Skip to content
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

Migrate issue-83112-incr-test-moved-file, type-mismatch-same-crate-name and issue-109934-lto-debuginfo run-make tests to rmake or ui #127538

Merged
merged 3 commits into from
Jul 11, 2024
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
3 changes: 0 additions & 3 deletions src/tools/tidy/src/allowed_run_make_makefiles.txt
Original file line number Diff line number Diff line change
Expand Up @@ -48,7 +48,6 @@ run-make/incr-add-rust-src-component/Makefile
run-make/incr-foreign-head-span/Makefile
run-make/interdependent-c-libraries/Makefile
run-make/issue-107094/Makefile
run-make/issue-109934-lto-debuginfo/Makefile
run-make/issue-14698/Makefile
run-make/issue-15460/Makefile
run-make/issue-18943/Makefile
Expand All @@ -62,7 +61,6 @@ run-make/issue-36710/Makefile
run-make/issue-47551/Makefile
run-make/issue-69368/Makefile
run-make/issue-83045/Makefile
run-make/issue-83112-incr-test-moved-file/Makefile
run-make/issue-84395-lto-embed-bitcode/Makefile
run-make/issue-85019-moved-src-dir/Makefile
run-make/issue-85401-static-mir/Makefile
Expand Down Expand Up @@ -143,6 +141,5 @@ run-make/test-benches/Makefile
run-make/thumb-none-cortex-m/Makefile
run-make/thumb-none-qemu/Makefile
run-make/translation/Makefile
run-make/type-mismatch-same-crate-name/Makefile
run-make/unstable-flag-required/Makefile
run-make/x86_64-fortanix-unknown-sgx-lvi/Makefile
27 changes: 27 additions & 0 deletions tests/run-make/incr-test-moved-file/rmake.rs
Original file line number Diff line number Diff line change
@@ -0,0 +1,27 @@
// The generated test harness code contains spans with a dummy location,
// but a non-dummy SyntaxContext. Previously, the incremental cache was encoding
// these spans as a full span (with a source file index), instead of skipping
// the encoding of the location information. If the file gest moved, the hash
// of the span will be unchanged (since it has a dummy location), so the incr
// cache would end up try to load a non-existent file using the previously
// enccoded source file id.
// This test reproduces the steps that used to trigger this bug, and checks
// for successful compilation.
// See https://github.com/rust-lang/rust/issues/83112

//@ ignore-none
// Reason: no-std is not supported
//@ ignore-nvptx64-nvidia-cuda
// FIXME: can't find crate for 'std'

use run_make_support::{fs_wrapper, rust_lib_name, rustc};

fn main() {
fs_wrapper::create_dir("incr");
fs_wrapper::create_dir("src");
fs_wrapper::create_dir("src/mydir");
fs_wrapper::copy("main.rs", "src/main.rs");
rustc().input("src/main.rs").incremental("incr").arg("--test").run();
fs_wrapper::rename("src/main.rs", "src/mydir/main.rs");
rustc().input("src/mydir/main.rs").incremental("incr").arg("--test").run();
}
12 changes: 0 additions & 12 deletions tests/run-make/issue-109934-lto-debuginfo/Makefile

This file was deleted.

9 changes: 0 additions & 9 deletions tests/run-make/issue-109934-lto-debuginfo/lib.rs

This file was deleted.

25 changes: 0 additions & 25 deletions tests/run-make/issue-83112-incr-test-moved-file/Makefile

This file was deleted.

19 changes: 0 additions & 19 deletions tests/run-make/type-mismatch-same-crate-name/Makefile

This file was deleted.

29 changes: 29 additions & 0 deletions tests/run-make/type-mismatch-same-crate-name/rmake.rs
Original file line number Diff line number Diff line change
@@ -0,0 +1,29 @@
// When a compilation failure deals with seemingly identical types, some helpful
// errors should be printed.
// The main use case of this error is when there are two crates
// (generally different versions of the same crate) with the same name
// causing a type mismatch. In this test, one of the crates
// is only introduced as an indirect dependency and the type is accessed via a reexport.
// See https://github.com/rust-lang/rust/pull/42826

use run_make_support::{rust_lib_name, rustc};

fn main() {
rustc().crate_type("rlib").input("crateA.rs").metadata("-1").extra_filename("-1").run();
rustc().crate_type("rlib").input("crateA.rs").metadata("-2").extra_filename("-2").run();
rustc()
.crate_type("rlib")
.input("crateB.rs")
.extern_("crateA", rust_lib_name("crateA-1"))
.run();
rustc()
.input("crateC.rs")
.extern_("crateA", rust_lib_name("crateA-2"))
.run_fail()
.assert_stderr_contains("mismatched types")
.assert_stderr_contains("crateB::try_foo(foo2);")
.assert_stderr_contains("different versions of crate `crateA`")
.assert_stderr_contains("crateB::try_bar(bar2);")
.assert_stderr_contains("expected trait `crateA::bar::Bar`, found trait `Bar`")
.assert_stderr_contains("different versions of crate `crateA`");
}
22 changes: 22 additions & 0 deletions tests/ui/lto/debuginfo-lto-alloc.rs
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I have noticed that the new UI test debuginfo-lto-alloc is outputting artifacts that aren't getting cleaned up because of its -C incremental. That might be the justification needed to keep it as a run-make test?

If that's the case, let's keep it as a rmake.rs test for now, can you please leave a remark about your finding in case someone tries to port it to ui tests? (We could try to fix ui tests not cleaning this up in another PR, but I'm also fine with leaving it as a rmake.rs test if it does something that's "more special" than ui tests, i.e. I don't want to shoehorn it into ui tests if rmake.rs is less friction).

Original file line number Diff line number Diff line change
@@ -0,0 +1,22 @@
// With the upgrade to LLVM 16, the following error appeared when using
// link-time-optimization (LTO) alloc and debug compilation mode simultaneously:
//
// error: Cannot represent a difference across sections
//
// The error stemmed from DI function definitions under type scopes, fixed by
// only declaring in type scope and defining the subprogram elsewhere.
// This test reproduces the circumstances that caused the error to appear, and checks
// that compilation is successful.

//@ check-pass
//@ compile-flags: --test -C debuginfo=2 -C lto=fat -C incremental=inc-fat

extern crate alloc;

#[cfg(test)]
mod tests {
#[test]
fn something_alloc() {
assert_eq!(Vec::<u32>::new(), Vec::<u32>::new());
}
}
Loading