[PATCH] D40632: Add flag to llvm-ar to test GNU64 format more efficently

[Rafael Avila de Espindola via llvm-commits]

Peter Collingbourne via Phabricator <reviews at reviews.llvm.org> writes:

> pcc added inline comments.
>
>
> ================
> Comment at: tools/llvm-ar/llvm-ar.cpp:102
>  
> +static cl::opt<int> Sym64PowerOfTwo("sym64-power-of-two", cl::Hidden);
> +
> ----------------
> ruiu wrote:
>> and change this to ForceGnu64 or something.
> If this flag is only meant to be used in one test, perhaps it would be better to move it into ArchiveWriter.cpp? That way, we can avoid extending the interface to the archive writer.

Agreed. Flags in libraries are normally to be avoided, but I think they
are the correct solution for test only flags. One day we will hopefully
have the build infrastructure to disable them in non-test builds.

Cheers,
Rafael