solid var(--border);
+57 -138
View File
@@ -46,162 +46,81 @@
}
});
// --- Address list shown only in list mode -----------------------------
// The "Addresses" field applies to list mode only; in wildcard mode the
// server ignores it, so hiding it removes a field that does nothing. The
// toggle runs on load too, because the edit form of an existing application
// may already be set to list mode.
function syncAddressField(select) {
var form = select.closest("form");
var field = form && form.querySelector("[data-addresses]");
if (!field) {
// --- Conditional field visibility ------------------------------------
// Several forms hide a block until a select, checkbox or file input says it
// applies. One rule table drives them all so the five near-identical helpers
// do not drift.
var showWhenRules = [
{
match: "select[data-list-mode]",
target: "[data-addresses]",
visible: function (el) { return el.value === el.dataset.listMode; }
},
{
match: "select[data-custom-mode]",
target: "[data-custom-address]",
visible: function (el) { return el.value === el.dataset.customMode; }
},
{
match: "select[data-global-role]",
target: "[data-domain-pick]",
visible: function (el) { return el.value !== el.dataset.globalRole; }
},
{
match: "input[data-encrypt-toggle]",
target: "[data-encrypt-fields]",
visible: function (el) { return el.checked; },
clearWhenHidden: true
},
{
match: "input[data-import-file]",
target: "[data-import-password-fields]",
visible: function (el) {
var name = (el.files && el.files[0] && el.files[0].name || "").toLowerCase();
return name !== "" && !/\.json$/.test(name);
},
clearWhenHidden: true
}
];
function syncShowWhen(control) {
var form = control.closest("form");
if (!form) {
return;
}
// The mode values come from the server (store.AddressModeList), so the
// select carries the one that means "list" rather than this script
// hard-coding it.
field.hidden = select.value !== select.dataset.listMode;
for (var i = 0; i < showWhenRules.length; i++) {
var rule = showWhenRules[i];
if (!control.matches(rule.match)) {
continue;
}
function initAddressFields(root) {
root.querySelectorAll("select[data-list-mode]").forEach(function (select) {
syncAddressField(select);
select.addEventListener("change", function () {
syncAddressField(select);
});
});
}
// --- Encryption password fields shown only when asked for --------------
// The backup, export and import forms carry an optional password block. It
// is hidden until the checkbox next to it is ticked, and cleared when it is
// unticked, so a password typed and then abandoned is never submitted. With
// JavaScript blocked the block stays visible and the forms behave exactly as
// the server reads them: the checkbox alone decides whether encryption
// happens.
function syncEncryptFields(box) {
var form = box.closest("form");
var fields = form && form.querySelector("[data-encrypt-fields]");
if (!fields) {
var target = form.querySelector(rule.target);
if (!target) {
return;
}
fields.hidden = !box.checked;
if (!box.checked) {
fields.querySelectorAll("input").forEach(function (input) {
var show = rule.visible(control);
target.hidden = !show;
if (!show && rule.clearWhenHidden) {
target.querySelectorAll("input").forEach(function (input) {
input.value = "";
});
}
}
function initEncryptFields(root) {
root.querySelectorAll("input[data-encrypt-toggle]").forEach(function (box) {
syncEncryptFields(box);
box.addEventListener("change", function () {
syncEncryptFields(box);
});
});
}
// --- Import password field shown based on the chosen file's extension ---
// The domain-import file decides for itself whether it is encrypted (the
// server checks the envelope magic, not a checkbox), so the panel offers
// the password field the same way: reveal it for a .spde file, hide and
// clear it for a plain .json one. With no file chosen yet there is nothing
// to ask a password for, so the field stays hidden until a file names it.
// An unrecognised name leaves the field visible rather than guessing wrong
// and hiding a password the file needs.
function syncImportPasswordField(input) {
var form = input.closest("form");
var fields = form && form.querySelector("[data-import-password-fields]");
if (!fields) {
return;
}
var name = (input.files && input.files[0] && input.files[0].name || "").toLowerCase();
var hide = name === "" || /\.json$/.test(name);
fields.hidden = hide;
if (hide) {
fields.querySelectorAll("input").forEach(function (pw) {
pw.value = "";
});
}
}
function initImportPasswordField(root) {
root.querySelectorAll("input[data-import-file]").forEach(function (input) {
syncImportPasswordField(input);
input.addEventListener("change", function () {
syncImportPasswordField(input);
function initShowWhen(root) {
showWhenRules.forEach(function (rule) {
root.querySelectorAll(rule.match).forEach(function (control) {
syncShowWhen(control);
control.addEventListener("change", function () {
syncShowWhen(control);
});
});
}
// --- Section index follows the page -----------------------------------
// The long pages list their own sections in the navigation column (the
// "sections" template). Marking the one currently in view turns that list
// from an index into a position, which is the whole point of it on a page
// nine cards tall. The links work without any of this; only the highlight
// depends on it.
//
// Each pass looks its targets up by id rather than holding on to elements
// found once: the status page replaces its cards wholesale every five
// seconds (adaptive polling on #status-body), and anything remembered here would be
// measuring boxes that had left the document.
var sectionLinks = [];
function markCurrentSection() {
var current = null;
sectionLinks.forEach(function (link) {
var target = document.getElementById(link.hash.slice(1));
// The section in view is the last one whose top has passed the reading
// line; the links are in document order, so the last match wins.
if (target && target.getBoundingClientRect().top <= 100) {
current = link;
}
});
if (window.innerHeight + window.scrollY >= document.documentElement.scrollHeight - 2) {
// At the foot of the page there is no scroll left to bring the last
// cards up to the reading line, so without this they could never be
// marked however far down you are — and the last card of the domain
// page is the one that deletes it.
current = sectionLinks[sectionLinks.length - 1];
} else if (!current) {
// Above the first heading nothing has been passed yet, and the page is
// still on its first section.
current = sectionLinks[0];
}
sectionLinks.forEach(function (link) {
link.classList.toggle("current", link === current);
});
}
function initSectionIndex() {
sectionLinks = Array.prototype.slice.call(
document.querySelectorAll(".sections a[href^='#']")
);
if (!sectionLinks.length) {
return;
}
var pending = false;
// Scroll fires far more often than the highlight can change, so the work
// is collapsed onto the next frame.
window.addEventListener("scroll", function () {
if (pending) {
return;
}
pending = true;
window.requestAnimationFrame(function () {
pending = false;
markCurrentSection();
});
}, { passive: true });
markCurrentSection();
}
document.addEventListener("DOMContentLoaded", function () {
initAddressFields(document);
initEncryptFields(document);
initImportPasswordField(document);
initSectionIndex();
initShowWhen(document);
});
// --- Adaptive monitoring polling ---------------------------------------
+27
View File
@@ -1,8 +1,10 @@
package view
import (
"io/fs"
"net/http"
"net/http/httptest"
"strings"
"testing"
)
@@ -26,6 +28,7 @@ func TestStaticAssetsCarryETag(t *testing.T) {
// The fonts are the assets this matters most for: they are the largest
// thing the panel serves and the ones a browser is most willing to keep.
"ibm-plex-sans.woff2", "ibm-plex-mono-400.woff2", "ibm-plex-mono-600.woff2",
"OFL.txt",
} {
rec := serveStatic("/static/"+name, nil)
if rec.Code != http.StatusOK {
@@ -77,3 +80,27 @@ func TestStaticETagsAreContentDerived(t *testing.T) {
t.Errorf("favicon.png and favicon.svg share the ETag %s", png)
}
}
// OFL condition 2: the licence text must travel with the Font Software. The
// WOFF2 files are embedded; OFL.txt sits next to them so a copy of the panel
// (source tree, image, or /static/OFL.txt) always has it.
func TestOFLTravelsWithFonts(t *testing.T) {
b, err := fs.ReadFile(assetsFS, "static/OFL.txt")
if err != nil {
t.Fatalf("OFL.txt is not embedded next to the Plex WOFF2 files: %v", err)
}
body := string(b)
if !strings.Contains(body, `Reserved Font Name "Plex"`) {
t.Error("OFL.txt is missing the IBM Plex reserved-font-name notice")
}
if !strings.Contains(body, "SIL OPEN FONT LICENSE Version 1.1") {
t.Error("OFL.txt is missing the SIL OFL 1.1 text")
}
rec := serveStatic("/static/OFL.txt", nil)
if rec.Code != http.StatusOK {
t.Fatalf("GET /static/OFL.txt: status %d, want 200", rec.Code)
}
if !strings.Contains(rec.Body.String(), "SIL OPEN FONT LICENSE Version 1.1") {
t.Error("GET /static/OFL.txt did not serve the OFL text")
}
}
-70
View File
@@ -1,70 +0,0 @@
{{define "content"}}
<h1>Settings</h1>
{{if .Flash}}<div class="flash">{{.Flash}}</div>{{end}}
<div class="card narrow">
<h2>Panel credentials</h2>
<p class="muted">These are the credentials for this control panel only.
Applications keep their own logins and passwords, which are not affected.</p>
{{if .Error}}<p class="error">{{.Error}}</p>{{end}}
<form method="post" action="/account">
<label for="username">Username</label>
<input id="username" name="username" autocomplete="username"
autocapitalize="none" spellcheck="false" value="{{.FormUsername}}" required>
<label for="current_password">Current password</label>
<input id="current_password" name="current_password" type="password"
autocomplete="current-password" required>
<label for="new_password">New password</label>
<input id="new_password" name="new_password" type="password" autocomplete="new-password">
<label for="new_password_confirm">Confirm new password</label>
<input id="new_password_confirm" name="new_password_confirm" type="password" autocomplete="new-password">
{{if .ShowDMARC}}
<h3>DMARC aggregate reports</h3>
<p class="muted">Optional default <code>rua=</code> address for every sending
domain (can be overridden per domain). Use a mailbox on a domain that
receives inbound mail. SelfPost is send-only today; a future release will
be able to receive reports in the panel itself.</p>
<label for="dmarc_report_email">Default report address</label>
<input id="dmarc_report_email" name="dmarc_report_email" type="email"
autocomplete="email" autocapitalize="none" spellcheck="false"
value="{{.FormDMARCEmail}}" placeholder="reports@your-mail-domain.com">
{{if .FormDMARCEmail}}
<p class="muted">When <code>rua=</code> points at another domain, that hub
domain must publish a report-authorisation record so receivers will deliver
the XML aggregates.</p>
<label>Report authorization — host / name</label>
<div class="code-row">
<span class="code">{{.ReportAuthName}}</span>
<button type="button" class="copy">Copy</button>
</div>
<label>Report authorization — value</label>
<div class="code-row">
<span class="code">{{.ReportAuthExample}}</span>
<button type="button" class="copy">Copy</button>
</div>
{{if .ReportAuthDNS.Status}}
<label>Report authorization DNS <span class="st st-{{.ReportAuthDNS.Status}}">{{.ReportAuthDNS.Status}}</span></label>
<p class="{{if eq .ReportAuthDNS.Status "ok"}}muted{{else}}error{{end}}">{{.ReportAuthDNS.Detail}}</p>
{{if .ReportAuthDNS.Records}}<span class="code">{{range .ReportAuthDNS.Records}}{{.}}
{{end}}</span>{{end}}
{{end}}
{{end}}
{{end}}
<button type="submit">Save changes</button>
</form>
<p class="muted">Leave both new-password fields empty to change the username
{{if .ShowDMARC}}or DMARC address {{end}}only. Changing the password signs out every other session;
this one stays signed in.</p>
</div>
{{end}}
+3 -3
View File
@@ -8,14 +8,14 @@
<h2>Add a sending domain</h2>
<form method="post" action="/domains">
<label for="name">Domain</label>
<div class="input-row">
<input id="name" name="name" type="text" placeholder="example.com"
autocomplete="off" autocapitalize="none" spellcheck="false"
value="{{.FormName}}" autofocus required>
{{if .Error}}<p class="error">{{.Error}}</p>{{end}}
<button type="submit">Add domain</button>
</div>
{{if .Error}}<p class="error">{{.Error}}</p>{{end}}
</form>
<p class="muted">A DKIM key is generated for the domain; you then publish the
shown DNS record. Adding a domain does not create an application.</p>
</div>
{{end}}
@@ -1,5 +1,6 @@
{{define "deliveries_rows"}}
<div id="deliveries-rows" data-poll hx-get="/deliveries/rows?domain={{.FilterDomain}}&app={{.FilterApp}}&p={{.Page}}"
<div id="deliveries-rows" data-poll aria-live="polite"
hx-get="/deliveries/rows?domain={{.FilterDomain | urlquery}}&app={{.FilterApp | urlquery}}&p={{.Page}}"
hx-trigger="load" hx-swap="outerHTML">
{{if .Rows}}
<table>
@@ -14,15 +15,15 @@
<td>{{.To}}</td>
<td class="subject"><span title="{{.Subject}}">{{.Subject}}</span></td>
<td class="status"><span class="st st-{{.Level}}">{{.Status}}</span></td>
<td class="actions"><a href="/deliveries/{{.ID}}?domain={{$.FilterDomain}}&app={{$.FilterApp}}&p={{$.Page}}">Details</a></td>
<td class="actions"><a href="/deliveries/{{.ID}}?domain={{$.FilterDomain | urlquery}}&app={{$.FilterApp | urlquery}}&p={{$.Page}}">Details</a></td>
</tr>
{{end}}
</tbody>
</table>
<p class="muted">
Page {{.Page}}{{if .LastPage}} of {{.LastPage}}{{end}}
{{if .HasPrev}} &middot; <a href="/deliveries?domain={{.FilterDomain}}&app={{.FilterApp}}&p={{.PrevPage}}">&larr; Newer</a>{{end}}
{{if .HasNext}} &middot; <a href="/deliveries?domain={{.FilterDomain}}&app={{.FilterApp}}&p={{.NextPage}}">Older &rarr;</a>{{end}}
{{if .HasPrev}} &middot; <a href="/deliveries?domain={{.FilterDomain | urlquery}}&app={{.FilterApp | urlquery}}&p={{.PrevPage}}">&larr; Newer</a>{{end}}
{{if .HasNext}} &middot; <a href="/deliveries?domain={{.FilterDomain | urlquery}}&app={{.FilterApp | urlquery}}&p={{.NextPage}}">Older &rarr;</a>{{end}}
</p>
{{else}}
<p class="muted">No messages logged yet.</p>
+1 -1
View File
@@ -18,7 +18,7 @@
<span class="st st-{{.Level}}">{{.Row.Status}}</span>
</p>
<a class="back" href="{{.BackURL}}">&larr; Back to deliveries</a>
{{template "back_link" (back .BackURL "Back to deliveries")}}
{{/* The two columns: what was recorded on the left, in what order it happened
on the right. They are a pair — the facts are only worth reading against
@@ -1,7 +1,7 @@
{{define "content"}}
<h1>Delete {{.Domain.Name}}</h1>
<a class="back" href="/domains/{{.Domain.ID}}">&larr; Back to {{.Domain.Name}}</a>
{{template "back_link" (back (printf "/domains/%d" .Domain.ID) (printf "Back to %s" .Domain.Name))}}
<div class="card">
<h2>Confirm deletion</h2>
+248 -214
View File
@@ -1,7 +1,66 @@
{{/* Wide so the .split pairs fill the column rather than the 48rem reading
measure (same pattern as Status). */}}
{{define "wide"}}wide{{end}}
{{/* The Host/name ‖ Type field-pair repeats for every DNS record this page
shows (DKIM, SPF, DMARC, report authorization) in both the status card
and the publishable-record cards below it — only the host and whether it
carries a Copy button change. Two variants rather than one templated
Copy flag: the DNS status card never offers Copy (its host is derived,
not something to paste), the record cards always do. */}}
{{define "host_type"}}
<div class="field-pair host-type">
<div>
<label>Host / name</label>
<span class="code">{{.}}</span>
</div>
<div class="field-type">
<label>Type</label>
<span class="code">TXT</span>
</div>
</div>
{{end}}
{{define "host_type_copy"}}
<div class="field-pair host-type">
<div>
<label>Host / name</label>
<div class="code-row">
<span class="code">{{.}}</span>
<button type="button" class="copy">Copy</button>
</div>
</div>
<div class="field-type">
<label>Type</label>
<span class="code">TXT</span>
</div>
</div>
{{end}}
{{/* Value partials pair with the two host_type variants above: field_values
is the DNS status card's raw record dump (no Copy — those values are for
comparison, not for pasting), field_value is the single publishable
value on the record cards below (always has Copy). Callers still guard
the empty case, since "no records yet" and "one blank record" read
differently. */}}
{{define "field_values"}}
<label>Value</label>
<span class="code">{{range .}}{{.}}
{{end}}</span>
{{end}}
{{define "field_value"}}
<label>Value</label>
<div class="code-row">
<span class="code">{{.}}</span>
<button type="button" class="copy">Copy</button>
</div>
{{end}}
{{define "content"}}
<h1>{{.Domain.Name}}</h1>
<a class="back" href="/domains">&larr; All domains</a>
{{template "back_link" (back "/domains" "All domains")}}
{{if .Flash}}<div class="flash">{{.Flash}}</div>{{end}}
{{if .RateLimitErr}}<div class="flash error">{{.RateLimitErr}}</div>{{end}}
@@ -9,8 +68,8 @@
{{if .NewCred}}
<div class="card credential" id="new-credential">
<h2>New application password</h2>
<p class="muted">This password is shown <strong>once only</strong> and is not
stored. Copy it now — if it is lost, regenerate a new one.</p>
<p class="muted">Shown <strong>once only</strong> and not stored. Copy it now
— if it is lost, regenerate a new one.</p>
<label>Login</label>
<div class="code-row">
<span class="code">{{.NewCred.Login}}</span>
@@ -24,77 +83,87 @@
</div>
{{end}}
<div class="card" id="dkim">
<h2>DKIM DNS record</h2>
<p class="muted">Publish this TXT record in the DNS for <strong>{{.Domain.Name}}</strong>.
It is not a secret and can be viewed at any time.</p>
{{/* Two rows of two checks (.check-cols). */}}
<div class="card" id="dns-status">
<h2>DNS status <span class="st st-{{.DNS.Overall}}">{{.DNS.Overall}}</span></h2>
<p class="muted">Cached a few minutes — use <em>Re-check</em> after
publishing.</p>
<label>Host / name</label>
<div class="code-row">
<span class="code">{{.Record.Name}}</span>
<button type="button" class="copy">Copy</button>
<div class="check-cols">
<div class="check-col">
<label>DKIM <span class="st st-{{.DNS.DKIM.Status}}">{{.DNS.DKIM.Status}}</span></label>
{{template "host_type" .Record.Name}}
{{if .DNS.DKIM.Records}}{{template "field_values" .DNS.DKIM.Records}}{{end}}
{{if ne .DNS.DKIM.Status "ok"}}
<p class="{{if eq .DNS.DKIM.Status "unknown"}}muted{{else}}error{{end}}">{{.DNS.DKIM.Detail}}</p>
{{end}}
</div>
<label>Type</label>
<span class="code">TXT</span>
<label>Value</label>
<div class="code-row">
<span class="code">{{.Record.Value}}</span>
<button type="button" class="copy">Copy</button>
<div class="check-col">
<label>SPF <span class="st st-{{.DNS.SPF.Status}}">{{.DNS.SPF.Status}}</span></label>
{{template "host_type" .Domain.Name}}
{{if .DNS.SPF.Records}}{{template "field_values" .DNS.SPF.Records}}{{end}}
{{if ne .DNS.SPF.Status "ok"}}
<p class="{{if eq .DNS.SPF.Status "unknown"}}muted{{else}}error{{end}}">{{.DNS.SPF.Detail}}</p>
{{end}}
<p class="muted">Shallow check: literal address only, no <code>include:</code> /
<code>redirect=</code>.</p>
</div>
<p class="muted">Mail is signed with selector <strong>{{.Domain.DKIMSelector}}</strong>.</p>
<div class="check-col">
<label>DMARC <span class="st st-{{.DNS.DMARC.Status}}">{{.DNS.DMARC.Status}}</span></label>
{{template "host_type" .DMARCName}}
{{if .DNS.DMARC.Records}}{{template "field_values" .DNS.DMARC.Records}}{{end}}
{{if ne .DNS.DMARC.Status "ok"}}
<p class="{{if eq .DNS.DMARC.Status "unknown"}}muted{{else}}error{{end}}">{{.DNS.DMARC.Detail}}</p>
{{else}}
<p class="muted">{{.DNS.DMARC.Detail}}</p>
{{end}}
</div>
<div class="check-col">
{{if .DNS.DMARCReportAuth.Status}}
<label>Report authorization <span class="st st-{{.DNS.DMARCReportAuth.Status}}">{{.DNS.DMARCReportAuth.Status}}</span></label>
{{template "host_type" .ReportAuthName}}
{{if .DNS.DMARCReportAuth.Records}}{{template "field_values" .DNS.DMARCReportAuth.Records}}{{end}}
<p class="{{if eq .DNS.DMARCReportAuth.Status "ok"}}muted{{else}}error{{end}}">{{.DNS.DMARCReportAuth.Detail}}</p>
{{else}}
<label>Report authorization</label>
<p class="muted">Not required (no external <code>rua=</code>).</p>
{{end}}
</div>
</div>
<form class="inline" method="post" action="/domains/{{.Domain.ID}}/dns-recheck">
<button type="submit">Re-check</button>
</form>
</div>
<div class="card" id="spf-dmarc">
<h2>SPF and DMARC records</h2>
<p class="muted">These two are not generated the way the DKIM record above is —
they are policy, and the domain may already publish an SPF record for other
senders. SelfPost is a send-only relay — most operators have no inbox on the
sending domain, so the suggested DMARC record omits <code>rua=</code> unless a
report address is configured below or in <a href="/account">Settings</a>.
Publish both as TXT records.</p>
<div class="split">
<div class="card" id="dkim-spf">
<h2>DKIM and SPF records</h2>
<label>SPF — host / name</label>
<div class="code-row">
<span class="code">{{.Domain.Name}}</span>
<button type="button" class="copy">Copy</button>
<p class="check-col-title">DKIM</p>
{{template "host_type_copy" .Record.Name}}
{{template "field_value" .Record.Value}}
<p class="muted">Not a secret. Signed with selector
<strong>{{.Domain.DKIMSelector}}</strong>.</p>
<p class="check-col-title">SPF</p>
{{template "host_type_copy" .Domain.Name}}
{{template "field_value" .SPFExample}}
<p class="muted">Merge into an existing SPF if the domain already has one —
do not publish a second record.</p>
</div>
<label>SPF — value</label>
<div class="code-row">
<span class="code">{{.SPFExample}}</span>
<button type="button" class="copy">Copy</button>
</div>
<div class="card" id="dmarc">
<h2>DMARC record</h2>
<p class="muted">A domain may have only one SPF record. If it already has one,
do not add a second — add this server's mechanism to the existing record
instead, before its <code>all</code> term.</p>
{{template "host_type_copy" .DMARCName}}
<form method="post" action="/domains/{{.Domain.ID}}/dmarc">
<label for="dmarc_rua_mode">DMARC aggregate reports (rua=)</label>
<select id="dmarc_rua_mode" name="dmarc_rua_mode">
<option value="inherit"{{if eq .DMARCRuaMode "inherit"}} selected{{end}}>Same as Settings{{if .ProfileDMARCEmail}} ({{.ProfileDMARCEmail}}){{end}}</option>
<option value="none"{{if eq .DMARCRuaMode "none"}} selected{{end}}>No aggregate reports</option>
<option value="custom"{{if eq .DMARCRuaMode "custom"}} selected{{end}}>Custom address</option>
</select>
<label for="dmarc_rua_email">Custom report address</label>
<input id="dmarc_rua_email" name="dmarc_rua_email" type="email"
autocapitalize="none" spellcheck="false" value="{{.DMARCRuaCustom}}"
placeholder="reports@your-mail-domain.com">
<button type="submit">Save DMARC report settings</button>
</form>
<label>DMARC — host / name</label>
<div class="code-row">
<span class="code">{{.DMARCName}}</span>
<button type="button" class="copy">Copy</button>
</div>
<label>DMARC — value{{if eq .DMARCSource "settings"}} <span class="muted">(from Settings)</span>{{else if eq .DMARCSource "custom"}} <span class="muted">(custom)</span>{{else if eq .DMARCSource "none"}} <span class="muted">(no reports)</span>{{end}}</label>
<label>Value{{if eq .DMARCSource "settings"}} <span class="muted">(from Settings)</span>{{else if eq .DMARCSource "custom"}} <span class="muted">(custom)</span>{{else if eq .DMARCSource "none"}} <span class="muted">(no reports)</span>{{end}}</label>
<div class="code-row">
<span class="code">{{.DMARCExample}}</span>
<button type="button" class="copy">Copy</button>
@@ -102,73 +171,26 @@
{{if .SameDomainRUA}}
<p class="error">The report address is on this sending domain. SelfPost does
not receive inbound mail — use a mailbox elsewhere or wait for in-panel report
reception in a future release.</p>
not receive inbound mail — use a mailbox on another domain.</p>
{{end}}
{{if .NeedsReportAuth}}
<label>Report authorization — host / name</label>
<div class="code-row">
<span class="code">{{.ReportAuthName}}</span>
<button type="button" class="copy">Copy</button>
</div>
<label>Report authorization — value</label>
<div class="code-row">
<span class="code">{{.ReportAuthValue}}</span>
<button type="button" class="copy">Copy</button>
</div>
<p class="check-col-title">Report authorization</p>
{{template "host_type_copy" .ReportAuthName}}
{{template "field_value" .ReportAuthValue}}
{{end}}
<p class="muted"><code>p=none</code> changes nothing about delivery. Omit
<code>rua=</code> on a send-only domain, or point it at a mailbox that
receives inbound mail. Tighten to <code>p=quarantine</code> and then
<code>p=reject</code> once aggregate reports look clean.</p>
<p class="muted"><code>p=none</code> does not affect delivery. Tighten to
<code>p=quarantine</code> then <code>p=reject</code> once reports look clean.
Report address is set under <a href="#domain-settings">Domain settings</a>.</p>
</div>
</div>
<div class="card" id="dns-status">
<h2>DNS status <span class="st st-{{.DNS.Overall}}">{{.DNS.Overall}}</span></h2>
<p class="muted">What DNS publishes for <strong>{{.Domain.Name}}</strong> right
now, checked against the key this server signs with. Results are cached for a
few minutes — after publishing a record, use <em>Re-check</em>.</p>
<label>DKIM <span class="st st-{{.DNS.DKIM.Status}}">{{.DNS.DKIM.Status}}</span></label>
<p class="{{if eq .DNS.DKIM.Status "ok"}}muted{{else}}error{{end}}">{{.DNS.DKIM.Detail}}</p>
{{if .DNS.DKIM.Records}}<span class="code">{{range .DNS.DKIM.Records}}{{.}}
{{end}}</span>{{end}}
<label>SPF <span class="st st-{{.DNS.SPF.Status}}">{{.DNS.SPF.Status}}</span></label>
<p class="{{if eq .DNS.SPF.Status "ok"}}muted{{else}}error{{end}}">{{.DNS.SPF.Detail}}</p>
{{if .DNS.SPF.Records}}<span class="code">{{range .DNS.SPF.Records}}{{.}}
{{end}}</span>{{end}}
<p class="muted">The SPF check is deliberately shallow: it looks for a
mechanism that literally covers this server's address and does not follow
<code>include:</code> or <code>redirect=</code>, so a record that authorises
the server through an include is reported as “cannot tell”, not as a failure.</p>
<label>DMARC <span class="st st-{{.DNS.DMARC.Status}}">{{.DNS.DMARC.Status}}</span></label>
<p class="{{if eq .DNS.DMARC.Status "ok"}}muted{{else}}error{{end}}">{{.DNS.DMARC.Detail}}</p>
{{if .DNS.DMARC.Records}}<span class="code">{{range .DNS.DMARC.Records}}{{.}}
{{end}}</span>{{end}}
{{if .DNS.DMARCReportAuth.Status}}
<label>Report authorization <span class="st st-{{.DNS.DMARCReportAuth.Status}}">{{.DNS.DMARCReportAuth.Status}}</span></label>
<p class="{{if eq .DNS.DMARCReportAuth.Status "ok"}}muted{{else}}error{{end}}">{{.DNS.DMARCReportAuth.Detail}}</p>
{{if .DNS.DMARCReportAuth.Records}}<span class="code">{{range .DNS.DMARCReportAuth.Records}}{{.}}
{{end}}</span>{{end}}
{{end}}
<form class="inline" method="post" action="/domains/{{.Domain.ID}}/dns-recheck">
<button type="submit">Re-check</button>
</form>
</div>
<div class="card" id="settings">
<h2>Sending server settings</h2>
<p class="muted">Point the mail client or script at these settings and
authenticate with an application login and password from the
<strong>Applications</strong> section below. They are the same for every
domain on this server.</p>
<div class="split">
<div class="card" id="connection">
<h2>Connection settings</h2>
<p class="muted">Same for every domain. Authenticate with an application
login from below.</p>
<label>Server</label>
<div class="code-row">
@@ -180,15 +202,13 @@
<span class="code">465 — SSL/TLS (implicit){{if .SubmissionEnabled}}
587 — STARTTLS (submission){{end}}</span>
<p class="muted">Authentication is required on every port. The username is the
application's login (see the table below) and the password is the one shown
once when that application was created or its password regenerated — if it was
lost, generate a new one.</p>
</div>
<p class="muted">Auth required on every port. The password is shown once at
create or regenerate.</p>
</div>
{{/* Create form above the list, the same order the domains page uses for
"Add a sending domain" above "Domains". */}}
<div class="card" id="add-application">
{{/* Create form beside connection settings, mirroring "Add a sending domain"
above the domains list. */}}
<div class="card" id="add-application">
<h2>Add an application</h2>
<form method="post" action="/domains/{{.Domain.ID}}/applications">
<label for="login">Login</label>
@@ -211,15 +231,15 @@
{{if .Error}}<p class="error">{{.Error}}</p>{{end}}
<button type="submit">Create application</button>
</form>
<p class="muted">A strong password is generated and shown once. The login must
be unique across all domains and may contain letters, digits, '.', '-' and '_'.</p>
<p class="muted">Password shown once. Login unique across domains; letters,
digits, '.', '-' and '_'.</p>
</div>
</div>
<div class="card" id="applications">
<h2>Applications</h2>
<p class="muted">Each application is a SASL login/password an app or script
uses to send mail as this domain. A login may send from any address of the
domain (<em>wildcard</em>) or only from a fixed list of addresses.</p>
<p class="muted">SASL logins for this domain — wildcard (*@domain) or a fixed
address list.</p>
{{if .Apps}}
<ul class="apps">
@@ -230,17 +250,12 @@
{{if eq .AddressMode $.Wildcard}}Any address of the domain — *@{{$.Domain.Name}}
{{else}}Fixed list — {{range $i, $a := .Addresses}}{{if $i}}, {{end}}{{$a}}{{end}}{{end}}
</p>
<!-- The two panels are opened by a checkbox and a label rather than by
<details>, so that the controls stay a single row of buttons and
what a panel reveals is laid out under all four of them instead of
splitting the row in two. The checkbox precedes both its label and
its panel, which is what lets the CSS alone show them (no script
involved) — see .panel-toggle in panel.css. -->
<!-- One Edit panel (mode ‖ rate limit) opened by a checkbox and label
rather than <details>, so the button row stays intact — see
.panel-toggle in panel.css. -->
<div class="actions">
<input class="panel-toggle t-mode" id="mode-{{.ID}}" type="checkbox">
<label class="toggle for-mode" for="mode-{{.ID}}">Edit mode</label>
<input class="panel-toggle t-limit" id="limit-{{.ID}}" type="checkbox">
<label class="toggle for-limit" for="limit-{{.ID}}">Rate limit{{if .HasLimit}} (active){{end}}</label>
<input class="panel-toggle t-edit" id="edit-{{.ID}}" type="checkbox">
<label class="toggle for-edit" for="edit-{{.ID}}">Edit{{if .HasLimit}} <span class="st st-ok">limit</span>{{end}}</label>
<form class="inline" method="post" action="/applications/{{.ID}}/password"
data-confirm="Regenerate the password for {{.Login}}? The current password stops working immediately.">
<button type="submit">New password</button>
@@ -249,10 +264,13 @@
data-confirm="Delete application {{.Login}}? Its credentials stop working immediately.">
<button type="submit" class="danger">Delete</button>
</form>
<div class="panel panel-mode">
<div class="panel panel-edit">
<div class="check-cols">
<div class="check-col">
<p class="check-col-title">Address mode</p>
<p class="muted">Which From addresses this application may use.</p>
<form method="post" action="/applications/{{.ID}}/mode">
<label>Address mode</label>
<select name="mode" data-list-mode="{{$.List}}">
<select name="mode" data-list-mode="{{$.List}}" aria-label="Address mode">
<option value="{{$.Wildcard}}" {{if eq .AddressMode $.Wildcard}}selected{{end}}>Any address of the domain</option>
<option value="{{$.List}}" {{if eq .AddressMode $.List}}selected{{end}}>Specific addresses (list)</option>
</select>
@@ -264,25 +282,36 @@
<button type="submit">Save mode</button>
</form>
</div>
<div class="panel panel-limit">
<div class="check-col">
<p class="check-col-title">Optional trusted-IP override
{{if .HasLimit}}<span class="st st-ok">active</span>{{else}}<span class="st st-unknown">inactive</span>{{end}}</p>
<p class="muted">One per line or comma-separated.</p>
<form id="rl-{{.ID}}" method="post" action="/applications/{{.ID}}/ratelimit">
<label>Expected client IPs (one per line or comma-separated)</label>
<textarea name="allowed_ips" rows="2" placeholder="203.0.113.10">{{.IPsText}}</textarea>
<label>Message limit</label>
<input name="max_messages" type="number" min="1" value="{{.MaxText}}" placeholder="500">
<textarea name="allowed_ips" rows="1" placeholder="203.0.113.10"
aria-label="Trusted client IPs">{{.IPsText}}</textarea>
<p class="muted">These IPs get a higher ceiling than the domain
(≤ level&nbsp;1) and skip the domain check; everyone else uses
the domain level-2 limit{{if $.DomainHasRL}} ({{$.DomainRLMaxNum}}){{end}}
if set, otherwise level&nbsp;1.</p>
<div class="field-pair">
<div>
<label>Message limit (max {{$.L1Messages}})</label>
<input name="max_messages" type="number" min="1" max="{{$.L1Messages}}"
value="{{.MaxText}}" placeholder="{{$.L1Messages}}">
</div>
<div>
<label>Window (seconds)</label>
<input name="window_seconds" type="number" min="1" value="{{.WindowVal}}">
</div>
</div>
</form>
<!-- Saving and removing the limit are two posts, so they are two
forms and could not share a line while the Save button sat
inside the block form above. It is bound to that form by id
instead (the form attribute, no script involved), which lets
both buttons stand together in one row under the fields. -->
<!-- Saving and removing the limit are two posts; the Save
button is bound by form= id so both buttons share a row. -->
<div class="panel-buttons">
<button type="submit" form="rl-{{.ID}}">Save limit</button>
{{if .HasLimit}}
<form class="inline" method="post" action="/applications/{{.ID}}/ratelimit"
data-confirm="Remove the rate limit for {{.Login}}? Only the global level-1 limit will apply.">
data-confirm="Remove the rate limit for {{.Login}}? The domain limit (or level 1) will apply.">
<input type="hidden" name="clear" value="1">
<button type="submit" class="danger">Remove limit</button>
</form>
@@ -290,6 +319,8 @@
</div>
</div>
</div>
</div>
</div>
</li>
{{end}}
</ul>
@@ -298,32 +329,59 @@
{{end}}
</div>
<div class="card" id="rate-limit">
<h2>Sending rate limit (domain)</h2>
<p class="muted">Optional level-2 limit (spec 7.4): cap how many messages this
domain may send from its expected client IP(s) within a time window, summed
across all its applications. It counts messages — one message to many
recipients counts once. Leave the IP list empty to disable it and rely only on
the global level-1 limit. Applications that send from changing IPs should be
left unbound here.</p>
<div class="card" id="domain-settings">
<h2>Domain settings</h2>
<p class="muted">Status:
{{if .DomainHasRL}}<strong>active</strong>{{else}}inactive (level-1 only){{end}}.</p>
<div class="check-cols check-cols-rows">
<div class="check-col">
<p class="check-col-title">DMARC reports</p>
<p class="muted">Default comes from <a href="/settings">Settings</a>;
override per domain here.</p>
<div class="check-col-fields">
<form id="dmarc-domain" method="post" action="/domains/{{.Domain.ID}}/dmarc">
<label for="dmarc_rua_mode">Aggregate reports (rua=)</label>
<select id="dmarc_rua_mode" name="dmarc_rua_mode" data-custom-mode="custom">
<option value="inherit"{{if eq .DMARCRuaMode "inherit"}} selected{{end}}>Same as Settings{{if .ProfileDMARCEmail}} ({{.ProfileDMARCEmail}}){{end}}</option>
<option value="none"{{if eq .DMARCRuaMode "none"}} selected{{end}}>No aggregate reports</option>
<option value="custom"{{if eq .DMARCRuaMode "custom"}} selected{{end}}>Custom address</option>
</select>
<form method="post" action="/domains/{{.Domain.ID}}/ratelimit">
<label for="d_ips">Expected client IPs (one per line or comma-separated)</label>
<textarea id="d_ips" name="allowed_ips" rows="2"
placeholder="203.0.113.10">{{.DomainRLIPs}}</textarea>
<label for="d_max">Message limit</label>
<input id="d_max" name="max_messages" type="number" min="1"
value="{{.DomainRLMax}}" placeholder="1000">
<div data-custom-address>
<label for="dmarc_rua_email">Custom report address</label>
<input id="dmarc_rua_email" name="dmarc_rua_email" type="email"
autocapitalize="none" spellcheck="false" value="{{.DMARCRuaCustom}}"
placeholder="reports@your-mail-domain.com">
</div>
</form>
</div>
<div class="check-col-actions">
<button type="submit" form="dmarc-domain">Save DMARC report settings</button>
</div>
</div>
<div class="check-col">
<p class="check-col-title">Level-2 rate limit
{{if .DomainHasRL}}<span class="st st-ok">active</span>{{else}}<span class="st st-unknown">inactive</span>{{end}}</p>
<p class="muted">Level&nbsp;1 backstop: {{.L1Messages}} messages / {{.L1Window}}s — <a href="/settings#rate-limits">Settings</a>.</p>
<div class="check-col-fields">
<form id="rl-domain" method="post" action="/domains/{{.Domain.ID}}/ratelimit">
<div class="field-pair">
<div>
<label for="d_max">Message limit (max {{.L1Messages}})</label>
<input id="d_max" name="max_messages" type="number" min="1" max="{{.L1Messages}}"
value="{{.DomainRLMax}}" placeholder="{{.L1Messages}}">
</div>
<div>
<label for="d_win">Window (seconds)</label>
<input id="d_win" name="window_seconds" type="number" min="1" value="{{.DomainRLWin}}">
<button type="submit">Save limit</button>
</div>
</div>
<p class="muted">Applies to every client IP on this domain. Leave the
message limit empty to use level&nbsp;1 only.</p>
</form>
</div>
<div class="check-col-actions">
<button type="submit" form="rl-domain">Save limit</button>
{{if .DomainHasRL}}
<form class="inline" method="post" action="/domains/{{.Domain.ID}}/ratelimit"
data-confirm="Remove the domain rate limit? Only the global level-1 limit will apply.">
@@ -331,51 +389,27 @@
<button type="submit" class="danger">Remove limit</button>
</form>
{{end}}
</div>
</div>
</div>
</div>
<div class="card" id="export">
<div class="split">
<div class="card" id="export">
<h2>Export domain</h2>
<p class="muted">Download this domain to move it to another SelfPost instance:
its DKIM key, selector and every application with its working password. On
import the DNS record stays the same, so no DNS change is needed.</p>
<p class="muted"><strong>The export file is a secret</strong> — it contains the
private DKIM key and application passwords. Transfer it securely and delete it
after the import, or encrypt it below and move a <code>.spde</code> file
(SelfPost domain export) instead; the import form asks for the password.</p>
<p class="muted"><strong>Secret file</strong> — transfer securely, or encrypt
below as <code>.spde</code>.</p>
{{if .ExportErr}}<p class="error">{{.ExportErr}}</p>{{end}}
<form method="post" action="/domains/{{.Domain.ID}}/export">
{{template "encryptfields" .}}
<button type="submit">Export domain</button>
</form>
</div>
</div>
<div class="card" id="danger">
<div class="card" id="danger">
<h2>Danger zone</h2>
<p class="muted">Deleting this domain also deletes its DKIM key and every
application bound to it.</p>
<p class="muted">Deletes the DKIM key and every application on this domain.</p>
<a class="danger" href="/domains/{{.Domain.ID}}/delete">Delete domain</a>
</div>
{{end}}
{{/* The domain page's section index, shown in the navigation column (see the
"sections" block in layout.html). This is the panel's longest page — the
DNS records to publish, the checks on them, the applications and two rate
limits — and setting a domain up means going back and forth between them.
The freshly generated password is only listed while it is on the page: it
is the one card that is not always there, and the one nothing should scroll
away from silently. */}}
{{define "sections"}}
<div class="sections">
<p class="sections-title">On this page</p>
{{if .NewCred}}<a href="#new-credential">New application password</a>{{end}}
<a href="#dkim">DKIM DNS record</a>
<a href="#spf-dmarc">SPF and DMARC records</a>
<a href="#dns-status">DNS status</a>
<a href="#settings">Sending server settings</a>
<a href="#add-application">Add an application</a>
<a href="#applications">Applications</a>
<a href="#rate-limit">Sending rate limit</a>
<a href="#export">Export domain</a>
<a href="#danger">Danger zone</a>
</div>
</div>
{{end}}
+28 -22
View File
@@ -61,9 +61,9 @@
{{define "nav"}}
<nav class="nav">
{{/* The order of the blocks is the order they are read in, top to bottom of
the column: the mark, the panel's pages, the current page's own sections,
and the session last. The mark goes to /status, the page the panel opens
on — named outright rather than as "/", which is only a redirect to it. */}}
the column: the mark, the panel's pages, and the session last. The mark
goes to /status, the page the panel opens on — named outright rather
than as "/", which is only a redirect to it. */}}
<a class="brand" href="{{if .IsGlobal}}/status{{else}}/domains{{end}}"><img src="/static/logo-compact.svg" width="220" height="100" alt="SelfPost"></a>
<div class="links">
{{if .IsGlobal}}
@@ -75,13 +75,12 @@
{{if eq .Active "mail_queue"}}<span aria-current="page">{{template "icon-mail-queue"}}Mail queue</span>{{else}}<a href="/mail-queue">{{template "icon-mail-queue"}}Mail queue</a>{{end}}
{{if eq .Active "system_log"}}<span aria-current="page">{{template "icon-system-log"}}System log</span>{{else}}<a href="/system-log">{{template "icon-system-log"}}System log</a>{{end}}
{{if eq .Active "backup"}}<span aria-current="page">{{template "icon-backup"}}Backup</span>{{else}}<a href="/backup">{{template "icon-backup"}}Backup</a>{{end}}
{{if eq .Active "users"}}<span aria-current="page">{{template "icon-account"}}Users</span>{{else}}<a href="/users">{{template "icon-account"}}Users</a>{{end}}
{{if eq .Active "users"}}<span aria-current="page">{{template "icon-users"}}Users</span>{{else}}<a href="/users">{{template "icon-users"}}Users</a>{{end}}
{{end}}
</div>
{{template "sections" .}}
<div class="session">
<span class="muted">User: {{.User}}</span>
{{if eq .Active "account"}}<span aria-current="page">{{template "icon-account"}}Settings</span>{{else}}<a href="/account">{{template "icon-account"}}Settings</a>{{end}}
<span class="session-user muted">{{template "icon-account"}}User: {{.User}}</span>
{{if eq .Active "settings"}}<span aria-current="page">{{template "icon-settings"}}Settings</span>{{else}}<a href="/settings">{{template "icon-settings"}}Settings</a>{{end}}
<form class="inline" method="post" action="/logout">
<button type="submit" class="danger">{{template "icon-sign-out"}}Sign out</button>
</form>
@@ -89,24 +88,22 @@
</nav>
{{end}}
{{/* The index of the current page's own sections, shown in the navigation
column under the page entries. Most pages are one or two cards and need no
index, so the block is empty here and only the pages that are genuinely
long — the domain page, the status page — redefine it with their own list
(see .sections in panel.css for what it looks like, and panel.js for the
marking of the section in view). Each page's template file is parsed after
this one, so its definition replaces this empty one; a page that defines
nothing keeps it and renders no index. */}}
{{define "sections"}}{{end}}
{{/* Whether the page takes the column whole. Empty here, so a page is held to
the reading measure unless it says otherwise; the three pages that are
tables of data or raw log lines redefine this as the word "wide", which
lands in <main>'s class list (see main.wide in panel.css). Same mechanism
as "sections" above: each page's template file is parsed after this one,
so its definition replaces the empty one. */}}
the reading measure unless it says otherwise; the pages that are tables of
data, raw log lines or side-by-side cards redefine this as the word "wide",
which lands in <main>'s class list (see main.wide in panel.css). Each
page's template file is parsed after this one, so its definition replaces
the empty one. */}}
{{define "wide"}}{{end}}
{{/* back_link — up-navigation on drill-down pages. Invoke with the "back"
function, e.g. {{template "back_link" (back "/users" "Back to users")}}.
Place it directly under the page <h1> and above the cards (see .back in
panel.css). TestDrillDownPagesPlaceBackLinkAboveContent guards this. */}}
{{define "back_link"}}
<a class="back" href="{{.Href}}">&larr; {{.Label}}</a>
{{end}}
{{/* Navigation icons. Inline SVG rather than an icon font or sprite file: they
inherit the link's colour through currentColor, cost no extra request, and
need no exemption from the panel's "default-src 'self'" policy. Each is
@@ -122,6 +119,15 @@
two drawers with handles is the picture of an archive. */}}
{{define "icon-backup"}}<svg class="icon" viewBox="0 0 16 16" fill="none" stroke="currentColor" stroke-width="1.5" stroke-linecap="round" stroke-linejoin="round" aria-hidden="true"><rect x="2.75" y="1.75" width="10.5" height="12.5" rx="1.15"/><path d="M2.75 8h10.5"/><path d="M6.4 4.85h3.2M6.4 11.15h3.2"/></svg>{{end}}
{{define "icon-account"}}<svg class="icon" viewBox="0 0 16 16" fill="none" stroke="currentColor" stroke-width="1.5" stroke-linecap="round" stroke-linejoin="round" aria-hidden="true"><circle cx="8" cy="5.4" r="2.75"/><path d="M2.9 14.25a5.1 5.1 0 0 1 10.2 0"/></svg>{{end}}
{{/* Two people: rear as a right-side crescent (head + shoulder) so it peeks
from behind; front is a full silhouette matching icon-account. Heads share
one baseline; shoulder arcs share one floor — otherwise the pair looks
lopsided at 16 px. */}}
{{define "icon-users"}}<svg class="icon" viewBox="0 0 16 16" fill="none" stroke="currentColor" stroke-width="1.5" stroke-linecap="round" stroke-linejoin="round" aria-hidden="true"><path d="M10.9 3.1a2.1 2.1 0 0 1 0 4.2"/><path d="M14.7 14.25a3.8 3.8 0 0 0-3.9-3.65"/><circle cx="5.5" cy="5.2" r="2.5"/><path d="M1.4 14.25a4.8 4.8 0 0 1 8.2 0"/></svg>{{end}}
{{/* A gear (Feather "settings"), not a sun-with-rays: at 16 px spokes alone
read as brightness, while the toothed rim is the picture of configuration.
viewBox 0 0 24 matches the source path; CSS sizes the icon to 1 rem. */}}
{{define "icon-settings"}}<svg class="icon" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2.25" stroke-linecap="round" stroke-linejoin="round" aria-hidden="true"><circle cx="12" cy="12" r="3"/><path d="M19.4 15a1.65 1.65 0 0 0 .33 1.82l.06.06a2 2 0 0 1 0 2.83 2 2 0 0 1-2.83 0l-.06-.06a1.65 1.65 0 0 0-1.82-.33 1.65 1.65 0 0 0-1 1.51V21a2 2 0 0 1-2 2 2 2 0 0 1-2-2v-.09A1.65 1.65 0 0 0 9 19.4a1.65 1.65 0 0 0-1.82.33l-.06.06a2 2 0 0 1-2.83 0 2 2 0 0 1 0-2.83l.06-.06A1.65 1.65 0 0 0 4.68 15a1.65 1.65 0 0 0-1.51-1H3a2 2 0 0 1-2-2 2 2 0 0 1 2-2h.09A1.65 1.65 0 0 0 4.6 9a1.65 1.65 0 0 0-.33-1.82l-.06-.06a2 2 0 0 1 0-2.83 2 2 0 0 1 2.83 0l.06.06A1.65 1.65 0 0 0 9 4.68a1.65 1.65 0 0 0 1-1.51V3a2 2 0 0 1 2-2 2 2 0 0 1 2 2v.09a1.65 1.65 0 0 0 1 1.51 1.65 1.65 0 0 0 1.82-.33l.06-.06a2 2 0 0 1 2.83 0 2 2 0 0 1 0 2.83l-.06.06A1.65 1.65 0 0 0 19.4 9a1.65 1.65 0 0 0 1.51 1H21a2 2 0 0 1 2 2 2 2 0 0 1-2 2h-.09a1.65 1.65 0 0 0-1.51 1z"/></svg>{{end}}
{{/* Sign out is the one icon here that marks an action rather than a page: an
open door with the arrow leading out of it. */}}
{{define "icon-sign-out"}}<svg class="icon" viewBox="0 0 16 16" fill="none" stroke="currentColor" stroke-width="1.5" stroke-linecap="round" stroke-linejoin="round" aria-hidden="true"><path d="M6.1 14.25H3.65a1.15 1.15 0 0 1-1.15-1.15V2.9a1.15 1.15 0 0 1 1.15-1.15H6.1"/><path d="M10.6 11.15 13.75 8 10.6 4.85"/><path d="M13.75 8H6.35"/></svg>{{end}}
@@ -1,5 +1,5 @@
{{define "mail_queue_body"}}
<div id="mail-queue-body" data-poll hx-get="/mail-queue/body" hx-trigger="load" hx-swap="outerHTML">
<div id="mail-queue-body" data-poll aria-live="polite" hx-get="/mail-queue/body" hx-trigger="load" hx-swap="outerHTML">
{{if .Error}}<p class="error">{{.Error}}</p>{{end}}
<span class="code">{{if .Output}}{{.Output}}{{else}}Queue is empty.{{end}}</span>
</div>
+122
View File
@@ -0,0 +1,122 @@
{{/* Wide enough for credentials and DMARC settings side by side (see .split
in panel.css, as on a delivery's page). */}}
{{define "wide"}}wide{{end}}
{{/* The username/password fields are identical for a global administrator
(split card, DMARC alongside) and a domain administrator (narrow card,
no DMARC card) — only the surrounding form and card differ. */}}
{{define "credentials_fields"}}
<label for="username">Username</label>
<input id="username" name="username" autocomplete="username"
autocapitalize="none" spellcheck="false" value="{{.FormUsername}}" required>
<label for="current_password">Current password</label>
<input id="current_password" name="current_password" type="password"
autocomplete="current-password" required>
<label for="new_password">New password</label>
<input id="new_password" name="new_password" type="password" autocomplete="new-password">
<label for="new_password_confirm">Confirm new password</label>
<input id="new_password_confirm" name="new_password_confirm" type="password" autocomplete="new-password">
{{end}}
{{define "content"}}
<h1>Settings</h1>
{{if .Flash}}<div class="flash">{{.Flash}}</div>{{end}}
{{if .Error}}<p class="error">{{.Error}}</p>{{end}}
{{if .ShowDMARC}}
<form method="post" action="/settings">
<div class="split">
<div class="card">
<h2>Panel credentials</h2>
<p class="muted">These are the credentials for this control panel only.
Applications keep their own logins and passwords, which are not affected.</p>
{{template "credentials_fields" .}}
</div>
<div class="card">
<h2>DMARC aggregate reports</h2>
<p class="muted">Optional default <code>rua=</code> address for every sending
domain (can be overridden per domain). Use a mailbox on a domain that
receives inbound mail. SelfPost does not receive inbound mail — point
<code>rua=</code> at a mailbox elsewhere.</p>
<label for="dmarc_report_email">Default report address</label>
<input id="dmarc_report_email" name="dmarc_report_email" type="email"
autocomplete="email" autocapitalize="none" spellcheck="false"
value="{{.FormDMARCEmail}}" placeholder="reports@your-mail-domain.com">
{{if .FormDMARCEmail}}
<p class="muted">When <code>rua=</code> points at another domain, that hub
domain must publish a report-authorisation record so receivers will deliver
the XML aggregates.</p>
<label>Report authorization — host / name</label>
<div class="code-row">
<span class="code">{{.ReportAuthName}}</span>
<button type="button" class="copy">Copy</button>
</div>
<label>Report authorization — value</label>
<div class="code-row">
<span class="code">{{.ReportAuthExample}}</span>
<button type="button" class="copy">Copy</button>
</div>
{{if .ReportAuthDNS.Status}}
<label>Report authorization DNS <span class="st st-{{.ReportAuthDNS.Status}}">{{.ReportAuthDNS.Status}}</span></label>
<p class="{{if eq .ReportAuthDNS.Status "ok"}}muted{{else}}error{{end}}">{{.ReportAuthDNS.Detail}}</p>
{{if .ReportAuthDNS.Records}}<span class="code">{{range .ReportAuthDNS.Records}}{{.}}
{{end}}</span>{{end}}
{{end}}
{{end}}
</div>
</div>
<button type="submit">Save changes</button>
<p class="muted">Leave both new-password fields empty to change the username
or DMARC address only. Changing the password signs out every other session;
this one stays signed in.</p>
</form>
{{else}}
<div class="card narrow">
<h2>Panel credentials</h2>
<p class="muted">These are the credentials for this control panel only.
Applications keep their own logins and passwords, which are not affected.</p>
<form method="post" action="/settings">
{{template "credentials_fields" .}}
<button type="submit">Save changes</button>
</form>
<p class="muted">Leave both new-password fields empty to change the username
only. Changing the password signs out every other session;
this one stays signed in.</p>
</div>
{{end}}
<div class="card" id="rate-limits">
<h2>Sending rate limits</h2>
<p class="muted">Configured in <code>.env</code> / Compose; restart the
container to change level&nbsp;1. Domain and application ceilings are set on
each domain's page.</p>
<label>Level 1 — per client IP (Postfix)</label>
<p class="code-row"><span class="code">{{.L1Messages}} messages / {{.L1Window}} seconds</span></p>
<p class="muted"><code>RATE_LIMIT_MESSAGES_PER_IP</code> /
<code>RATE_LIMIT_WINDOW_SECONDS</code>. Hard ceiling for every connecting IP;
the panel cannot raise a domain or application limit above this.</p>
<label>Level 2 — domain</label>
<p class="muted">Optional ceiling for <em>all</em> senders on a domain. When
unset, only level&nbsp;1 applies. Must be ≤ level&nbsp;1.</p>
<label>Level 2 — application (trusted IPs)</label>
<p class="muted">Optional override: list client IPs and a ceiling
<em>strictly above</em> the domain limit (still ≤ level&nbsp;1). Those IPs
skip the domain check; everyone else stays under the domain (or level&nbsp;1).</p>
</div>
{{end}}
+5 -44
View File
@@ -1,34 +1,14 @@
{{/* Wide so Machine|Processes and the other .split pairs fill the column
rather than the 48rem reading measure. */}}
{{define "wide"}}wide{{end}}
{{define "content"}}
<h1>Server status</h1>
<h1>Status</h1>
{{if .Flash}}<div class="flash">{{.Flash}}</div>{{end}}
{{template "status_body" .}}
<div class="card" id="hostname">
<h2>Hostname and reverse DNS <span class="st st-{{.PTR.Status}}">{{.PTR.Status}}</span></h2>
<p class="muted">Receiving servers check that the name this server announces
resolves to its address <em>and</em> that the address resolves back to the same
name (forward-confirmed reverse DNS). A missing or mismatched reverse record is
the most common reason self-hosted mail is rejected or scored as spam. The
reverse record is set at the hosting provider, not in the domain's DNS zone.</p>
<label>Server hostname</label>
<span class="code">{{if .Hostname}}{{.Hostname}}{{else}}(SELFPOST_HOSTNAME is not set){{end}}</span>
{{if .PTR.Records}}
<label>Forward and reverse lookup</label>
<span class="code">{{range .PTR.Records}}{{.}}
{{end}}</span>
{{end}}
<p class="{{if eq .PTR.Status "ok"}}muted{{else}}error{{end}}">{{.PTR.Detail}}</p>
<form class="inline" method="post" action="/status/recheck">
<button type="submit">Re-check DNS</button>
</form>
</div>
<div class="card" id="configuration">
<h2>Configuration</h2>
<p class="muted">Regenerates the OpenDKIM and Postfix configuration from the
@@ -41,22 +21,3 @@
</form>
</div>
{{end}}
{{/* The status page's section index, shown in the navigation column (see the
"sections" block in layout.html). The first six cards are the ones the
polling fragment replaces on a timer; their ids are part of
status_body.html and do not change with the reading, so the links here hold
across a refresh. */}}
{{define "sections"}}
<div class="sections">
<p class="sections-title">On this page</p>
<a href="#overall">Overall</a>
<a href="#processes">Processes</a>
<a href="#machine">Machine</a>
<a href="#queue">Mail queue</a>
<a href="#certificate">TLS certificate</a>
<a href="#sockets">Milter sockets</a>
<a href="#hostname">Hostname and reverse DNS</a>
<a href="#configuration">Configuration</a>
</div>
{{end}}
+49 -32
View File
@@ -1,36 +1,13 @@
{{define "status_body"}}
<div id="status-body" data-poll hx-get="/status/fragment" hx-trigger="load" hx-swap="outerHTML">
<div id="status-body" data-poll aria-live="polite" hx-get="/status/fragment" hx-trigger="load" hx-swap="outerHTML">
<div class="card" id="overall">
<h2>Overall <span class="st st-{{.OverallStatus}}">{{.OverallStatus}}</span></h2>
<p class="muted">{{.OverallHeading}}</p>
</div>
<div class="card" id="processes">
<h2>Processes <span class="st st-{{.ProcessStatus}}">{{.ProcessStatus}}</span></h2>
{{if .ProcessError}}
<p class="error">Could not ask supervisord for the process list.</p>
{{else}}
<table>
<thead><tr><th>Program</th><th>State</th><th>Detail</th></tr></thead>
<tbody>
{{range .Processes}}
<tr>
<td>{{.Name}}</td>
<td><span class="st st-{{.Status}}">{{.State}}</span></td>
<td class="muted">{{.Detail}}</td>
</tr>
{{end}}
</tbody>
</table>
{{end}}
</div>
<div class="split">
<div class="card" id="machine">
<h2>Machine <span class="st st-{{.Machine.Status}}">{{.Machine.Status}}</span></h2>
<p class="muted">Processor, memory and network of the machine this container
runs on, read from the kernel's counters. CPU and throughput are rates, so
they describe{{if .Machine.WindowText}} the {{.Machine.WindowText}}{{end}}
since the previous refresh rather than this instant.</p>
<table>
<thead><tr><th class="metric">Resource</th><th>Usage</th><th>Detail</th></tr></thead>
<tbody>
@@ -63,7 +40,7 @@
</td>
<td class="muted">
{{range .Machine.Network.Interfaces}}
<div>{{.Name}}: {{.InText}} in, {{.OutText}} out{{if .Measured}} (↓ {{.InRateText}} ↑ {{.OutRateText}}){{end}}</div>
<div>{{.Name}}: {{.InText}} in, {{.OutText}} out</div>
{{end}}
{{if .Machine.Network.Detail}}<div>{{.Machine.Network.Detail}}</div>{{end}}
</td>
@@ -72,41 +49,81 @@
</table>
</div>
<div class="card" id="processes">
<h2>Processes <span class="st st-{{.ProcessStatus}}">{{.ProcessStatus}}</span></h2>
{{if .ProcessError}}
<p class="error">Could not ask supervisord for the process list.</p>
{{else}}
<table>
<thead><tr><th>Program</th><th>State</th><th>Detail</th></tr></thead>
<tbody>
{{range .Processes}}
<tr>
<td>{{.Name}}</td>
<td><span class="st st-{{.Status}}">{{.State}}</span></td>
<td class="muted">{{.Detail}}</td>
</tr>
{{end}}
</tbody>
</table>
{{end}}
</div>
</div>
<div class="split">
<div class="card" id="queue">
<h2>Mail queue <span class="st st-{{.QueueStatus}}">{{.QueueStatus}}</span></h2>
{{if .QueueError}}
<p class="error">{{.QueueError}}</p>
{{else}}
<p>{{if .QueueSummary}}{{.QueueSummary}}{{else}}Mail queue is empty.{{end}}</p>
<a class="btn" href="/mail-queue">Full queue</a>
<a class="btn" href="/mail-queue">View queue</a>
{{end}}
</div>
<div class="card" id="certificate">
<h2>TLS certificate <span class="st st-{{.Cert.Status}}">{{.Cert.Status}}</span></h2>
<p class="muted">The certificate Postfix serves on port 465{{if .Cert.Subject}} ({{.Cert.Subject}}){{end}}.
It is supplied by the reverse proxy through a read-only mount; SelfPost only reads it.</p>
{{if not .Cert.NotAfter.IsZero}}
<label>Expires</label>
<span class="code">{{.Cert.NotAfter.UTC.Format "2006-01-02 15:04 UTC"}}</span>
{{end}}
<p class="{{if eq .Cert.Status "ok"}}muted{{else}}error{{end}}">{{.Cert.Detail}}</p>
</div>
</div>
<div class="split">
<div class="card" id="sockets">
<h2>Milter sockets <span class="st st-{{.SocketStatus}}">{{.SocketStatus}}</span></h2>
<table>
<thead><tr><th>Milter</th><th>Socket</th><th>State</th></tr></thead>
<thead><tr><th>Milter</th><th>State</th><th>Detail</th></tr></thead>
<tbody>
{{range .Sockets}}
<tr>
<td>{{.Name}}</td>
<td class="muted">{{.Path}}</td>
<td><span class="st st-{{.Status}}">{{.Status}}</span> {{.Detail}}</td>
<td><span class="st st-{{.Status}}">{{.Status}}</span></td>
<td class="muted">{{.Detail}}</td>
</tr>
{{end}}
</tbody>
</table>
</div>
<div class="card" id="hostname">
<h2>Hostname and reverse DNS <span class="st st-{{.PTR.Status}}">{{.PTR.Status}}</span></h2>
<label>Server hostname</label>
<span class="code">{{if .Hostname}}{{.Hostname}}{{else}}(SELFPOST_HOSTNAME is not set){{end}}</span>
{{if .PTR.Records}}
<label>Forward and reverse lookup</label>
<span class="code">{{range .PTR.Records}}{{.}}
{{end}}</span>
{{end}}
<form class="inline" method="post" action="/status/recheck">
<button type="submit">Re-check DNS</button>
</form>
</div>
</div>
</div>
{{end}}
@@ -1,5 +1,5 @@
{{define "system_log_body"}}
<div id="system-log-body" data-poll hx-get="/system-log/body" hx-trigger="load" hx-swap="outerHTML">
<div id="system-log-body" data-poll aria-live="polite" hx-get="/system-log/body" hx-trigger="load" hx-swap="outerHTML">
{{if .Error}}<p class="error">{{.Error}}</p>{{end}}
<span class="code">{{if .Lines}}{{range .Lines}}{{.}}
{{end}}{{else}}No log lines yet.{{end}}</span>
@@ -0,0 +1,16 @@
{{define "content"}}
<h1>Delete {{.TargetUsername}}</h1>
{{template "back_link" (back (printf "/users/%d" .TargetID) (printf "Back to %s" .TargetUsername))}}
<div class="card narrow">
<h2>Confirm deletion</h2>
<p>You are about to delete the panel user <strong>{{.TargetUsername}}</strong>.
This cannot be undone; a signed-in session for this user stops working
immediately.</p>
<form method="post" action="/users/{{.TargetID}}/delete">
<button type="submit" class="danger">Delete {{.TargetUsername}}</button>
</form>
</div>
{{end}}
+14 -7
View File
@@ -1,6 +1,8 @@
{{define "content"}}
<h1>{{if .IsEdit}}Edit user{{else}}Create user{{end}}</h1>
{{template "back_link" (back "/users" "Back to users")}}
<div class="card narrow">
{{if .Error}}<p class="error">{{.Error}}</p>{{end}}
<form method="post" action="{{if .IsEdit}}/users/{{.UserID}}{{else}}/users/new{{end}}">
@@ -12,16 +14,18 @@
<input id="password" name="password" type="password" autocomplete="new-password" {{if not .IsEdit}}required{{end}}>
<label for="role">Role</label>
<select id="role" name="role">
{{if .LastGlobalLocked}}<input type="hidden" name="role" value="{{.FormRole}}">{{end}}
<select id="role" name="role" data-global-role="{{.GlobalRole}}"{{if .LastGlobalLocked}} disabled{{end}}>
<option value="domain_admin" {{if eq .FormRole "domain_admin"}}selected{{end}}>Domain administrator</option>
<option value="global" {{if eq .FormRole "global"}}selected{{end}}>Global administrator</option>
</select>
{{if .LastGlobalLocked}}<p class="muted">The only global administrator cannot be demoted.</p>{{end}}
<fieldset id="domain-pick">
<fieldset id="domain-pick" data-domain-pick>
<legend>Assigned domains</legend>
<p class="muted">Required for domain administrators.</p>
{{range .Domains}}
<label class="checkbox">
<label class="check">
<input type="checkbox" name="domain_ids" value="{{.ID}}" {{if index $.FormDomains .ID}}checked{{end}}>
{{.Name}}
</label>
@@ -29,10 +33,13 @@
</fieldset>
<button type="submit">{{if .IsEdit}}Save{{else}}Create{{end}}</button>
{{if .IsEdit}}
<button type="submit" name="action" value="delete" class="danger">Delete user</button>
{{end}}
</form>
<p class="muted"><a href="/users">Back to users</a></p>
{{if .IsEdit}}
{{if .LastGlobalLocked}}
<p class="muted">The only global administrator cannot be deleted.</p>
{{else}}
<a class="danger" href="/users/{{.UserID}}/delete">Delete user</a>
{{end}}
{{end}}
</div>
{{end}}
+1 -1
View File
@@ -1,5 +1,5 @@
{{define "content"}}
<h1>Panel users</h1>
<h1>Users</h1>
{{if .Flash}}<div class="flash">{{.Flash}}</div>{{end}}
+123 -79
View File
@@ -29,79 +29,6 @@ func TestEveryPageResolvesNav(t *testing.T) {
}
}
// The section index each long page shows in the navigation column works by
// overriding an empty "sections" block defined in the layout, which only holds
// as long as the layout is parsed before the page's own files (see pageFiles).
// Reverse that order and every index would silently disappear — the empty
// definition would win and no page would fail to render — so the two ends are
// asserted here: the long pages produce a list, and a page that defines nothing
// produces nothing at all.
func TestSectionIndexIsOnTheLongPagesOnly(t *testing.T) {
engine, err := New("test")
if err != nil {
t.Fatalf("New: %v", err)
}
// Anchors the index links to, taken from the page's own cards.
wantAnchors := map[string]string{
"status": `href="#certificate"`,
"domain_detail": `href="#danger"`,
}
for name, page := range engine.Pages() {
var buf bytes.Buffer
// The domain page's index hides the freshly generated credential entry
// unless one is on the page, so the data map carries the key it reads.
if err := page.ExecuteTemplate(&buf, "sections", map[string]any{"NewCred": nil}); err != nil {
t.Fatalf("execute sections for %q: %v", name, err)
}
out := buf.String()
anchor, wanted := wantAnchors[name]
switch {
case wanted && !strings.Contains(out, anchor):
t.Errorf("page %q shows no section index (expected %s):\n%s", name, anchor, out)
case !wanted && strings.TrimSpace(out) != "":
t.Errorf("page %q is not long enough to carry a section index:\n%s", name, out)
}
}
}
// A section link that points at no card is a link that does nothing, and
// nothing about rendering the page says so. Every anchor the index offers must
// name an element the same page defines an id for.
func TestSectionLinksPointAtCardsThatExist(t *testing.T) {
engine, err := New("test")
if err != nil {
t.Fatalf("New: %v", err)
}
// The pages that carry an index; both are checked with a credential shown,
// which is the domain page's one conditional entry.
for _, name := range []string{"status", "domain_detail"} {
var index bytes.Buffer
if err := engine.Page(name).ExecuteTemplate(&index, "sections", map[string]any{"NewCred": true}); err != nil {
t.Fatalf("execute sections for %q: %v", name, err)
}
// The cards are spread over the page's template files, so the ids are
// collected from the files rather than from a rendered page — rendering
// one would need the whole of a handler's data map.
ids := map[string]bool{}
for _, file := range pageFiles[name] {
body, err := fs.ReadFile(assetsFS, file)
if err != nil {
t.Fatalf("read %s: %v", file, err)
}
// Cards only: a form field's id is not somewhere a section link may
// land, so matching those too would weaken the check.
for _, m := range regexp.MustCompile(`class="card[^"]*" id="([a-z-]+)"`).FindAllStringSubmatch(string(body), -1) {
ids[m[1]] = true
}
}
for _, m := range regexp.MustCompile(`href="#([a-z-]+)"`).FindAllStringSubmatch(index.String(), -1) {
if !ids[m[1]] {
t.Errorf("page %q indexes #%s, which no card on it carries", name, m[1])
}
}
}
}
// The version comes from render(), not from each handler's data map, so the
// footer is only correct as long as every page composes with the layout and
// render keeps supplying the key. Both are asserted here rather than trusted.
@@ -263,7 +190,7 @@ func TestOnlyThePagesMadeOfDataDeclareThemselvesWide(t *testing.T) {
if err != nil {
t.Fatalf("New: %v", err)
}
wide := map[string]bool{"deliveries": true, "delivery": true, "mail_queue": true, "system_log": true}
wide := map[string]bool{"settings": true, "deliveries": true, "delivery": true, "mail_queue": true, "status": true, "system_log": true, "domain_detail": true}
for name, page := range engine.Pages() {
var buf bytes.Buffer
if err := page.ExecuteTemplate(&buf, "wide", nil); err != nil {
@@ -279,6 +206,102 @@ func TestOnlyThePagesMadeOfDataDeclareThemselvesWide(t *testing.T) {
}
}
// The domain page pairs cards the same way Status does: three .split rows
// (DKIM+SPF|DMARC, connection|add-app, export|danger). DNS status, Applications
// and Domain settings are full-width; DNS status and Domain settings (and the
// application Edit panel) use .check-cols. Losing a row silently stacks again.
func TestDomainDetailPageHasPairedCards(t *testing.T) {
body, err := fs.ReadFile(assetsFS, "templates/domain_detail.html")
if err != nil {
t.Fatalf("read domain_detail: %v", err)
}
src := string(body)
if got := strings.Count(src, `class="split"`); got != 3 {
t.Errorf("domain detail has %d .split rows, want 3", got)
}
if !strings.Contains(src, `class="check-cols"`) {
t.Error("domain detail is missing the check-cols grid")
}
if !strings.Contains(src, `class="panel-toggle t-edit"`) {
t.Error("application Edit should be a single panel-toggle")
}
if strings.Contains(src, `panel-toggle t-mode`) || strings.Contains(src, `panel-toggle t-limit`) ||
strings.Contains(src, `panel-mode`) || strings.Contains(src, `panel-limit`) {
t.Error("application Edit mode and Rate limit should be one Edit button")
}
for _, id := range []string{
`id="dkim-spf"`, `id="dns-status"`, `id="dmarc"`,
`id="connection"`, `id="add-application"`, `id="applications"`,
`id="domain-settings"`, `id="export"`, `id="danger"`,
} {
if !strings.Contains(src, id) {
t.Errorf("domain detail is missing %s", id)
}
}
if strings.Contains(src, `id="rate-limit"`) {
t.Error("domain rate limit should live inside domain-settings, not its own card")
}
if strings.Contains(src, `id="d_ips"`) {
t.Error("domain rate limit must not ask for client IPs")
}
if !strings.Contains(src, "{{.L1Messages}}") && !strings.Contains(src, "{{$.L1Messages}}") {
t.Error("domain rate limit should show the L1 message count")
}
if !strings.Contains(src, "Level&nbsp;1 backstop") {
t.Error("domain rate limit should show a Level 1 backstop line")
}
if !strings.Contains(src, "Trusted client IPs") {
t.Error("application override should ask for trusted client IPs")
}
if strings.Contains(src, `id="spf-dmarc"`) {
t.Error("SPF should sit with DKIM, not with DMARC")
}
}
func TestSettingsPageDocumentsRateLimits(t *testing.T) {
body, err := fs.ReadFile(assetsFS, "templates/settings.html")
if err != nil {
t.Fatalf("read settings: %v", err)
}
src := string(body)
if !strings.Contains(src, `id="rate-limits"`) {
t.Error("settings should include a sending rate limits card")
}
for _, want := range []string{
"RATE_LIMIT_MESSAGES_PER_IP",
"Level 2 — domain",
"trusted IPs",
"{{.L1Messages}} messages / {{.L1Window}} seconds",
} {
if !strings.Contains(src, want) {
t.Errorf("settings rate limits card missing %q", want)
}
}
}
func TestDrillDownPagesPlaceBackLinkAboveContent(t *testing.T) {
drillDown := map[string]bool{
"user_form.html": true,
"user_delete.html": true,
"domain_detail.html": true,
"domain_delete.html": true,
"delivery.html": true,
}
forEachTemplate(t, func(name, body string) {
if !drillDown[name] {
return
}
if !strings.Contains(body, `template "back_link"`) {
t.Errorf("%s is a drill-down page but does not use the shared back_link template", name)
}
backIdx := strings.Index(body, `template "back_link"`)
cardIdx := strings.Index(body, `class="card`)
if cardIdx >= 0 && backIdx > cardIdx {
t.Errorf("%s places the back link after the first card", name)
}
})
}
// Since the panel root redirects to the status page, a link left pointing at
// "/" silently lands on the wrong screen instead of failing — so no template may
// contain one.
@@ -353,17 +376,38 @@ func TestStatusPageRendersEveryCheck(t *testing.T) {
"opendkim", "FATAL", "Mail queue is empty", "mail.example.com",
"203.0.113.10 → no PTR record", `action="/reload"`,
`hx-get="/status/fragment"`, `class="st st-error"`,
// Three .split rows inside the polled fragment: machine|processes,
// queue|certificate, and sockets|hostname. Ids stay on the cards.
`id="processes"`, `id="machine"`, `id="queue"`, `id="certificate"`, `id="sockets"`, `id="hostname"`,
`action="/status/recheck"`,
// The machine card: the bars carry their reading in an attribute
// (the CSP rules out sizing them with a style), and the figures are
// printed beside them for anything that does not render a meter.
`<meter value="12"`, `<meter value="50"`,
"load average 0.31, 0.24, 0.19", "2.0 GiB used of 4.0 GiB",
"4 cores · 4 threads", "2.0 GiB used of 4.0 GiB",
"eth0: 1.0 MiB in, 512.0 KiB out",
} {
if !strings.Contains(out, want) {
t.Errorf("status page is missing %q", want)
}
}
if got := strings.Count(out, `class="split"`); got != 3 {
t.Errorf("status page has %d .split rows, want 3", got)
}
// Hostname must live inside the fragment so a poll refresh keeps it beside
// sockets; Configuration stays outside (static reload control).
body := strings.Index(out, `id="status-body"`)
conf := strings.Index(out, `id="configuration"`)
if body < 0 || conf < 0 || conf < body {
t.Fatal("status-body or configuration card missing or out of order")
}
frag := out[body:conf]
if !strings.Contains(frag, `id="hostname"`) {
t.Error("hostname card is outside the polled status-body fragment")
}
if strings.Contains(frag, `id="configuration"`) || strings.Contains(frag, `action="/reload"`) {
t.Error("configuration reload must stay outside the polled fragment")
}
}
// A machine whose counters could not be read — no /proc, or a first reading
@@ -427,13 +471,13 @@ func statusPageData() map[string]any {
},
"Machine": health.Machine{
CPU: health.CPU{
Measured: true, BusyPct: 12.4, Cores: 4,
Measured: true, BusyPct: 12.4, Cores: 4, Threads: 4,
Load: [3]float64{0.31, 0.24, 0.19}, HasLoad: true,
Status: health.StatusOK, Detail: "4 core(s) · load average 0.31, 0.24, 0.19",
Status: health.StatusOK, Detail: "4 cores · 4 threads",
},
Memory: health.Memory{
Measured: true, TotalBytes: 4 << 30, UsedBytes: 2 << 30, UsedPct: 50,
Status: health.StatusOK, Detail: "2.0 GiB used of 4.0 GiB; 2.0 GiB available to new work.",
Status: health.StatusOK, Detail: "2.0 GiB used of 4.0 GiB.",
},
Network: health.Network{
Measured: true, RxRate: 2048, TxRate: 1024,
@@ -446,7 +490,7 @@ func statusPageData() map[string]any {
Status: health.StatusOK,
},
"Sockets": []health.Socket{
{Name: "OpenDKIM", Path: "/run/opendkim/opendkim.sock", Present: true, Status: health.StatusOK, Detail: "Listening."},
{Name: "OpenDKIM", Path: "/run/opendkim/opendkim.sock", Present: true, Status: health.StatusOK, Detail: "Listening"},
},
"SocketStatus": health.StatusOK,
"OverallStatus": health.StatusError,
+14 -2
View File
@@ -34,9 +34,10 @@ var pageFiles = map[string][]string{
"setup": {"templates/setup.html"},
"login": {"templates/login.html"},
"dashboard": {"templates/dashboard.html"},
"account": {"templates/account.html"},
"settings": {"templates/settings.html"},
"users": {"templates/users.html"},
"user_form": {"templates/user_form.html"},
"user_delete": {"templates/user_delete.html"},
"backup": {"templates/backup.html", "templates/encrypt_fields.html"},
"domain_detail": {"templates/domain_detail.html", "templates/encrypt_fields.html"},
"domain_delete": {"templates/domain_delete.html"},
@@ -65,7 +66,7 @@ func New(version string) (*Engine, error) {
}
for name, files := range pageFiles {
patterns := append([]string{"templates/layout.html"}, files...)
tmpl, err := template.New("layout.html").ParseFS(assetsFS, patterns...)
tmpl, err := template.New("layout.html").Funcs(templateFuncs()).ParseFS(assetsFS, patterns...)
if err != nil {
return nil, fmt.Errorf("parse template %s: %w", name, err)
}
@@ -81,6 +82,17 @@ func New(version string) (*Engine, error) {
return e, nil
}
// templateFuncs supplies helpers shared across page templates.
func templateFuncs() template.FuncMap {
return template.FuncMap{
// back builds the map back_link reads; keeps href and label paired at
// the call site instead of repeating the <a class="back"> markup.
"back": func(href, label string) map[string]string {
return map[string]string{"Href": href, "Label": label}
},
}
}
// Page returns a parsed page template by logical name. It is exported for
// template guard tests that assert structural properties across all pages.
func (e *Engine) Page(name string) *template.Template {
+21 -2
View File
@@ -1,4 +1,4 @@
// Package web implements the SelfPost control panel's HTTP surface: the
// Package web implements the SelfPost control panel's HTTP surface: the
// one-time administrator setup flow (security.md), login/session handling
// (security.md) and the authenticated shell the later phases build on.
package web
@@ -70,6 +70,11 @@ type Config struct {
// checks must not go through the system resolver — see dnscheck's
// externalResolver — so this is how a closed network points them at its own.
DNSResolvers []string
// RateLimitMessagesPerIP and RateLimitWindowSeconds are the level-1
// Postfix anvil backstop (env RATE_LIMIT_*), mirrored into the panel for
// display and to cap domain/app level-2 ceilings (guide § Rate limiting).
RateLimitMessagesPerIP int
RateLimitWindowSeconds int
}
// Server is the panel HTTP application.
@@ -105,6 +110,8 @@ func New(st *store.Store, domains *domain.Service, apps *app.Service, cfg Config
TLSCertFile: cfg.TLSCertFile,
OpenDKIMSocket: cfg.OpenDKIMSocket,
JournalSocket: cfg.JournalSocket,
RateLimitMessagesPerIP: cfg.RateLimitMessagesPerIP,
RateLimitWindowSeconds: cfg.RateLimitWindowSeconds,
}, v, dnscheck.New(cfg.DNSResolvers), &health.MachineSampler{}, a)
return &Server{cfg: cfg, auth: a, handlers: h}, nil
}
@@ -151,13 +158,16 @@ func (s *Server) Handler() http.Handler {
authed.HandleFunc("POST /applications/{aid}/delete", h.HandleDeleteApplication)
authed.HandleFunc("POST /reload", h.HandleReload)
authed.HandleFunc("/account", h.HandleAccount)
authed.HandleFunc("/settings", h.HandleSettings)
authed.HandleFunc("/account", redirectSettings)
authed.HandleFunc("GET /users", h.HandleUsers)
authed.HandleFunc("GET /users/new", h.HandleUserNew)
authed.HandleFunc("POST /users/new", h.HandleUserNew)
authed.HandleFunc("GET /users/{uid}", h.HandleUserEdit)
authed.HandleFunc("POST /users/{uid}", h.HandleUserEdit)
authed.HandleFunc("GET /users/{uid}/delete", h.HandleUserDeleteConfirm)
authed.HandleFunc("POST /users/{uid}/delete", h.HandleUserDelete)
authed.HandleFunc("GET /backup", h.HandleBackupPage)
authed.HandleFunc("POST /backup", h.HandleBackup)
@@ -174,6 +184,15 @@ func (s *Server) Handler() http.Handler {
return s.secure(mux)
}
// redirectSettings sends legacy /account bookmarks to /settings (308 preserves POST).
func redirectSettings(w http.ResponseWriter, r *http.Request) {
target := "/settings"
if r.URL.RawQuery != "" {
target += "?" + r.URL.RawQuery
}
http.Redirect(w, r, target, http.StatusPermanentRedirect)
}
func redirectHome(w http.ResponseWriter, r *http.Request) {
p, ok := auth.PrincipalFromRequest(r)
if ok && !p.IsGlobal() {
+6 -1
View File
@@ -60,8 +60,13 @@ services:
# records the test harness publishes, so Postfix's outbound MX lookup for
# the sink and the panel's own dnscheck resolve inside the isolated network
# instead of hitting the real internet.
#
# Pinned to a release tag rather than `latest`, so a CoreDNS upload cannot
# change what the release gate runs against between two runs of the same
# commit. Not a digest: the tag is a multi-arch manifest and the stand has to
# come up on arm64 developer machines as well as amd64 CI.
coredns:
image: coredns/coredns:latest
image: coredns/coredns:1.14.6
command: ["-conf", "/dns/Corefile"]
volumes:
- ./dns-stage:/dns
+98
View File
@@ -0,0 +1,98 @@
package e2e
import (
"fmt"
"os"
"os/exec"
"path/filepath"
"strings"
"testing"
)
// checkLogrotateConfigMode verifies the image pins /etc/logrotate.d/mail at 0644
// so logrotate will not silently ignore it (docs/development.md § Building
// binaries and the image).
func checkLogrotateConfigMode(s *stack) error {
mode, err := s.execIn("selfpost", "stat", "-c", "%a", "/etc/logrotate.d/mail")
if err != nil {
return fmt.Errorf("stat logrotate config: %w", err)
}
mode = strings.TrimSpace(mode)
if mode != "644" {
return fmt.Errorf("/etc/logrotate.d/mail mode is %q, want 644", mode)
}
return nil
}
// checkLogrotateRotation forces a rotation and checks the recreated mail.log is
// panel-readable (0640 postfix:selfpost per build/logrotate-mail.conf).
func checkLogrotateRotation(s *stack) error {
const logPath = "/data/log/mail.log"
marker := "e2e-logrotate-marker\n"
if _, err := s.execIn("selfpost", "sh", "-c",
fmt.Sprintf("printf %q >> %s", marker, logPath)); err != nil {
return fmt.Errorf("write mail.log: %w", err)
}
if _, err := s.execIn("selfpost", "logrotate", "-f", "/etc/logrotate.d/mail"); err != nil {
return fmt.Errorf("logrotate -f: %w", err)
}
rotated, err := s.execIn("selfpost", "sh", "-c", "test -f /data/log/mail.log.1 && echo yes")
if err != nil || strings.TrimSpace(rotated) != "yes" {
return fmt.Errorf("expected /data/log/mail.log.1 after forced rotation (out=%q err=%v)", rotated, err)
}
mode, err := s.execIn("selfpost", "stat", "-c", "%a", logPath)
if err != nil {
return fmt.Errorf("stat rotated mail.log: %w", err)
}
if strings.TrimSpace(mode) != "640" {
return fmt.Errorf("new %s mode is %q, want 640", logPath, strings.TrimSpace(mode))
}
body, err := s.execIn("selfpost", "grep", "-F", strings.TrimSpace(marker), "/data/log/mail.log.1")
if err != nil || !strings.Contains(body, strings.TrimSpace(marker)) {
return fmt.Errorf("rotated file does not contain marker (out=%q err=%v)", body, err)
}
return nil
}
// TestImageBuildPreservesLogrotateMode builds from a context where
// logrotate-mail.conf is group-writable and checks the image still ships 0644.
func TestImageBuildPreservesLogrotateMode(t *testing.T) {
conf := filepath.Join(h.repoRoot, "build", "logrotate-mail.conf")
info, err := os.Stat(conf)
if err != nil {
t.Fatalf("stat source config: %v", err)
}
origMode := info.Mode().Perm()
if err := os.Chmod(conf, origMode|0o020); err != nil {
t.Fatalf("chmod g+w source config: %v", err)
}
t.Cleanup(func() {
_ = os.Chmod(conf, origMode)
})
tag := "selfpost:e2e-logrotate-mode"
build := exec.Command("docker", "build",
"-f", filepath.Join(h.repoRoot, "build", "Dockerfile"),
"-t", tag,
"--build-arg", "VERSION=e2e",
h.repoRoot,
)
build.Env = os.Environ()
if out, err := build.CombinedOutput(); err != nil {
t.Fatalf("docker build with group-writable context config: %v\n%s", err, out)
}
t.Cleanup(func() {
_, _ = exec.Command("docker", "rmi", "-f", tag).CombinedOutput()
})
run := exec.Command("docker", "run", "--rm", tag, "stat", "-c", "%a", "/etc/logrotate.d/mail")
run.Env = os.Environ()
out, err := run.CombinedOutput()
if err != nil {
t.Fatalf("stat in image: %v\n%s", err, out)
}
if strings.TrimSpace(string(out)) != "644" {
t.Fatalf("image logrotate config mode is %q, want 644", strings.TrimSpace(string(out)))
}
}
+8 -2
View File
@@ -19,7 +19,7 @@ import (
"time"
)
var statusCellPattern = regexp.MustCompile(`<td>(queued|sent|deferred|bounced|rejected)</td>`)
var statusCellPattern = regexp.MustCompile(`class="st st-[^"]+">(queued|sent|deferred|bounced|rejected)</span>`)
// h is the single shared stand for the whole ordered scenario in TestE2E.
// TestHostnameGate does not use it — it spins its own disposable container.
@@ -105,6 +105,12 @@ func TestE2E(t *testing.T) {
if err := checkSupervisorProcesses(h); err != nil {
t.Fatal(err)
}
if err := checkLogrotateConfigMode(h); err != nil {
t.Fatal(err)
}
if err := checkLogrotateRotation(h); err != nil {
t.Fatal(err)
}
if err := waitForPanelReady(); err != nil {
t.Fatal(err)
}
@@ -241,7 +247,7 @@ func uniqueToken(label string) string {
}
func containsCell(html, needle string) bool {
return strings.Contains(html, "<td>"+needle+"</td>")
return strings.Contains(html, `<span class="st st-`) && strings.Contains(html, `">`+needle+`</span>`)
}
func firstStatusCell(html string) string {
+1 -1
View File
@@ -219,5 +219,5 @@ func testLevel1RateLimit(t *testing.T, sc *scenario) {
return
}
}
t.Fatalf("level-1 rate limit (RATE_LIMIT_MESSAGES_PER_IP=5) never tripped after %d sends", maxAttempts)
t.Fatalf("level-1 rate limit (RATE_LIMIT_MESSAGES_PER_IP=50) never tripped after %d sends", maxAttempts)
}
+9 -6
View File
@@ -194,15 +194,18 @@ func (c *panelClient) addApplication(domainID, login, mode, addresses string) (a
return appLogin, password, nil
}
// setRateLimit saves a level-2 differentiated limit (guide § Rate limiting)
// on either a domain (/domains/{id}/ratelimit) or an application
// (/applications/{id}/ratelimit).
// setRateLimit saves a level-2 limit (guide § Rate limiting) on an application
// (/applications/{id}/ratelimit). allowedIP is required for the trusted-IP
// override; domain ceilings are posted without IPs.
func (c *panelClient) setRateLimit(path, allowedIP string, maxMessages, windowSeconds int) error {
resp, body, err := c.postForm(path, url.Values{
"allowed_ips": {allowedIP},
vals := url.Values{
"max_messages": {fmt.Sprintf("%d", maxMessages)},
"window_seconds": {fmt.Sprintf("%d", windowSeconds)},
})
}
if allowedIP != "" {
vals.Set("allowed_ips", allowedIP)
}
resp, body, err := c.postForm(path, vals)
if err != nil {
return err
}