[PATCH] Add a fallback mechanism for undefined atom.

Rui Ueyama ruiu at google.com
Thu Sep 12 12:36:04 PDT 2013


On Thu, Sep 12, 2013 at 12:29 PM, Michael Spencer <bigcheesegs at gmail.com>wrote:

>
>
> ================
> Comment at: lib/ReaderWriter/Native/NativeFileFormat.h:171-178
> @@ -169,1 +170,10 @@
>
> +//
> +// The NCS_UndefinedAtomsV2 chunk contains an array of these structs
> +//
> +struct NativeUndefinedAtomIvarsV2 {
> +  uint32_t  nameOffset;
> +  uint32_t  flags;
> +  uint32_t  fallbackNameOffset;
> +};
> +
> ----------------
> Rui Ueyama wrote:
> > Shankar Kalpathi Easwaran wrote:
> > > We should just add the fallbackNameOffset to V1 (or) if you want to
> still do this, you could derive the second struct from the first one.
> > >
> > > struct NativeUndefinedAtomIvarsV2 : public NativeUndefinedAtomIvarsV1
> > done.
> We are still modifying V1.
>

Shankar said that we should add the field to V1? If we shouldn't we could
define V2 and remove the field from V1.

================
> Comment at: test/core/undef-fallback.objtxt:13-14
> @@ +12,4 @@
> +  - name: foo
> +    fallback:
> +      name: bar
> +...
> ----------------
> Why isn't this just fallback: bar ?
>

Because fallback is not a string but an undefined atom.


>
>
> http://llvm-reviews.chandlerc.com/D1550
>
> COMMIT
>   http://llvm-reviews.chandlerc.com/rL190625
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20130912/d06c489c/attachment.html>


More information about the llvm-commits mailing list