Phoronix News Bot
GStreamer 1.0 Is Looking To Finally Be Released Soon
Phoronix: GStreamer 1.0 Is Looking To Finally Be Released Soon
Keynoting the GStreamer 2012 Conference in San Diego was Wim Taymans of Collabora. Taymans was talking about GStreamer 1.0, which should be officially released very soon -- perhaps before GNOME 3.6 ships...
http://www.phoronix.com/vr.php?view=MTE2OTQ
Phoronix Member
I don't see why the new gnome can't simply depend on at least gstreamer-0.11. Distributions can push the stable version, once it is released. Why rushing out a stable version, which isn't stable? The 1.0 stamp doesn't fix all the remaining bugs magically...
Senior Member
And what difference would it make? Either it's a pre-release 0.11 version or a not-so-stable 1.0 version. GNOME 3.6 would depend on it anyway including all the bugs the users would encounter. It's not like not calling it 1.0 would make things better.Originally Posted by oleid
The real question is: Why did the GNOME devs once again rush to unstable dependencies?
Phoronix Member
The difference is, that this makes version numbers meaningless (again), as seen e.g. with KDE 4.0. One should be able to trust, that a 1.0 is (more) stable (than a 0.11). Also calling it 0.11 might encourage distributions to backport a stable 1.0 once it is released.
The things these bindings are needed for seem to be stable (or stable enough). Yet, your question is still a good one.
Senior Member
Hence why I think SemVer/GemVer is a good idea. After all, +1.0 might mean that it's stable, or it might mean that it's a new development cycle.
Senior Member
Great you bring KDE 4.0 up, because not even KDE 4.0 relied on unstable 3rd party dependencies! KDE 4.0 was built on Qt 4.3 and not Qt 4.0 or even 4.0beta2 or whatever.
Are you kidding me? Of course lots of people noticed when beta-quality dbus was introduced into GNOME, just as everybody noticed when Telepathy was added which couldn't even send files (= not feature-complete and therefore alpha quality), and the GNOME devs thought that replacing esd with PulseAudio was a great idea…There was no rush. We made a concious decision to speed up the development of GStreamer 1.0 by porting applications to it. A library cannot become 1.0 without that happening. With applications being ported, you find problems in the API and because there are programs using the new library, people can discover bugs in them. Meaning: it becomes more stable.
When the decision was made, the schedule of Gstreamer 1.0 was supposed to match with GNOME 3.6. Obviously there is risk involved with such a large library. Something that is actually considered.
GNOME has been doing this with loads of libraries for the last 10 years or so, but as that all went ok I guess it was not noticed
These are just the three top examples that I could come up with without even thinking.
This once again shows the beauty of KDE’s Phonon. Want to test GStreamer 1.0? Great, just fork the existing GStreamer 0.10 wrapper, modify and compile it against GStreamer 1.0 and all Phonon applications use GStreamer 1.0.
Why didn't you time GStreamer 1.0 to be ready for beta1 releases of GNOME? Then you'd have the entire beta period of GNOME to also fix GStreamer bugs and have a perfect point release ready for GNOME 3.x.0.
I wouldn't even bitch at all if the Gnomes didn't always shout how enterprise-ready they are and how unstable everything else supposedly is (usually pointing fingers to KDE).
Phoronix Member
Lots of people? Citation needed.
All your other arguments are a bit.. strange.
E.g. not having a particular feature = alpha quality? That is not the definition of alpha quality.
PulseAudio bit is laughable as well. We made that optional for loads of various releases.
About dbus you don't really say anything.
Suggest to bring better examples.These are just the three top examples that I could come up with without even thinking.
You seem to lack experience on how software is released.This once again shows the beauty of KDE’s Phonon. Want to test GStreamer 1.0? Great, just fork the existing GStreamer 0.10 wrapper, modify and compile it against GStreamer 1.0 and all Phonon applications use GStreamer 1.0.
Why didn't you time GStreamer 1.0 to be ready for beta1 releases of GNOME? Then you'd have the entire beta period of GNOME to also fix GStreamer bugs and have a perfect point release ready for GNOME 3.x.0.
I wouldn't even bitch at all if the Gnomes didn't always shout how enterprise-ready they are and how unstable everything else supposedly is (usually pointing fingers to KDE).
Phoronix Member
There was no rush. We made a concious decision to speed up the development of GStreamer 1.0 by porting applications to it. A library cannot become 1.0 without that happening. With applications being ported, you find problems in the API and because there are programs using the new library, people can discover bugs in them. Meaning: it becomes more stable.
When the decision was made, the schedule of Gstreamer 1.0 was supposed to match with GNOME 3.6. Obviously there is risk involved with such a large library. Something that is actually considered.
GNOME has been doing this with loads of libraries for the last 10 years or so, but as that all went ok I guess it was not noticed
Posting Permissions
Reply With Quote