Just thought I'd give a 'status' update on where the CVS version of gaim-vv is, and what I'd like to see the next few steps to be. Marv may well object, but I'll leave that up to him to follow up.
The difference between 1.2.0 and CVS are largely renaming 'webcam' routines to 'media'. The general idea will be that we won't have to actually use a webcam as a broadcast source, but any V4L, or for that matter GST aware, subsystem. Thus relying on 'webcam' names for routines would just be confusing. This work is probably 3/4 complete (so far, a basic search/replace, filerename and a few routine tweaks have been performed). We may still need to abstract this out further, but that'll take a bit more analysis of code.
Yahoo status: We've just plugged a leak in the yahoo cam decoder routine (technically, gst-j2k), so cam reception will work for longer. But, there's still a major problem (see below). Still no broadcast, or audio.
MSN status: Libmimic is complete, we're waiting for some time to be emailed to us to implement it.
Major bugs at the moment:
1) CPU driven to 100% utilisation on Y! receive.
See bug tracker.
Website needs tidying up, better contrast of colours and this post to be put in the 'news' (news should probably be spun off as a seperate page).
Personally, I'm still very sceptical that it's actually gaim-vv or the gst-plugin code causing the high CPU utilisation, and think it's gstreamer itself. Hopefully someone can track this down, and patch accordingly. The lack of thorough gstreamer docs does not help.
We're aware of various third party programs that can be launched to do -vv stuff, by why take that route when it'd just be dropped when the proper way of doing things is implemented. Please don't ask us about it, unless you intend to 'port' to the gaim-vv structure.
I think that's about it for now. For what it's worth, I do have some time on my hands over the next few days, however I'll be utilising this to fix up deeper Y! code within gaim, and tidy the website/info for gaim-vv up before I start looking at further code. Please don't hassle me about this.
Oh, yeah. Skype. It's a closed source protocol. Unless someone can reverse engineer it, the most we'll be able to do is to launch the skype program to contact a buddy (whose skype ID is stored with the buddy contact details). Feel free to submit patches.
SIP? Feel free to submit patches.
In both cases, patches will only be accepted against current CVS, and will have to conform and expand the 'media' abstraction concept.
ANother update... I'm importing gaim 1.4.0 as it has some nifty Y! improvements that I could use on a daily basis. Also as an experience points for myself on SF's cvs. Don't expect an immediate release.
You may have noticed my recent commits. Here's a brief rundown on what I've done.
(From the new file: ChangeLog.gaim-vv)
Gaim-vv: The friendly video/voice fork of Gaim
* New configure option of --disable-gst, with appropriate #ifdefs.
* Import gaim v1.4.0
* Created ChangeLog.gaim-vv to track gaim-vv changes
* Gaim-vv now occupies seperate space to gaim itself, aiding in side-by-side development
* Disable building by default those protocols which do not yet have voice or video capability, as it just slows down build time
* Disable perl/tcl/tk plugins. Not really needed for gaim-vv development, and just slows down build time
* Seperate man page entries
* Commenced appending details to README.gaim-vv on how to hack gaim-vv sections on to any 1.(>=2).x gaim build
Note no explicitly new -vv functionality is added, apart from the ability to turn off -vv features. I mainly commited all this as experience for myself, and for those who actually use gaim-vv and want yahoo profiles working again ;-) Note the 'new' manpage entries are just copies of the gaim man pages, and nothing specific to -vv. Disabling -vv functions via --disable-gst should help us potentially get onto the main tree, however I'm still unsure whether I want --disable-gst, or --disable-vv. I guess I'll just suck it and see.
I'm happy to take direct comments/reports etc. from gaim devs or cpw's, anyone else please use the proper forums.
Although I've previously stated that I wasn't going to change the API, I'm heavily leaning towards renaming a bunch of stuff so it's even less 'webcam' specific. It's messing with my head too much when I try to think of video sinks when objects refer to cams, and is probably just bad mojo doing things like that anyway.
That's about it for now. More news later.
Log in to post a comment.