On Wed, Mar 20, 2013 at 2:22 PM, Diego Biurrun <[email protected]> wrote:
> On Sun, Mar 17, 2013 at 07:59:35AM +0100, Reinhard Tartler wrote:
>> The suppression file has been originally proposed by Reimar Doeffinger.
>> The idea of using set_defaults is from Luca Barbato.
>> --- /dev/null
>> +++ b/tests/fate-valgrind.supp
>> @@ -0,0 +1,31 @@
>> +# seems fixed in newer versions
>> +# http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=577135
>> +{
>> + zlib-inflate
>> + Memcheck:Cond
>> + fun:inflateReset2
>> + fun:inflateInit2_
>> +}
>> +# libc overreads on purpose
>> +# http://sourceware.org/bugzilla/show_bug.cgi?id=12424
>> +{
>> + eval-strtod
>> + Memcheck:Addr8
>> + fun:__GI___strncasecmp_l
>> + fun:____strtod_l_internal
>> + fun:av_strtod
>> +}
>> +{
>> + eval-strtod
>> + Memcheck:Value8
>> + fun:__GI___strncasecmp_l
>> + fun:____strtod_l_internal
>> + fun:av_strtod
>> +}
>> +{
>> + eval-strtod
>> + Memcheck:Cond
>> + fun:__GI___strncasecmp_l
>> + fun:____strtod_l_internal
>> + fun:av_strtod
>> +}
>
> The suppression file is highly specific to library versions and
> architecture. We could have a collection of standard suppression
> files for certain distros or so somewhere, but our repo is not
> the right place.
The proposed patch makes setting up a valgrind enabled fate instance
easier, and would allow to document it in a straight-forward way. Now
you say that the suppression file may not not apply to all, espc.
modern system. While true, I do not see this as a problem because
suppressing memory allocation mistakes that do not happen is not going
to invalidate the results of valgrind. This means that the patch still
has significant merits.
I would like to hear what you suggest to achieve the same goal. At the
moment I read only a "reject, please go away". Can we do better?
--
regards,
Reinhard
_______________________________________________
libav-devel mailing list
[email protected]
https://lists.libav.org/mailman/listinfo/libav-devel
