danil,
I have "special" access to the try-bot, so if you like what you see in
sergio patch, I will prepare a CL for you, and trigger a dry run, let me
know.
On Wed, Mar 29, 2017 at 12:40 PM, Sergio Garcia Murillo <
sergio.garcia.murillo@xxxxxxxxx> wrote:
I have ported my patch to the chrome release 57
It is untested yet, but here it is just for reference
https://github.com/agouaillard/perc-webrtc/commit/
358bff3c36981395292f3b841258d12b0e2d79da
Best regards
Sergio
El 29/3/2017 19:19, "Alexandre GOUAILLARD" <agouaillard@xxxxxxxxx>
escribió:
I think Sergio has a patch already. Let me confirm with him and if he
has, I ll make a cl
Do you have a bug number for us to reference ?
Sent from my iPhone
On 29 Mar 2017, at 12:07, Danil Chapovalov <danilchap@xxxxxxxxxx> wrote:
Actually, I need dynamic size one-byte header extensions too, so can
implement them in webrtc shortly myself.
On Wed, Mar 29, 2017 at 4:43 PM, Erik Språng <sprang@xxxxxxxxxx> wrote:
When you say variable length extension, you're still talking about
elements of the one-byte header extension right?
https://tools.ietf.org/html/rfc5285#section-4.2
But with len= set at runtime rather than fixed at compile time? That
should be doable.
In the initial sketches of RtpPacket and header extension parsing I made
a long time ago, I had an extremely flexible solution that introduced an
insane amount of code complexity, so I scaled it way do to what we actually
used and optimized for readable usage of the api instead. I believe Danil
based the current implementation RtpPacket on those drafts, so I'll take
some blame for that decision.
I think support for variable length one-byte header extension elements
would be good, just don't go overboard in flexibility like I did :)
E
On Wed, Mar 29, 2017 at 10:39 AM Danil Chapovalov <danilchap@xxxxxxxxxx>
wrote:
I would welcome support for dynamic size header extensions if they are
needed.
Lack of their support is more an overlook from my side than a conscious
choice.
On Tue, Mar 28, 2017 at 8:14 PM, Taylor Brandstetter <
deadbeef@xxxxxxxxxx> wrote:
Adding danilchap@ and sprang@, who may know more about that code
judging from git history.
On Tue, Mar 28, 2017 at 7:31 AM, Alexandre GOUAILLARD <
agouaillard@xxxxxxxxx> wrote:
hi guys,
we definitely miss you here, so many SFU guys around hacking chrome
code.
Today we almost finished the implementation of the frame marking
extension (chrome compilation pending), and the OHB extension is next. We
realised that the code in
modules/rtp_rtcp/source/rtp_header_extension(s).(h|cc)
only accept fixed length extension which will likely be a problem. Is it ok
if we go ahead and implement support for dynamic size header extensions?
regards,
alex,
--
Alex. Gouaillard, PhD, PhD, MBA
------------------------------------------------------------
------------------------
President - CoSMo Software Consulting, Singapore
------------------------------------------------------------
------------------------
sg.linkedin.com/agouaillard
-
