On Thu, Feb 28, 2008 at 1:31 PM, The Rasterman Carsten Haitzler
<[EMAIL PROTECTED]> wrote:
> On Wed, 27 Feb 2008 19:57:04 -0300 "Gustavo Sverzut Barbieri"
> <[EMAIL PROTECTED]> babbled:
>
>
> > Guys, I did a quick test with emotion_test (c version) and it's
> > leaking... I'm running home, but the relevant part is:
> >
> > ==19235== LEAK SUMMARY:
> > ==19235== definitely lost: 112 bytes in 6 blocks.
> > ==19235== indirectly lost: 240 bytes in 20 blocks.
> > ==19235== possibly lost: 37,053 bytes in 906 blocks.
> > ==19235== still reachable: 26,833,797 bytes in 8,821 blocks.
> > ==19235== suppressed: 0 bytes in 0 blocks.
> >
> > almost 27mb of a simple video is really leaking something :-)
>
> still reachable is just irrelevant. never even bother with that. it's totally
> bogus. use memprof if anything.
man, 2-5mb I'd say it's bullshit, like list node caches, pre
calculated tables, ...
but 27mb?!
I did yet another test, but unfortunatelly I was recompiling efl
during that so the last part is "corrupted" and just show 0x2F, but it
might bring you attention to:
==28413== 4,942,080 bytes in 15 blocks are still reachable in loss
record 142 of 143
==28413== at 0x4005426: malloc (in
/usr/lib/valgrind/x86-linux/vgpreload_memcheck.so)
==28413== by 0x42AFB47: _emotion_frame_format_update
(emotion_xine_vo_out.c:374)
==28413== by 0x42E7CE3: (within /usr/lib/libxine.so.1.19.0)
==28413==
==28413==
==28413== 18,115,480 bytes in 2,537 blocks are still reachable in loss
record 143 of 143
==28413== at 0x400487B: calloc (in
/usr/lib/valgrind/x86-linux/vgpreload_memcheck.so)
==28413== by 0x4303FD3: xine_xmalloc (in /usr/lib/libxine.so.1.19.0)
==28413== by 0x2F: ???
==28413==
==28413== LEAK SUMMARY:
==28413== definitely lost: 40 bytes in 4 blocks.
==28413== possibly lost: 37,085 bytes in 908 blocks.
==28413== still reachable: 26,848,760 bytes in 8,402 blocks.
==28413== suppressed: 0 bytes in 0 blocks.
so of 27mb, 18mb is from xine_xmalloc() and 5mb is from emotion-xine
in _emotion_frame_format_update():
fr->vo_frame.base[0] = malloc(y_size + (2 * uv_size));
which does call _emotion_frame_data_free(), so it's not due calling
this function more than once (that would override the pointer and that
would also issue a "definitely lost").
I'm investigating where it's being lost... but I guess "fr" itself is
being lost somewhere.
--
Gustavo Sverzut Barbieri
--------------------------------------
Jabber: [EMAIL PROTECTED]
MSN: [EMAIL PROTECTED]
ICQ#: 17249123
Skype: gsbarbieri
Mobile: +55 (81) 9927 0010
-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2008.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
_______________________________________________
enlightenment-devel mailing list
enlightenment-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/enlightenment-devel
