Rollup of 15 pull requests#157673
Closed
GuillaumeGomez wants to merge 40 commits into
Closed
Conversation
The `TypeOutlives` handler in `evaluate_predicate_recursively` means to check whether a type in a `T: 'a` predicate has free regions, bound regions, non-region inference variables, or non-region generic parameters -- and if so, to return `EvaluatedToOkModuloRegions` rather than `EvaluatedToOk`. Correspondingly, the comment says "no free lifetimes or generic parameters". But the code was mistakenly checking `has_non_region_infer()` twice instead of checking both `has_non_region_infer()` and `has_non_region_param()`. This meant that `TypeOutlives(T, 'static)` where `T` is a type parameter returned `EvaluatedToOk` -- claiming the result holds unconditionally -- when the correct answer is `EvaluatedToOkModuloRegions`. The distinction matters during marker trait winnowing in `prefer_lhs_over_victim`, which uses `must_apply_considering_regions()` (true only for `EvaluatedToOk`) to decide whether one impl beats another when there are multiple candidates. With the bug, a `T: 'static`-bounded impl appeared equally as strong as an unrestricted impl, making the winner depend on source order. This caused spurious E0310 errors when the more-constrained impl happened to appear after the less-constrained one. This fixes Rust issue 109481. See PR 153847. This same symptom was originally filed as issue 84917 and fixed in PR 88139. Then PR 102472 rewrote the `TypeOutlives` handler, introducing the duplicate `has_non_region_infer()` and losing the param check, regressing this. Around this same time, issue 109481 was filed. It noted the connection to PR 102472 -- that the bug only appeared after it -- but the duplicate condition was not noticed.
It's harder than one would imagine to demonstrate that the `has_non_region_infer()` check in the `TypeOutlives` handler is load bearing (even though the check seems right analytically). Fortunately, we did find a way to show this. Let's add that test. In working that out, we found two other interesting ways of showing that the `has_non_region_param()` check matters. Let's add those too. Though we're concerned here with code in the old solver, we test against both.
Since approx. Windows SDK 20348, the corresponding cdb (and/or its underlying WinDbg engine) changed or regressed the `OsStr`/`OsString` visualization, and no longer renders the emoji. Since approx. SDK/cdb 26100, the output formatting of the string containing UTF-8 (i.e. the multi-byte emoji grapheme) seems to have further regressed (e.g. the end quotation mark is no longer shown and command output becomes garbled). Relevant issues: * RUST-88840 * RUST-148743 * RUST-88796
The hand-written reader bootstrap.py uses to find the stage0 cargo and rustc only recognized `[table]` headers, so dotted keys like `build.cargo = "..."` were ignored. Parse the dotted prefix on each key as the table it belongs to, resolved relative to the enclosing section, and scope the cargo/rustc lookup to `[build]` so the Python and Rust sides agree.
…oad it from the per-owner info instead
This makes a few changes to rustc_public to make it a little easier to analyze ADT types. * It implements `CrateDef` and `CrateDefType` for `FieldDef`, which allows easy access to the underlying `DefId`, names, and tool annotations. * It adds `Crate::adts` to simplify stepping through all the ADTs in the crate. Note that I did use Gemini to assist with writing this patch, but I wrote most of it, reviewed all the vode, and verified the tests pass locally.
Moreover, flesh out the rustdoc book section about code block attributes.
* Fix async drop glue for Box<T> * Adress reviewer concerns * Address more concerns
Incorrect types can't cause unsafety and making it untyped will make it easier to get rid of the ProcMacro enum in the proc-macro ABI.
Instead encode this information in the crate metadata. This helps shrink the proc-macro ABI to just the parts necessary for transmitting raw bytes to perform RPC on top of.
…keys, r=clubby789 bootstrap: Handle dotted table keys when parsing bootstrap.toml Fixes rust-lang#156948 `bootstrap.py` has a small hand-rolled reader that pulls the stage0 cargo and rustc out of `bootstrap.toml` before the real TOML parser is built. It only understood `[table]` headers, so the dotted form that `bootstrap.example.toml` now uses (`build.cargo = "..."`) was silently dropped, and the cargo/rustc lookup matched those keys in any table rather than only `[build]`. Teach the reader to strip an optional dotted prefix and treat it as the key's table, the same way `configure.py` does when it writes the file, and scope the cargo/rustc lookup to `[build]`. New tests cover the dotted form, equivalence with the section form, wrong-table rejection, and dotted target keys. *Disclosure: the code was written by AI (Claude); reviewed and tested by me.*
…db, r=Kobzol Disable `tests/debuginfo/pretty-std.rs` `OsString` cdb check ## Summary Since approx. Windows SDK 20348, the corresponding cdb (and/or its underlying WinDbg engine) changed or regressed the `OsStr`/`OsString` visualization, and no longer renders the emoji. Since approx. SDK/cdb 26100, the output formatting of the string containing UTF-8 (i.e. the multi-byte emoji grapheme) seems to have further regressed (e.g. the end quotation mark is no longer shown and command output becomes garbled). Relevant issues: * rust-lang#88840 * rust-lang#148743 * rust-lang#88796 This was failing for me locally on host Windows MSVC under `./x test tests/debuginfo`. And the issues re. this case failing has been open for a long while and have not been addressed, so I propose disabling this case. I believe this is either a regression or a change in cdb and/or the underlying WinDbg. I have not bothered filing such a cdb bug report, I don't know enough about this. The check was actually relaxed to not require the emoji to be rendered properly in the output in rust-lang#88842, but I feel like that regressed the original intention of the check (that multi-byte Unicode graphemes are properly rendered), so I just restored the original checks but commented them out. Maybe r? @wesleywiser (or compiler)
…trochenkov Remove ProcMacro enum from proc macro ABI Instead encode this information in the crate metadata. This helps shrink the proc-macro ABI to just the parts necessary for transmitting raw bytes to perform RPC on top of. r? @petrochenkov cc @cyrgani
…test-harness, r=GuillaumeGomez rustdoc: Test & document `test_harness` code block attribute They were fully untested and undocumented previously despite being stable. Context: [#t-rustdoc > &rust-lang#96;test_harness&rust-lang#96; langstr @ 💬](https://rust-lang.zulipchat.com/#narrow/channel/266220-t-rustdoc/topic/.60test_harness.60.20langstr/near/546748038). While at it, I've also improved the documentation for code block attributes in general. I was only inspired to do so because there was no good place for `test_harness` and because it got hard to skim due to a lack of subsubsections.
…ssing-param-check, r=lcnr Fix marker trait winnowing depending on impl order Story: I was looking through this code for an unrelated reason and happened to notice the duplicate condition. That seemed obviously wrong, especially in light of the comment, so I worked backward to what it affected, wrote a test that failed, and then found that the test matched an existing `known-bug` test, which pointed me to rust-lang#109481. --- The `TypeOutlives` handler in `evaluate_predicate_recursively` means to check whether a type in a `T: 'a` predicate has free regions, bound regions, non-region inference variables, or non-region generic parameters -- and if so, to return `EvaluatedToOkModuloRegions` rather than `EvaluatedToOk`. Correspondingly, the comment says "no free lifetimes or generic parameters". But the code was mistakenly checking `has_non_region_infer()` twice instead of checking both `has_non_region_infer()` and `has_non_region_param()`. This meant that `TypeOutlives(T, 'static)` where `T` is a type parameter returned `EvaluatedToOk` -- claiming the result holds unconditionally -- when the correct answer is `EvaluatedToOkModuloRegions`. The distinction matters during marker trait winnowing in `prefer_lhs_over_victim`, which uses `must_apply_considering_regions()` (true only for `EvaluatedToOk`) to decide whether one impl beats another when there are multiple candidates. With the bug, a `T: 'static`-bounded impl appeared equally as strong as an unrestricted impl, making the winner depend on source order. This caused spurious E0310 errors when the more-constrained impl happened to appear after the less-constrained one. Fixes rust-lang#109481. This same symptom was originally filed as rust-lang#84917 and fixed in PR rust-lang#88139. Then PR rust-lang#102472 rewrote the `TypeOutlives` handler, introducing the duplicate `has_non_region_infer()` and losing the param check, regressing this. Around this same time, rust-lang#109481 was filed. It noted the connection to rust-lang#102472 -- the bug only appeared after it -- but the duplicate condition was not noticed. r? @lcnr cc @oli-obk @nikomatsakis
…nder-fix, r=GuillaumeGomez Reorder `impl` restriction rendering and add bottom margin This PR moves the `impl` restriction rendering above the trait description. It also adds an ⓘ marker, italicizes the text, and adds a bottom margin. Tracking issue: rust-lang#105077 before: <img width="1108" height="651" alt="image" src="https://github.com/user-attachments/assets/520d95cf-a648-44e7-8036-e453cfbdf569" /> after: <img width="1085" height="671" alt="image" src="https://github.com/user-attachments/assets/84493e56-6227-433b-93d0-4fcbe28ac00c" /> r? @GuillaumeGomez cc @Urgau @jhpratt
…ering, r=fmease Report duplicate relaxed bounds during ast lowering Instead of during hir-ty-lowering. Not sure it is worth erroring on at all, but it's better to do it directly during lowering instead of collecting bounds after the fact.
fix doc for unicode normalization faq on `casefold` APIs APIs added in rust-lang#154742 have an outbound doc link to https://www.unicode.org/faq/normalization, but this URL appears to not exist. The correct URL appears to be https://www.unicode.org/faq/normalization.html cc @Jules-Bertholet
…=jieyouxu Update to ar_archive_writer v0.5.2 This includes rust-lang/ar_archive_writer#32 and rust-lang/ar_archive_writer#33. The former fixes a panic when there are more than 64k archive members.
…haustively, r=lqd Add test for matches in `rustc_must_match_exhaustively` r? @lcnr
rustbot
added
A-attributes
Member
Author
|
@bors r+ rollup=never p=5 |
Contributor
rust-bors
Bot
added
S-waiting-on-bors
Contributor
|
Trying commonly failed jobs |
Contributor
|
⌛ Trying commit cc98143 with merge 0060cc7… To cancel the try build, run the command Workflow: https://github.com/rust-lang/rust/actions/runs/27221255971 |
rust-bors Bot
pushed a commit
that referenced
this pull request
Jun 9, 2026
Rollup of 15 pull requests try-job: dist-various-1 try-job: test-various try-job: x86_64-gnu-aux try-job: x86_64-gnu-llvm-21-3 try-job: x86_64-msvc-1 try-job: aarch64-apple try-job: x86_64-mingw-1 try-job: i686-msvc-2
Member
|
added a couple more PRs to the rollup to hopefully clear the queue sooner |
rust-bors
Bot
added
S-waiting-on-author
Contributor
|
This pull request was unapproved due to being closed. |
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
18 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.
Successful merges:
tests/debuginfo/pretty-std.rsOsStringcdb check #157503 (Disabletests/debuginfo/pretty-std.rsOsStringcdb check)test_harnesscode block attribute #148183 (rustdoc: Test & documenttest_harnesscode block attribute)implrestriction rendering and add bottom margin #157633 (Reorderimplrestriction rendering and add bottom margin)casefoldAPIs #157652 (fix doc for unicode normalization faq oncasefoldAPIs)rustc_must_match_exhaustively#157668 (Add test for matches inrustc_must_match_exhaustively)r? @ghost
Create a similar rollup