This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: query re compile/961203-1.c
- To: Zack Weinberg <zack at rabi dot columbia dot edu>
- Subject: Re: query re compile/961203-1.c
- From: Jeffrey A Law <law at upchuck dot cygnus dot com>
- Date: Sat, 15 May 1999 03:06:03 -0600
- cc: egcs at egcs dot cygnus dot com
- Reply-To: law at cygnus dot com
In message <199905141600.MAA01977@blastula.phys.columbia.edu>you write:
> On Thu, 13 May 1999 16:49:30 -0600, Jeffrey A Law wrote:
> >
> > In message <199905131206.IAA10386@blastula.phys.columbia.edu>you write:
> > > I'm confused. I have a 32 bit platform, the test passes, and the
> > > generated assembly is correct on inspection. Are you saying you think
> > > it'll still fail or still consume too much memory on 64 bit platforms?
> >Precisely.
>
> That seems a bit odd, but I'll take your word for it. In that case,
> shall we reenable the test for 32 bit platforms only?
It's dependent on both the host & target if I remember correctly. So
it would be difficult to conditionally enable it in that manner.
My memory is fuzzy, but I believe the reason it doesn't blow up the compiler
anymore is because we check TREE_OVERFLOW in a critical place which catches
the overflow when the size of the structure overflows (in bits).
When you move to a 64bit host, you do not necessarily overflow anymore.
I'm more than happy to be proved wrong, but until then I'd suggest that
test remain unchanged because of it's nasty behavior.
jeff