[17:53] OvenWerks: studio-controls crashes in the live session if it has no config file. I'll throw a bug report as soon as I can.' [17:54] Eickmeyer: can anything write to the home directory when running this way? [17:54] Yes, we should be able to. [17:55] Granted, it gets erased on reboot (goes to ram). [17:55] ram is fine [17:56] I'm filing a bug report because we're beyond beta freeze. [17:56] Beta is tomorrow. [17:56] on the iso does .config exist? I am guessing ./config/autojack does not but it should be created in convert [17:57] ~/.config does exist for any and all users. [17:57] mkdir -p (or equivalent) is good for this use case. [17:57] I am wondering if running convert on startup before autojack runs would help. [17:58] You mean, put it in the autojack-start I made? [18:02] yes, if it runs before anything else, it checks to make sure ~/.config/autojack exists and creats it if not. [18:02] then it writes a config file [18:04] That would work. Bug is filed, bug 1944607 [18:04] Bug 1944607 in studio-controls (Ubuntu) "Studio Controls crashes if run in live session" [High, New] https://launchpad.net/bugs/1944607 [18:11] Eickmeyer: pushed, do you want to look at autojack-start to make sure it makes sense? [18:17] I'll take a look at it in a sec. [18:18] release as 2.2.3 if it looks right [18:21] OvenWerks: It didn't work. [18:22] can you run convert from commandline? [18:22] Yep. Let me get you the output. [18:24] $ convert-studio-controls... (full message at https://libera.ems.host/_matrix/media/r0/download/libera.chat/a7ed038ab80a8c516ce2efeac4b11af3d33e3af1) [18:24] OvenWerks: ^ [18:25] (Matrix makes for a good pastebin) [18:25] I am just looking at the autojack code and it looks like we log stdout but not stderr [18:26] Interesting. [18:26] this has been why pastes people have left in issues have not shown the error [18:27] That makes sense. [18:33] This is tricky. The obvious fix is to test for 'none' at that line... but there should never be a device called none and to get to that line there would have to be [18:34] Eickmeyer: if there is a USB device plugged in at boot and no config file, that USB device will become jack master by default. [18:34] is that good or bad? [18:35] Shouldn't it be internal audio as master by default? [18:35] I mean, unless that's too complicated to do. [18:39] Eickmeyer: sorry, it only does that if it can't find anythng else to use as default [18:39] So that is fine [18:39] Yes, that's 100% OK. [18:40] so if someone has their on board audio disabled in bios... [18:40] Imagine... [18:40] But it looks like we add the usbdev even if it is 'none' [18:41] I actually used to disable the internal audio... I only turned it back on so I could test things. [18:41] Well, I mean, if that's what it takes and doesn't fundamentally break anything. [18:43] Eickmeyer: I have: if def_config['USBDEV'] not in devicelist and def_config['USBDEV'] != 'none': [18:43] That *should* work. [18:44] I am wondering if that is effectively: if def_config['USBDEV'] not (in devicelist and def_config['USBDEV'] != 'none'): [18:45] changing to: if def_config['USBDEV'] != 'none' and def_config['USBDEV'] not in devicelist: [18:55] Eickmeyer: it works as is in 20.04 :P [18:55] * OvenWerks thinks python may have changed [18:56] I'm talking about the live session prior to any config being made. [18:56] that is what I am testing [19:01] That's very possible. We've got 3.9 in Impish. [19:04] Ok braces to check coming up :P [19:09] Sounds good. [19:10] uploaded can you try that one? [19:11] Eickmeyer: ^^ [19:13] OvenWerks: Working on it now. [19:18] OvenWerks: I directly ran it, no good. Same issue. [19:18] $ convert-studio-controls... (full message at https://libera.ems.host/_matrix/media/r0/download/libera.chat/73a602032dcd8d7ac51069ebaa4cf19033e16bd6) [19:21] ok. so something else is adding a 'none' device [19:28] Eickmeyer[m]: another try. [19:28] Ok [19:31] OvenWerks: Still no good.... (full message at https://libera.ems.host/_matrix/media/r0/download/libera.chat/cbbd2f5720c4fac429d420d8c5c08816d6a53bc6) [19:48] Eickmeyer[m]: something is happening I don't understand, I have added some debugging lines [19:49] Eickmeyer[m]: do you actually have a device called none,0,0? [19:49] OvenWerks: Looking. What command would show that? [19:49] (I'm running it in a VM) [19:50] aplay -l [19:50] **** List of PLAYBACK Hardware Devices **** [19:50] card 0: I82801AAICH [Intel 82801AA-ICH], device 0: Intel ICH [Intel 82801AA-ICH] [19:50] Subdevices: 1/1 [19:50] Subdevice #0: subdevice #0 [19:51] That should be ok [19:53] however, on the off chance there is such a thing, I guess I should check for that. But before that. I would like to see the output of what I just did [19:53] Ok. I'll look at your latest commit. [19:56] OvenWerks: Using what you did: [19:56] $ convert-studio-controls... (full message at https://libera.ems.host/_matrix/media/r0/download/libera.chat/a3ffccf88d14b3d6cec8a7314ad61d9642bd53ac) [19:59] it says it think you have a device named none... I need to check that more. another upload coming soon [20:00] Ok. [20:00] I've got this one as release blocking. The release team wasn't happy when I told them that. [20:06] Eickmeyer[m]: what does tyhis get? cat /proc/asound/card*/id [20:07] I82801AAICH [20:08] that should give I82801AAICH,0,0 as default device [20:09] Right. [20:18] Eickmeyer[m]: another one to test (bug should be gone but printout is still needed [20:25] Also a paste of: ls /proc/asound/I82801AAICH or /proc/asound/card0 [20:26] if either of those error, maybe: ls /proc/asound [20:29] OvenWerks: Ok. Sorry, had a phone call. [20:32] OvenWerks: $ ls /proc/asound/I82801AAICH [20:32] codec97#0 id intel8x0 pcm0c pcm0p pcm1c [20:32] card0 returns same [20:32] that is different [20:32] Is that an AC97 card then? [20:33] If so, alsa has changed the way they deal with them [20:33] Well, it's virtual. [20:34] even virtual that would be a change [20:34] It's basically a Virtualbox passthrough masquerading as an Intel chipset. [20:35] But that is ok. it should still be treated the same as a PCI audio device [20:38] Agreed. Pulseaudio is OK, so perhaps some changes kept up with Alsa there. The question then is, how do we solve this in studio-controls? [20:39] I need a new paste of my latest upload [20:41] Ok [20:41] I have fixed two things and need to know which one fixed your error [20:41] (or if both failed ... sigh) [20:42] I think it worked this time: [20:42] convert-studio-controls... (full message at https://libera.ems.host/_matrix/media/r0/download/libera.chat/4a27b7d6d68028fecaade375ebf0d2e848c9ac1c) [20:42] Trying one more time for good measure. [20:43] We're good. \o/ [20:43] $ convert-studio-controls... (full message at https://libera.ems.host/_matrix/media/r0/download/libera.chat/0634de3e1806233c8d8604cc4534a6a42865e633) [20:43] I have also fixed things for an audio card with the id 'none' [20:43] Sweet. [20:43] ok, I should remove some of the debug messages [20:44] Feel free to tag/release when ready, I'll get this uploaded and tell the release team we need it expedited. [20:53] Eickmeyer[m]: released [20:54] OvenWerks: On it. [20:57] Eickmeyer[m]: that would have effected anyone without a standard hda device (if those are still shown the same way in 21.10) [20:57] OvenWerks: Either way, it will prevent an embarrassing situation where one of our core components is utterly broken for people trying it. [20:58] it did not help that I can not reproduce on this machine [20:59] Yeah, I get that. This is why it's useful to have people like me around to do these smoke tests. [21:00] I had switched two operations around and only when there is an HDA device does it work. [21:01] It is what comes from adding complexity to existing code [21:01] anyway, I am likely not available for the rest of the day to more than answer [21:15] OvenWerks: Ok, sounds good. I think I've got it from here. [21:34] OvenWerks: Uploaded, once I get the migration email from proposed I'll be able to respin the iso. [21:40] Eickmeyer: I have heard some good talk about PW on Ardour lately. It seems latencies are much closer to expected. [21:40] Oh, that's great! [21:42] BTW, 21.10 kernel seems to actually load the alsa module for the fireworks chip again, though performance is still unacceptable [21:42] That's disappointing. [21:44] So far as I can tell, PW still only conects to jack if jack asks for device access, so using jack for ffado and pw for everything else doesn't work [21:46] wtay says the jack bridge can't be added manually but I would guess it can be just looking at some of the cli options for it's controller utility [21:47] Possibly. Sounds like he's got some feature completeness to work on.