[PlanetCCRMA] Must I upgrade to build Ardour? (and more on jack and friends)

Mark Knecht markknecht@attbi.com
Sat Apr 5 11:09:01 2003

On Sat, 2003-04-05 at 10:37, Fernando Pablo Lopez-Lezcano wrote:
> I see. I got another message off the list from a Planet CCRMA user
> concerned about the upcoming jack change (and the stability of the
> overall planet). It is a tough call. 


> Ardour is not the only reason for the update. The newer jack supposedly
> handles clients that do not meet the hard deadlines much better. 

While I understand the attraction for some folks, it's of no interest to
me really. Either these tools are going to work right, or I'm not going
to use them. I don't have much interest in personally debugging
someone's app and under what situation it causes a glitch, nor do I want
a glitch in my audio recorded in Pro Tools.

> I see
> this problem of qjackconnect sometimes dying (no, not all the time) when
> clients exit, but to be fair I have seen this before with clients
> crashing and taking down not only qjackconnect but other clients as
> well, something which I have not (yet) seen on the new jack. 

I think this may be a bigger problem for me. Some soft synths, like
ZynAddSubFx, will not connect their audio outputs to anything (as for as
I know) without qjc. If qjc is seg-faulting during operation I will be
an unhappy camper if it brings down Zyn. (If it can kill itself without
hurting other processes and without causing glitches, I'll be

> Another option I'm considering is temporarily using the "test"
> planetedge respository to host this bunch of upgrades so that users can
> test them and report back their experiences. And depending on what
> happens I'll move the stuff to the normal planetccrma repository.
> Obviously I don't want to keep the split open for a long time. I also
> don't want to rebuild things twice (one for the old jack and another
> time for the new one) so that any changes in versions will only go to
> the "test" repository (for example I already have ams 1.5.8b built).
> Thoughts?

I think this is a good idea _when_ you have some folks interested and
willing to do the testing. Without that I think it would be sort of hard
to close the loop. I've been running your test kernel, but frankly only
because I hoped it had fixes for problems I was having with the standard
kernel. (Disk order, Broadcom Ethernet, Radeon 9000 at 1280 * 1024, RME
HDSP 9652 - and it did, other than the RME which Thomas will likely
commit in 1-2 weeks) Without a reason to do use the test kernel, I never
got there on my previous box. Will others? Maybe some of your machines
at Stanford are candidates?
> [also, I'm starting work on rebuilding for 9 so there's plenty to do...]

Really? I'm sure you're up to date on all the problems RH people are
seeing with 9 so far. Think I'll be waiting a while, or again, until I
have a pressing need.

That said, now you're supporting 7.2, 7.3, 8.0 and 9.0? That quite a
menu. Maybe it's time for Leland to buy you some big fat machines just
for compiling all this stuff! ;-)

Thanks much,