draining the Maemo/Meego patch queue

Colin Walters walters at verbum.org
Tue Mar 16 11:04:03 PDT 2010


So as part of preparing a new release I did a bit of bugzilla
processing, and while looking at an older Nokia bug there was a patch
that was far distant from dbus-1.2.  A bit of searching revealed


Which has a huge patch list.  Let me try to go over this:

* 00-install-doxygen-docs.patch
No upstream bug.  OS vendors should have submitted a patch for this
long ago.  I find it annoying equivalent code exists in the Fedora
dbus.spec.  But we probably need --docdir for configure.  If someone
motivated would dive into the autotools required for this I'd be

* 01-add-initgroups-call.patch
No upstream bug.  Need explanation - I think this is trying to
preserve the groups for the system bus user?  But why?

* 02-allow-all-in-session.patch
Maemo specific - however, I'd totally take a patch which added

* 04-no-fork.patch

* 07-open-match-detection.patch
Ewwww!  This one should really be rewritten as a systemtap script.

* 08-create-new-process-group.patch
Has upstream bug (hooray!), I commented there.

* 09-remove-max-match-limit.patch
I think this one could also be done in bus/maemo-session.conf

* 10-up-connection-limits.patch

* 11-fix-dirent_buf_size.patch
Hmmm....I think this is http://bugs.freedesktop.org/show_bug.cgi?id=15922

* 17-do-not-inherit-oom-adj.patch
No upstream bug.  But a reasonable patch, would like to see it upstream.

* 18-fix-remove-bogus-filter.patch
Looks like this one was committed as 93023eb6a447bdb640

* 21-add-missing-va_end.patch
No upstream bug.  Is obviously correct, so I just committed it to dbus-1.2 now.

* 22-use-signed-return-value.patch
No upstream bug.  Is obviously correct, so I  just committed it to dbus-1.2 now.

* 32-add-oom-check-to-main.patch
I cleaned this up and just committed it to dbus-1.2 now.

* 35-fixed-socket-dir.patch
Maemo specific, could go in maemo-session.conf again.

* 36-bug-896-ignore-epipe-on-writes.patch
This was committed earlier as 87ddff6b

* 40-prevent-autolaunch.patch
No upstream bug.  Not clear why this patch exists?

* 45-dbus-clock-monotonic-1.patch
This is in dbus-1.2.20.

* 50-ignore-epipe-on-write.patch
Heh, a patch for the earlier patch.  Anyways, irrelevant now.

* 55-rewrite-rulematching.patch
This one is on master now (but not dbus-1.2).  If someone wants to
backport it to 1.2, I'd consider accepting that patch.

* 60-no-match-acks-when-not-needed.patch
No upstream bug.  A good idea though!  I cleaned this up so that we
use it consistently, and put it here:
http://bugs.freedesktop.org/show_bug.cgi?id=27112  Someone please

More information about the dbus mailing list