[FFmpeg-devel] The state of FATE

Måns Rullgård mans
Sun Nov 1 21:58:50 CET 2009


Reimar D?ffinger <Reimar.Doeffinger at gmx.de> writes:

> On Sun, Nov 01, 2009 at 07:34:41PM +0000, matthieu castet wrote:
>> M?ns Rullg?rd <mans <at> mansr.com> writes:
>> 
>> > 
>> > matthieu castet <castet.matthieu <at> free.fr> writes:
>> > 
>> > 
>> > That test is meant to fail if sizeof(char*) < 8.  The exact reason for
>> > failing is irrelevant.
>> > 
>> But the problem is that it doesn't fail on x86_32 with clang!
>> It allocate an array of 0xfffffffd
>
> Uh, and how should that be possible? It would have to create a .bss
> segment of that size without overlapping with .text etc., I can't
> see how that's possible unless llvm optimizes the array away completely...
> Or do we check only compilation and not linking?

We do only test compiling, but that's irrelevant.  The test even
manages to link.  Trying to access the last element in the array does
crash with a SIGSEGV.

-- 
M?ns Rullg?rd
mans at mansr.com



More information about the ffmpeg-devel mailing list