refactor(lint): read the HIR through gcx - #16621
Draft
DaniPopes wants to merge 16 commits into
Draft
Conversation
`Gcx` already exposes the HIR as `gcx.hir`, so helpers and visitors that carried both a `gcx` and a `hir` were redundant. Drop the `hir` parameter and field from `costly-loop`, `cache-array-length` and `incorrect-using-for` and read the HIR through `gcx` instead. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Drop redundant hir parameters and fields from helpers, analyzers and visitors that already have a Gcx, reading the HIR through gcx.hir. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
`Gcx` already exposes the HIR as `gcx.hir`, so the low-severity lints no longer pass or store a separate `hir` next to `gcx`. Analyzer and walker structs keep only `gcx` and implement `Visit::hir` as `&self.gcx.hir`; helpers that were hir-only but only ever called with a `gcx` in scope (`resolved_super_function_ids`, `is_deprecated_oz`, `is_unchecked_token_op`, `event_mentions_state_var`) now take `gcx` so call sites are uniform. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Analyzer and visitor structs in the high-severity lints stored both the global context and a separate HIR reference, and helpers took both as parameters. Drop the redundant `hir` fields and parameters and read the HIR through `gcx.hir` instead; `hir::Visit` impls return `&self.gcx.hir`. Helpers that only need the HIR and have no `gcx` in scope are unchanged. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Contributor
✅ Changelog foundThe deterministic check will validate the changed entry. |
Bump Solar to paradigmxyz/solar#1376, whose LateLintPass hooks receive only the global context, and update every pass accordingly. The bump also brings the new for-loop lowering that keeps the update statement on LoopSource::For instead of inside the body, so the loop-walking analyses now run the update after each completed iteration through the shared loop_update and loop_stmts helpers, and the lints that reverse-engineered the old lowering (ecrecover, function-selector-collision, cache-array-length, enumerable-loop-removal) read it directly.
The lifetime is the global context's, of which the HIR is one field; name it accordingly, matching the Solar side.
DaniPopes
added a commit
to paradigmxyz/solar
that referenced
this pull request
Sep 4, 2026
`Gcx<'hir>` already exposes the HIR as `gcx.hir`, so every `LateLintPass` hook now receives only the global context: the hooks that took both `gcx` and `hir` drop the `hir` parameter, and the hooks that took only `hir` take `gcx` instead, giving the whole trait one uniform shape. `LateLintVisitor` no longer stores a separate HIR handle and its constructor loses that argument. This is a breaking change for downstream lint passes (Foundry's lints are being updated alongside in foundry-rs/foundry#16621). Written with AI assistance (Claude Code). 🤖 Generated with [Claude Code](https://claude.com/claude-code)
Pin the merged paradigmxyz/solar#1376 instead of its branch head.
Drop the local hir bindings that only aliased gcx.hir.
* chore(ci): enforce Cargo cooldown * fix(ci): annotate Cargo action pins * fix(ci): harden cargo cooldown Invoke the checksum-pinned verifier directly with isolated policy state and all-target coverage. Run baseline checks before candidate code and pin Docker build tools outside the workspace lockfile. * refactor(ci): use shared cargo cooldown Pin the merged tempoxyz action with strict project policy for both the primary dependency gate and benchmark baseline. Remove the duplicated installer, verifier wrapper, and local behavioral tests now owned upstream. * fix(ci): allow pre-policy benchmark baseline Enable strict cooldown configuration only when the checked-out master baseline contains cooldown.toml. Older baselines remain valid while every baseline that carries policy is still validated strictly.
* refactor(symbolic): consolidate call state machines * style nits --------- Co-authored-by: Mablr <59505383+mablr@users.noreply.github.com>
Drop a stale duplicate doc comment and import Gcx instead of spelling its path inline.
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
4 participants
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.
Follow-up to #16615.
Gcx<'hir>already carries the HIR asgcx.hir, yet most lints kept passing a separatehir: &Hiralongsidegcxto their helpers and storing both in their analyzer and visitor structs. This drops that redundancy everywhere: helpers and analyzers read&gcx.hir(or bind it once locally),hir::Visitimpls return&self.gcx.hir, and no function or struct in the crate takes both any more.The
LateLintPassAPI itself is fixed on the Solar side in paradigmxyz/solar#1376, whose hooks receive only the global context; this PR bumps the Solar pin to its merge commit and migrates every pass. The bump also picks up Solar's newfor-loop lowering, which keeps the update statement onLoopSource::For { update }instead of splicing it into the body, so the loop-walking analyses now run the update after each completed iteration through the sharedloop_update/loop_stmtshelpers, and the lints that used to reverse-engineer the old lowering (ecrecover,function-selector-collision,cache-array-length,enumerable-loop-removal) read it directly.No lint behavior changes; the UI fixtures are untouched and pass unchanged. Written with AI assistance (Claude Code).
🤖 Generated with Claude Code