[PATCH] Expose all of the Value hierarchy to the C API

Rafael EspĂ­ndola rafael.espindola at gmail.com
Tue May 7 21:56:19 PDT 2013


We maintain way more backward compatibility in the C api than in the
C++ one, so it is probably best if we only add to it as needed.

On 6 May 2013 20:11, Ben Longbons <brlongbons at gmail.com> wrote:
> I was motivated to do this after I discovered that it's possible to
> create a ConstantDataArray with the C API, but not determine if a
> Value is a ConstantDataArray.
>
> I created this patch to expose the entire hierarchy, which led me to
> discover a couple of odd cases:
>  - ConstantExpr having private subclasses is not ugly.
>  - PseudoSourceValue being in CodeGen is only slightly ugly.
>  - IntrinsicInst being a fake subclass is fairly ugly, but it was
> useful, and besides removing it might break existing callers.
>  - Operator is really ugly and deserves to die, so I did not expose it.
>
> I'm o11c on IRC btw. I gave a whitespace-only patch to baldrick that
> he already committed to keep this one cleaner.
>
> Thanks,
> -Ben
>
> _______________________________________________
> llvm-commits mailing list
> llvm-commits at cs.uiuc.edu
> http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits
>



More information about the llvm-commits mailing list