Adding missing @Override annotations (part 5). Adding RRA/RWA/RUI annotations. Some localization and refactoring

[unv-unv]

No description provided.

[unv-unv]

@claude review pr with inline comments

[github-actions]

Claude finished @unv-unv's task in 2m 6s —— View job

---

PR Review — completed

• Read the full diff (50 files, +914/−760)
• Analyze changes for correctness / style issues
• Post inline review comment with a one-click suggestion

Summary

This is a large, mostly-mechanical cleanup PR: adding @Override, @RequiredReadAction/@RequiredWriteAction/@RequiredUIAccess annotations, migrating *Bundle.message(...) → *Localize calls, and modernizing syntax (pattern-matching instanceof, diamond operators, lambdas). The vast majority of it is correct and a nice readability improvement. 👍

I verified the non-trivial rewrites (the negated ContentIterator lambda in GroovycStubGenerator, the switch in GroovyElementPattern.methodCallParameter, the calculateAnchorToInsertBefore restructuring in GroovyExtractMethodHandler, and the many instanceof conversions) — those are all behavior-preserving.

Findings

🔴 1 bug — GroovyElseUnwrapper.findTopmostIfStatement (posted inline)

The pattern-matching rewrite changed the loop from climbing one ancestor level per iteration to two:

// before — parent moves up ONE level
while (parent.getParent() instanceof GrIfStatement) {
    parent = parent.getParent();
}
// after — ifStmt already *is* parent.getParent(), so this moves up TWO levels
while (parent.getParent() instanceof GrIfStatement ifStmt) {
    parent = ifStmt.getParent();
}

For nested else if chains this skips every other if and returns the wrong topmost element, breaking the "Unwrap else" action. The fix is parent = ifStmt; — see the inline suggestion on the file. (Contrast with GroovyElseUnwrapperBase.collectElementsToIgnore, where the same idiom is correct because the instanceof there matches parent itself.)

Minor nit (not blocking): In GroovyDocMethodHandler.shortenParameterReferences, the reworked if (... instanceof GrDocMethodReference methodRef) { line and its for body are indented inconsistently (6 spaces vs. the surrounding 8/10). Cosmetic only.

Nice work on this batch — once the GroovyElseUnwrapper one-liner is applied, this looks good to merge.
· branch overrides-5

[github-actions]

Behavioral regression from the pattern-matching rewrite.

The original loop climbed one level per iteration:

while (parent.getParent() instanceof GrIfStatement) {
    parent = parent.getParent();
}

Here parent = parent.getParent() is exactly the matched GrIfStatement. In the rewrite, the pattern variable ifStmt already is parent.getParent(), so parent = ifStmt.getParent() climbs two levels per iteration — skipping every other if and returning the wrong element for nested else if chains.

Suggested change

	parent = ifStmt.getParent();
	parent = ifStmt;

[unv-unv]

@claude review pr with inline comments

[github-actions]

Claude finished @unv-unv's task in 2m 42s —— View job

---

PR Review — completed

• Read the full diff (50 files, +916/−760)
• Analyze changes for correctness / behavior-preservation
• Post inline review comment with a one-click suggestion

Summary

This is a large, mostly-mechanical modernization pass: adding @Override, @RequiredReadAction/@RequiredWriteAction/@RequiredUIAccess, migrating *Bundle.message(...) → *Localize calls, pattern-matching instanceof, diamond operators, lambdas, and GROOVY_FILE_TYPE → INSTANCE. I reviewed every file and the vast majority is correct and a genuine readability win. 👍

I paid particular attention to the non-trivial rewrites and confirmed they are behavior-preserving:

• GroovycStubGenerator — the negated ContentIterator → lambda (fileOrDir.isDirectory() || JavaFileType.INSTANCE != fileOrDir.getFileType()) is a correct De Morgan transform of the original !isDirectory() && == JavaFileType → return false. ✅
• GroovyElementPattern.methodCallParameter — the switch (call) over GrMethodCall/GrNewExpression/default → null matches the original ternary chain, and the collapsed && guard preserves the original ordering/short-circuit. ✅
• GroovyExtractMethodHandler.calculateAnchorToInsertBefore — restructured getParent() check is equivalent. ✅
• GroovyConstantExpressionEvaluator / GroovyExpressionPattern / GroovyFindUsagesProvider — the many instanceof binding-pattern conversions are all faithful. ✅

Re: the previously-flagged GroovyElseUnwrapper bug

A prior automated review flagged findTopmostIfStatement as a two-levels-per-iteration regression. That does not apply to the current code — the committed version is already:

while (parent.getParent() instanceof GrIfStatement ifStmt) {
    parent = ifStmt;   // ifStmt == parent.getParent(), so this climbs exactly ONE level
}

Since ifStmt is bound to parent.getParent(), parent = ifStmt is identical to the original parent = parent.getParent(). No behavioral change — this one is correct. ✅ (GroovyElseUnwrapperBase.collectElementsToIgnore uses the same idiom correctly as well.)

Findings

No correctness issues. One cosmetic nit, posted inline as a one-click suggestion:

• GroovyElementType.java:48 — public String toString() switched to 4-space indentation while the @Override directly above it (and the rest of the file) uses tabs. Inline suggestion realigns it.

One thing worth a sanity check (not blocking): in GroovyConfigUtils, ModuleUtilCore.findModuleForPsiElement(psiElement) was replaced with psiElement.getModule() in isVersionAtLeast / getSDKVersion(PsiElement). That's fine as long as getModule() resolves the module the same way for these PSI elements — assuming it's the intended platform-API swap, no action needed.

Nice, clean batch — good to merge once the one-line indentation nit is applied.
· branch overrides-5

[github-actions]

Nit (cosmetic): this line switched to 4-space indentation, but the rest of the file — including the @Override right above it — is tab-indented. Realign to a tab for consistency.

Suggested change

	public String toString()
	public String toString()