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 563543 - Don't do umask
Don't do umask
Status: RESOLVED FIXED
Product: gnome-settings-daemon
Classification: Core
Component: general
unspecified
Other Linux
: Normal normal
: ---
Assigned To: gnome-settings-daemon-maint
gnome-settings-daemon-maint
Depends on:
Blocks:
 
 
Reported: 2008-12-07 09:01 UTC by Behdad Esfahbod
Modified: 2008-12-11 08:15 UTC
See Also:
GNOME target: ---
GNOME version: ---



Description Behdad Esfahbod 2008-12-07 09:01:09 UTC
When I converted g-s-d to have its own daemonization code instead of calling daemon(), I added umask(0177), which is what I found in some other daemon code...

Well, umask(0177) is a disaster if we end up creating a directory.  Which happens if the systemwide fontconfig cache is not up to date and it's a fresh user so ~/.fontconfig doesn't exist.  We create that dir, but it's not executable, so creating caches in it fail.

Incidentally, when I hit this today, the reason the systemwide cache was not up to date was that some broken fedora font package also does umask 0177 before calling fc-cache!  Very weird coincidence.

Anyway, the umask should go.  daemon() doesn't do it.  We don't need to do either.

/me goes cleaning up the mess in fedora
Comment 1 Behdad Esfahbod 2008-12-07 09:17:22 UTC
2008-12-07  Behdad Esfahbod  <behdad@gnome.org>

        * gnome-settings-daemon/main.c (daemon_detach): Don't call umask (bug
        #563543)

Comment 2 Pacho Ramos 2008-12-11 08:07:24 UTC
I have just read your blog post in planet.gnome.org and I would ask you if the fixes are also being pushed to 2.24 branch or only to trunk

Thanks for info :-)
Comment 3 Behdad Esfahbod 2008-12-11 08:15:25 UTC
The gnome-settings-daemon bug was not pushed to 2.24 branch, except that we backported it to Fedora 10.  So, the bug does not exist in g-s-d 2.24 branch, and I've already fixed it in trunk.