Remove the null-MinimumZoom branch and ExplicitZoomLevel #209
Loading…
Reference in a new issue
No description provided.
Delete branch "retire-explicit-zoom"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Closes task 266. With tasks 251 and 254 both landed, the only surviving producer of a
MapZoomRequest(Home.HandleCurrentLocation) always names a floor, soMinimumZoomwas nevernullandLeafletMap.ExplicitZoomLevelwas never reached. Confirmed by search, not assumption — no other constructor call exists anywhere in the tree.MapZoomRequest.MinimumZoomis now a requiredint.ExplicitZoomLeveland thesetViewbranch it fed are removed. Tests that existed only to cover that branch are deleted, not inverted; the tests covering the survivingRequestIdmechanism are kept and updated to the new required parameter.ADR-0172 records the removal and partly supersedes ADR-0132 and ADR-0166, which are the reason the branch was still there. Both, plus ADR-0167, carry
Correctedfields and inline blockquotes at the claims this change falsifies, rather than being rewritten.6bb082602577a656d3ddVerdict: mergeable
Verified independently rather than taking the PR body's claims on trust:
new MapZoomRequest(...):Home.razor.cs:313is the only constructor call anywhere insrc/ortests/, and it always suppliesCurrentLocationMinimumZoom. No JS interop or external parameter set reachesZoomRequestanother way.ExplicitZoomLevelno longer exists as an identifier anywhere — the two remaining hits are prose (<c>ExplicitZoomLevel</c>) in doc comments. The dead-code claim holds.setView/null branch; nothing live lost coverage. The five kept/renamed tests still assert real behaviour (RequestId reconciliation, no redundant calls, interop-failure handling, never-zoom-out).dotnet build -c Releaseis clean (0 warnings/errors) anddotnet testonPlaceMark.Architecture.Tests(23/23, includingAdrFiles_Corrected_HasAMatchingInlineMarkerInBothDirectionsand the reciprocal supersession check) andPlaceMark.WebUI.Tests(914 passed, 4 pre-existing unrelated skips) both pass, all under the pinned 10.0.100 SDK.Partially supersedesfield lists both 0132 and 0166; 0132'sPartially superseded byfield's continuation prose names 0172 (picked up by the architecture test's field-continuation parser, confirmed by reading its logic); 0166 gets its own dedicatedPartially superseded by: [ADR-0172]field. Corrections are honest — original claims (16 as the chosen level, the "kept rather than deleted" judgement) stay legible, with dated inline blockquotes at the falsified sentences rather than rewrites.LeafletMap.OnParametersSetAsync'sZoomRequestbranch is gated behind_map is not null, which only becomes true afterOnAfterRenderAsync(firstRender: true)runs — the same method #210's readiness wait hooks into viacreateMap. The two changes don't interact: this PR touches neitherOnAfterRenderAsyncnorcreateMap.center,_animateZoom— are pre-existing vendored Leaflet internals untouched by this diff).CI: build job green as of this review; e2e and container-images were still running (SHA-verified via
get_workflow_runon run 885, cross-checked against the run's own commit SHA — the run-793-by-URL-id lookup returns a stale unrelated run, exactly the trap called out, so don't trust that id directly). Worth a final glance at the e2e result before merging, per the note that main's suite is now healthy and a failure here would be meaningful.