On Thu, 2010-01-21 at 14:55 +0100, Daniel Stenberg wrote:
> On Wed, 20 Jan 2010, Daniel Stenberg wrote:
>
> > Cool. I've edited it a bit more now, and I'm prepared to commit what I have
> > in a short while and then we should be able to continue smoothen out the
> > remaining quirks against CVS.
>
> Ok, I've just now done the intial commit. Please check out the minor changes I
> did and see that things still run fine for you. I removed some minor stuff and
> moved some code.
The changes all seem good to me, and the rtsp-test still works.

> I still want to rename CURLOPT_RTPFUNCTION and CURLOPT_RTPDATA to something
> more agnostic to make the names less specific for RTSP and more generic for an
> alternative data stream. I haven't yet figured out a name I like since I don't
> want it to include the word 'data' since we by tradition call the option for
> the customer callback pointer *DATA so it would make a strange double-data
> name. I'm thinking perhaps CURLOPT_INTERLEAVEFUNCTION, CURLOPT_STREAMFUNCTION
> or CURLOPT_EXTRAFUNCTION... Anyone with a particular opinion?
My vote is for CURLOPT_INTERLEAVEFUNCTION/INTERLEAVEDATA since that
captures the rationale for splitting it out into a separate function. If
we're going to make this function name more agnostic, then we should
probably have some sort of agnostic mechanism of setting the
CURLOPT_RTSPREQ_RECEIVE since other protocols will need similar
functionality.

However, I'm not aware of any other protocols that follow this model
since, frankly, it's not the best way to do things.

>
> Oh, and it would be great if you could convert your test code into a somewhat
> decent-looking example we can include in the package!
Definitely!

--Chris
-------------------------------------------------------------------
List admin: http://cool.haxx.se/list/listinfo/curl-library
Etiquette: http://curl.haxx.se/mail/etiquette.html
Received on 2010-01-21