| OvenWerks | arraybolt3[m]: I have copy pasted all that to a file in case some bits didn't get logged. | 01:23 |
|---|---|---|
| * OvenWerks has had family in town since thursday | 01:24 | |
| Eickmeyer[m] | arraybolt3: It didn't register in Matrix for whatever reason, but OvenWerks took everything you wrote and copied it into a text file locally for himself, and he's had relatives in town, so hence radio silence on his side. | 15:54 |
| Eickmeyer[m] | Looks like it was in the middle of a bridge restart. | 15:54 |
| arraybolt3[m] | Nice! Glad I was actually able to be helpful! | 15:55 |
| OvenWerks | arraybolt3[m]: the last time I looked into things was a while ago. PW development has been rapid and extensive both before and after. | 15:57 |
| * arraybolt3[m] decides to try getting PipeWire working in Arch Linux so I can steal the config and make things work in Ubuntu Studio, wish me luck | 15:57 | |
| arraybolt3[m] | OvenWerks: OK. Maybe all that will need to happen is we'll need an updated PipeWire. I think the only reason we didn't get total success with our plan is because of a WirePlumber bug. | 15:58 |
| OvenWerks | arraybolt3[m]: basically, PW as a replacement for pulse and jack or just a replacement for pulse already works. | 15:58 |
| arraybolt3[m] | That's what I'm hopefully about to find out. | 15:58 |
| arraybolt3[m] | Yeah, we just need to bridge PW Jack and real Jack. | 15:59 |
| OvenWerks | There is (or was) a pw repo on launch pad that has the latest version of pw. | 16:00 |
| arraybolt3[m] | The most recent news in Arch Linux is that they're using pipewire-media-session instead of WirePlumber, and in the link I found, someone had what we were looking for working using pipewire-media-session (I believe) under Fedora. | 16:00 |
| arraybolt3[m] | I remember it. Maybe that will shake WirePlumber into working. I'll check it before I fight with Arch, that's a good idea. | 16:01 |
| arraybolt3[m] | Hmm. The guy running the PipeWire PPA is reasonably, but annoyingly, only packaging for Jammy and earlier, not Kinetic. I guess I could kludge the packages in. | 16:05 |
| Eickmeyer[m] | Use caution. | 16:05 |
| arraybolt3[m] | Eickmeyer: We call that a VM. | 16:05 |
| Eickmeyer[m] | hehe | 16:05 |
| arraybolt3[m] | (I do use Ubuntu Studio as my daily driver, but I do ALL development work in VMs or dedicated testing machines, so I should be safe.) | 16:06 |
| Eickmeyer[m] | Actually, same. | 16:06 |
| arraybolt3[m] | OK, crud. There's so many binary packages built from the source packages in the PipeWire PPA that it would be a nightmare to try to manually install them all. I think I should either build my own PPA or fight with Arch to try to get the perfect config files. Which do y'all think I should tackle? | 16:19 |
| arraybolt3[m] | Actually, just found a way to force an old PPA into a newer system. OK, this should be messy. | 16:21 |
| Eickmeyer[m] | Been there, done that. It's as easy as changing the /etc/apt/sources.list.d file. | 16:21 |
| arraybolt3[m] | Oh nice. Didn't know that trick. | 16:21 |
| arraybolt3[m] | Crud, did a jackd update come through just recently into Kinetic? I just updated my Ubuntu Studio VM, installed the PipeWire PPA, now jackd only recognizes the "dummy" backend and refuses to start even if you use "jackd -d dummy", and attempting to start JACK via Studio Controls hangs on "Configuring...", which kinda throws a wrench in the works of the test I was about to do. | 16:34 |
| arraybolt3[m] | (jackd is also acting very weird, displaying diagnostic messages I don't remember seeing before.) | 16:35 |
| OvenWerks | jackd -V | 16:35 |
| Eickmeyer[m] | arraybolt3: As far as I know, there was a new jackd (jack 1) that came through due to a new upstream import in Debian, but we don't seed that. | 16:35 |
| arraybolt3[m] | OvenWerks: jackdmp version 1.9.21 | 16:36 |
| OvenWerks | I see: jackdmp version 1.9.21 tmpdir /dev/shm protocol 9 | 16:36 |
| Eickmeyer[m] | That's jackd2. | 16:36 |
| arraybolt3[m] | Same here. | 16:36 |
| arraybolt3[m] | The jackd on my physical system allows "dummy" to be used as a master backend, and it shows many more backends. It's 1.9.20. | 16:37 |
| arraybolt3[m] | OvenWerks: What backends are you told exist if you do "jackd -d"? | 16:37 |
| Eickmeyer[m] | I backported 1.9.21 into the backports PPA and am having no issues. | 16:38 |
| arraybolt3[m] | Crud, maybe my PPA mangled it then. | 16:38 |
| Eickmeyer[m] | arraybolt3: What's that PPA so I can inspect? | 16:39 |
| arraybolt3[m] | https://launchpad.net/~pipewire-debian/+archive/ubuntu/pipewire-upstream?field.series_filter=jammy | 16:39 |
| arraybolt3[m] | Also used the WirePlumber PPA it mentions in the installation instructions. | 16:39 |
| Eickmeyer[m] | There's no jackd package, so likely not the culpret. | 16:39 |
| arraybolt3[m] | Well, hey, VM installs are cheap, I'll just trash this one and install from scratch with a zsync'd ISO. | 16:39 |
| Eickmeyer[m] | Snapshots are cheap too. | 16:40 |
| arraybolt3[m] | (Actually, I'll keep it around just in case it comes in handy.) | 16:40 |
| OvenWerks | arraybolt3[m]: I looked at my jackd log for the last start which was 1.9.20 so I restarted (ffado backend) and it came up again ok. | 16:40 |
| arraybolt3[m] | LOL I suppose so, but you get this feeling of invulnerability when you use a VM and don't think to do wise things like that. | 16:40 |
| arraybolt3[m] | Looks like the PipeWire PPA must overwrite the default JACK libraries or something. Reinstalled Ubuntu Studio, everything was fine. Updated, fine. Added PPA and installed, things immediately started going weird with JACK (carla would only open PipeWire JACK, not real JACK), and upon rebooting the VM, jackd was borked. | 18:46 |
| Eickmeyer | FYI, I heard about a package conflict with pipewire and qjackctl. snd1 has an upstream bug in Debian he's been working on with that, so that *might* be your culpret, but I'm not 100% that's the case. | 18:48 |
| Eickmeyer | You might have to uninstall qjackctl first. | 18:49 |
| arraybolt3[m] | OK, that's a possibility. I still have the VM here. Odd that apt would just let the thing install if it really is overwriting files - isn't it supposed to protect from that sort of thing happening? | 19:02 |
| arraybolt3[m] | (Oh, and I made a snapshot this time.) | 19:02 |
| Eickmeyer[m] | That means there's some package conflicts. | 19:02 |
| Eickmeyer[m] | And that PPA doesn't take into account packages and the files they own. | 19:03 |
| arraybolt3[m] | Ah. Crummy. Well, I'll try reverting the snapshot, uninstalling qjackctl, installing the PPA, and seeing how that goes. | 19:04 |
| Eickmeyer[m] | If I were you, I'd go with the pipewire that's in Kinetic now, not something from a PPA. | 19:04 |
| arraybolt3[m] | I did that and got the results above (unless there's been a very recent pipewire/wireplumber update in Kinetic - is there?). | 19:04 |
| Eickmeyer[m] | Anything that's in Kinetic is likely sync'd from Debian Unstable (sid). Unfortunately, that's what we're going to have to use because we can't use outside sources. | 19:05 |
| arraybolt3[m] | OK. | 19:05 |
| Eickmeyer[m] | arraybolt3 (@arraybolt3:matrix.org): It's one of the burdens of being an official flavor: the ISO build process can only use the main and universe repositories, it cannot use any other sources. | 19:19 |
| arraybolt3[m] | Well, on the bright side, we can patch whatever we want, right? | 20:20 |
| Eickmeyer[m] | Well, yes, so long as it's in the universe repository. I'm a MOTU, so I can upload with patches. | 21:20 |
| Eickmeyer[m] | But then it loses sync with Debian, which puts the onus on us to maintain, and the less we maintain the better arraybolt3 (@arraybolt3:matrix.org). | 21:21 |
| arraybolt3[m] | Eickmeyer: Possibly silly question, but can we "cheat" and upload a package to universe that contains a script that, if the user triggers it, it enables a PPA? Something like a button in Studio Controls labeled "Enable FFADO on PipeWire" that, when you clicked it, would display a warning "This will pull in Personal Package Archive data into your system, are you sure you want to do this?" Seems similar in concept to Arch | 21:30 |
| arraybolt3[m] | Linux's AUR, and I wonder if Ubuntu would allow that. It would get around the "can't use anything outside main and universe" rule, but require user intervention to do so, and user intervention can already work around that rule with add-apt-repository. | 21:30 |
| Eickmeyer[m] | Well, ubuntustudio-installer has something like that, but it makes the user very aware that's what's going on (enables the Ubuntu Studio Backports PPA). However, the Ubuntu Studio Backports PPA is considered a trusted source. We can't consider every PPA out there a trusted source, which is why it wouldn't be allowed. | 21:32 |
| arraybolt3[m] | (OK, my thing about the AUR was totally off, but you get what I'm trying to say.) | 21:32 |
| Eickmeyer[m] | And the AUR is scary, scary, scary. | 21:32 |
| arraybolt3[m] | True. But we could upload any build of PipeWire we wanted to into a PPA. Though... that would put us back in the role of maintainer... gah, they don't make this easy, do they? | 21:32 |
| arraybolt3[m] | (The most I ever used AUR for during my brief stint with Arch Linux was to install 2048-qt, so I guess I don't have a whole lot of experience with it.) | 21:33 |
| Eickmeyer[m] | Nope. I'd rather go with what's in Kinetic. Pipewire is there, and Ubuntu (proper) is going with Pipewire audio by default. | 21:33 |
| arraybolt3[m] | Yeah, OK. If I can just get pipewire-media-session to make the little box that does the FFADO thingy, we're golden. But if it requires a bug fix, and that bug fix doesn't land in Debian, and therefore doesn't land in Kinetic, then... I guess at that point it's a matter of how much maintenance work we want to do to get the feature. | 21:34 |
| Eickmeyer[m] | A tiny patch isn't a deal breaker, especially if we can just do a sync during the next cycle. | 21:35 |
| arraybolt3[m] | OK, great. One last question - is there any particular reason that we would need to use WirePlumber rather than pipewire-media-session? | 21:36 |
| arraybolt3[m] | (I ask because people actually got the JACK-to-PipeWire bridge working on pipewire-media-session, but not on WirePlumber that I saw.) | 21:37 |
| arraybolt3[m] | (I have to go afk for a while, but I'll see your response later and hopefully be ready to start working on stuff again. Thank you!) | 21:37 |
| Eickmeyer[m] | Yeah, thanks for your work! OvenWerks would have the answer to your question. | 21:38 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!