On Thu, 2009-05-28 at 00:33 -0400, Pierre-Luc Beaudoin wrote:
> 
> I've spent a lot of time (on many occasions) trying to pinpoint the
> source of a *huge* memory leak in libchamplain (both when using
> Clutter 0.8.8 or 0.9).

Hum, never mind me posting very late at night... The valgrind results
were wrong because I wasn't using G_SLICE=always-malloc
G_DEBUG=gc-friendly.

I am using Gnome's system monitor to observe libchamplain's demo apps,
and browsing around the world can quickly take up to 200 Mb of ram (and
it keeps growing).  That seems huge considering that at any point in
time there is a maximum on 32 256x256 images.

I guess I'll continue to investigate. Tell me if such a behavior is to
be expected ;)

Pierre-Luc

PS: here are the more sensible results I got:

==11883== 3,856,203 bytes in 433 blocks are still reachable in loss
record 308 of 308
==11883==    at 0x4C21540: memalign (vg_replace_malloc.c:460)
==11883==    by 0x4C215FA: posix_memalign (vg_replace_malloc.c:569)
==11883==    by 0xECB2763: _mesa_align_malloc
(in /usr/lib/dri/i965_dri.so)
==11883==    by 0xECF30A2: _math_matrix_ctr
(in /usr/lib/dri/i965_dri.so)
==11883==    by 0xECB66DF: (within /usr/lib/dri/i965_dri.so)
==11883==    by 0xECB6731: _mesa_init_matrix
(in /usr/lib/dri/i965_dri.so)
==11883==    by 0xEC81107: _mesa_initialize_context
(in /usr/lib/dri/i965_dri.so)
==11883==    by 0xEC49974: intelInitContext
(in /usr/lib/dri/i965_dri.so)
==11883==    by 0xEC5A6C1: brwCreateContext
(in /usr/lib/dri/i965_dri.so)
==11883==    by 0xEC3F560: (within /usr/lib/dri/i965_dri.so)
==11883==    by 0x928804A: (within /usr/lib/libGL.so.1.2)
==11883==    by 0x9288532: glXCreateContext (in /usr/lib/libGL.so.1.2)

Attachment: signature.asc
Description: This is a digitally signed message part

Reply via email to