Don't error when adding a staticlib with bitcode files compiled by newer LLVM

[bjorn3]

cc #128955 (comment)

[rustbot]

r? @oli-obk

rustbot has assigned @oli-obk.
They will have a look at your PR within the next two weeks and either review your PR or reassign to another reviewer.

Use r? to explicitly pick a reviewer

[bjorn3]

For testing this how I can create a minimal LLVM bitcode file which is guaranteed to be interpreted as having an invalid LLVM version? Something obviously bogus like BC\xC0\xDEaaaa doesn't make rustc emit an error and nm only says "bfd plugin: LLVM gold plugin has failed to create LTO module: Bitcode file does not contain any modules", which seems to suggest that the bitcode file is technically valid, just missing any contents.

[bjorn3]

The test uncovered two memory corruption bugs due to accidentally trying to use both an io::Error and a Cow<str> as String on the other side of the FFI. These are now fixed.

[pcc]

For testing this how I can create a minimal LLVM bitcode file which is guaranteed to be interpreted as having an invalid LLVM version? Something obviously bogus like BC\xC0\xDEaaaa doesn't make rustc emit an error and nm only says "bfd plugin: LLVM gold plugin has failed to create LTO module: Bitcode file does not contain any modules", which seems to suggest that the bitcode file is technically valid, just missing any contents.

There are some invalid bitcode files in LLVM's test suite, e.g. https://github.com/llvm/llvm-project/blob/24830ea7108197c7880939aa2e32db3aa4bc6284/llvm/test/Bindings/llvm-c/Inputs/invalid.ll.bc

[bjorn3]

Looks like telling llvm-ar to not attempt to produce a symbol table is enough to make the test pass on CI.

@rustbot ready

[oli-obk]

I had an inline question above

@rustbot reviewed 😆

[bjorn3]

@rustbot ready

[oli-obk]

@bors r+

[bors]

📌 Commit 9f190d7 has been approved by oli-obk

It is now in the queue for this repository.