Bug 2399 - gdk-pixbuf: hundreds of warnings on boot
Summary: gdk-pixbuf: hundreds of warnings on boot
Status: RESOLVED FIXED
Alias: None
Product: OE-Core
Classification: Build System, Metadata & Runtime
Component: core (show other bugs)
Version: 1.2.1
Hardware: x86 Multiple
: Medium+ normal
Target Milestone: 1.2.2
Assignee: Scott Garman
QA Contact:
URL:
Whiteboard:
: 2663 (view as bug list)
Depends on:
Blocks:
 
Reported: 2012-05-03 04:03 UTC by Tom Zanussi
Modified: 2012-12-06 16:10 UTC (History)
8 users (show)

See Also:
OS type for building Yocto: ---
Type of Regression: ---
Verified:
Documentation change: ---


Attachments
new bootlog (serial capture) of an emenlow boot (176.60 KB, application/octet-stream)
2012-10-02 18:24 UTC, Tom Zanussi
no flags Details
Potential fix (1.42 KB, patch)
2012-10-03 15:58 UTC, Ross Burton
no flags Details | Diff

Note You need to log in before you can comment on or make changes to this bug.
Description Tom Zanussi 2012-05-03 04:03:32 UTC
On boot on all the meta-intel BSPs (I assume the same would be true for qemu), there are hundreds[1] of these warning messages:

(gtk-update-icon-cache:1234): GdkPixbuf-WARNING **: Cannot open pixbuf loader module file '/usr/lib/gdk-pixbuf-2.0/2.10.0/loaders.cache': No such file or directory


[1] $ cat bootlog | egrep GdkPixbuf-WARNING | wc -l
921
Comment 1 Tom Zanussi 2012-06-26 20:24:04 UTC
*** Bug 2663 has been marked as a duplicate of this bug. ***
Comment 2 Laurentiu Palcu 2012-09-20 14:04:55 UTC
The patches for postinst improvements (bug #1621) will fix this issue too. The problem was the the loaders.cache was not generated for target when the target rootfs was constructed. With the fixes for #1621, this issue will dissapear.
Comment 3 Saul Wold 2012-09-27 20:48:18 UTC
Since we are not pulling in the rootfs gen patches, is there a way to fix this when running directly on the target?
Comment 4 Ross Burton 2012-10-02 15:44:45 UTC
Tom, how would I get a bootlog easily?  I don't have a bootlog when I boot up a qemux86 image.
Comment 5 Tom Zanussi 2012-10-02 16:06:38 UTC
Not sure with qemu - the bootlog referred to was just a serial capture.  Maybe /var/log/messages...
Comment 6 Ross Burton 2012-10-02 16:16:52 UTC
Nothing in messages.  I'll dig out a serial cable...
Comment 7 Tom Zanussi 2012-10-02 16:19:57 UTC
Just maybe an obvious note that I think this will only be seen on the first boot.
Comment 8 Ross Burton 2012-10-02 16:24:29 UTC
Yep, this was first boot.  gdkpixuf loaders and icon themes all postinit'd without a warning.

The X server taking over meant I missed the last few lines.  If you can replicate easily with a recent image that would be much appreciated.
Comment 9 Tom Zanussi 2012-10-02 16:27:30 UTC
OK, I'll hopefully be booting a fresh emenlow image ;-) and will capture the serial if I still see the problem.
Comment 10 Tom Zanussi 2012-10-02 18:24:48 UTC
Created attachment 864 [details]
new bootlog (serial capture) of an emenlow boot

meta-intel/master: 50ac6e8785c167ea4aa4601fd690ef783151853d
poky/master: 0ca8b6248e5d00351e8a35c87037cc47945e436b

poky/master had the following revert in order to get a bootable test image:

    Revert "initrd: Spawn an emergency shell when something goes wrong"
    
    This reverts commit ffb6928f5783e5202d9849c3a185e29be1d41c63.
Comment 11 Ross Burton 2012-10-02 21:30:01 UTC
Heh, so I don't get any warnings but I do see what packages are being postinstalled.  You get warnings but no hint as to what is failing.
Comment 12 Ross Burton 2012-10-03 15:46:26 UTC
Okay, interesting that all of those warnings come from a single PID.  I wish there was an easy way to find out what package installed /etc/rpm-postinsts/102.
Comment 13 Ross Burton 2012-10-03 15:58:05 UTC
Created attachment 867 [details]
Potential fix

I've a hunch this is sato-icon-theme, so try this.
Comment 14 Tom Zanussi 2012-10-03 21:42:42 UTC
That seems to have done the trick - no more gdkpixbuf warnings seen here on emenlow.
Comment 15 Ross Burton 2012-10-04 11:42:28 UTC
Revised patch sent to oe-core.
Comment 16 Ross Burton 2012-10-08 10:49:59 UTC
Merged in OE-Core rev: 8848ea6793ddaab61c9dad250ec578d68d7d087d.
Comment 17 Ross Burton 2012-10-08 10:50:48 UTC
I meant oe-core 9d98dbdae4c05fcf50d546f554a04dc3f0bd66c3.
Comment 18 Stoicescu Cornel 2012-11-02 11:37:46 UTC
Issue fixed in master and danny, but still present in denzil, commit: 846b7c388738f382e83576e77b705aa547c1dd9b
Comment 19 Ross Burton 2012-11-20 15:49:56 UTC
Re-assigning to Scott, as it's his call if this gets fixed in Denzil or not.
Comment 20 Richard Purdie 2012-12-06 16:10:36 UTC
It doesn't make sense to backport this to 1.2 at this point unless there is significant community demand.