Remove crate from salsa `EditionedFileId` by Veykril · Pull Request #21767 · rust-lang/rust-analyzer

Veykril · 2026-03-06T12:40:06Z

This broke the contract of `HashEqLike`, causing interning to return different IDs

ChayimFriedman2 · 2026-03-09T08:05:25Z

@Veykril did you measure the memory impact?

Copilot

Pull request overview

This PR refactors EditionedFileId to no longer carry a Crate, and threads crate context explicitly where required (notably item-tree lowering), aiming to fix a regression around span/origin handling in large workspaces (rust-analyzer issue #21721).

Changes:

Simplify base_db::EditionedFileId to wrap only span::EditionedFileId and remove “guess origin” APIs.
Update item-tree queries/APIs to accept an explicit Crate parameter and adjust call sites accordingly.
Refresh affected tests/fixtures and golden test outputs to match new IDs/behavior.

Reviewed changes

Copilot reviewed 38 out of 38 changed files in this pull request and generated 4 comments.

Show a summary per file

File	Description
crates/test-fixture/src/lib.rs	Switch fixture helpers to construct `EditionedFileId` from `span::EditionedFileId` without crate guessing.
crates/rust-analyzer/src/cli/ssr.rs	Update SSR CLI to use `EditionedFileId::current_edition`.
crates/rust-analyzer/src/cli/analysis_stats.rs	Thread `Crate` into item-tree calls; adjust source-root iteration.
crates/load-cargo/src/lib.rs	Use `span_file_id` accessors when producing proc-macro span responses.
crates/ide/src/view_item_tree.rs	Pass an explicit crate into `file_item_tree` for debug item-tree rendering.
crates/ide/src/typing/on_enter.rs	Replace `current_edition_guess_origin` usage with `current_edition`.
crates/ide/src/typing.rs	Remove crate-guessing wrapper creation; pick a crate and construct `EditionedFileId` directly.
crates/ide/src/signature_help.rs	Update expected signature-help output ordering.
crates/ide/src/references.rs	Update reference-search scope construction to new `EditionedFileId` API.
crates/ide/src/lib.rs	Replace `current_edition_guess_origin` usage with `current_edition` in multiple APIs.
crates/ide-ssr/src/from_comment.rs	Use `current_edition` when parsing comment-based SSR patterns.
crates/ide-db/src/traits.rs	Update test fixture `EditionedFileId` construction to `from_span_file_id`.
crates/ide-db/src/test_data/test_symbols_with_imports.txt	Update golden IDs after interned-ID changes.
crates/ide-db/src/test_data/test_symbols_exclude_imports.txt	Update golden IDs after interned-ID changes.
crates/ide-db/src/test_data/test_symbol_index_collection.txt	Update golden salsa IDs after query/input changes.
crates/ide-db/src/test_data/test_doc_alias.txt	Update golden salsa IDs after query/input changes.
crates/ide-db/src/search.rs	Construct `EditionedFileId` without embedding crate context.
crates/hir/src/semantics/child_by_source.rs	Adjust derive-macro mapping to obtain a crate without `HirFileId::krate`.
crates/hir/src/semantics.rs	Update edition attachment fallbacks to `current_edition` and remove crate from `EditionedFileId::new`.
crates/hir/src/lib.rs	Update macro diagnostic span/file-id comparisons to new accessor naming.
crates/hir-ty/src/tests/macros.rs	Update cfg-test fixture to use a non-`test` cfg key.
crates/hir-ty/src/tests/incremental.rs	Update incremental event expectations (new shim/query names).
crates/hir-expand/src/span_map.rs	Use `span_file_id` when building real span maps.
crates/hir-expand/src/lib.rs	Remove `HirFileId::krate` helper and adjust edition retrieval.
crates/hir-expand/src/db.rs	Resolve spans via `from_span_file_id` rather than guessing origin crate.
crates/hir-expand/src/builtin/fn_macro.rs	Construct `EditionedFileId` without crate in include/relative-file logic.
crates/hir-def/src/nameres/tests/incremental.rs	Update item-tree query call sites to pass `db.test_crate()`.
crates/hir-def/src/nameres/mod_resolution.rs	Drop `Crate` threading from module-file resolution and adjust `EditionedFileId` construction.
crates/hir-def/src/nameres/collector.rs	Thread crate context into item-tree lowering and adjust tests using `test_crate`.
crates/hir-def/src/macro_expansion_tests/mod.rs	Use `db.test_crate()` instead of extracting crate from file id.
crates/hir-def/src/macro_expansion_tests/mbe.rs	Update golden expansion spans/IDs after interned-ID changes.
crates/hir-def/src/item_tree/tests.rs	Pass crate explicitly into `file_item_tree` in tests.
crates/hir-def/src/item_tree/lower.rs	Store crate in lowering context to obtain cfg options without `HirFileId::krate`.
crates/hir-def/src/item_tree.rs	Change tracked item-tree queries to accept `Crate`; plumb crate through `TreeId::item_tree`.
crates/hir-def/src/expr_store/tests/body/block.rs	Update golden `BlockId` after interned-ID changes.
crates/hir-def/src/db.rs	Change `DefDatabase::file_item_tree` signature to include `Crate`.
crates/base-db/src/input.rs	Update `Crate::root_file_id` to construct `EditionedFileId` without crate.
crates/base-db/src/editioned_file_id.rs	Replace custom interned struct (with crate) with a simpler `#[salsa::interned]` wrapper.

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

You can also share your feedback on Copilot code review. Take the survey.

View changes since this review

Copilot · 2026-03-09T09:08:15Z

crates/ide/src/typing.rs

+    let krate = db
+        .relevant_crates(position.file_id)
        .first()
-        .map_or(Edition::CURRENT, |crates| crates.data(db).edition);
-    // FIXME: We are hitting the database here, if we are unlucky this call might block momentarily
-    // causing the editor to feel sluggish! We need to make this bail if it would block too long?
-    let editioned_file_id_wrapper = EditionedFileId::from_span_guess_origin(
-        db,
-        span::EditionedFileId::new(position.file_id, edition),
-    );
+        .copied()
+        .unwrap_or_else(|| *db.all_crates().first().unwrap());
+    let edition = krate.data(db).edition;
+    let editioned_file_id_wrapper = EditionedFileId::new(db, position.file_id, edition);


unwrap_or_else(|| *db.all_crates().first().unwrap()) can panic when the crate graph hasn’t been populated yet (e.g., during startup), crashing on-type assists. Consider falling back to Edition::CURRENT (and/or early-return None) when there is no relevant crate, instead of unwrapping all_crates().first().

Copilot · 2026-03-09T09:08:15Z

crates/rust-analyzer/src/cli/analysis_stats.rs

        let mut item_tree_sw = self.stop_watch();
        let source_roots = krates
            .iter()
            .cloned()
-            .map(|krate| db.file_source_root(krate.root_file(db)).source_root_id(db))
+            .map(|krate| (db.file_source_root(krate.root_file(db)).source_root_id(db), krate))
            .unique();

        let mut dep_loc = 0;
        let mut workspace_loc = 0;
        let mut dep_item_trees = 0;
        let mut workspace_item_trees = 0;

        let mut workspace_item_stats = PrettyItemStats::default();
        let mut dep_item_stats = PrettyItemStats::default();

-        for source_root_id in source_roots {
+        for (source_root_id, krate) in source_roots {
            let source_root = db.source_root(source_root_id).source_root(db);


source_roots is now deduped with .unique() over (SourceRootId, Crate), which means the same SourceRootId will be visited once per crate (common for workspace roots). This can dramatically increase the amount of work (re-scanning/parsing/item-tree building the same files repeatedly) and skew LOC/item-tree counts. Consider keeping uniqueness by SourceRootId and choosing a representative crate per root (or determining a per-file crate inside the loop via relevant_crates(file_id) as needed).

Copilot · 2026-03-09T09:08:15Z

crates/base-db/src/editioned_file_id.rs

@@ -1,305 +1,46 @@
 //! Defines [`EditionedFileId`], an interned wrapper around [`span::EditionedFileId`] that
 //! is interned (so queries can take it) and remembers its crate.


The module-level doc comment still claims EditionedFileId “remembers its crate”, but the type no longer stores a crate. Please update this comment to reflect the new semantics (interned wrapper around span::EditionedFileId only).

Suggested change

//! is interned (so queries can take it) and remembers its crate.

//! is interned (so queries can take it) and stores only the underlying `span::EditionedFileId`.

crates/base-db/src/editioned_file_id.rs

rustbot added the S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. label Mar 6, 2026

Veykril enabled auto-merge March 6, 2026 12:40

Veykril force-pushed the push-xxyxurxwwmvy branch 2 times, most recently from 86ae3d8 to b784fb3 Compare March 6, 2026 12:56

Veykril added this pull request to the merge queue Mar 6, 2026

Veykril removed this pull request from the merge queue due to a manual request Mar 6, 2026

fix: Fix invalid PartialEq impl for EditionedFileIdData

99fa130

This broke the contract of `HashEqLike`, causing interning to return different IDs

Veykril force-pushed the push-xxyxurxwwmvy branch from b784fb3 to d9560f8 Compare March 7, 2026 07:01

Veykril changed the title ~~fix: Fix invalid PartialEq impl for EditionedFileIdData~~ Remove crate-based EditionedFileIdData Mar 7, 2026

Veykril force-pushed the push-xxyxurxwwmvy branch 4 times, most recently from b33ecf3 to 6d08c43 Compare March 8, 2026 18:43

Remove crate from salsa EditionedFileId

5abc9af

Veykril force-pushed the push-xxyxurxwwmvy branch from 6d08c43 to 5abc9af Compare March 8, 2026 18:45

Veykril changed the title ~~Remove crate-based EditionedFileIdData~~ Remove crate from salsa EditionedFileId Mar 8, 2026

Veykril changed the title ~~Remove crate from salsa EditionedFileId~~ Remove crate from salsa EditionedFileId Mar 8, 2026

Veykril added this pull request to the merge queue Mar 9, 2026

Merged via the queue into rust-lang:master with commit 018847f Mar 9, 2026
17 checks passed

Veykril deleted the push-xxyxurxwwmvy branch March 9, 2026 08:07

rustbot removed the S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. label Mar 9, 2026

Veykril requested a review from Copilot March 9, 2026 09:02

Copilot started reviewing on behalf of Veykril March 9, 2026 09:03 View session

Copilot AI reviewed Mar 9, 2026

View reviewed changes

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Remove crate from salsa `EditionedFileId`#21767

Remove crate from salsa `EditionedFileId`#21767
Veykril merged 2 commits intorust-lang:masterfrom
Veykril:push-xxyxurxwwmvy

Veykril commented Mar 6, 2026 •

edited

Loading

Uh oh!

ChayimFriedman2 commented Mar 9, 2026

Uh oh!

Uh oh!

Copilot AI left a comment •

edited by rustbot

Loading

Uh oh!

Copilot AI Mar 9, 2026

Uh oh!

Copilot AI Mar 9, 2026

Uh oh!

Copilot AI Mar 9, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

		@@ -1,305 +1,46 @@
		//! Defines [`EditionedFileId`], an interned wrapper around [`span::EditionedFileId`] that
		//! is interned (so queries can take it) and remembers its crate.

Conversation

Veykril commented Mar 6, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

ChayimFriedman2 commented Mar 9, 2026

Uh oh!

Uh oh!

Copilot AI left a comment • edited by rustbot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

Copilot AI Mar 9, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Mar 9, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Mar 9, 2026

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

Veykril commented Mar 6, 2026 •

edited

Loading

Copilot AI left a comment •

edited by rustbot

Loading