about summary refs log tree commit diff
path: root/compiler/rustc_llvm/llvm-wrapper/RustWrapper.cpp
diff options
context:
space:
mode:
authorJacob Pratt <jacob@jhpratt.dev>2025-01-31 00:26:33 -0500
committerGitHub <noreply@github.com>2025-01-31 00:26:33 -0500
commit4e64180917bbb5a2a1d7eecb23072b0b420b67af (patch)
tree9cc75f7925506e111b85af3532d9b43f3743a894 /compiler/rustc_llvm/llvm-wrapper/RustWrapper.cpp
parente2a73ab7ade83ad68a69d1f5e072e2cea6833a33 (diff)
parent4fcae667d6ff477a274eaecc6437d8dc828b9265 (diff)
downloadrust-4e64180917bbb5a2a1d7eecb23072b0b420b67af.tar.gz
rust-4e64180917bbb5a2a1d7eecb23072b0b420b67af.zip
Rollup merge of #136310 - lqd:long-types, r=compiler-errors
normalize `*.long-type.txt` paths for compare-mode tests

When using a compare mode, the name of the test + compare-mode is embedded in some of rustc's output, like the location where a long type `bla.long-type(-some-hash)?.txt` is written to. That generally makes these tests fail under all compare-modes.

This PR fixes this by normalizing the compare-mode suffix away in the stderr output. We can also see some remnants of the long-removed `nll` compare mode being normalized away ^^.

I did this to fix some failures with `--compare-mode next-solver` (but it also fixes them with e.g. `--compare-mode polonius` of course):
- it makes 9 new tests pass with the new solver
- however, 3 tests I changed here still don't pass with the new solver (IIRC there were 2 ICEs, and some duplicate errors for the 3rd one)

(There was also one that triggered slowness in the new solver while triggering the long type failure, I'll mention this on zulip. )
Diffstat (limited to 'compiler/rustc_llvm/llvm-wrapper/RustWrapper.cpp')
0 files changed, 0 insertions, 0 deletions