Run ui tests on the parallel frontend by ywxt · Pull Request #153801 · rust-lang/rust

ywxt · 2026-03-13T02:24:03Z

This PR adds two arguments for tests:

--parallel-frontend-threads: specify -Zthread to compile test case (currently UI tests only)
--iteration-count: the number of times to run each test

Also, due to the non-deterministic diagnostic orders and cycle errors, this PR adds the directive //@ ignore-parallel-frontend to ignore tests with cycle error when the parallel-frontend is enabled (by --parallel-frontend-threads) and enables //@ compare-output-by-lines by default.

Context: #t-compiler/parallel-rustc > Add the parallel front-end test suite @ 💬

This PR should work with #153797 together.

ywxt · 2026-03-13T08:11:00Z

cc @jieyouxu @Kobzol @Zoxc @zetanumbers @Ozzy1423 @teor2345

petrochenkov · 2026-03-13T11:16:50Z

This may be a bit silly, but to avoid many unnecessary changes in stderr files you could put the //@ ignore-parallel-frontend into the first empty lines in tests, like I did with //@ check-pass in #151130 (the directives shouldn't necessarily be at the top).
Hopefully the ignores will be removed away sooner or later, so this way we'll avoid thrashing the files on both directive addition and removal. (Now majority of the diff is meaningless stderr file changes.)

petrochenkov · 2026-03-13T11:19:12Z

What exactly failures are ignored by ignore-parallel-frontend in practice?
Are they .stderr file comparison errors, or there are //~ ERROR annotation mismatches too?

Right now ignore-parallel-frontend ignores the test entirely, but maybe we could ignore only some of the checks.

ywxt · 2026-03-13T11:35:02Z

What exactly failures are ignored by ignore-parallel-frontend in practice? Are they .stderr file comparison errors, or there are //~ ERROR annotation mismatches too?

Right now ignore-parallel-frontend ignores the test entirely, but maybe we could ignore only some of the checks.

failures would produce different errors so that //~ ERROR annotation mismatches

src/doc/rustc-dev-guide/src/tests/compiletest.md

tests/ui/async-await/mutually-recursive-async-impl-trait-type.rs

src/tools/compiletest/src/common.rs

ywxt · 2026-03-13T13:10:15Z

diff of stderr:

105     LL | pub static R1: &[()] = unsafe { from_ptr_range(ptr::null()..ptr::null()) }; // errors inside libcore
106        |                                 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ evaluation of `R1` failed here
107
-       error[E0080]: in-bounds pointer arithmetic failed: attempting to offset pointer by 8 bytes, but got ALLOC10 which is only 4 bytes from the end of the allocation
109       --> $DIR/forbidden_slices.rs:54:25
110        |
111     LL |     from_ptr_range(ptr..ptr.add(2)) // errors inside libcore

157                    HEX_DUMP
158                }
159
-       error[E0080]: in-bounds pointer arithmetic failed: attempting to offset pointer by 8 bytes, but got ALLOC11+0x1 which is only 7 bytes from the end of the allocation
161       --> $DIR/forbidden_slices.rs:79:25
162        |
163     LL |     from_ptr_range(ptr..ptr.add(1))

actual:

error[E0080]: in-bounds pointer arithmetic failed: attempting to offset pointer by 8 bytes, but got alloc66+0x1 which is only 7 bytes from the end of the allocation
  --> /home/ywxt/Projects/Rust/rust/tests/ui/const-ptr/forbidden_slices.rs:79:25
   |
LL |     from_ptr_range(ptr..ptr.add(1))
   |                         ^^^^^^^^^^ evaluation of `R8` failed here

error[E0080]: in-bounds pointer arithmetic failed: attempting to offset pointer by 8 bytes, but got alloc69 which is only 4 bytes from the end of the allocation
  --> /home/ywxt/Projects/Rust/rust/tests/ui/const-ptr/forbidden_slices.rs:54:25
   |
LL |     from_ptr_range(ptr..ptr.add(2)) // errors inside libcore
   |                         ^^^^^^^^^^ evaluation of `R2` failed here

The difference is from alloc ids. Should we mask all of them?

petrochenkov · 2026-03-13T13:38:04Z

It may be possible to normalize them instead of ignoring.
I'd ping the author of the test and ask their opinion.

ywxt · 2026-03-13T16:46:46Z

tests/ui/const-generics/issues/issue-100313.rs

@@ -1,5 +1,5 @@
 //@ dont-require-annotations: NOTE
-
+//@ ignore-parallel-frontend different alloc ids


I will add normalization rules if it's confirmed.

ywxt · 2026-03-13T16:48:30Z

tests/ui/const-generics/min_const_generics/invalid-patterns.rs

 //@ stderr-per-bitwidth
 //@ dont-require-annotations: NOTE
-
+//@ ignore-parallel-frontend different alloc ids


Although some consts tests didn't get failed in my environment, I marked them as ignored due to alloc ids in stderr.

ywxt · 2026-03-13T16:50:38Z

tests/ui/lint/rfc-2383-lint-reason/force_warn_expected_lints_fulfilled.rs

 //@ compile-flags: --force-warn unused_mut
 //@ check-pass
-
+//@ ignore-parallel-frontend the message `requested on the ...` appears in different lines


The message is reported on two different source lines. Maybe we can solve it by comparison approaches?

ywxt · 2026-03-13T16:57:56Z

tests/ui/dep-graph/dep-graph-caller-callee.rs

 //@ compile-flags: -Z query-dep-graph
-
+//@ ignore-parallel-frontend dep graph
 #![feature(rustc_attrs)]


The dep graph tests failed on some rustc_then_this_would_need attributes.

Do you know the detail? @zetanumbers @Zoxc

I was able to reproduce a failure on main. It seems to be non-deterministic. Perhaps #152621 helps. I haven't looked into the issues with -Z query-dep-graph.

Kobzol · 2026-03-13T21:00:49Z

src/doc/rustc-dev-guide/src/tests/compiletest.md

+
+Also, when running with `--parallel-frontend-threads`, the `compare-output-by-lines` directive would be implied for all tests, since the output from the parallel frontend can be non-deterministic in terms of the order of lines.
+
+The parallel frontend is avaliable in UI tests only at the moment, and is not currently supported in other test suites.


Suggested change

The parallel frontend is avaliable in UI tests only at the moment, and is not currently supported in other test suites.

The parallel frontend is available in UI tests only at the moment, and is not currently supported in other test suites.

Kobzol · 2026-03-13T21:01:14Z

src/doc/rustc-dev-guide/src/tests/compiletest.md

+
+The parallel frontend is avaliable in UI tests only at the moment, and is not currently supported in other test suites.
+
+If you run with `--parallel-frontend-threads` and `--bless`, then `--parallel-frontend-threads` will be valied.


What does this sentence mean? :) I didn't understand it due to the last word probably being typoed.

Kobzol · 2026-03-13T21:03:32Z

src/tools/compiletest/src/rustdoc_gui_test.rs

        bypass_ignore_backends: Default::default(),
        jobs: Default::default(),
+        parallel_frontend_threads: Default::default(),
+        iteration_count: Default::default(),


This should be 1, I guess?

This comment has been minimized.

Sign in to view

petrochenkov self-assigned this Mar 13, 2026

ywxt force-pushed the parallel-test branch from e95bfb4 to 6d401a3 Compare March 13, 2026 06:43

petrochenkov added S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. and removed S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. labels Mar 13, 2026

petrochenkov reviewed Mar 13, 2026

View reviewed changes

src/doc/rustc-dev-guide/src/tests/compiletest.md Outdated Show resolved Hide resolved

tests/ui/async-await/mutually-recursive-async-impl-trait-type.rs Outdated Show resolved Hide resolved

src/tools/compiletest/src/common.rs Outdated Show resolved Hide resolved

petrochenkov added S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. and removed S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. labels Mar 13, 2026

ywxt force-pushed the parallel-test branch from 6d401a3 to 67ebcd0 Compare March 13, 2026 16:41

ywxt added 2 commits March 14, 2026 00:43

Run ui tests on the parallel frontend

328efb5

Ignore tests for the parallel frontend

414d715

ywxt force-pushed the parallel-test branch from 67ebcd0 to 414d715 Compare March 13, 2026 16:44

ywxt commented Mar 13, 2026

View reviewed changes

Kobzol reviewed Mar 13, 2026

View reviewed changes

		@@ -1,5 +1,5 @@
		//@ dont-require-annotations: NOTE

		//@ ignore-parallel-frontend different alloc ids


		Also, when running with `--parallel-frontend-threads`, the `compare-output-by-lines` directive would be implied for all tests, since the output from the parallel frontend can be non-deterministic in terms of the order of lines.

		The parallel frontend is avaliable in UI tests only at the moment, and is not currently supported in other test suites.

	The parallel frontend is avaliable in UI tests only at the moment, and is not currently supported in other test suites.
	The parallel frontend is available in UI tests only at the moment, and is not currently supported in other test suites.


		The parallel frontend is avaliable in UI tests only at the moment, and is not currently supported in other test suites.

		If you run with `--parallel-frontend-threads` and `--bless`, then `--parallel-frontend-threads` will be valied.

Uh oh!

Conversation

ywxt commented Mar 13, 2026

Uh oh!

This comment has been minimized.

ywxt commented Mar 13, 2026

Uh oh!

petrochenkov commented Mar 13, 2026

Uh oh!

petrochenkov commented Mar 13, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

ywxt commented Mar 13, 2026

Uh oh!

Uh oh!

Uh oh!

Uh oh!

ywxt commented Mar 13, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

petrochenkov commented Mar 13, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

6 participants

petrochenkov commented Mar 13, 2026 •

edited

Loading

ywxt commented Mar 13, 2026 •

edited

Loading

petrochenkov commented Mar 13, 2026 •

edited

Loading