| OvenWerks | Eickmeyer: Carla 2.2 rc2 seems to have fixed the segfault problem with zita-ajbridge | 15:33 |
|---|---|---|
| Eickmeyer[m] | OvenWerks: Good to know. Should I yeet it into backports? | 16:13 |
| OvenWerks | Eickmeyer[m]: scratch that | 16:19 |
| Eickmeyer[m] | Uhhhh.... | 16:19 |
| OvenWerks | it worked fine for 4 or 5 times | 16:19 |
| Eickmeyer[m] | Yikes. | 16:19 |
| OvenWerks | soon as I told falktx it seems to be working I get the same problem. | 16:19 |
| Eickmeyer[m] | Welp... | 16:20 |
| OvenWerks | It seems that with the system having been running for a few days it worked fine for a few tries. but when I rebooted it it fails first time every time. | 16:21 |
| Eickmeyer[m] | That is odd. | 16:21 |
| Eickmeyer[m] | I mean, why would a reboot affect it? | 16:22 |
| OvenWerks | general system activity higher? I don't know. I am running htop and it doesn't seem to be high. | 16:24 |
| OvenWerks | the problem does show up within 5 tries even after system has been up a while | 16:25 |
| Eickmeyer[m] | How spaced-out are these tries? Like, what's the interval? | 16:25 |
| OvenWerks | close. within 2 or 3 minutes | 16:26 |
| OvenWerks | I am not using controls to start/stop zita-ajbridge | 16:26 |
| Eickmeyer[m] | Ok, so not like spamming it. | 16:26 |
| OvenWerks | I am doing it manually, so two terminals one for a2j and one for j2a | 16:26 |
| OvenWerks | So plug in, start one then start 2. check that jack_lsp shows ports | 16:27 |
| OvenWerks | (or Carla shows ports) | 16:27 |
| OvenWerks | unplug, make input terminal focus, ctrl-c make output terminal focus, ctrl-c | 16:28 |
| Eickmeyer[m] | Is it just having an error monitoring dbus, you think? | 16:30 |
| OvenWerks | I can do this as many times as I like with Carla closed. zita-ajbridge stops normally. but with Carla open, zita-ajbridge segfaults on ctrl-c (sigint) jack_lsp fails to finish (hangs after showing ports) | 16:30 |
| OvenWerks | dbus should have nothing to do with this | 16:30 |
| Eickmeyer[m] | Ok, just spitballing. | 16:30 |
| OvenWerks | this is all jack api stuff | 16:31 |
| Eickmeyer[m] | Maybe it's because you have two items looking at the connections at once (jack_lsp and Carla)? | 16:31 |
| Eickmeyer[m] | Basically, making the same jack api calls? | 16:32 |
| OvenWerks | That would be a jack bug, but carla 2.1 does not have this problem, I can use qjackctl and jack_lsp | 16:32 |
| OvenWerks | I get the same problem even if I don't run jack_lsp and just try to move one of carlas jack port items | 16:33 |
| OvenWerks | Eickmeyer[m]: it should be ok to have 5 clients making those api calls | 16:36 |
| Eickmeyer[m] | Yeah, good point. | 16:36 |
| OvenWerks | Alsa_in has high cpu on device unplug, Alsa_out just quits with "Write error: No such device | 16:39 |
| OvenWerks | So Alsa_in is perhaps broken too. | 16:40 |
| Eickmeyer[m] | Wouldn't surprise me. | 16:55 |
| * Eickmeyer[m] uploaded an image: 4e8x5b.jpg (89KiB) < https://matrix.org/_matrix/media/r0/download/matrix.org/BSgyyjLltEkgWWqTifvLogyc/4e8x5b.jpg > | 16:56 | |
| Eickmeyer[m] | Just a little fun. :) | 16:56 |
| OvenWerks | Eickmeyer[m]: there was a while where it was pretty much for some things. It is unfortunate that people just added the whole thing rather than cherry picking. It is also unfortunate that ladish was auto started by all the kx tools | 17:06 |
| Eickmeyer[m] | Yep, indeed. Now ladish is a thing of the past. | 17:07 |
| OvenWerks | Studio was in a rather bad state just before you stepped in | 17:07 |
| Eickmeyer[m] | Yeah, I know, which is part of why I try to steer people away from 18.04. | 17:07 |
| * OvenWerks doesn't have a choice on that one... last thing to support his laptop | 17:08 | |
| Eickmeyer[m] | Yeah, unfortunately. | 17:10 |
| OvenWerks | That laptop will probably move to debian next | 17:14 |
| OvenWerks | It will be sort of Studio-ized with ardour at least and a good kernel. | 17:17 |
| OvenWerks | Eickmeyer[m]: the problem with Carla 2.2 is also present in falktx' working Catia split as well but not in catia that comes with latest cadence. | 20:20 |
| OvenWerks | falktx and I have been able to isolate this to one portion of code but because Fons is well known to use only jackd1 (and is even possibly working on a "Jackd3" extended from jackd1, not jackd2, it could well be a jackd1/2 difference | 20:23 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!
