This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: Patch installed for PR7227 gcc.dg/uninit-C.c
- From: "John David Anglin" <dave at hiauly1 dot hia dot nrc dot ca>
- To: ghazi at caip dot rutgers dot edu (Kaveh R. Ghazi)
- Cc: gcc-patches at gcc dot gnu dot org
- Date: Thu, 17 Apr 2003 12:41:33 -0400 (EDT)
- Subject: Re: Patch installed for PR7227 gcc.dg/uninit-C.c
> > I had the impression that the test was checking a problem specific
> > to TImode, so I wasn't sure it was reasonable to reduce the mode
> > size that was used. There is a test for the number of bits in a
> > long long in c99-intconst-1.c that might be useful.
>
> I don't remember if smaller modes triggered the bug. (Falling back to
> a smaller mode couldn't hurt though.) What I was suggesting was to
> ifdef out the entire test if __LONG_LONG_MAX__ was less than a certain
> size. But it seems that __LONG_LONG_MAX__ isn't exactly on point. If
> we need to check for a 128-bit mode, testing long long won't work on
> most systems either since that type is usually 64-bits, even on
> systems that support a TI mode, right?
It would take some research to determine why TI mode works on
hppa64-hp-hpux11.11 and not on hppa2.0w-hp-hpux11.11. Long longs
are 64 bits on both, although in the former case instructions
exist for direct manipulation of DImode integers.
Dave
--
J. David Anglin dave dot anglin at nrc-cnrc dot gc dot ca
National Research Council of Canada (613) 990-0752 (FAX: 952-6602)