On Mon, Apr 8, 2013 at 1:03 PM, Derek Schuff <[email protected]> wrote:
> I think it might be better to just throw an error on regparm > attributes (i.e. by setting regParmMax=0 in PNaClTargetInfo) rather > than ignoring them. That way if someone tries to build against glibc > headers (the reason for this code's existence in the first place), it > will fail loudly rather than just silently generating incompatible > code. > > Good point. Attaching updated patch. PNaClTargetInfo now sets RegParmMax to 0 explicitly to disallow it, and NaClTargetInfo inherits it from the underlying architecture. As for tests, I noticed that with the removal of regparm support, le32-arguments.c already covers the interesting cases so what remains for le32-regparm.c is to verify the Clang error. Eli
le32-simplify-regparm-abi.2.patch
Description: Binary data
_______________________________________________ cfe-commits mailing list [email protected] http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
