Add CL and CMD into to pdb debug info

[nebulark]

Partial fix for #96475

The Arg0 and CommandLineArgs of the MCTargetOptions cpp class are not set within

rust/compiler/rustc_llvm/llvm-wrapper/PassWrapper.cpp

Line 378 in bb548f9

extern "C" LLVMTargetMachineRef LLVMRustCreateTargetMachine(

This causes LLVM to not neither output any compiler path (cl) nor the arguments that were used when invoking it (cmd) in the PDB file.

This fix adds the missing information to the target machine so LLVM can use it.

[rustbot]

Thanks for the pull request, and welcome! The Rust team is excited to review your changes, and you should hear from @b-naber (or someone else) soon.

Please see the contribution instructions for more information. Namely, in order to ensure the minimum review times lag, PR authors and assigned reviewers should ensure that the review label (S-waiting-on-review and S-waiting-on-author) stays updated, invoking these commands when appropriate:

• @rustbot author: the review is finished, PR author should check the comments and take action accordingly
• @rustbot review: the author is ready for a review, this PR will be queued again in the reviewer's queue

[nebulark]

I really have trouble understanding what caused the check to fail here. The thing I most likely see that could have caused it is adding extra members to rustc_session::Session and rustc_interface::interface::Config, to be able to pass the expanded through some functions.
But I really don't understand why this occurs.

Note that I am developing on Windows.

[jyn514]

this is an interesting error. i'm going to go through it step-by-step to see if i can help you figure out how i found the root cause :)

the first thing i did was look at the rust-log-analyzer output and decide it was too long, so i looked at the original github actions workflow in case RLA got confused. that shows the following error at the end (along with ~16 or so others):

error[E0277]: the trait bound `syn::ItemImpl: quote::ToTokens` is not satisfied
Error:   --> tests/ui/auxiliary/proc_macro_attr.rs:96:23
   |
LL |     TokenStream::from(quote!(#item))
   |                       ^^^^^^^^^^^^^
   |                       |
   |                       the trait `quote::ToTokens` is not implemented for `syn::ItemImpl`
   |                       required by a bound introduced by this call
   |
   = help: the following other types implement trait `quote::ToTokens`:
             &'a T
             &'a mut T
             bool
             char
             f32
             f64
             i128
             i16
           and 23 others
   = note: this error originates in the macro `quote` (in Nightly builds, run with -Z macro-backtrace for more info)

error: aborting due to 6 previous errors

Some errors have detailed explanations: E0277, E0308.
For more information about an error, try `rustc --explain E0277`.


FAILURES:
    tests/ui/empty_line_after_doc_comments.rs
    tests/ui/empty_line_after_outer_attribute.rs
    tests/ui/needless_arbitrary_self_type_unfixable.rs

test result: FAIL. 3 tests failed, 868 tests passed, 4 ignored, 0 filtered out
thread 'main' panicked at 'called `Result::unwrap()` on an `Err` value: tests failed

Location:
    /cargo/registry/src/index.crates.io-6f17d22bba15001f/ui_test-0.11.6/src/lib.rs:308:13', src/tools/clippy/tests/compile-test.rs:99:6
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace
error: test failed, to rerun pass `--test compile-test`
Build completed unsuccessfully in 0:02:40

there are some important clues here. first, the panic happens in src/tools/clippy/tests/compile-test.rs:99:6. that tells me that this is

• a compiletest testsuite
• for clippy
• and that rustc and clippy compiled successfully, and this only broke when running clippy on a test. so your hypothesis about new struct fields breaking the code seems unlikely, since we compiled clippy without errors.

now i know how to replicate the failure locally: x test src/tools/clippy. that takes a while to run, but unfortunately does not show me the error from CI, it succeeds instead.

there are more clues here.

• the error happens in tests/ui/auxiliary/proc_macro_attr.rs:39:26. auxiliary is what the compiletest test suite uses instead of cargo; proc_macro_attr is a dependency that will be used in the test. auxiliary crates are always supposed to build, so the problem here is that we're getting the "trait bound for syn::ItemImpl: quote::ToTokens not satisfied" error where we weren't before.
• all three tests mention proc_macro_attr so we know there is only one thing that is breaking, all the failures have the same cause.

at this point i look at the diff in your PR and notice you didn't modify clippy at all, so i think it's likely this is due to breakage in either syn or quote. i run git log --stat on your PR which shows that you didn't update Cargo.lock either. i now suspect that clippy's fork of compile-test does a bad job of pinning dependency versions. looking at the --extern args it passes, i see "--extern" "quote=/checkout/obj/build/x86_64-unknown-linux-gnu/stage2-tools/ui_test/x86_64-unknown-linux-gnu/debug/deps/libquote-36054bcec486f3c0.rmeta" which doesn't tell me which version it used.

https://crates.io/crates/syn shows that the latest release was 2.0.24, 8 hours ago, just after CI ran. https://crates.io/crates/quote was last released 10 days ago. the timing for syn is suspicious, which makes me think there was some bug fix published just after the workflow ran.

i think this is upstream breakage in syn that was already fixed last night after the workflow ran, so i retriggered CI in hopes that will used the updated version on crates.io.

[jyn514]

yeah @oli-obk i don't think this is stable because it only uses Cargo.toml without allowing clippy to specify a lockfile

rust/src/tools/clippy/tests/compile-test.rs

Line 32 in de69d55

dependencies_crate_manifest_path: Some("clippy_test_deps/Cargo.toml".into()),

[jyn514]

hmm. ui_test is using --manifest-path though, which i expect would discover that it's in a workspace and use the top-level Cargo.lock from rust-lang/rust? it doesn't really have enough logging for me to tell ....
https://github.com/oli-obk/ui_test/blob/5f391d3f734c16f26e1220737a2a923b5190fdb5/src/dependencies.rs#L51-L53

[nebulark]

@rustbot author
going to fix the memory leak issue, probably does not make sense to review before then

[nebulark]

Now storing the cpp args in an object instead of global. I put it into ModuleLlvm so it should outlive the targetmachine that has reference to it, and allows me to dispose of it to avoid memory leaks.
Also added a windows only test using findstr.

[bors]

⌛ Testing commit 7509c34 with merge cb956d913092ea9e2c7f25f11969512d77563758...

[bors]

💔 Test failed - checks-actions

[nebulark]

Changed unit test to use to run only on windows-msvc instead for all windows variants.

@rustbot review

[petrochenkov]

@bors r+

[bors]

📌 Commit 4cdc633 has been approved by petrochenkov

It is now in the queue for this repository.

[bors]

⌛ Testing commit 4cdc633 with merge 9be4eac...

[bors]

☀️ Test successful - checks-actions
Approved by: petrochenkov
Pushing 9be4eac to master...

[rust-timer]

Finished benchmarking commit (9be4eac): comparison URL.

Overall result: ✅ improvements - no action needed

@rustbot label: -perf-regression

Instruction count

This is a highly reliable metric that was used to determine the overall result at the top of this comment.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-0.6%	[-0.6%, -0.6%]	1
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-0.6%	[-0.6%, -0.6%]	1

Max RSS (memory usage)

Results

This is a less reliable metric that may be of interest but was not used to determine the overall result at the top of this comment.

	mean	range	count
Regressions ❌ (primary)	1.3%	[1.0%, 1.5%]	2
Regressions ❌ (secondary)	2.3%	[1.4%, 3.0%]	4
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-2.4%	[-3.9%, -1.0%]	3
All ❌✅ (primary)	1.3%	[1.0%, 1.5%]	2

Cycles

Results

This is a less reliable metric that may be of interest but was not used to determine the overall result at the top of this comment.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	2.7%	[2.6%, 2.8%]	2
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-	-	0

Binary size

This benchmark run did not return any relevant results for this metric.

Bootstrap: 629.137s -> 629.695s (0.09%)
Artifact size: 318.13 MiB -> 318.12 MiB (-0.00%)