ID |
|
|---|---|
Status |
Backlog |
Bucket |
structural |
Priority |
5 |
Theme |
docs |
Created |
2026-07-08 |
Updated |
2026-07-08 |
Post-R438 stale-reference residue: ConditionResolution javadoc + fkjoin-alias-dead-storage item
R438’s In Review to Done review surfaced two staleness residues its own stale-reference sweep did not cover; captured here as a small cleanup since R438 is closed.
-
BuildContext.ConditionResolutionjavadoc drift. The record’s javadoc still asserts "exactly one ofrefanderroris non-null," butresolveConditionRefreturns(null, null)in two cases: whenclassName/methodName/svcis absent, and whenreflectTableMethodfails. That both-null case is exactly what R438’s review-fix (b0ab513) added caller guards for (res.ref() == nullchecks in the threeparsePathElementbranches). Correct the javadoc to state that a(null, null)result means "unresolved" and that callers must guardres.ref() == nullin addition tores.error() != null. -
fkjoin-alias-dead-storageBacklog item is stale.roadmap/fkjoin-alias-dead-storage.mdreferences the deleted typesFkJoin.alias/ConditionJoin.aliasand a staleBuildContext.java:694line number; R438’s reshape moved that concern ontoHop.alias. Refresh the item’s prose againstHop.alias(the dead-storage question survives the reshape, just under a new component) or retire it if the alias is now consumed.
Both are doc/comment-only; neither changes generated output or behavior.