Migrate to farsight 0.2* / gstreamer 1.0

Bug #1295207 reported by Bryan Quigley
22
This bug affects 4 people
Affects Status Importance Assigned to Milestone
Pidgin
Unknown
Unknown
pidgin (Debian)
Fix Released
Unknown
pidgin (Fedora)
Fix Released
Undecided
pidgin (Ubuntu)
Fix Released
Undecided
Unassigned

Bug Description

1) Ubuntu 14.04
2) 1:2.10.9-0ubuntu2
3) I expect it to depend against the latest farsight/gstreamer 1.0
4) Instead, it's one of the reasons we still have gstreamer0.10 on the trusty cd, through libpurple.

It seems upstream already supports building against farstream-0.2 and gstreamer1.0, see:
https://developer.pidgin.im/wiki/Dependencies

Tags: 14.04 14.10
Revision history for this message
In , Matthias (matthias-redhat-bugs) wrote :

We would like to complete the gstreamer 1.0 transition and get rid of gstreamer 0.10 from the live cd entirely. libpurple is one of the remaining obstacles to reaching this goal.

https://developer.pidgin.im/ticket/15386 has a patch to port pidgin to 1.0

Revision history for this message
In , Matthias (matthias-redhat-bugs) wrote :

Ping, action on this would be greatly appreciated. We'd like to stop shipping two gstreamer stacks on the live image where space is limited.

Revision history for this message
In , Jan (jan-redhat-bugs) wrote :

I don't see any gstreamer-1.0 built anywhere. Also, the upstream patch is completely untested and the actual committed patch is totally different.

Revision history for this message
In , Jan (jan-redhat-bugs) wrote :

Created attachment 794596
[Patch] backport of gstreamer-1.0 patch

Backport of https://developer.pidgin.im/ticket/15386.

I have *not* tested the patch. Also, there is no gstreamer-1.0 built in rawhide so I didn't check if it actually builds correctly with gstreamer-1.0. It seems to build ok with gstreamer-0.10.

Revision history for this message
In , Peter (peter-redhat-bugs) wrote :

(In reply to Jan Synacek from comment #2)
> I don't see any gstreamer-1.0 built anywhere. Also, the upstream patch is
> completely untested and the actual committed patch is totally different.

gstreamer1-devel and friends. basically you just need to change the build requires by 's/gstreamer/gstreamer1/g'.

Revision history for this message
In , Kalev (kalev-redhat-bugs) wrote :

I looked at the openSUSE piding package a few weeks ago and they have a gst-1.0 patch there. Might be worth using the same patch to be able the share the maintenance efforts.

https://build.opensuse.org/package/show/openSUSE:Factory/pidgin

Revision history for this message
In , Jan (jan-redhat-bugs) wrote :

(In reply to Kalev Lember from comment #5)
> https://build.opensuse.org/package/show/openSUSE:Factory/pidgin

Good to know, thanks.

So, I took the patch, appended additional configure.ac diff that checks for farstream-0.2 and attempted to build the package.

I'm getting this:

Making all in example
make[4]: Entering directory `/home/jsynacek/pidgin-gstreamer1/pidgin/pidgin-2.10.7/libpurple/example'
  CC nullclient.o
  CCLD nullclient
../../libpurple/.libs/libpurple.so: undefined reference to `gst_video_overlay_set_window_handle'
../../libpurple/.libs/libpurple.so: undefined reference to `gst_video_overlay_get_type'
collect2: error: ld returned 1 exit status
make[4]: *** [nullclient] Error 1

I don't get it. The library is now correctly linked against gstreamer-1.

# ldd pidgin-2.10.7/libpurple/.libs/libpurple.so
 linux-vdso.so.1 => (0x00007fffd532b000)
 libdbus-glib-1.so.2 => /lib64/libdbus-glib-1.so.2 (0x00007f0eb8e9a000)
 libdbus-1.so.3 => /lib64/libdbus-1.so.3 (0x00007f0eb8c54000)
 libgmodule-2.0.so.0 => /lib64/libgmodule-2.0.so.0 (0x00007f0eb8a4f000)
 libgthread-2.0.so.0 => /lib64/libgthread-2.0.so.0 (0x00007f0eb884d000)
 libxml2.so.2 => /lib64/libxml2.so.2 (0x00007f0eb84e4000)
 libfarstream-0.2.so.2 => /lib64/libfarstream-0.2.so.2 (0x00007f0eb82ce000)
 libgstbase-1.0.so.0 => /lib64/libgstbase-1.0.so.0 (0x00007f0eb8075000)
 libgstreamer-1.0.so.0 => /lib64/libgstreamer-1.0.so.0 (0x00007f0eb7d70000)
 libgobject-2.0.so.0 => /lib64/libgobject-2.0.so.0 (0x00007f0eb7b1d000)
 libglib-2.0.so.0 => /lib64/libglib-2.0.so.0 (0x00007f0eb77f1000)
 libidn.so.11 => /lib64/libidn.so.11 (0x00007f0eb75be000)
 libutil.so.1 => /lib64/libutil.so.1 (0x00007f0eb73ba000)
 libpthread.so.0 => /lib64/libpthread.so.0 (0x00007f0eb719c000)
 libm.so.6 => /lib64/libm.so.6 (0x00007f0eb6e95000)
 libnsl.so.1 => /lib64/libnsl.so.1 (0x00007f0eb6c7a000)
 libresolv.so.2 => /lib64/libresolv.so.2 (0x00007f0eb6a5f000)
 libc.so.6 => /lib64/libc.so.6 (0x00007f0eb6695000)
 libgio-2.0.so.0 => /lib64/libgio-2.0.so.0 (0x00007f0eb6324000)
 librt.so.1 => /lib64/librt.so.1 (0x00007f0eb611c000)
 libdl.so.2 => /lib64/libdl.so.2 (0x00007f0eb5f18000)
 libz.so.1 => /lib64/libz.so.1 (0x00007f0eb5d01000)
 liblzma.so.5 => /lib64/liblzma.so.5 (0x00007f0eb5adc000)
 libffi.so.6 => /lib64/libffi.so.6 (0x00007f0eb58d3000)
 /lib64/ld-linux-x86-64.so.2 (0x00000035d6200000)
 libselinux.so.1 => /lib64/libselinux.so.1 (0x00007f0eb56af000)
 libpcre.so.1 => /lib64/libpcre.so.1 (0x00007f0eb5448000)

The only thing that comes to mind is that those symbols are simply missing in libgstreamer-1.0.so.

I spent the whole day trying to get it to compile. I *didn't* know about the farstream02 and I think that it's a huge mess. It's not enough to link against gstreamer-1 only, because farstream-0.1 is still linked against gstreamer-0.1.
So, there is gstreamer1, farstream02, gstreamer1-plugins-.... It's a mess and it's frustrating.

Revision history for this message
In , Jan (jan-redhat-bugs) wrote :

Created attachment 795901
[Patch] gstreamer-1.0 patch

Revision history for this message
In , Jan (jan-redhat-bugs) wrote :

Created attachment 795902
[Patch] spec

This is how the spec is patched, I'm not sure if I miss anything else.

Revision history for this message
In , Fedora (fedora-redhat-bugs) wrote :

This bug appears to have been reported against 'rawhide' during the Fedora 20 development cycle.
Changing version to '20'.

More information and reason for this action is here:
https://fedoraproject.org/wiki/BugZappers/HouseKeeping/Fedora20

Revision history for this message
In , Kalev (kalev-redhat-bugs) wrote :

(In reply to Jan Synacek from comment #6)
> So, I took the patch, appended additional configure.ac diff that checks for
> farstream-0.2 and attempted to build the package.

It would appear that the openSUSE package is built with the the voice and video support disabled (--disable-vv) and this is why they can get away without the farstream-0.2 fix.

> ../../libpurple/.libs/libpurple.so: undefined reference to
> `gst_video_overlay_set_window_handle'
> ../../libpurple/.libs/libpurple.so: undefined reference to
> `gst_video_overlay_get_type'
> collect2: error: ld returned 1 exit status
> make[4]: *** [nullclient] Error 1
>
> I don't get it. The library is now correctly linked against gstreamer-1.

These symbols are from the libgstreamer-video-1.0 library. The easiest way get it to link with it would be to change
PKG_CHECK_MODULES(GSTREAMER, [gstreamer-1.0] ...
to
PKG_CHECK_MODULES(GSTREAMER, [gstreamer-1.0 gstreamer-video-1.0] ...

Revision history for this message
In , Fedora (fedora-redhat-bugs) wrote :

This package has changed ownership in the Fedora Package Database. Reassigning to the new owner of this component.

Revision history for this message
Launchpad Janitor (janitor) wrote :

Status changed to 'Confirmed' because the bug affects multiple users.

Changed in pidgin (Ubuntu):
status: New → Confirmed
Revision history for this message
Pander (pander) wrote :
tags: added: 14.04 14.10
Revision history for this message
vgeloven (vgeloven) wrote :
description: updated
Revision history for this message
Bryan Quigley (bryanquigley) wrote :

The Pidgin version in upstream git (seems to be for version 3.x), can be built --with-gstreamer=1.0

The released (2.10.9) version always wants gstremer0.10. (I'm not sure if there are other changes between them that makes gstreamer1.0 work better..)

If you grep configure.ac and yours has "gst10_pkg" that indicates you're on the gstreamer1.0 compatible one.

Revision history for this message
In , Kevin (kevin-redhat-bugs) wrote :

Ping? More than one year later, libpurple still pulls in GStreamer 0.10. This now also affects the KDE spin (which has moved to GStreamer 1 for Fedora 21), through @kde-telepathy → telepathy-haze → libpurple → gstreamer and through … → libpurple → farstream → gstreamer (so yes, it also needs to use the new farstream). In addition, this dependency chain can result in mixed GStreamer 0.10 and 1.0 in kde-telepathy's ktp-call-ui, causing crashes.

Revision history for this message
In , Kevin (kevin-redhat-bugs) wrote :

One month later, still waiting for an answer.

Shall we start the nonresponsive maintainer process?

Revision history for this message
In , Jan (jan-redhat-bugs) wrote :

(In reply to Kevin Kofler from comment #14)
> Shall we start the nonresponsive maintainer process?

Am I nonresponsive? You should check the changelogs, really...

I still keep pidgin alive and running, but don't have time to fight the build system anymore, sorry.

Revision history for this message
In , Kevin (kevin-redhat-bugs) wrote :

The issue is that it took 1 month to get this answer, and we missed the F21 Final freeze because of that. :-(

I can help out (I'm a provenpackager), but it's probably too late to get GStreamer 0.10 off the F21 Final live images.

Revision history for this message
In , Peter (peter-redhat-bugs) wrote :

What is the status on this? This bug has been open coming up 2 years :-(

Revision history for this message
In , Jan (jan-redhat-bugs) wrote :

To make comment 15 even clearer: I took the package to keep it alive and update it to the latest versions, and perhaps even to fix urgent bugs that keep it from functioning, which I still do. I did it because the original maintainer hadn't been updating it for some time. I spent some time on this bug, found out that it really wasn't easy and then moved on to more prioritized things. This bug is in no way critical and trivial to get right and I don't have time to solve it anymore. It's not even ASSIGNED.

So the status is: this bug needs to be fixed by somebody willing to spend their time to fight the build system and different gstreamer versions.

Revision history for this message
In , David (david-redhat-bugs) wrote :

This patch should fix it. I've tested this with pidgin-sipe, with voice calls as well as the file transfer and screen sharing added in later patches.

https://pidgin.im/pipermail/devel/2015-March/023645.html

Revision history for this message
Unit 193 (unit193) wrote :

Fedora's report shows that they are now using http://pkgs.fedoraproject.org/cgit/pidgin.git/commit/?id=b6ad37e47b91773314637cdd2b3f25a88f5ef70b (https://pidgin.im/pipermail/devel/2015-March/023645.html) and I've partially merged Debian's Pidgin and applied it as a quilt patch with minor refreshing. It all built and runs fine.

Changed in pidgin (Debian):
status: Unknown → Confirmed
Revision history for this message
Bryan Quigley (bryanquigley) wrote :

Just to be clear, it made it into the repo for the 2.x releases: https://hg.pidgin.im/pidgin/main/rev/2415067473ba

Is this (and related dropping gstreamer) something to do a FFE for?

Revision history for this message
Launchpad Janitor (janitor) wrote :

This bug was fixed in the package pidgin - 1:2.10.11-0ubuntu2

---------------
pidgin (1:2.10.11-0ubuntu2) wily; urgency=medium

  * debian/control:
    - Build-depend on libgstreamer1.0-dev, libgstreamer-plugins-base1.0-dev,
      libfarstream-0.2
    - Recommend gstreamer1.0-plugins-base, gstreamer1.0-plugins-good
  * debian/patches/gstreamer1.patch:
    - Use gstreamer 1.0 (LP: #1295207)

 -- Robert Ancell <email address hidden> Fri, 29 May 2015 11:28:51 +1200

Changed in pidgin (Ubuntu):
status: Confirmed → Fix Released
Changed in pidgin (Debian):
status: Confirmed → Fix Released
Changed in pidgin (Fedora):
importance: Unknown → Undecided
status: Unknown → Fix Released
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.