Fwd: how do we prepare for jack1 and jack2?

[Adrian Knoth]

On Wed, Dec 16, 2009 at 07:23:59PM +0000, Christophe Mutricy wrote:

> > Actually the plan is to replace jack1 with jack2 once we are happy
> > enough with the latter. Note that the two are fully API/ABI compatible,
> > so the other packages won't be really affected.
> >
> 
> Hmm that remind me of #532339
> I don't know jack at all. could one of you comment on this bug and patch?

I can, though I might be a bit biased, because I wrote the patch ;)


To quote from the bug report:

  Unlike jack1, jack2 expects the process callback function to return 0.


The return code of the callback function has never been checked in
jackd1, it was undefined. jackd2 now requires this function to return 0
on success.

And yes, this is a slightly difference between those two, but jackd1
could still jump the bandwaggon and also place such a requirement.

In other words, the application developers where left alone in the old
days when it came to return codes, and it's now just specified. The fix
is easy, and once fixed, the application runs on both, jackd1 and
jackd2.

Take fmit: cannot run on jackd2 at the moment, because it uses some old,
deprecated, even undocumented and long removed function. It still works
with jackd1, because there's the deprecated legacy code, but jackd2
doesn't include this ancient stuff.

So it's also a one line patch to fmit, then, it's ready for jackd2, it's
up to date with the API spec and switching between jackd1 and jackd2
will just work.


I'll send a bug report with a patch to fmit.


HTH

-- 
mail: adi at thur.de  	http://adi.thur.de	PGP/GPG: key via keyserver