[00:01] Ok, just looking now. ... [00:01] more problems than jackd not starting [00:01] which desktop? [00:01] is pulse installed? [00:02] The calls using pactl are failing [00:02] " Connection refused" [00:03] what does .log/jack/jackdbus.log look like? I expect it says no access to the device [00:04] if pactl can't connect to pulse via dbus, it is a good guess jack can't either [00:05] Eickmeyer: ^^^ [00:06] that may be a logind/systemd difference in dbus instance [00:07] I believe it was Gnome, but I'll check with Plasma as well. [00:07] And, yes, Pulse is definitely installed. [00:10] if the gnome desktop on fedora is the same as on ubuntu, logout and in will not work. a reboot each time probably. [00:12] gnome never realy shuts down their logind session properly... or maybe opens a second one on top of the first. either way pulse ends up on another instance of dbus. also make sure that pulse is set up as a user process not system [00:12] * OvenWerks goes to spend time with wife [00:20] OK, maybe that's it. Will keep investigating. [00:21] Just a lot on my plate at the moment. [00:44] OvenWerks: It has the same problem in Plasma. It seems as though it's shutting-down Pulse, but then Pulse isn't coming back for whatever reason. [00:46] Bingo: W: [pulseaudio] module-jack-sink.c: JACK error >Cannot use real-time scheduling (RR/15) (1: Operation not permitted)< [01:14] Eickmeyer: thats pretty much normal and should not stop things from working [01:21] Eickmeyer: do you have autospawn = yes in client.conf in /etc/pulse/? [01:22] because pulseaudio -k kills pulse and expects it to respawn [01:22] cadence turns that off) [01:31] Eickmeyer: something to try: [01:31] killall autojack [01:31] ps ax |grep autojack should show no autojack running [01:32] then use -controls to start jack again. [01:33] this should start another instance of autojack. If this works... we may have to go back to /etc/xdg/autostart/autojack.desktop [01:34] anyway, it seems dbus related [02:01] OvenWerks: Interesting... autospawn=no is set in Fedora by default. :S [02:01] Never had this issue before. [02:42] OvenWerks: That method did not work. I'm guessing we've got some new dbus stuff that's messing up our flow. Unfortunately, if it's in Fedora now, then that means it'll land in Ubuntu in the future. [15:10] Eickmeyer: it seems carla is to blame for things freezing up... but I am not sure why. I think there is also a problem with jack itself because once carla messes things up, restarting jack does not clear the problem. [15:11] so if carla freezes, nothing that reads or changes the graph works (qjackctl, jack_lsp, -controls, etc) [15:11] even after restarting jack. [15:13] so there is something left in memory... This looks like a problem with either jack itself or libjack. [15:16] that is aside from the problem with carla :) I actually think if the problem in jack/libjack was taken care of, the problem with carla would at least be less of a problem. [15:17] Eickmeyer: with regard to the different way pulse may be set up from one system to the next, I feel that is a small problem. [15:18] it means we go from: kill pulse and expect it to reset itself [15:19] to: kill pulse, check if it is running, start it if not. [15:21] OvenWerks: Yeah, it's more complicated than that, unfortunately. [15:23] have they also gone from user based to system based? [15:23] I'm not following. [15:24] pulse can be run as a system service so that if there is more than one user logged in at the same time they all have access to the audio device. [15:24] it does limit pulse's operation which is why it is normally not done. [15:25] it may also be that the respawn setting is done by systemd rather than pulse itself. [15:25] Hmmm... not sure. Seems to be a user service, but I'm not 100% sure. [15:26] you should find it either in /lib/systemd/ system or user [15:27] ls /lib/systemd/user/pulse* [15:30] I'm rebooting into Fedora to find out. That seems to have had to be a systemwide change proposal, and I saw nothing. [15:32] Fedora doesn't make changes like that without making it a big deal. [15:55] so far carla 2.1 has shown no problems.... [15:55] * OvenWerks notes that carla 2.2 uses juce and maybe uses juce libs to talk to jack [15:56] certainly juce is the big difference from 2.1 to 2.2 [15:56] Hmmmm..... [15:57] Perhaps I should package Cadence if only for catia, i.e. not making the cadence binary available. [15:58] I like that 2.2 handles jack restarts but I can't recommend carla as a replacement for patchage/qjackctl if we ship 2.2 [15:58] catia _should_ be out on it's own "soon". [15:59] but falktx is pretty busy these days [16:00] I think patchage has sprouted python3 in the mean time too. but it may not be in debian yet. (and I could be wrong) [16:01] if catia has the same trouble as carla... it does use the same patchbay code after all, all bets are off [16:01] Yeah, but you've got to remember that Carla is primarily a plugin host, so it makes sense for it to use juce. [16:05] New patchage is not yet in Debian. [16:05] Looks like it's being worked on. [16:10] with so many things using juce, finding bugs is worth while. [16:10] Also remember that Carla 2.2 is just an RC at this point, he's still working on it. [16:11] yup. [16:24] OvenWerks: in Fedora, pulseaudio is a user service. [16:24] good. [16:25] https://paste.centos.org/view/3b3522ac [16:25] That's the log. [16:26] maybe we can use systemd commands to stop start. [16:27] Also this: https://paste.centos.org/view/61979ad1 [16:27] I'm wondering if Jack is the issue. [16:30] Either way, that's two separate machines I've tried now with Fedora that are experiencing the exact same issue with studio-controls. [16:30] assume starting jack with -controls so max ports should be 2k or so [16:32] Jack::JackGraphManager::AssertPort(jack_port_id_t): Assertion `port_index < fPortMax' failed. [16:33] Is that the issue right there? [16:34] but I see other problems anyway. from the stuff you shared with me before, pactl is not able to communicate with pulse. [16:34] yes an assert is always a fail and dump. [16:35] but there are other things: [pulseaudio] module-alsa-card.c: Failed to find a working profile. [16:35] module alsacard should have already been unloaded [16:36] from your autojack logfile that did not happen because pactl got a permision denied when it tried to do things with pulse. It looks like a dbus permission denied [16:37] I have seen reports of similar things in the ardour forums [16:37] Hmmm.... [16:40] we may need to stop the systemd pulse and directly start our own, but I think others have found a solution without [16:40] I will do some research... but it may take a while [16:42] ok [16:57] in the mean time maybe remove carla 2.2 from backports [16:57] Ok, can do. [16:58] I am not sure what to do with groovy... will 2.2 see a fix and release? [16:59] I can't do a bug from here for that, maybe later I can reboot to groovy and do it there. [17:00] Well, you can always report your findings upstream to https://github.com/falktx/carla [17:00] I suspect that it is not just unplugging USB devices but other graph changes as well [17:00] Because this issue isn't a packaging error, so it's completely beyond my control. Falktx doesn't watch launchpad for bugs. [17:01] And while launchpad can watch github issues, github cannot watch launchpad bugs. [17:01] So I'd rather any bug reporting go straight to falktx's github. [17:01] 314 issues :P [17:02] Yep, but you've got to remember that a lot of those are support questions and feature requests, not actual bugs. [17:02] Github issues don't differentiate that like Launchpad and Bugzilla do. [17:03] It's up to the repo owner on github to tag issues as actual bugs, feature requests, support questions, or discussions. [17:07] why can't github put dates on things? (26days ago...) a proper time stamp date time would be better. [17:07] Hover over it and it will give you the date. [17:10] hmm, there is a commit 12 hours ago, "Alternative/better way to find all audio ports in juce plugins" [17:11] That might be the fix, but we won't know for a while. [17:11] This actually might be a good discussion to have in #lad. [17:22] https://github.com/falkTX/Carla/issues/1237 [17:22] I will now clone master and build to see if it works better [17:23] Ok. [17:47] Problem still exists [19:59] this is painful... I have to reboot every time I kick this bug. [21:39] OvenWerks: Yikes. [21:39] Well, I've deleted it from the PPA, so if you "sudo apt autoremove carla && sudo apt update && sudo apt clean && sudo apt install carla" it should downgrade you. [23:06] it is too bad it doesn't happen every time. [23:07] I managed to downgrade thankyou. [23:08] I have 2.1 installed and 2.2 built that I can run from the source directory. So I can compare [23:59] Eickmeyer: i need a list of packages you need updated. [23:59] never got a ping from you that I can see [23:59] unless the Internet fubar'd [23:59] teward: On telegram. [23:59] Only two so far, no updates at this time since I can handle most of those myself. [23:59] oh NOW it arrived [23:59] never arrived before heh [23:59] Hahaha