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: law at cygnus dot com
- Subject: Re: query re compile/961203-1.c
- From: Zack Weinberg <zack at rabi dot columbia dot edu>
- Date: Tue, 04 May 1999 10:28:50 -0400
- cc: egcs at egcs dot cygnus dot com
On Mon, 03 May 1999 23:55:59 -0600, Jeffrey A Law wrote:
>
> In message <19990503201643.W738@rjlhome.sco.com>you write:
> > > 961203-1.c has to do with the handling of ridiculously large structs.
> > [ munch ]
> > > However, it appears to compile just fine with current CVS and with 1.1.2
> > > on i586-pc-linux-gnu, which is 32 bit; cc1 completes in 0.04 sec and
> >
> > OpenServer was one of the targets that was crippled by this test. I
> > just checked the trunk and it no longer depletes VM so I'm guessing that
> > things are indeed different.
> >
> > When you say that the code works correctly, you mean that it does
> > "correctly" take a SIGSEGV when dereferencing the pointer 'p' that is
> > set to zero, right?
>Note, when you run the testsuite, it will not actually try to compile
>this test. It just marks it as failed (expected) and goes to the next test.
>
>If it's being run on via the testsuite, then something is wrong.
>
>As it turns out I believe the compiler may know enough to not blow its
>cookies on this test anymore (due to some unrelated changes). But I haven't
>tested it.
I removed the .x file in my local tree; the test passed at all
optimization levels. strace indicates the compiler wants about 100k
of heap for this file. But my platform is rather vanilla...
zw