Hi everyone,
It's been a few months since some of us met in Copenhagen, and since
then we've got a new version of PulseAudio out the door. So I think it's
time to look back and see if there's any progress on any of these
issues, and if not, how we can refuel and get started again :-)
Based on Arun's PulseConf notes [1]:
=== Patch review capacity ===
* So far, I haven't seen one brought up yet. It looks like we're doing
well right now with Tanu doing the bulk of the review work, but it
wouldn't hurt if we could share the load a bit better. Colin, as you
agreed to setting up the first one, what do you think?
=== Routing infrastructure ===
* It looks like there was some progress, but then work seems to have
stalled with the last commit being three months ago [2]. Janos/Jaska,
how are things going right now, is there something digestible that we
can merge into the main tree at this point?
=== Low latency ===
* I feel work has been a bit sporadic here.
- I once tried to track down a 10 ms+ delay in the kernel which I
think was in the i915 driver, and asked for help on LKML but didn't get
much. Maybe just doing another try here would be the right thing?
- Arun worked a little on changing the buffering refill values to
better suit low latency scenarios, but I don't think anything was
sufficiently tested and merged. What is your experience?
- I suggested to test the maxlength parameter better, and I should
probably do that, but haven't got to it yet. Also, I think this is
somewhat related to the buffer parameters Arun was working with, so
maybe wait until that is merged?
=== Improving (laptop) surround sound ===
* No progress, and nobody who signed up for anything AFAIK?
=== Unified cross-distribution ALSA configuration ===
* No progress, and nobody who signed up for anything AFAIK?
=== Base volumes ===
* No progress, and nobody who signed up for anything AFAIK?
=== Devices with dynamic capabilities (HDMI) ===
* I've written some patches that read the HDMI ELD info and that
recently have been merged. It also required some kernel patches which
will be merged for kernel 3.9 (to support hotplug better). This would at
least enable UIs to show the monitor name, which is nice. I'm hoping
that Arun will write something that would take the ELD formats and turn
into PulseAudio's native format.
The question is whether or when to proceed with the dynamic card
thing. I somewhat fear regressions since Pierre's email [3]. The
question is if the kernel is ready to give us correct information. One
strategy could be to wait for GNOME et al to see if we get many bugs in
terms of "When I plug in my fancy monitor, GNOME says its name is
'!%&#&&#!#'", and perhaps proceed only if there are (almost) no such
bugs. What do you think?
=== Splitting of configuration ===
* I think we're currently waiting for Tanu to do some work on this
issue and come up with a more concrete proposal, is that correct?
=== Better drain/underrun reporting ===
* No work so far; and I can reiterate that it feels like it's quite
far down on the priority list. Do you agree or should I try to start
working on this (compared to other stuff)?
=== User experience ===
* No progress AFAIK. Did anybody sign up for anything in this area?
--
David Henningsson, Canonical Ltd.
https://launchpad.net/~diwic
[1] http://arunraghavan.net/2012/11/pulseconf-2012-report/
[2] https://github.com/otcshare/pulseaudio/tree/routing
[3]
http://lists.freedesktop.org/archives/pulseaudio-discuss/2013-January/015943.html
