mix
295d1afb7d
The roadmap links straight into these files, so a reader following an item landed in Russian one click after an English page. Translated in full; goals, boundaries, done-when criteria and risks are unchanged in substance. The model assigned to inbound-relay is dropped, as it was from the roadmap item: model routing lives in development.md. Figures and code references were checked against the tree while translating rather than copied forward — internal/web is still 50 files, 25 .go, ~4300 lines, and requireAuth(authed) and sendLogData are still where domain-admin.md says they are. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
2.1 KiB
Plan: web-split (splitting internal/web)
Status: agreed
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).
- 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. - domain-admin — authorisation in every handler builds on a package layout already chosen.
- 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.
Done when
The decision is made deliberately when the work starts — either the package is
split along the chosen scheme, or it is settled that it stays flat. After a
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.