Mark Wielaard <[email protected]> writes:

> Hi Sam.
>
> On Tue, 2026-02-17 at 10:37 +0000, Sam James wrote:
>> Mark Wielaard <[email protected]> writes:
>> 
>> > clang with -flto produces object files that contain bitcode. This
>> > breaks various (self) testcases that expect those .o files are ELF
>> > files. readelf, nm, elfcheck and elfcompress won't work on such files.
>> 
>> This isn't specific to Clang. GCC does exactly the same thing. You only
>> get 'real code' in them if you pass -ffat-lto-objects too.
>
> As far as I know GCC does still produces an ELF file. While clang seems
> to generate something that isn't an ELF object file.

You're right, I was convinced they still used an ELF container, but they
don't! Sorry.

>
> The various elf tests don't really care what exactly is inside the ELF
> object file, just that it is still valid after whatever transformation
> has occurred on them. But if it isn't an ELF file to begin with there
> is little to test.
>
> Is there a configuration where GCC generates non-ELF .o files?
>

No, I checked, we're all good.

> Thanks,
>
> Mark

thanks,
sam

Attachment: signature.asc
Description: PGP signature

Reply via email to