Bugzilla Bug 260063

This bug will be the stabilization bug for GNOME-2.24. Filing it before the
stabilization list, so that we can track the true blocker bugs for
stabilization purposes, and have other maintainers be able to ask here
beforehand packages to be included together with the rest for possible
compatibility reasons.

Please look into (the non exhaustive) bug list depending on this one which
should be fixed before proceeding to stable if possible.

Also (note to self mostly) don't forget that due do API/ABI breakage, some
packages really need to get stabilized together.

Created an attachment (id=182951) [details]
bump.list

first tentative list, please poke me for anything I would have forgotten.

> first tentative list, please poke me for anything I would have forgotten.

I need the following packages as well on my amd64 installation:

>=x11-libs/pixman-0.12.0
~dev-python/libgnome-python-2.22.3
~dev-python/libbonobo-python-2.22.3
~dev-python/gconf-python-2.22.3
~dev-python/libgnomecanvas-python-2.22.3
~dev-python/gnome-python-base-2.22.3
~dev-python/gnome-vfs-python-2.22.3

You missed the lines talking about the stabilization of split python bindings
it seems. Wrt to pixman, I supposed it's pulled by cairo, I was told cairo-1.8
was good to go, will check specific pixman status. Thanks for reporting.

Created an attachment (id=183628) [details]
bump.list

Updated list.

enough with the wait & fix, it's time for some action.

Hi arches, please find attached to this bug the list of packages and their
respective target keywords for gnome 2.24 stabilization. Please do go through
the list of bugs depending on this one, they need to be completed as well.

If you find any problem with this list, please come and see us on
#gentoo-desktop or send us a mail.

Thanks in advance.

mips done.

(there is no stable mips ;)

*** Bug 253934 has been marked as a duplicate of this bug. ***

FTR, that should be =media-video/totem-2.24.3

No -r1, which is masked anyway.

And media-libs/swfdec-0.8.4 should be in the list.

Any chance we could get a new list to work with (ala comment #4) that includes
comments 8 & 9 AND gstreamer?  Would be mucho appreciated.

I'll make sure to do that by tomorrow. About gstreamer, I can do that as well
but I really need a decision on what's in and what's out.

I'd suggest to leave out the gstreamer stuff if no individual apps actually
need the newer versions and all we'd have to do is just downgrade the dep in
the meta package.
That's to not have arches needing to handle an even bigger set of things at
once here.
gstreamer herd stuff I'm hoping to get back on track during March, with
hopefully looking into gst stabilizations later this week.

Yeah, I only meant that the new master list would include required gstreamer
packages.  As for across the board stabilization of gstreamer, I would assume
another bug would be file as has been suggested.  So what I looking for is a
more complete list from the earlier comment.  Is that fair?  

I've used the following command to verify what was requiring 0.10.21

find -name "*.ebuild" -exec egrep -H "gst.*0\.10\.21" {} \+ |cut -f1 -d: |awk
-F / '{print $2"/"$4}' |sed "s/.ebuild//g" |sort -u

besides banshee and gnome meta, it seems we can go ahead with just 0.10.20 so
I'll adjust meta and post the updated list later today.

Created an attachment (id=183938) [details]
build.list

Ok here it is. I've downgraded gstreamer-0.10.21 requirement per previous
comment.
Enjoy.

ppc64 done; please add us back in if we missed something. Now onto ppc32.

For the record, I asked rangerpb to test and keyword evolution-2.24.5,
evolution-data-server-2.24.5 and gtkhtml-3.24.5 since it provides at least two
important fixes over 2.24.4. Not crashing on print and working imap flag sync
(which resulted in read mail being unread when consulted from another client or
another place). I'll provide an updated list to match this asap.

ppc(32) done now too; please readd if I missed something

Created an attachment (id=184223) [details]
bump.list

Here is an updated list from cvs up 1 hour ago.

Right now, all libcanberra versions in portage depends on
>=media-sound/pulseaudio-0.9.11 if the user activates the pulse USE flag. But
pulseaudio-0.9.11 requieres baselayout-2.0 and openrc in order to stabilize it.
There are some ways of solving the problem:

 - lowering the libcanberra's pulseaudio requirement to the current stable
version (0.9.9)
 - deactivating the pulse USE flag in libcanberra; it still should be able to
play sounds using the other available backends (alsa, gstreamer, oss)
 - proving a new pulseaudio-0.9.11 ebuild which doesn't depend on
openrc/baselayout-2.0, and make it stable
 - waiting for openrc/baselayout-2.0 to stabilize, in order to stabilize GNOME
2.24

I don't see a problem in use.masking pulseaudio, solution 1 is out of the
question it's a requirement from libcanberra's configure, 3 is afaik not wanted
by pa maintainer, and 4 is out of the question too, openrc/baselayout2
stabilization is taking a long time and we don't need to delay gnome 2.24 any
longer that it already has been.

The only baselayout 2 stuff in the pa ebuild is parsing the pa config file to
create the rc deps.. We can pretty easily make a baselayout-1 version by doing
what the current ebuild is doing, since openrc/baselayout-2 is taking
centuries..

Created an attachment (id=184644) [details]
split python keywords for 2.24

These are the 2.24 split python keywords.  Anyone keywording the gnome-2.24.1
meta needs these too.

For libcanberra, I've removed the pulseaudio flag and dep from 0.11, and added
a 0.11-r1 that's ~arch with the flag.  As soon as pulse goes stable, that can
go stable too.

amd64 done.

Would be important get bug 262182 fixed before getting evolution-2.24
stabilized as it will cause users to get evolution crashing at startup. I
attach the patch used in fedora currently and that seems to work ok for me

(sorry for amd64 users for sending this a bit late, but I discovered this just
after updating)

gedit and vinagre are failing tests here on x86:

make[3]: Nothing to be done for `check-am'.
make[3]: Leaving directory
`/var/tmp/portage/net-misc/vinagre-2.24.2-r1/work/vinagre-2.24.2/src'
make[2]: Leaving directory
`/var/tmp/portage/net-misc/vinagre-2.24.2-r1/work/vinagre-2.24.2/src'
make[1]: Leaving directory
`/var/tmp/portage/net-misc/vinagre-2.24.2-r1/work/vinagre-2.24.2/src'
Making check in help
make[1]: Entering directory
`/var/tmp/portage/net-misc/vinagre-2.24.2-r1/work/vinagre-2.24.2/help'
xmllint --noout --noent --path C:./C --xinclude --postvalid ./C/vinagre.xml
xmllint --noout --noent --path de:./de --xinclude --postvalid ./de/vinagre.xml
./de/vinagre.xml:256: element para: validity error : Element shortcut is not
declared in para list of possible children
Document ./de/vinagre.xml does not validate
make[1]: *** [check-doc-docs] Error 3
make[1]: Leaving directory
`/var/tmp/portage/net-misc/vinagre-2.24.2-r1/work/vinagre-2.24.2/help'
make: *** [check-recursive] Error 1
 * 
 * ERROR: net-misc/vinagre-2.24.2-r1 failed.
 * Call stack:
 *               ebuild.sh, line   49:  Called src_test
 *             environment, line 2865:  Called _eapi0_src_test
 *               ebuild.sh, line  616:  Called die
 * The specific snippet of code:
 *                      hasq test $FEATURES && die "Make check failed. See
above for details."
 *  The die message:
 *   Make check failed. See above for details.

make[2]: Nothing to be done for `check-am'.
make[2]: Leaving directory
`/var/tmp/portage/app-editors/gedit-2.24.3/work/gedit-2.24.3/help'
xmllint --noout --noent --path C:./C --xinclude --postvalid ./C/gedit.xml
xmllint --noout --noent --path de:./de --xinclude --postvalid ./de/gedit.xml
./de/gedit.xml:446: element menuchoice: validity error : Element menuchoice
content does not follow the DTD, expecting (shortcut? , (guibutton | guiicon |
guilabel | guimenu | guimenuitem | guisubmenu | interface)+), got (CDATA
guimenu guimenuitem)
Document ./de/gedit.xml does not validate
make[1]: *** [check-doc-docs] Error 3
make[1]: Leaving directory
`/var/tmp/portage/app-editors/gedit-2.24.3/work/gedit-2.24.3/help'
make: *** [check-recursive] Error 1
 * 
 * ERROR: app-editors/gedit-2.24.3 failed.
 * Call stack:
 *               ebuild.sh, line   49:  Called src_test
 *             environment, line 3094:  Called _eapi0_src_test
 *               ebuild.sh, line  616:  Called die
 * The specific snippet of code:
 *                      hasq test $FEATURES && die "Make check failed. See
above for details."
 *  The die message:
 *   Make check failed. See above for details.

x86 stable

alpha/ia64 stable, i think

i've tried to upgrade gnome to 2.24 (marked as stable for x86) and it fails on
x11-libs/pixman which cannot be compiled with sse2 USE flag with gcc older than
4.2.4. And gcc-4.2.4 is not available in x86.

Shouldn't it be fixed?

But the ebuild already tells you to do 'echo "x11-libs/pixman -sse2" >>
/etc/portage/package.use' if you are using earlier GCC version.

first, i did add -sse2 to package.use of course and i'm merging gnome, the
question is if it is ok to have unstable dependency (this way gcc-4.2 is
something like dependency isn't it?) in stable package?

second merging pixman fails with something which looks like big problem (ERROR:
setup blah blah...) not just INFO: please type echo "blahblah" >> package.use

last but not least, it's not a problem for me and probably 99% of users, i just
wanted to tell you that something like this happens and maybe it should be
resolved in more user friendly way.

(In reply to comment #32)

1. It does not depend on gcc-4.2.
2. It fails with the error and not the warning because the emerge process
cannot be continued.

And it seems pretty user-friendly to me :-)

This bug is not for chatting about your own bugs, open a new bug please
(although it seems familiar to me so search through bugzilla should help).

Hi arches, please don't stabilize libxklavier 3.8.

We will open a separate stablization bug once xorg-server 1.5.3 and friends go
stable.

Thanks

sparc stable

s390 should be done

arm/m68k should be done, sh is waiting for totem

sh waits for bug 196852

This bug is obsoleted by bug #281427, see you there.