Re: Bug#236291: FWD: [asac@jwsdot.com: Bug#236291: Help needed for hppa Segmentation fault on buildd during configure]
On Fri, Mar 12, 2004 at 09:34:27AM +0100, Joel Soete wrote:
> Ok, I can so just replace conftest.h by the few #define and I can reproduce
> de segv with "gcc conftest.c" and even isolate the pb which is here:
> [snip]
> p = (char**) ccp;
> ccp = (char const *const *) p;
> { /* SCO 3.2v4 cc rejects this. */
> char *t;
> char const *s = 0 ? (char *) 0 : (char const *) 0;
>
>
> *t++ = 0; <========== THIS IS THE PB
> }
> [snip]
>
> IMHO it should be a pb on all arch as far as "t" is not initialized?
>
> replacing "char *t;" by "char *t="anotherstring"; make it works but I have
> no clue on how this code is supposed to failed in this case.
As a datapoint, I'm able to build, without problems:
dpkg-buildpackage -rfakeroot
mozilla-thunderbird_0.4-1_hppa.deb
...
Using my up-to-date unstable debian hppa box. Is that buildd okay?
c.
Reply to: