Avoid oscillating line wrapping due to symbol indices - #162608

aytey · 2026-09-10T18:58:25Z

A comment against PR #151817 mentioned that the test file added in this PR kept causing "arbitrary" test file changes.

This is because one line of the expanded output sits just at 78 columns, which is the pretty-printer's margin. Adding or removing a symbol shifts every index after it, and when one of those gains or loses a digit, the wrapping moves.

We cannot use normalize-stdout as the layout is decided before normalization runs.

We fix this by avoiding a strict golden comparison, and instead using a regex to directly test that the expected syntax contexts exist in the expanded output.

LLM disclosure: I used an LLM to help me understand the problem and how the regex would look, which the LLM then actioned. I reviewed the diff and the commit (along with the message) was performed by hand with me (with the LLM checking the content of the commit message, but the commit message was, indeed, written by a human).

rustbot · 2026-09-10T18:58:30Z

r? @TaKO8Ki

rustbot has assigned @TaKO8Ki.
They will have a look at your PR within the next two weeks and either review your PR or reassign to another reviewer.

Use r? to explicitly pick a reviewer

Why was this reviewer chosen?

The reviewer was selected based on:

Owners of files modified in this PR: compiler
compiler expanded to 76 candidates
Random selection from 21 candidates

RalfJung · 2026-09-10T20:53:48Z

Thanks!

An alternative may be to normalize the newlines as well. I don't know if that is practical.

aytey · 2026-09-10T21:54:56Z

Thanks!

An alternative may be to normalize the newlines as well. I don't know if that is practical.

I think you'd end-up needing to flatten the whole file, which would the make the golden comparison "a bit rough".

I am not attached to what I've done in this PR at all though, if there's another approach people prefer.

A comment against PR 151817 mentioned that the test file added in this PR kept causing "arbitrary" test file changes. This is because one line of the expanded output sits *just* at 78 columns, which is the pretty-printer's margin. Adding or removing a symbol shifts every index after it, and when one of those gains or loses a digit, the wrapping moves. We cannot use `normalize-stdout` as the layout is decided before normalization runs. We fix this by avoiding a strict golden comparison, and instead using a regex to directly test that the expected syntax contexts exist in the expanded output.

rustbot · 2026-11-02T23:39:32Z

This PR was rebased onto a different main commit. Here's a range-diff highlighting what actually changed.

Rebasing is a normal part of keeping PRs up to date, so no action is needed—this note is just to help reviewers.

fmease

Thanks! Rebased.

@bors rollup
@bors r+ force

View changes since this review

fmease · 2026-11-02T23:44:00Z

+// `'a` share their text but carry different syntax contexts.
+//@ dont-check-compiler-stdout
+//@ check-stdout
+//@ regex-error-pattern: f\s*/\*\s*\d+#0\s*\*/<'a\s*/\*\s*\d+#0\s*\*/,\s*'a\s*/\*\s*\d+#1\s*\*/


Unfortunately, compiletest directives are so inflexible; if there was a way to break up directives into multiple lines I would turn this into a regex with modifier x set / where whitespace is ignored using (?x) and space out the various parts of the pattern because right now it's super illegible.

rust-bors · 2026-11-02T23:44:41Z

📌 Commit 79520af has been approved by fmease

It is now in the queue for this repository.

… r=fmease Avoid oscillating line wrapping due to symbol indices A comment against PR rust-lang#151817 mentioned that the test file added in this PR kept causing "arbitrary" test file changes. This is because one line of the expanded output sits *just* at 78 columns, which is the pretty-printer's margin. Adding or removing a symbol shifts every index after it, and when one of those gains or loses a digit, the wrapping moves. We cannot use `normalize-stdout` as the layout is decided before normalization runs. We fix this by avoiding a strict golden comparison, and instead using a regex to directly test that the expected syntax contexts exist in the expanded output. **LLM disclosure**: I used an LLM to help me understand the problem and how the regex would look, which the LLM then actioned. I reviewed the diff and the commit (along with the message) was performed by hand with me (with the LLM checking the content of the commit message, but the commit message was, indeed, written by a human).

…uwer Rollup of 15 pull requests Successful merges: - #165378 (rustfmt subtree update (v1.12.0)) - #149245 (Make inline attributes apply to the generated `poll` in async fn) - #164304 ([perf] Reuse CanonicalVarKinds if no new variables were added) - #165109 (Avoid building a region map when erasing bound regions) - #163393 (std: Use getexecpath() for current_exe() on OpenBSD) - #164896 (offload: pass the linked device image to the host pass) - #165358 (Reenable some cg_gcc tests that were disabled) - #165371 (Don't use `llvm_type` as the type of `static`s.) - #162232 (Replace `builtin #` with `k#`) - #162608 (Avoid oscillating line wrapping due to symbol indices) - #165029 (Use `MalformedLintAttributeInput` in more places) - #165203 (AIX: Implement TLS destructor guard in std and update dynamic-loading-cdylib test) - #165230 (Fix rigidity assertion ICE) - #165354 (Remove `AbiAlign`) - #165374 (Refactorings in preparation for the move trait)

… r=fmease Avoid oscillating line wrapping due to symbol indices A comment against PR rust-lang#151817 mentioned that the test file added in this PR kept causing "arbitrary" test file changes. This is because one line of the expanded output sits *just* at 78 columns, which is the pretty-printer's margin. Adding or removing a symbol shifts every index after it, and when one of those gains or loses a digit, the wrapping moves. We cannot use `normalize-stdout` as the layout is decided before normalization runs. We fix this by avoiding a strict golden comparison, and instead using a regex to directly test that the expected syntax contexts exist in the expanded output. **LLM disclosure**: I used an LLM to help me understand the problem and how the regex would look, which the LLM then actioned. I reviewed the diff and the commit (along with the message) was performed by hand with me (with the LLM checking the content of the commit message, but the commit message was, indeed, written by a human).

…uwer Rollup of 14 pull requests Successful merges: - #165378 (rustfmt subtree update (v1.12.0)) - #149245 (Make inline attributes apply to the generated `poll` in async fn) - #164304 ([perf] Reuse CanonicalVarKinds if no new variables were added) - #165109 (Avoid building a region map when erasing bound regions) - #163393 (std: Use getexecpath() for current_exe() on OpenBSD) - #164896 (offload: pass the linked device image to the host pass) - #165358 (Reenable some cg_gcc tests that were disabled) - #165371 (Don't use `llvm_type` as the type of `static`s.) - #162232 (Replace `builtin #` with `k#`) - #162608 (Avoid oscillating line wrapping due to symbol indices) - #165029 (Use `MalformedLintAttributeInput` in more places) - #165203 (AIX: Implement TLS destructor guard in std and update dynamic-loading-cdylib test) - #165230 (Fix rigidity assertion ICE) - #165374 (Refactorings in preparation for the move trait)

… r=fmease Avoid oscillating line wrapping due to symbol indices A comment against PR rust-lang#151817 mentioned that the test file added in this PR kept causing "arbitrary" test file changes. This is because one line of the expanded output sits *just* at 78 columns, which is the pretty-printer's margin. Adding or removing a symbol shifts every index after it, and when one of those gains or loses a digit, the wrapping moves. We cannot use `normalize-stdout` as the layout is decided before normalization runs. We fix this by avoiding a strict golden comparison, and instead using a regex to directly test that the expected syntax contexts exist in the expanded output. **LLM disclosure**: I used an LLM to help me understand the problem and how the regex would look, which the LLM then actioned. I reviewed the diff and the commit (along with the message) was performed by hand with me (with the LLM checking the content of the commit message, but the commit message was, indeed, written by a human).

…uwer Rollup of 14 pull requests Successful merges: - #149245 (Make inline attributes apply to the generated `poll` in async fn) - #164304 ([perf] Reuse CanonicalVarKinds if no new variables were added) - #165109 (Avoid building a region map when erasing bound regions) - #163393 (std: Use getexecpath() for current_exe() on OpenBSD) - #164896 (offload: pass the linked device image to the host pass) - #165358 (Reenable some cg_gcc tests that were disabled) - #165371 (Don't use `llvm_type` as the type of `static`s.) - #162232 (Replace `builtin #` with `k#`) - #162608 (Avoid oscillating line wrapping due to symbol indices) - #164241 (make the safety documentation consistent in Ptr module) - #165029 (Use `MalformedLintAttributeInput` in more places) - #165203 (AIX: Implement TLS destructor guard in std and update dynamic-loading-cdylib test) - #165230 (Fix rigidity assertion ICE) - #165374 (Refactorings in preparation for the move trait)

… r=fmease Avoid oscillating line wrapping due to symbol indices A comment against PR rust-lang#151817 mentioned that the test file added in this PR kept causing "arbitrary" test file changes. This is because one line of the expanded output sits *just* at 78 columns, which is the pretty-printer's margin. Adding or removing a symbol shifts every index after it, and when one of those gains or loses a digit, the wrapping moves. We cannot use `normalize-stdout` as the layout is decided before normalization runs. We fix this by avoiding a strict golden comparison, and instead using a regex to directly test that the expected syntax contexts exist in the expanded output. **LLM disclosure**: I used an LLM to help me understand the problem and how the regex would look, which the LLM then actioned. I reviewed the diff and the commit (along with the message) was performed by hand with me (with the LLM checking the content of the commit message, but the commit message was, indeed, written by a human).

…uwer Rollup of 13 pull requests Successful merges: - #149245 (Make inline attributes apply to the generated `poll` in async fn) - #164304 ([perf] Reuse CanonicalVarKinds if no new variables were added) - #165109 (Avoid building a region map when erasing bound regions) - #163393 (std: Use getexecpath() for current_exe() on OpenBSD) - #165358 (Reenable some cg_gcc tests that were disabled) - #165371 (Don't use `llvm_type` as the type of `static`s.) - #162232 (Replace `builtin #` with `k#`) - #162608 (Avoid oscillating line wrapping due to symbol indices) - #164241 (make the safety documentation consistent in Ptr module) - #165029 (Use `MalformedLintAttributeInput` in more places) - #165203 (AIX: Implement TLS destructor guard in std and update dynamic-loading-cdylib test) - #165230 (Fix rigidity assertion ICE) - #165374 (Refactorings in preparation for the move trait)

… r=fmease Avoid oscillating line wrapping due to symbol indices A comment against PR rust-lang#151817 mentioned that the test file added in this PR kept causing "arbitrary" test file changes. This is because one line of the expanded output sits *just* at 78 columns, which is the pretty-printer's margin. Adding or removing a symbol shifts every index after it, and when one of those gains or loses a digit, the wrapping moves. We cannot use `normalize-stdout` as the layout is decided before normalization runs. We fix this by avoiding a strict golden comparison, and instead using a regex to directly test that the expected syntax contexts exist in the expanded output. **LLM disclosure**: I used an LLM to help me understand the problem and how the regex would look, which the LLM then actioned. I reviewed the diff and the commit (along with the message) was performed by hand with me (with the LLM checking the content of the commit message, but the commit message was, indeed, written by a human).

…uwer Rollup of 20 pull requests Successful merges: - #149245 (Make inline attributes apply to the generated `poll` in async fn) - #163495 (Parallelize AST -> HIR lowering) - #164304 ([perf] Reuse CanonicalVarKinds if no new variables were added) - #165109 (Avoid building a region map when erasing bound regions) - #163393 (std: Use getexecpath() for current_exe() on OpenBSD) - #163625 (Add a lint helper attribute namespace ) - #164414 (Add `let_underscore_drop_guard` lint.) - #165342 (Avoid unnecessary MIR traversals in CheckPackedRef lint) - #165351 (delegation: insert probing adjustments before args check adjustments) - #165358 (Reenable some cg_gcc tests that were disabled) - #165377 (Make `no_implicit_prelude` not suppress tool attributes) - #165397 (Add codegen test for unused match arms removed after inlining) - #165398 (Rework parallel reproducibility tests) - #162608 (Avoid oscillating line wrapping due to symbol indices) - #164241 (make the safety documentation consistent in Ptr module) - #165029 (Use `MalformedLintAttributeInput` in more places) - #165203 (AIX: Implement TLS destructor guard in std and update dynamic-loading-cdylib test) - #165230 (Fix rigidity assertion ICE) - #165293 (unstable-book: update default-visibility.md to link to tracking issue) - #165374 (Refactorings in preparation for the move trait)

… r=fmease Avoid oscillating line wrapping due to symbol indices A comment against PR rust-lang#151817 mentioned that the test file added in this PR kept causing "arbitrary" test file changes. This is because one line of the expanded output sits *just* at 78 columns, which is the pretty-printer's margin. Adding or removing a symbol shifts every index after it, and when one of those gains or loses a digit, the wrapping moves. We cannot use `normalize-stdout` as the layout is decided before normalization runs. We fix this by avoiding a strict golden comparison, and instead using a regex to directly test that the expected syntax contexts exist in the expanded output. **LLM disclosure**: I used an LLM to help me understand the problem and how the regex would look, which the LLM then actioned. I reviewed the diff and the commit (along with the message) was performed by hand with me (with the LLM checking the content of the commit message, but the commit message was, indeed, written by a human).

…uwer Rollup of 21 pull requests Successful merges: - #165416 (fix unsoundness when compiling `[diverging_const; 0]`) - #149245 (Make inline attributes apply to the generated `poll` in async fn) - #163495 (Parallelize AST -> HIR lowering) - #164304 ([perf] Reuse CanonicalVarKinds if no new variables were added) - #165109 (Avoid building a region map when erasing bound regions) - #163393 (std: Use getexecpath() for current_exe() on OpenBSD) - #163625 (Add a lint helper attribute namespace ) - #164414 (Add `let_underscore_drop_guard` lint.) - #165342 (Avoid unnecessary MIR traversals in CheckPackedRef lint) - #165351 (delegation: insert probing adjustments before args check adjustments) - #165358 (Reenable some cg_gcc tests that were disabled) - #165377 (Make `no_implicit_prelude` not suppress tool attributes) - #165397 (Add codegen test for unused match arms removed after inlining) - #165398 (Rework parallel reproducibility tests) - #162608 (Avoid oscillating line wrapping due to symbol indices) - #164241 (make the safety documentation consistent in Ptr module) - #165029 (Use `MalformedLintAttributeInput` in more places) - #165203 (AIX: Implement TLS destructor guard in std and update dynamic-loading-cdylib test) - #165230 (Fix rigidity assertion ICE) - #165293 (unstable-book: update default-visibility.md to link to tracking issue) - #165374 (Refactorings in preparation for the move trait)

… r=fmease Avoid oscillating line wrapping due to symbol indices A comment against PR rust-lang#151817 mentioned that the test file added in this PR kept causing "arbitrary" test file changes. This is because one line of the expanded output sits *just* at 78 columns, which is the pretty-printer's margin. Adding or removing a symbol shifts every index after it, and when one of those gains or loses a digit, the wrapping moves. We cannot use `normalize-stdout` as the layout is decided before normalization runs. We fix this by avoiding a strict golden comparison, and instead using a regex to directly test that the expected syntax contexts exist in the expanded output. **LLM disclosure**: I used an LLM to help me understand the problem and how the regex would look, which the LLM then actioned. I reviewed the diff and the commit (along with the message) was performed by hand with me (with the LLM checking the content of the commit message, but the commit message was, indeed, written by a human).

…uwer Rollup of 20 pull requests Successful merges: - #149245 (Make inline attributes apply to the generated `poll` in async fn) - #163495 (Parallelize AST -> HIR lowering) - #164304 ([perf] Reuse CanonicalVarKinds if no new variables were added) - #165109 (Avoid building a region map when erasing bound regions) - #163393 (std: Use getexecpath() for current_exe() on OpenBSD) - #163625 (Add a lint helper attribute namespace ) - #164414 (Add `let_underscore_drop_guard` lint.) - #165342 (Avoid unnecessary MIR traversals in CheckPackedRef lint) - #165351 (delegation: insert probing adjustments before args check adjustments) - #165358 (Reenable some cg_gcc tests that were disabled) - #165377 (Make `no_implicit_prelude` not suppress tool attributes) - #165397 (Add codegen test for unused match arms removed after inlining) - #165398 (Rework parallel reproducibility tests) - #162608 (Avoid oscillating line wrapping due to symbol indices) - #164241 (make the safety documentation consistent in Ptr module) - #165029 (Use `MalformedLintAttributeInput` in more places) - #165203 (AIX: Implement TLS destructor guard in std and update dynamic-loading-cdylib test) - #165230 (Fix rigidity assertion ICE) - #165293 (unstable-book: update default-visibility.md to link to tracking issue) - #165374 (Refactorings in preparation for the move trait)

…uwer Rollup of 22 pull requests Successful merges: - #149245 (Make inline attributes apply to the generated `poll` in async fn) - #164304 ([perf] Reuse CanonicalVarKinds if no new variables were added) - #165109 (Avoid building a region map when erasing bound regions) - #163393 (std: Use getexecpath() for current_exe() on OpenBSD) - #163625 (Add a lint helper attribute namespace ) - #164414 (Add `let_underscore_drop_guard` lint.) - #165342 (Avoid unnecessary MIR traversals in CheckPackedRef lint) - #165351 (delegation: insert probing adjustments before args check adjustments) - #165358 (Reenable some cg_gcc tests that were disabled) - #165377 (Make `no_implicit_prelude` not suppress tool attributes) - #165397 (Add codegen test for unused match arms removed after inlining) - #165398 (Rework parallel reproducibility tests) - #162232 (Replace `builtin #` with `k#`) - #162608 (Avoid oscillating line wrapping due to symbol indices) - #164241 (make the safety documentation consistent in Ptr module) - #164258 (clarify when `Vec::into_boxed_slice()` is in-place) - #165029 (Use `MalformedLintAttributeInput` in more places) - #165203 (AIX: Implement TLS destructor guard in std and update dynamic-loading-cdylib test) - #165230 (Fix rigidity assertion ICE) - #165293 (unstable-book: update default-visibility.md to link to tracking issue) - #165374 (Refactorings in preparation for the move trait) - #165421 (another Enzyme submodule update for nested autodiff)

Rollup merge of #162608 - aytey:unpretty-hygiene-test-churn, r=fmease Avoid oscillating line wrapping due to symbol indices A comment against PR #151817 mentioned that the test file added in this PR kept causing "arbitrary" test file changes. This is because one line of the expanded output sits *just* at 78 columns, which is the pretty-printer's margin. Adding or removing a symbol shifts every index after it, and when one of those gains or loses a digit, the wrapping moves. We cannot use `normalize-stdout` as the layout is decided before normalization runs. We fix this by avoiding a strict golden comparison, and instead using a regex to directly test that the expected syntax contexts exist in the expanded output. **LLM disclosure**: I used an LLM to help me understand the problem and how the regex would look, which the LLM then actioned. I reviewed the diff and the commit (along with the message) was performed by hand with me (with the LLM checking the content of the commit message, but the commit message was, indeed, written by a human).

…uwer Rollup of 22 pull requests Successful merges: - rust-lang/rust#149245 (Make inline attributes apply to the generated `poll` in async fn) - rust-lang/rust#164304 ([perf] Reuse CanonicalVarKinds if no new variables were added) - rust-lang/rust#165109 (Avoid building a region map when erasing bound regions) - rust-lang/rust#163393 (std: Use getexecpath() for current_exe() on OpenBSD) - rust-lang/rust#163625 (Add a lint helper attribute namespace ) - rust-lang/rust#164414 (Add `let_underscore_drop_guard` lint.) - rust-lang/rust#165342 (Avoid unnecessary MIR traversals in CheckPackedRef lint) - rust-lang/rust#165351 (delegation: insert probing adjustments before args check adjustments) - rust-lang/rust#165358 (Reenable some cg_gcc tests that were disabled) - rust-lang/rust#165377 (Make `no_implicit_prelude` not suppress tool attributes) - rust-lang/rust#165397 (Add codegen test for unused match arms removed after inlining) - rust-lang/rust#165398 (Rework parallel reproducibility tests) - rust-lang/rust#162232 (Replace `builtin #` with `k#`) - rust-lang/rust#162608 (Avoid oscillating line wrapping due to symbol indices) - rust-lang/rust#164241 (make the safety documentation consistent in Ptr module) - rust-lang/rust#164258 (clarify when `Vec::into_boxed_slice()` is in-place) - rust-lang/rust#165029 (Use `MalformedLintAttributeInput` in more places) - rust-lang/rust#165203 (AIX: Implement TLS destructor guard in std and update dynamic-loading-cdylib test) - rust-lang/rust#165230 (Fix rigidity assertion ICE) - rust-lang/rust#165293 (unstable-book: update default-visibility.md to link to tracking issue) - rust-lang/rust#165374 (Refactorings in preparation for the move trait) - rust-lang/rust#165421 (another Enzyme submodule update for nested autodiff)

rustbot assigned TaKO8Ki Sep 10, 2026

rustbot added S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. T-compiler Relevant to the compiler team, which will review and decide on the PR/issue. labels Sep 10, 2026

aytey mentioned this pull request Sep 10, 2026

Fix missing syntax context in lifetime hygiene debug output #151817

Merged

This comment has been minimized.

Sign in to view

fmease force-pushed the unpretty-hygiene-test-churn branch from 75bf1aa to 79520af Compare November 2, 2026 23:39

fmease assigned fmease and unassigned TaKO8Ki Nov 2, 2026

fmease approved these changes Nov 2, 2026

View reviewed changes

rust-bors Bot added S-waiting-on-bors Status: Waiting on bors to run and complete tests. Bors will change the label on completion. and removed S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. labels Nov 2, 2026

JonathanBrouwer mentioned this pull request Nov 3, 2026

Rollup of 14 pull requests #165392

Closed

JonathanBrouwer mentioned this pull request Nov 3, 2026

Rollup of 15 pull requests #165393

Closed

JonathanBrouwer mentioned this pull request Nov 3, 2026

Rollup of 14 pull requests #165394

Closed

JonathanBrouwer mentioned this pull request Nov 3, 2026

Rollup of 14 pull requests #165400

Closed

JonathanBrouwer mentioned this pull request Nov 3, 2026

Rollup of 13 pull requests #165403

Closed

JonathanBrouwer mentioned this pull request Nov 3, 2026

Rollup of 20 pull requests #165411

Closed

JonathanBrouwer mentioned this pull request Nov 3, 2026

Rollup of 21 pull requests #165418

Closed

JonathanBrouwer mentioned this pull request Nov 3, 2026

Rollup of 20 pull requests #165426

Closed

JonathanBrouwer mentioned this pull request Nov 3, 2026

Rollup of 22 pull requests #165436

Merged

rust-bors Bot merged commit e51bfcf into rust-lang:main Nov 4, 2026
14 checks passed

rustbot added this to the 1.101.0 milestone Nov 4, 2026

Navigation Menu

Repository navigation

Avoid oscillating line wrapping due to symbol indices - #162608

Avoid oscillating line wrapping due to symbol indices#162608
rust-bors[bot] merged 1 commit into
rust-lang:mainfrom
aytey:unpretty-hygiene-test-churn

aytey commented Sep 10, 2026 •

edited by rustbot

Loading

Uh oh!

rustbot commented Sep 10, 2026

Uh oh!

RalfJung commented Sep 10, 2026

Uh oh!

aytey commented Sep 10, 2026

Uh oh!

This comment has been minimized.

rustbot commented Nov 2, 2026

Uh oh!

fmease left a comment •

edited by rustbot

Loading

Uh oh!

fmease Nov 2, 2026 •

edited

Loading

Uh oh!

rust-bors Bot commented Nov 2, 2026

Uh oh!

Uh oh!

Reviewers

Uh oh!

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

Conversation

aytey commented Sep 10, 2026 • edited by rustbot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

rustbot commented Sep 10, 2026

Uh oh!

RalfJung commented Sep 10, 2026

Uh oh!

aytey commented Sep 10, 2026

Uh oh!

This comment has been minimized.

rustbot commented Nov 2, 2026

Uh oh!

fmease 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

Uh oh!

fmease Nov 2, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

rust-bors Bot commented Nov 2, 2026

Uh oh!

Uh oh!

Reviewers

Uh oh!

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

aytey commented Sep 10, 2026 •

edited by rustbot

Loading

fmease left a comment •

edited by rustbot

Loading

fmease Nov 2, 2026 •

edited

Loading