Hi everyone,

It's been a while since last update, and things calmed down a bit. This is
due to yet another burst of real-life
activities (Hey, I haven't a boring life ;) ).
But we know, transcode is hard to kill, and things are still evolving behind
the scenes.

We're still stuck on the plan and I still want to put 1.1.0 final out as
soon as is possible. On this topic,
we have both good and bad news.

The bad news is that framebuffer and thread core are under rewriting. Again.
The good news is that should squash a large pile of still open bugs (hangup
and stuff).

New code is more regular, clean, has less quirks and has better
documentation.
I have only a major regret in doing this: the task should be started before,
alpha5 is a bit late.
Anyway, better late than never.

Anyone interested can see the various progress by downloading the
preliminary patches
from http://fromani.exit1.org/tc-1.1.0-alpha5-preX.patch.gz
where X is current revision.
Code will be merged when it becomes more stable AND it passes cleanly
regression tests
and stress tests on my box. In the while, brave users can fetch the patches
and apply against
1.1.0-alpha4. Every patch is against clean 1.1.0-alpha4 (they are NOT
incremental)

Those patches are HUGE beasts and they have some known problems. It's still
work in progress.
But they are supposed to be the only change between alpha4 and alpha5.

It's also worth to say that fixed those bugs and tested new sequential
decoder, the only
known bugs left are those reported on our mantis, so beta is definitely not
too far.

I'd also like to say that, after a bit of thinking, I've chosen to keep
1.0.x branch alive
(as Phil and others suggested) for some months more. Bigger issues will not
be
addressed (since that's is the scope 1.1.0 and beyond), but codebase will be
mantained,
and updated if needed to track libraries API changes (es: imagemagick,
libavcodec and more)
in order to provide a transition path smooth for users.

Future 1.0.5 will feature (a variant of) the patch currently included in
debian, so
ASM acceleration will not printed until requested.

That's all :)

-- 
Francesco Romani // Ikitt

Reply via email to