-
Notifications
You must be signed in to change notification settings - Fork 12.8k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Add regression test for a spurious import #88482
Conversation
Thanks for the pull request, and welcome! The Rust team is excited to review your changes, and you should hear from @jyn514 (or someone else) soon. Please see the contribution instructions for more information. |
This comment has been minimized.
This comment has been minimized.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Hmm, this makes me nervous that it will fail the same way #88269 did. Could you maybe try adding wasm to the targets in config.toml so you can test locally?
I tried to understand what you mean by reading the linked issue but I don't get it. Sorry I am a complete noob when it comes to the rustc code base.
This is the hint I was looking for. I guess I need to change this line?: # Which triples to build libraries (core/alloc/std/test/proc_macro) for. Each of
# these triples will be bootstrapped from the build triple themselves.
#
# Defaults to `host`. If you set this explicitly, you likely want to add all
# host triples to this list as well in order for those host toolchains to be
# able to compile programs for their native target.
#target = ["x86_64-unknown-linux-gnu"] (as an example) Doesn't the CI run wasm tests? 😱 |
Yes, but only on merge, not for PRs. So the three green checks it currently shows don't mean anything. |
Yes, that looks right. |
@the8472 suggested running |
Tried that one. It ran for hours. It ignored the wasm test :(
I also tried running this
with this # Includes one of the default files in src/bootstrap/defaults
profile = "compiler"
changelog-seen = 2
[rust]
lld = true
[build]
target = ["wasm32-unknown-unknown"] It can't build lld:
|
You'll have to install additional dependencies. https://rustc-dev-guide.rust-lang.org/building/prerequisites.html#dependencies |
I can't find the information how to use a OS installed LLVM. I tried this way:
But this still yields the Another thing that puzzles me is that in the |
r? @the8472 - I don't have much experience with cross compiling |
Oh you're on osx. I don't know if the dependencies are different there. Normally the build tools equivalent to what you can find in the docker file will do the job. If llvm build acts up then removing the build directory can help too. Anyway, I checked out the commit, ran Since that was @jyn514's main concern @bors r+ |
📌 Commit 14cbb4b has been approved by |
Add regression test for a spurious import This PR adds a test that verifies that the bug described in the linked issue does not creep back into the code. In essence it checks that compiling some specific code (that uses 128 bit multiplication) with a specific set of compiler options does not lead to a spurious import of a panic function. I noticed that other wasm tests use `# only-wasm32-bare` in their `Makefile`. This will skip the test for me. I did not find out how to run this test locally. Maybe someone can help. closes rust-lang#78744 r? `@jyn514`
@the8472 Thanks for helping me out here. I find it concerning that the docker run did skip the test. That should be independent of macOS. Do you have an explanation for that? |
Ah sorry, that was the wrong job. That CI job tests |
☀️ Test successful - checks-actions |
This PR adds a test that verifies that the bug described in the linked issue does not creep back into the code. In essence it checks that compiling some specific code (that uses 128 bit multiplication) with a specific set of compiler options does not lead to a spurious import of a panic function.
I noticed that other wasm tests use
# only-wasm32-bare
in theirMakefile
. This will skip the test for me. I did not find out how to run this test locally. Maybe someone can help.closes #78744
r? @jyn514