Bug#677778: libbonobo: Please add multiarch support

Goswin von Brederlow goswin-v-b at web.de
Sun Jun 17 10:41:40 UTC 2012


Michael Biebl <biebl at debian.org> writes:

> On 17.06.2012 01:39, Goswin von Brederlow wrote:
>> Michael Biebl <biebl at debian.org> writes:
>> 
>>> On 17.06.2012 00:25, Goswin von Brederlow wrote:
>>>> Michael Biebl <biebl at debian.org> writes:
>>>
>>>>> But, to get rid of ia32-libs, we don't need to convert libbonobo to ma,
>>>>> right? Or is there another pressing need to convert libbonobo?
>>>>
>>>> We do need libbonobo:i386 for ia32-libs to keep 32bit software,
>>>> esspecially 3rd party 32bit apps, working. 
>>>
>>> I don't see libbonobo in ia32-libs_20120102, nor is liborbit converted
>>> to ma [1].
>> 
>> Sorry, should have been specific and said ia32-libs-gtk.
>
> Okay.
>
>> And yes, orbit2 also isn't converted and you could set your bug as being
>> blocked by it.
>> 
>> But the orbit2 includes a patch so you can build a local multiarch
>> orbit2 package and test a multiarch libbonobo. That way they can go in
>> close together. No need to wait with fixing libbonobo till orbit2 is
>> fixed in the archive.
>
> I no longer use any package which uses liborbit, which makes testing a
> bit harder. This is also why I'm a bit reluctant to change the package
> at this point of the release cycle.
>
> If there is a realistic chance that ia32-lib(-gtk) will be converted to
> multiarch, then I'll try to have a look.
>
>> Ia32-libs and ia32-libs-gtk have to go multiarch in wheezy or be
>> removed. So there is a high pressure to fix the remaining 10 out of ~200
>
> Are there tracking bugs for ia32-libs and ia32-libs-gtk (or a
> corresponding usertag)?

http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=677741
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=677762

Note that there are a number of NMUs in the delayed queue so the list
looks longer than it actualy is.

>> packages that block this. Orbit2 will probably be NMUed this weekend
>> given how long the patch has been sitting in the BTS.
>
> I guess that includes libidl0, a rdep of liborbit2?

http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=641614

Note that libidl0 does not need to be multiarch for liborbit2 to be
multiarch unless I'm mistaken. Libbonobo only has this requirement
because it provides a plugin for orbit2 and orbit2 has to learn to
search for plugins in the multiarch dir before libbonobo can move its
plugin there.

MfG
        Goswin





More information about the pkg-gnome-maintainers mailing list