[pulseaudio-discuss] [PATCH v3 18/24] echo-cancel: Improve webrtc canceller error handling a bit

Arun Raghavan arun at accosted.net
Wed Feb 17 12:52:53 UTC 2016


On Wed, 2016-02-17 at 13:12 +0200, Tanu Kaskinen wrote:
> On Wed, 2016-02-17 at 16:23 +0530, Arun Raghavan wrote:
> > On Thu, 2016-02-11 at 12:17 +0200, Tanu Kaskinen wrote:
> > > On Wed, 2016-02-10 at 09:56 +0530, Arun Raghavan wrote:
> > > > This particular case is extremely low probability, so I'd like
> > > > to
> > > > not
> > > > change it if possible. I don't think the extra code to add a
> > > > noop
> > > > mode
> > > > is justified since we're very unlikely to hit this.
> > > > 
> > > > Maybe we could/should change the run/play/record vmethods to
> > > > return
> > > > an
> > > > error code, but I'd prefer to do that separately if it's done.
> > > 
> > > So what are the extremely rare conditions where this error is
> > > triggered? I'll assume that this error doesn't indicate a bug in
> > > our
> > > code or in the webrtc library, and can really happen, so an
> > > assertion
> > > is not appropriate here.
> > 
> > The only failure case that I see currently is where we send a
> > different
> > block size fro mwhat we promised.
> 
> That should never happen, right? So an assertion is appropriate after
> all.

I picked the log message over the assert since it's based on external
code, so we don't want to crash if the webrtc code changes. That said,
it's enough of corner case for this to not be a big deal, so I'll just
make it an assert so we can move on.

-- Arun


More information about the pulseaudio-discuss mailing list