Page 1 of 1

Pipewire [Audio]

Posted: Thu Feb 17, 2022 4:10 am
by oafunv

Pipewire is awesome and allows users easy access to jack tools like calf and no latency for AV sync unlike pulse.

Will we see it in future puppies?


Re: Pipewire [Audio]

Posted: Thu Feb 17, 2022 4:23 am
by Grey
oafunv wrote: Thu Feb 17, 2022 4:10 am

Will we see it in future puppies?

I think @dimkr has already been working in this direction :) I saw something on the Git.


Re: Pipewire [Audio]

Posted: Thu Feb 17, 2022 5:40 am
by williwaw

Re: Pipewire [Audio]

Posted: Thu Feb 17, 2022 6:23 am
by dimkr

woof-CE supports PipeWire.

Vanilla Dpup 10.x development builds (https://github.com/vanilla-dpup/unstable/releases) already use it, and I've been able to build an Ubuntu 22.04 based Puppy with it (https://github.com/puppylinux-woof-CE/woof-CE/pull/2857). It works nicely, and everything that works with PulseAudio (pavucontrol, pa-applet, etc') keeps working.


Re: Pipewire [Audio]

Posted: Thu Feb 17, 2022 7:10 am
by wiak

If you are on an Arch Linux based system (I am with WDL_Arch64) or similar to below via apt with Debian/Ubuntu base.

Code: Select all

pacman -Sy pipewire

To configure PipeWire, you can copy files from /usr/share/pipewire to the alternate system-wide location /etc/pipewire, or to the user location ~/.config/pipewire

(at least on nice Arch Linux filesystem standard) or if you want more facility:

Code: Select all

pacman -Sy pipewire pipewire-alsa pipewire-jack pipewire-pulse

and remove the likes of pulseaudio-alsa since no longer needed and would conflict.

Arch recommends to choose session manager WirePlumber rather than simple PipeWire Media Session when asked during installation of pipewire.

No idea how ready for 'prime time' pipewire now is. I'm in no hurry - trivial, as Arch Wiki decribes to install anyway.

https://wiki.archlinux.org/title/PipeWire

I tried full Wayland (with appropriate dm and apps) on my WDL_Arch64 long time back and that also just worked in trivial to install fashion (per Arch Linux wiki again), so I just wait on the upstream distros dictating when they move their support to that - no hurry... pipewire is available, as is full Wayland (with or without Xwayland) if you want or need it, and coming as default probably soon enough anyway. Mind you, Puppy has a history of rejecting the new (still uses old sysVinit for example - if against systemd, runit, as used in Void Linux and derivitaves, is a better choice for sure - I've used all of these - decades using sysVinit - thank goodness most have moved on from that). Puppy used to try to be leading edge, but seems to have a lot of stick-to-tradition issues (of course I am well aware of the war been going on and on for years re: systemd); though good to see official package managers being adopted in some newer somewhat unofficial Pup variants at last.

Despite size being more than a little unimportant for the most part disto-wise nowadays, Puppy's strength nevertheless remains in trying to keep things smallish (otherwise might as well use larger distro with JWM or Openbox for responsiveness and resource usage pretty similar to Puppy) - Even if Puppy wasn't used so much by most people as their main desktop distro, being kept smallish will always be useful - especially for use in virtual machines and so on, where you might want to build a network of virtual machines (with whole OS loaded into RAM on each maybe), or even just for the absolute convenience of small download (well... ultra fast fibre makes that no big deal either once most everyone probably has it one day soonish).

But keeping up with best technology (which may be pipewire and Wayland soon enough if not now) is vital for Puppy to attract users - but, even then, the question that troubles my mind is 'why use a dpkg/apt capable Puppy, especially if you accept it 'growing in size' when you could instead use a dpkg/apt capable small actual Debian/Ubuntu system? What would be the advantage is what I wish I knew the answer to (tradition/nostalgia/familiarity??? or is it something else?). Fortunately I have no doubts about this forum, which has much going on and much going for it as a place to discuss ideas and experiment accordingly - might not get thousands of actual Kennel-created distro users though (but does that matter really? - anyway, such matters are for a different thread to answer - I'm sure pipewire will become default in most Pups if it becomes default every where else).

One thing I think for sure though - it is not enough to simply incorporate 'new' (and not so 'new' really) technologies into Puppy. There has to be something better about Puppy design that makes it preferable to use Puppy than something else. I don't think enough thought is given to that. Wayland, Xwayland, pipewire - nothing special there - lots of distros use or can use these - so why use Puppy is the question even if future Puppy's include pipewire support via Woof-CE or whatever. To increase success (if measured by usage) of Puppy, characteristics need to be determined that would make a new Puppy somehow 'better' at least in some circumstances, than the many 'rest of distros' available here and out there. And somehow, 'pipewire' and single-user OS doesn't seem to match to me (and pseudo multi-user with spot or whatever doesn't quite do it for me nowadays - all the hacking around to even make Puppy work with browsers that don't like running as 'root' - life is too short).


Re: Pipewire [Audio]

Posted: Thu Feb 17, 2022 7:53 am
by dimkr
wiak wrote: Thu Feb 17, 2022 7:10 am

Puppy used to try to be leading edge, but seems to have a lot of stick-to-tradition issues

Look at the "Done" column: most items there are counter-evidence for this statement.

https://github.com/puppylinux-woof-CE/w ... projects/1

Puppy with all Wayland-related features, including PipeWire, is super lightweight compared to today's Puppy and its distant relatives (even RAM consumption is lower), and it's much better in terms of utilizing the GPU.

wiak wrote: Thu Feb 17, 2022 7:10 am

why use Puppy is the question even if future Puppy's include pipewire support

You never miss an opportunity to bash Puppy, don't you?


Re: Pipewire [Audio]

Posted: Thu Feb 17, 2022 10:12 am
by wiak
dimkr wrote: Thu Feb 17, 2022 7:53 am
wiak wrote: Thu Feb 17, 2022 7:10 am

why use Puppy is the question even if future Puppy's include pipewire support

You never miss an opportunity to bash Puppy, don't you?

By taking less than one sentence of the paragraph I wrote, you disgracefully take it completely out of context. In fact, my intention was not at all to 'bash Puppy' but was clearly asking the Puppy community to think carefully what it is that should be in Puppy to make it remain competitive. My whole comment was about the fact that it is not enough to simply throw in a lot of new work into woof-CE (which certainly should be done), new technologies and so on; because all distros are incorporating these new technologies. What we need, for Puppy, is to establish a 'point of difference' that makes it something people will be drawn towards in a special way. No matter how good or great anyone's work beavering away up at woof-CE is, that alone will be to no avail if specific thought isn't given to that 'point of difference' - not just the 'whats' to include, but also the 'whys' in terms of what to include and the 'point of difference' being aimed at in the final design. Here was my actual paragraph and not the out-of-context attempt to smear my actual purpose:

wiak wrote: Thu Feb 17, 2022 7:10 am

One thing I think for sure though - it is not enough to simply incorporate 'new' (and not so 'new' really) technologies into Puppy. There has to be something better about Puppy design that makes it preferable to use Puppy than something else. I don't think enough thought is given to that. Wayland, Xwayland, pipewire - nothing special there - lots of distros use or can use these - so why use Puppy is the question even if future Puppy's include pipewire support via Woof-CE or whatever. To increase success (if measured by usage) of Puppy, characteristics need to be determined that would make a new Puppy somehow 'better' at least in some circumstances, than the many 'rest of distros' available here and out there. And somehow, 'pipewire' and single-user OS doesn't seem to match to me (and pseudo multi-user with spot or whatever doesn't quite do it for me nowadays - all the hacking around to even make Puppy work with browsers that don't like running as 'root' - life is too short).

I certainly stand by my actual comment, which moreover applies to all distros discussed on this forum to some extent (many are true multiuser so that much doesn't apply). Perhaps a lot should be being learned from EasyOS; maybe EasyPup incorporated the container innovations and so on employed there (I don't know), but alas support for EasyPup did not continue anyway.


Re: Pipewire [Audio]

Posted: Thu Feb 17, 2022 3:45 pm
by rockedge

I can easily incorporate Pipewire into KLV-Airedale but I would provide that version as a separate branch. I do not see a pressing need for it. Since the effort does not match the demand for it in Puppy Linux's user base, now or in the future, I am perplexed as to why this is important at this time.

How about somebody make it easier to "program" a "recipe" to guide woof-CE to build more of what I want. That seems to me way more important than Pipewire. I have had a fork for years of woof-CE and still can't reliably work with it. On the other hand I have like 20 PLUG files that will build a variety of different versions of WDL or KLV.

That's fact. I do not see where the audio component of Puppy Linux is making it any more interesting. @taersh was the only one I know of who used Puppy to produce actual music and video, and he's disappeared.

I used to build kernels like a mad man with the kernel-kit. Like full real time kernels......now these days I am lucky if anything even will compile. I have had successful builds do nothing but kernel panic. So now I don't since the failure rate made it no fun and a waste of time.


Re: Pipewire [Audio]

Posted: Thu Feb 17, 2022 8:35 pm
by Clarity

Not to be taken the wrong way as I sometime are: But, one good reason for Pipewire is its intuitive approach to "connecting" multimedia units. It just makes "plain" sense.

I dont see @wiak as bashing; rather more on the practical side of the obvious.

I think @dimkr sometimes does a dig and it is hoped that we can come to some middle ground where we "try" as best we can to present good behavior to each other in hopes of developing a better product outcome. I hope he doesn't anger at my comment. He has in the past, but I am NOT taking pot-shots at him!

Wayland has also come a LONG way in PUppyland, thank to the excellent work of @dimkr and his contributions along with the other developers of WoofCE.

There are some very excellent work being done on multiple fronts across the forum, in much part to the efforts which starts with a minor mention(s) as @bigpup brings to the table. Thanks to him, several projects have sprang into real meaningful advancements.

The current technologies of Pipewire, Wayland, GRUB2, SystemD has PUPs running neck and neck (even ahead of most) of the findings on Distrowatch. The benefit is that, In Puppy, they accomodate the rapid movements in hardware. This keeps PUPs running on old as well as new Hardware which continues to emerge.

Thanks to all of the various approaches that have been demonstrated across Puppyland, especailly, over the last year.