Fix semantic classification errors for generics, slices, CEs, and open type#19939
Open
T-Gro wants to merge 7 commits into
Open
Fix semantic classification errors for generics, slices, CEs, and open type#19939T-Gro wants to merge 7 commits into
T-Gro wants to merge 7 commits into
Conversation
Pin current behaviour of all 7 items reported in #19905. Items that already pass on HEAD (item 1 via #19813, item 7 with the minimal repro) become real regression tests. Items that still reproduce are marked [<Fact(Skip = "Tracked in #19905 ...")>] and will be un-skipped (and inverted) by their fix sprints. No production code changes in this commit. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
…ot classified as Function (#19905) In TcMethodItemThen, TcTypeItemThen, and SynExpr.New, CallNameResolutionSink was emitting Item.MethodGroup / Item.Types / Item.CtorGroup resolutions with the wide mExprAndTypeArgs / synObjTy.Range range covering 'Method<int>' / 'MailboxProcessor<int*int>' / 'MailboxProcessor<int>.Start'. The downstream classifier in SemanticClassification.fs then painted the entire span as Function/DisposableType. Narrow the CNR ranges to the identifier-only ranges (mItem/mItemIdent for methods, the type's leading SynLongIdent for ctors via the new synTypeLeadingIdentRange helper, and rangeOfLid longId as the wholem for the recursive dot-lookup) so the classifier paints only the identifier and leaves the type-arg punctuation and inner type names to their own resolutions. Updates Project11/17/18 symbol-use baselines and one neg/error-range baseline to reflect the narrower (more correct) ranges that flow from the narrower CNRs. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
…`]` as Function (#19905) TcIndexingThen synthesized the GetSlice/SetSlice identifier with mWholeExpr as its range, causing the op_GetSlice CNR to span the entire `list[...]` expression including the closing `]`. The semantic classifier then painted that whole span as Method/Function. Mark the synthesized identifier's range as synthetic for slice expressions (not regular indexing) so NotifyNameResolution drops the CNR. Regular `x.[n]` indexing keeps the visible range to preserve find-all-references on the implicit Item indexer (Project17 baselines unchanged). Also swap the symmetric `range1` to `range2` in ExpandIndexArgs for the missing upper-bound branch (no behavioural change for classification - mkSynNoneExpr already makes the range synthetic - but matches the `Some` branch using range2 for the upper bound). Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
…19905) `UnusedOpens.getUnusedOpens` only recorded symbol uses whose declaring entity was a namespace or module. Members imported by `open type T` have T as their declaring entity, so when the match between an open declaration's target and a symbol use's declaring entity is needed, the type was missing from the lookup dictionary. Now records symbol uses against any declaring entity (namespace, module, or type) for both partitions returned by `splitSymbolUses`. The change is purely additive: `symbolUses2` still flows through the existing `RevealedSymbolsContains` path, while the dictionary lookup now also matches type entities for `open type`. Entity equality goes through `IsEffectivelySameAs` so type entities match themselves correctly. Adds two new regression tests in `SemanticClassificationRegressions.fs`: one confirming `open type T` with no usage is still flagged unused, and one confirming plain `open System` is still flagged when unused. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
…s ComputationExpression (#19905) takeCustomBuilder in SemanticClassification.fs only handled the 2-element CNR-group case (Value + CustomBuilder). Inside a list/array comprehension, name resolution can record 3+ CNRs at the same range for the builder identifier; the existing code fell through, the first Item.Value hit won the dedup race, and the builder name was painted as a local value. Generalize takeCustomBuilder to prefer any Item.CustomBuilder or Item.CustomOperation over Item.Value regardless of group size. Also apply the same grouping in the whole-file (range = None) code path so the basic `async { .. }` case is unaffected. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Add one bullet per fixed item from the #19905 cluster under docs/release-notes/.FSharp.Compiler.Service/11.0.100.md, ### Fixed. Issue-comment draft is in .tools/ralph/issue-comment.md (not tracked); the user will post it manually after pushing the fix commits. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Contributor
✅ No release notes required |
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
github-actions
Bot
added
the
AI-Tooling-Check-Bypassed
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Fixes #19905
Type.Method<int>()/new T<args>(...)/T<args>.M(...)no longer classify<…>as Function/Type; span narrows to the identifier.list[0..]trailing]no longer classified as Function.[ for .. do builder { } ]classified as ComputationExpression, not Value.open type Tno longer always reported as unused.