After an evaluation, GNOME has moved from Bugzilla to GitLab. Learn more about GitLab.
No new issues can be reported in GNOME Bugzilla anymore.
To report an issue in a GNOME project, go to GNOME GitLab.
Do not go to GNOME Gitlab for: Bluefish, Doxygen, GnuCash, GStreamer, java-gnome, LDTP, NetworkManager, Tomboy.
Bug 104837 - [PATCH] Problems with GdkPixbuf on PowerPC
[PATCH] Problems with GdkPixbuf on PowerPC
Status: RESOLVED DUPLICATE of bug 71597
Product: gdk-pixbuf
Classification: Platform
Component: general
git master
Other Linux
: High major
: ---
Assigned To: gtk-bugs
Control-Center Maintainers
Depends on:
Blocks:
 
 
Reported: 2003-01-30 18:31 UTC by Matt Brubeck
Modified: 2010-07-10 04:05 UTC
See Also:
GNOME target: ---
GNOME version: ---


Attachments
Screenshot of the problem (31.22 KB, image/png)
2003-01-31 02:43 UTC, Matt Brubeck
  Details
gnome-control-center, background, shows pixbuf(?) bug (39.53 KB, image/png)
2003-01-31 10:03 UTC, Christof Petig
  Details
fix for endianness issue? (553 bytes, patch)
2003-02-02 22:39 UTC, Jon Nall
none Details | Review

Description Matt Brubeck 2003-01-30 18:31:46 UTC
In gnome-theme-manager, the metacity theme preview image is currupted, as
when viewing an invalid image file.  The image is just a scattering of
apparently random colored pixels.  I will attach a screenshot later.  The
icon portion of the preview is okay; only the window image behind it is
corrupt.

I run Linux on powerpc, so it's possible this is an endianness bug.

Using control-center 2.2.0 and metacity 2.4.21 from Debian sid.  I don't
know whether this bug is in gnome-theme-manager or libmetacity (or elsewhere).
Comment 1 Havoc Pennington 2003-01-30 19:26:09 UTC
Do metacity themes work if you select them?
It's the same code drawing the preview and the actual window frame...
Comment 2 Matt Brubeck 2003-01-30 19:34:10 UTC
Yes, the actual window frames draw correctly when I select a theme. 
Only the preview images are corrupt.
Comment 3 Matt Brubeck 2003-01-31 02:43:02 UTC
Created attachment 13979 [details]
Screenshot of the problem
Comment 4 Matt Brubeck 2003-01-31 02:47:39 UTC
Another powerpc user was able to reproduce this bug.  Definitely seems
to be ppc-specific:

http://lists.debian.org/debian-gtk-gnome/2003/debian-gtk-gnome-200301/msg00669.html
Comment 5 Michele Campeotto 2003-01-31 08:52:01 UTC
I can confirm this bug and I use PPC too.
Comment 6 Christof Petig 2003-01-31 10:02:24 UTC
Perhaps this is related to a gdk-pixbuf bug which shows up for .xpm
and .jpg images on buttons. Once the button is highlighted (or active)
the image gets corrupt.

I have also debian/ppc, and yes, I can confirm the bug exists since 2.0.0.

I'll attach an image of my gnome-control-center background selection.
Once I move the mouse over a button it's contents are scrambled.
Comment 7 Christof Petig 2003-01-31 10:03:52 UTC
Created attachment 13986 [details]
gnome-control-center, background, shows pixbuf(?) bug
Comment 8 Christof Petig 2003-01-31 10:08:18 UTC
I'd vote for this being an endianness issue since it seems to mirror
the image in four pixel wide stripes.
Comment 9 Jon Nall 2003-02-02 13:54:00 UTC
just wanted to add a "me too". my screenshot looks exactly like matt's
attachment #13979 [details]. i'm running on gentoo linux on PPC with gnome-2.2_rc2.
Comment 10 Jon Nall 2003-02-02 20:10:46 UTC
just another observation. the folder icon looks just great. it's the
part of the pixbuf that's generated from
gdk_pixbuf_get_from_drawable() (around theme-thumbnail.c:197) that
looks like it has endian issues.
Comment 11 Jon Nall 2003-02-02 22:37:13 UTC
attaching a patch against gtk+-2.2.0. this fixes the problem observed
in matt's screenshot. i believe this is the culprit, however it's a
bit strange that the big and little endian code would be identical.

anyway, here's a shot at it...
Comment 12 Jon Nall 2003-02-02 22:39:46 UTC
Created attachment 14045 [details] [review]
fix for endianness issue?
Comment 13 Andrew Sobala 2003-02-02 22:44:20 UTC
-> gtk, please move back to gnome-control-center if it turns out not
to be a Gtk bug.
Comment 14 Owen Taylor 2003-02-03 02:39:46 UTC

*** This bug has been marked as a duplicate of 71579 ***
Comment 15 Owen Taylor 2003-02-03 02:40:20 UTC
Wrong number
Comment 16 Owen Taylor 2003-02-03 02:40:43 UTC

*** This bug has been marked as a duplicate of 71597 ***