[cairo-bugs] [Bug 91967] Assertion "(_cairo_atomic_int_get (&(&surface->ref_count)->ref_count) > 0)"

bugzilla-daemon at freedesktop.org bugzilla-daemon at freedesktop.org
Sat May 7 08:42:29 UTC 2016


https://bugs.freedesktop.org/show_bug.cgi?id=91967

--- Comment #14 from Wolfgang Ulbrich <chat-to-me at raveit.de> ---
Ok, i will try to explain what did crash in MATE desktop (gtk3 version).
Maybe this helps to get your attention.
Steps to reproduce:
1. Installing Mate desktop gtk3 version from repos for fedora or archlinux.
https://copr.fedorainfracloud.org/coprs/raveit65/Mate-GTK3/
Don't update cairo here because this repo use a patched cairo version to avoid
this annoying issue.

2. Load random applets into mate-panel.

3. enable compositor in marco (our WM) if it isn't already enabled.

4. Disable compositor via GUI or gsettings.
gsettings set org.mate.Marco.general compositing-manager false

5 Now you will see that several panel applets are crashing and the panel tries
to reload them.

This isn't a randomly failure and is reproducible all the time.

Crash in stacktrace from our clock-applet (full version i will attach):
Thread 1 (Thread 0x7f9fff7ae980 (LWP 27490)):
#0  0x00007f9ffbd629c8 in __GI_raise (sig=sig at entry=6) at
../sysdeps/unix/sysv/linux/raise.c:55
        resultvar = 0
        pid = 27490
        selftid = 27490
#1  0x00007f9ffbd6465a in __GI_abort () at abort.c:89
        save_stage = 2
        act = {__sigaction_handler = {sa_handler = 0x4, sa_sigaction = 0x4},
sa_mask = {__val = {140731922981920, 18793946, 140731922982000, 47270444160, 0,
0, 0, 21474836480, 140325103031070, 140731922982152, 140325162516673, 22530432,
140325162516673, 24234832, 140324717641728, 140325121230504}}, sa_flags = 953,
sa_restorer = 0x7f9ffd010af0 <__PRETTY_FUNCTION__.11261>}
        sigs = {__val = {32, 0 <repeats 15 times>}}
#2  0x00007f9ffbd5b187 in __assert_fail_base (fmt=<optimized out>,
assertion=assertion at entry=0x7f9ffd0106a8 "((*&(&surface->ref_count)->ref_count)
> 0)", file=file at entry=0x7f9ffd0105a0 "cairo-surface.c", line=line at entry=953,
function=function at entry=0x7f9ffd010af0 <__PRETTY_FUNCTION__.11261>
"cairo_surface_destroy") at assert.c:92
        str = 0x171bed0 " \b\212\001"
        total = 4096
> cut

With the patch from reporter this does not happen and we can can happily switch
the compositor off/on.
I use the fix since several month for myself and for the Mat gtk3 repo, and i
didn't saw any collateral damage.

Dear cairo devs,
hope this is enough info and please take a look at it again.

-- 
You are receiving this mail because:
You are the QA Contact for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cairographics.org/archives/cairo-bugs/attachments/20160507/9efa45e8/attachment.html>


More information about the cairo-bugs mailing list