Hacker News new | past | comments | ask | show | jobs | submit login
Non Session Manager vs New Session Manager (tuxfamily.org)
328 points by beprogrammed on July 10, 2021 | hide | past | favorite | 168 comments



That's one side of the story, if anyone cares to read the other, this appears to be the mailing list thread that is being referred to: https://lists.linuxaudio.org/archives/linux-audio-dev/2021-F...

Then there's also this reddit comment: https://old.reddit.com/r/linuxaudio/comments/lk3u74/non_sess...

> New Session Manager was created as a fork because Non Session Manager grew stale and Jonathan was very abrasive to work with [..] getting PRs/patches merged didn't go anywhere.

> Also Jonathan (the Non suite creator) removed all his repos [..], which is something that's obviously not acceptable behavior.

The "new session manager" appears to be on GitHub: https://github.com/jackaudio/new-session-manager

Please also note that this drama is months old at this point.


The truly surprising behavior to me is choosing to name the forked project nsm.

Fork the heck out of anything for any reason you like. Say critical things on mailing lists. Fight over who's right and who's a jerk. But actions which lead to less clarity -- like, say, polluting a claimed name in the process of making a fork -- doesn't look like good faith to me.


> The truly surprising behavior to me is choosing to name the forked project nsm.

I think the point was having the binary be of the same name, so it could work as a drop-in replacement that is compatible with existing software.


New's bogus insinuation that Non had ads or spyware makes me disinclined to buy this innocent justification.


IIRC (anyone else in LAD-proper correct me) the insinuation came from using some standard boilerplate in a release announcement for the fork. The fork's release tagline outlined what open source was and how it is beneficial to users. Specifically there was mention of the fork not having any spyware/ads in it.

I agree with the posted article that saying "X doesn't have Y" from the fork implies that it is a differentiating factor, but it doesn't strike me as an intentionally malicious act. More of just not being familiar how such release blurbs would be read when applied to a forked project.


This seems like a very reasonable comment and I had exactly the same thought, yet I found this comment dead and had to vouch for it. Can any of the downvoters explain why? Or was this somehow automatic?


Some new accounts are automatically dead on arrival or if they post too quickly, and have to be vouched. Nothing sinister about it.


Didn't mean to imply sinisterness. Was wondering if an automatic process like that existed, so thanks for the confirmation.


Wouldn’t a compatibility symlink or shell script have been enough?


Perhaps, though you'd end up with the same 'nsmd', 'nsm-proxy', and 'nsm-proxy-gui' commands with the same 'NSM_URL' environmental variable. Ideally there'd be more separation, but then you do lose the drop in compatibility. For a longer lived fork it would make sense (IMO) to use the proposed symlinks, though initially it was basically the same exact software as upstream + a few modest patches.


The front-end GUI /usr/bin/non-session-manager is replaced by a symlink to nsm-legacy-gui in a New Season Manager install. One doesn't really run the daemon directly anyway.


> I think the point was having the binary be of the same name, so it could work as a drop-in replacement that is compatible with existing software.

This is a long solved problem, you don’t use the same name. That’s very bad joojoo.


I'm not sure that's a good justification for naming something the same as an existing project. What if someone forked Nginx or PostgreSQL and did the same thing? Would that be acceptable behavior?


Can you clarify what features did the new NWS add that old NWS was refusing to add?

That seems like the crux of the entire tiff, and without that knowledge it's very difficult to discern which side is more reasonable.

Edit: Honestly, some of the initial commits to the new repository are more in line with Liles' account: https://github.com/jackaudio/new-session-manager/commits/mas...

Starting in April 2020, changes include:

- Dumping the timeline, sequencer, and mixer: https://github.com/jackaudio/new-session-manager/commit/80e9...

- Removing any thing with logos or icons, e.g., "filing the serial number off": https://github.com/jackaudio/new-session-manager/commit/7992...

- Enable debug mode / verbose mode by default: https://github.com/jackaudio/new-session-manager/commit/3975...

- Change project build system over to meson: https://github.com/jackaudio/new-session-manager/commit/e431... (plus others)

- Update build system to exclude current GUI / make it optional: https://github.com/jackaudio/new-session-manager/commit/63e6...

- Set up symlinking by default in install: https://github.com/jackaudio/new-session-manager/commit/3681...

- Force the project onto GPL3, more serial numbers filing: https://github.com/jackaudio/new-session-manager/commit/2731...

After that is a number of fixes. Unfortuantely, from this list of "immediate" changes, it reads like the Argodejo developers decided they would be the preferred GUI, and decided to fork the project to abandon the current GUI. The README was even modified to this end: https://github.com/jackaudio/new-session-manager/commit/7f23... They also forced the project onto a different build system, and made new (and possibly incorrect) assumptions about installation. If the fixes came first (build fixes, etc.) and the other work came later, I would have a more-charitable view of this fork. Unfortunately, the immediate changelog sort of speaks for itself: they forked NSM, filed off all the names, changed over the licensing and build system, and ripped out the GUI. Then, a few months later, they actually started fixing bugs. The motivation here really seems to be subsuming ownership and default UI. This really just looks like a classic internet slap fight.

Of course, this is only conjecture from the git commit history, but... it's supported by the git commit history.


>- Force the project onto GPL3

Pretty sure this isn't allowed. The terms of GPLv2 allows the recipient of the software to modify and distribute the software under terms of GPLv2 or later. The software isn't public domain, someone owns the IP. You can distribute your own changes as you please. You cannot strip the IP owner's license and replace it wholesale.


They changed from GPL-2.0-or-later to GPL-3.0-or-later. The latter is a subset, so any content licensed under the former can be relicensed under the latter without consulting the original author.


The license sets the terms by which you can redistribute the software, it doesn't allow you to set the terms for the people you distribute it to. The only person/entity who can change the terms of the license is the copyright holder. People who receive this software will always have the right to redistribute under GPLv2.


As soon as you do any copyrightable change, you have the right to declare that change GPLv3 or later and that makes the overall project GPLv3 or later.

The previously existing code will still be GPLv2 or later, but you’ll have to trawl the commit history to find out what’s also available under GPLv2 or later.

So it is perfectly legal to relicense under GPLv3 or later, it’s just rude: https://www.draketo.de/english/politics/relicensing-gpl2plus...


If I have a program that consists of 5 files each of which is exactly 2000 lines lone and I distribute under GPLv2 or later. You like my program but realize that you have a novel new algorithm that will speed up the core functionality of my program by 300% and your algorithm only needs 100 new lines in one of the source files. You are totally free to add your hundred lines under terms that are compatible with my original license. You cannot proceed to remove or change my headers that specify my copyright and my license terms to the 10,000 lines of code of which I am the sole copyright holder.


As someone who's not too familiar with IP law/GPL versions: is GPL3 not "GPLv2 or later"?


They meant that you cannot once a piece of code is released by GPLv2 you cannot unrelease under GPLv2. So even if every new version and contribution is required to specify a GPLv3-or-later license all previous contributions and versions are still in GPLv2-or-later, which is more permissive.

I think the comment was addressing a misconception around how the project as a whole is licensed and its individual contributions or past versions are licensed and the fact that GPLv2-or-later works cannot be revoked as the user can always freely chose between the GPLv2 and the "later".


Rebranding a fork is absolutely normal. not "filing the serial number off". You realize that all git commit history is kept? They are not "scrubbing" Jonathan's work.

Do you even know what NSM is?


I believe the point was: “Their initial work was not fixes or changes they had tried to submit to NSM, but a rebranding effort.”

And no, I don’t know what NSM is, and I might be totally wrong on GPS intention.


From https://lists.linuxaudio.org/archives/linux-audio-dev/2021-F...

“We already had history with Jonathan; when we tried to do a few things that were useful to us but not within his design, sometimes even minor things like allowing to use 64x64 icons or having an option to export the whole session as a tar file, we were met with insults.”


The full pull request history for non is available on the Internet Archive. The only one I could find from Filipe is this one: https://web.archive.org/web/20201122235832/https://github.co...

I'm not seeing any insults in there, just a detailed explanation from Jonathan for why he didn't want to merge it. I don't understand the full technical details, but it's clear he gave it a lot of thought and there was a difference in project philosophy.


"Every time I look at LV2, I am repulsed. LADSPA is KISS, one header file dependency and relatively straight-forward. LV2 is like the opposite of KISS. It's like FART: Forget About Representing Terseness. It reminds me of something from IBM. The dependency is like 6 different little libraries that no one will ever use for anything else. And all that this complexity really buys you is shitty in-process GUIs that crash your host program."

Lol. That's what made his software so great, he kept it non. But I can see the problem when others start wanting to add features via bloat. Two contrasting desires, does he stick with the thing that made his software great, or add any feature someone wants?


falkTX maintains a fork of Non-Mixer that adds LV2 support;

https://aur.archlinux.org/packages/non-mixer-lv2-git


Just FYI, I don't think any of the claimed heated threads were captured by the internet archive's sweeps. There were a couple linked to post-fork which were either archived too early, too late, or not at all.

disclaimer - I am one of the github users in the linked thread.


Yeah, although if I read the thread right his first comment was also 3 years after the pull request was opened.


The nature of what different people consider insulting seems to vary greatly. I'd like to judge these supposed insults for myself. Have they been archived?


To the best of my knowledge, no.


Why would they remove the timeline, mixer, and sequencer?


The original upstream was more or less a monorepo of several subprojects. The fork only involved the session manager, so the other pieces were removed from the forked repository.


I figured something was up, the original author’s post reeked of bitterness.


Jonathan Moore Liles of non software has been a master craftsman of Linux software. Some of the most beautiful pieces I've had the pleasure of using.

This pains me greatly.



Thanks. The replies to the OP are interesting.

"OMG, so silly.

I've read the Mail vom J. Lilies with interest. It reflects to some degree similar issues I've encountered. Banning the OP because of this mail is insane. It gives hints on that Filipe Coelho is disliked also by others, because of several actions and deficiencies.

So that's why I ask you David Runge to undo the ban. The OP is not the core of the problem. By removing OP, you remove symptoms only."


NSM is a decent session management API. Currently recommended for any kind of session management in PipeWire.

Unfortunate that JML imploded like this, his work is legendary and software elegant.

All smear campaigns aside it's really crap to see this continuing hostility. Free Software also means moving on after a conflict and learning from past mistakes, and fork when necessary. Not to throw a tantrum and destroy all your work.

Btw most NSM stuff works pretty good under PipeWire. RaySession is a nice tool for that: https://github.com/Houston4444/RaySession


"NSM" meaning "Non"? Or "New"? Not saying just adds to the confusion.


The API is NSM, there is no difference between the two. 'New' is a fork of 'Non' implementation, but they are compatible.

Heck, now that Jonathan removed _all_ Non project repositories basically no longer exists. Only in the git history of the fork. At this point in time we do not have a Non-SM other than legacy packages and forked repositories, but the NSM API lives on.

The NSM design is sound, people are using it. Thank you Jonathan


You make yourself clear.


Agordejo is also a good NSM front-end.

https://www.laborejo.org/agordejo

RaySession kinda unilaterally pushes the spec for advanced features, but there was stagnation in the core NSM because the main dev essentially refused to work with most other developers.

Now there is the community NSM and the spec is slightly tighter and the reference client is a bit more reliable.


I find RaySession very reliable in my experience so far. And we have an awesome patchbay with grouped ports (stereo patching!)


The fact of using the same binary names makes it totally shady.

They seem to have taken advantage of the fact that the original author did not create official packages to basically steal the name in Ubuntu launchpad.

If the forkers are sincere they should not use the same binary names, eg "nsmd". That is totally unethical, and is technically a trojan as it misleadingly replaces on program with another without making it clear to the existing users that is a new package by a new team altogether.

Even the page at https://libreav.org/software/new-session-manager has broken links to files dated 2017 clearly named "non-session-manager"


libreav.org is my site, and that's an edge-case artefact of the github release feed for the new-sm project and/or the way I have configured the aggregation so far


I'm not up on my open source community politics, but I'm intrigued by the motivations. Let's take it as a given that everything in this post is 100% correct. It certainly sounds like a common enough occurrence. Those 3 antagonists are acting out of malice and greed.

What's the play here? What does one do once one has acquired control of a linux tool or a mailing list or a consortium? Is it just to have power over something? Or is that power useful to some further goal?


>What's the play here?

If Freenode is anything to go by...

Greed, ego, self importance, the need for control... and cocaine.


Fame, resume padding, and consulting gags, especially recruiters and companies looking at GitHub projects as some kind of metrics.


Or maybe there is just a varied group of enthusiastic geeks that just want to collaborate on common tooling like APIs, libraries, frameworks, packaging, and distribution.

It may "seem" that there's some sort of "take-over", it's just that people are actually coordinating on these common goals of free software audio tools.


care to share any of your reasons for optimism?


Because I use Linux Pro Audio tools in my studio and daily life and am involved with testing builds, helping with ports, generally messing around (, break stuff), and occasionally make a tune or two.

I am very happy with the free and open tools I have at my disposal. Even with the numerous rough edges, duct-tape and hacks to keep some things together.

This is Hacker News, right?


it's hacker news so I expect some actual reasons not hand waving generalities that don't amount to a "hill of beans"


I am not really involved in the open source community, but are these kinds of fights and coups as common as Hacker News makes them seem?


Years ago I wrote a p2p file sharing program and open-sourced under GPL. Soon a group contacted me for forking the project; I gave them my blessing and wished them luck.

Later I looked at their fork. All they did was adding their names to the copyright and the new project name to the comment section for all the hundreds of files. I was frustrated as this was a clear case of taking copyright ownership for the software. I asked them to remove their names from the copyright unless it's new changes by them but they said GPL allowed it. I contacted FSF for clarification and FSF replied that it's allowed as changing the comment constituted a new change. It's a bunch of BS. It took the rosy glasses off my naive self and soured my view on open sourcing my software in the future.


IANAL, and this is not legal advice, but I am surprised at the FSF's view here. I believed (and believe) that copyright requires a substantive change to form a derivative work, and I would rather doubt that changing the comments counts. (Now, I think that precedents in the court have set the bar for "substantive" pretty low, but not that low.)

I would also say that "legal" is not always the same as "ethical", particularly in this circumstance.


That's a dirty move.


If you asked me 5 years ago if I thought something like this could happen I'd have said no. Then I relocated to SiValley and experienced Machiavellian engineers for the first time in my 20+ year career. It's good to remember that not everyone puts technology and community first. Some people will do whatever it takes to make a name for themselves, even if that just means being slightly less forgotten than they would have been previously.


Politician engineers, ugh.


are these the same people who chest on exams in undergrad?


*cheat


It happens.

As is said of academic politics: the battles are so fierce because the stakes are so small.[1]

Back at the dawn of the dot-com age, just as Linux was starting to boom, and the first IPOs (Red Hat, VA Linux) were occurring, a company appeared out of nowhere^WReno, Nevada (same thing), called "Linux One", claiming to have their own novel Linux distro. Oh, and plans for its own IPO.

It proved to be little more than a search-and-replace of "Red Hat" with "Linux one", though there were also some contributions from MandrakeSoft, another distro at the time. Stealing is copying from one source, curation is copying from several....

The thing is that, at least so far as the software is concerned, that's pretty much perfectly kosher under the GNU GPL and other Free Software / Open Source licences which comprised the scope of Red Hat's offering. In terms of copyright and licensing ... there was nothing actually wrong with this. Skeevy as all getout, yes. But not a GPL (or BSD, or MIT, or Apache, or ...) violation. (The company may have failed its source provision obligations for the GPL, however.)

The robustness of the IPO failing, er, filing, might have raised a few eyebrows elsewhere though. I seem to recall it being cancelled rather quietly, there's

The Register published an investigation of the episode at the time: https://www.theregister.com/1999/11/02/linuxone_takes_more_t...

________________________________

Notes:

1. See: https://quoteinvestigator.com/2013/08/18/acad-politics/


Lost in edits, 'theres..." should read "there's no record of the IPO itself, though I also didn't find any mention of its cancellation."


I'm interested in audio on Linux, especially recently and I know neither who those people are nor what NSM / Non does. It's a bizarre post that could use a lot more context.

The title is also weird - not only is Linux audio not dead, it's doing better than ever with the recent pipewire work.


NSM is an API to launch and manage JACK audio applications to store connections and program state. NON is the "groupware" DAW project of Jonathan Moore Liles from which NSM originated.


As a guess, the past year or eighteen months has probably been worse than usual. With a global pandemic on, people were stressed and people who feel helpless about very real and large problems often get controlling and nitpicky about anything they feel they actually can control.

When all parties to a drama are equally stressed and handling it similarly, it's a recipe for escalation with little hope of real solutions.


As of late it seems to be becoming more of a problem.

Where we used to have the good will of others we now have a tide of ill intents it would seem.


It seems like people fighting over fiefdoms. I followed the freenode debacle and that seemed like comic book villain level of insanity.


For as long as I can recall, it seems that every couple of years someone decides the current state of audio on linux is terrible, but instead of trying to fix the existing available code, they start a new project. Usually there's a node to maintaining compatibility with other libraries, but that interoperability stops right where the new project thinks the old projects are Doing It Wrong.

The X vs Wayland fight is polite to a fault in comparison.


NSM can potentially be used with JACK on win and macos. It also works with JACK implementation of PipeWire for Linux.

I expect Linux audio to converge on PW (unifying alsa, jack, and pulse in the process).

Linux video has this split with Wayland which is a whole different problem (I want my fluxbox :/). PW originates as Pulse Video, and designed to work with gstreamer pipelines and wayland compositing. aiming for better userspace separation of hardware and applications. I really love the idea of having JACK, but then for video. (no more vloopback hacks, named pipes or random socket servers to bounce framebuffers between applications)

JACK will then live on for dedicated Pro Audio applications and hopefully on win and macos as the old-skool "this also works" audio API that really only a small niche uses (and can break at the snip of silicon valley's fingers).


> I expect Linux audio to converge

That's what alsa, JACK, and pulse said. That's what they all said, except perhaps OSS, because nothing proceeded it.


Whilst such things have been common in large corporation office political culture for decades, too see it play out in the wild like this in open source, is just shocking.

Sadly it does seem to be becoming more and more, more since some view open source/projects as potential revenue streams and be it irc, or some carefully nurtured project that has grown over a lifetime, to be rug pulled by such `political games` all in the end for what I dare say be some agenda greed. It just saddens us all.


Shit happens all over the place.

I can point at it happening in the scuba diving industry.

High school never ended for some people.


At least in corporations there is money at stake, so I get (do not excuse, but understand) shitty behaviour. But all that seems to be at stake here is clout.


Kind of. Big scraps? No, however it is worth pointing out that small open source communities can be a melting pot of really clever often contrarian/polemical developers who are kind of bound to clash at one point or another.


no they are not - but the bigger the stakes, the bigger the audience, the bigger the lever.. Faust comes to mind


Yes.


Here's a link to the controversial fork: https://github.com/jackaudio/new-session-manager


I understand that this title is just verbatim taken from the blogpost, but could we switch it to something more suitable like "Linux Audio Moderation Accused of Misconduct"? The current one just seems silly given how trivial the contents of the article are.


No.


It appears that Fedora dumped this guy's version in favor of the fork some time back: they package new-session-manager, not NOM. Seems the fork was years ago.


> Seems the fork was years ago.

The article says this all started in "early 2020"; about a year and a half ago. Indeed, looking at the new-session-manager commit history, the fork was started in April 2020[1], and first announced in June 2020[2].

> It appears that Fedora dumped this guy's version in favor of the fork some time back

So did Arch Linux. Notably, the Arch Linux contributor who did it is David Runge, one of the 3 people called out in the article.

[1]: https://github.com/jackaudio/new-session-manager/commit/6352... (the first commit after the last commit from Jonathan Moore Liles)

[2]: https://lists.linuxaudio.org/archives/linux-audio-user/2020-... (1.3 is the initial version of the fork, as at the time 1.2.0 was the latest version of non-session-manager)


Jonathan Moore Liles removed all NON repositories from the internet. Only backups and forks exist.


Well, I guess that means the new-session-manager project has won out. If the NON stuff doesn't exist anywhere, then people / distros don't have much of a choice to include it.


omg - this is dirty tricks playlist.. What must civil people do in the face of this Machiavellian marching band


Forking a project is fine. Stealing the name is not. It creates confusion as to which is 'real'.

The original is the real NSM, the fork, if they really want to be seen as legitimate, should do a full scale rename.

It really is that simple. All the rest of the comments are noise.


Personal issues are at least as important as technical issues in large software projects.


This article is why desktop Linux is so broken. Nobody firmly in control to root out misbehavior like this.

This article broke my heart, because it’s indicative of a greater problem in our beloved Linux community.


Sounds like it might be time to pull a Libera and migrate to a new mailing list.


Maybe rasengan can pull some funds out of his ass to host the server.


To be clear, NSM is one project, and doesn't represent all of Linux Audio.


The allegation is that the "Linux Audio" consortium (linuxaudio.org, github.com/linuxaudio) is compromised, and therefore "dead". The new-session-manager announcement emails come from software@linuxaudio.org. new-session-manager lives at github.com/linuxaudio . The moderators of the linuxaudio.org mailing lists are banning non-session-manager messages. new-session-manager may be one project, but its leaders have infiltrated the Linux Audio organization.


Oh, no. That's just the NON meta repository mirror (because Jonathan removed his repos from the internet).

The New-SM repo is hosted here: https://github.com/jackaudio/new-session-manager


Half-right.

https://github.com/jackaudio/new-session-manager was originally named https://github.com/linuxaudio/new-session-manager and has since been renamed (you can tell because of the redirect).

The github.com/linuxaudio URL is the one they were using as recently as January this year: https://lists.linuxaudio.org/archives/linux-audio-user/2021-...

When I wrote the parent comment, I did not notice that I was being redirected from github.com/linuxaudio to github.com/jackaudio; I apologize for that mistake.


The title is misleading, this is just one application for linux audio, which I never heard about.

is this similar to JUCE?


NSM is an API to launch and manage JACK audio applications to store connections and program state. NON is the "groupware" DAW project of Jonathan Moore Liles from which NSM originated.


Misleading title. This is personal drama about a group of people. Nothing to do with audio on Linux.


they keep changing the audio server and every time there is more problems if you switch... the latest being Pipewire, I think...


still on freenode IRC?


for people as confused as I am, "Linux Audio" here refers to some sort of libre software group centered around audio applications on linux, not the actual audio stack of linux systems.


Yes, if this has said 'Linux Audio Consortium' it would have been a lot clearer. The described behavior sounds appalling, but the author is also not doing himself any favors in how he explains it.

I have never heard of NON before and beginning the story in media res while omitting any details of what the 3 antagonists asked him to do with his project makes the whole thing confusing. Until he got to talking about the LAD mailing list, which I am somewhat familiar with, I thought it was some sort of scam post.


thank you - neither the article title nor the article text make that terribly clear without specific prior knowledge of both.



Submitted title was "Linux Audio is dead!". That's obviously linkbait, so a mod changed it, in keeping with the site guidelines.

"Please use the original title, unless it is misleading or linkbait."

https://news.ycombinator.com/newsguidelines.html


Thank you. Had a bit of confusion mixed with a mini heart attack there.


As a linguistic point, I note that title case is unhelpful in this instance: “Linux Audio Is Dead” here on HN, “Linux Audio is Dead” in the original article, both are ambiguous. Sentence case, which is quite rapidly supplanting title case in most English locales, would have made it clearer, though still subtle: “Linux Audio is dead”.


Good point. My first language is German and I've never been able able to warm up to English title case. There's no such thing in German that I'm aware of.


FWIW the actual audio stack of linux systems is pretty broken too. Ever try to use some pretty standard bluetooth headsets with Linux? Never works. I always had to delete and re-pair for every conference call. And you have to go through this nonsense

https://github.com/dheera/scripts/blob/master/config/install...

if you want your audio to sound better than a snake in a swimming pool, and nobody will ever tell you that. On iOS, Android, Windows, MacOS, bluetooth headsets pretty much "just work".

I'm a Linux fan and all my computers run Linux but Bluetooth audio IS broken.


On Arch, my Bluetooth headphones are working great. MPOW-x3 and sony wh-1000xm4

My biggest gripe is that I can't use the microphone at the same time as the headphones, without switching to tin can mode.


This is a limitation of Bluetooth, not Linux. It's crazy, but no decent 2-way Bluetooth protocols exist.



What's the limitation? My wild-ass guess would be that rx and tx have to be on the same radio frequency, so you can't do full duplex without packet collisions.



Some sort of hardware limitation preventing two protocols(profiles) going through CODEC or something. “Back/return channel” features are emerging to encapsulate profile over profile so it might change in the future


It's that the handset profile which allows two way audio but only with crappy phone style codecs is different from the audio profile that only has one direction for an audio stream but has decent audio quality.


What I don't get is that why the H/W companies don't just solve this by including a 2nd Bluetooth chip. Give up on running duplex, run high quality audio out to the headphones, and the mic as a separate device using whatever is best there.


Half the battery life.

And there are fixes at the protocol layer already, you just have to use nonstandard extensions. This is how airpods work.


And beats (at least my studio 3s) on Mac/iOS/Android.

If your going to do wireless on windows, either get a separate mic, use a corded headset, or a headset with a USB-wireless adapter unless you want it sounding like a tin can.


You could just stuff two headsets into one same headphone, sharing only speaker output and power. Battery life isn’t a problem, a random BL-5C from parts bin should work just fine.


My XM4s used to pair fine on Arch too, but not anymore. I think the mic issue isn't specific to Linux but the headphones themselves.


Have you switched to pipewire? I haven't pulled the trigger yet so I'm wondering if bluetooth gets any better.


I have a new laptop running Arch, and Bluetooth headphones which nominally can act as a headset as well, but I actively disabled that functionality under Windows, because if something tried to use the atrocious-quality headset microphone, the high-quality headphones output would silently (literally!) not work, and only the low-quality headset audio output would work.

The headset modes haven’t appeared under Linux at all, which happens to suit me just fine, so I haven’t investigated the lack. A couple of times I’ve had issues with connecting at the BlueZ level, but putting the laptop to sleep and waking it up again has resolved it. (Just restarting bluetoothd doesn’t help.)

I was running PulseAudio for a few weeks, then I switched to PipeWire. I no longer get the Bluetooth indicator on it in waybar, but other than that the switch was fairly uneventful. One point is an improvement: it now seems to remember which devices I like to use, so that when I plug in my Yeti microphone it doesn’t switch audio output to it (it has a 3.5mm monitor port and can feed audio from the computer through that too), which it had always done under PulseAudio and I hadn’t yet gone to the trouble of figuring out how to stop it from doing that. One point in the Bluetooth device handling is potentially a slight regression: when silence should be being fed through the connection, I observe extremely quiet whine which I didn’t under PulseAudio. Not sure what’s at fault there.


yes much better and greater reliable with codec switching. i encourage switch. also mSBC better 2way sound


And LDAC support.


pipewire better in this, use mSBC having better quality 2way sounds


> On iOS, Android, Windows, MacOS, bluetooth headsets pretty much "just work".

I wouldn't go that far. On windows my headset shows up as two cryptically named devices, and using the wrong one or touching microphone input destroys the audio quality.


I can second this. I have a couple different pairs of Bluetooth headsets with microphones that work flawlessly on android, iOS, and Mac.

Windows 10 gets decent audio quality until you try to engage the microphone at which point it switches to a different audio driver with the quality of bad cellular phone call. It sees the headphones AND a generic “Bluetooth headset”. I’ve seen the same behaviour across Sony, Beats, and AirPods.


That's standard for Bluetooth. A2DP doesn't support the mic.


FWIW, my headset works great out of the box, defaults to stereo, and the kde audio settings let's me pick headset to switch it to crappy audio, but mic mode for calls. No need to disconnect.

NixOS 21.05; with last year's release, it did default to headset mode, and the audio profile switching was hidden behind an "advanced" button, so things are improving fast.


I have to do this with windows as well. And Mac. So, this isn't necessarily different across platforms.


I haven’t had any problems with my two headsets and Fedora. Not sure what the difference is.


Really depends on the hardware. My XPS 17 didn’t have fully functional sound (speakers, mic, headset) until kernel 5.11. And even then, if I restart, I lose everything except speakers. The solution is to shutdown and then boot up.

An older latitude works perfectly.


Nothing Blue Tooth "just works"... maybe in another 100 years or so.


Shouldn't those Bluetooth issues be fixed by PulseAudio 15?


Yeah my Jabra headset works flawlessly on Ubuntu.


Mine doesn't. I have a Jabra 75t elite, 20.04, and after trying to unpair and repair 3 times I have to run this goddamn script to keep the microphone volume up or it always auto-adjusts it to 0.

https://github.com/dheera/scripts/blob/master/jabra-stop-cha...

Also, if you didn't do what my parent post's script did, you probably have shitty audio on your Jabra.


I have the same pair, they work flawlessly on my Manjaro desktop/laptop. The audio sounds the same as it does on Windows/Android/Apple devices, at least on my machines.


I'm not very familiar with using Bluetooth with Linux, but looking at the script this appears to be a pulseaudio thing more than a Linux thing.


Yeah well ... I'm not an audio driver engineer, my Linux distro is the most popular (Ubuntu), came with PulseAudio, and I just want my audio to "just work" so I can have my damn meeting and work on the other Linux-based things I need to work on.


How do you kill that which has no life?


the linux community never fails to deliver drama, it's like watching a bunch of nerds have a slap fight.


If there is such a thing as a non-trivial commercial software project without office drama, I have never seen nor heard of it. I think drama occurs whenever and wherever there are a non-trivial number of people working together. It has nothing to do with linux specifically.


I never heard of either project before today. Searching Debian for anything that seems related reveals only LADI Session Handler. It is not at all clear how Non Session Manager fits into the Linux audio milieu. It does not seem as if I have needed a session manager. But I might be mistaken. It has been a long time since I had any complaints about audio on Linux, basically since PulseAudio and things that talk to it got working.

Who uses this? And, why would anybody do something as shady as to make off with somebody's project name? What benefit do they imagine they could get from that?

I would rather see, instead, an announcement of a new discussion group with decent governance, and clean packages for Debian, Fedora, Arch, Nix, etc., renamed if necessary. Needing to rename an old project to work around malfeasance is inconvenient, but nobody is married to the old name.


This kind of audio session management is used by musicians and audio engineers. The idea is to be able to launch multiple applications that have audio/midi input and output, and manage the signal routing between them. Think of it as being kind of like Unix pipes for sound.


A really one sided take.

The Non dev was being abusive in public threads and even nastier in private mails to various Linux Audio devs.

The full story is on the archives of the Linux Audio Development and Consortium mailing lists.

Folk should come visit #lad (Linux/Libre) on Libera.Chat.


If they had picked a different name for the forked project, none of the drama would exist. By picking a very similar and the same acronym, the gang is doing something underhanded.

In the business world, this is a clear case of trademark infringement and they would be sued out of existence.


I have no context here, but from reading the post, it sounds like they did in fact pick a different name, but they picked one that let them reuse the same executable name for API compatibility.

In the business world, picking a different name for marketing and using functional elements for compatibility is a good way to get sued but win. See, for instance, Sega v. Accolade - Accolade sold software for the Sega Genesis without Sega's approval. Sega added something to the Genesis game loader which checked for the string "SEGA" and displayed the message "Produced by or under license from Sega Enterprises LTD." Accolade's games caused that message to appear, even though there was no license. The courts ruled that Accolade was entitled to do this, because the point was compatibility.


Using the same binary name for a daemon is not necessary for compatibility.


As it happens, the first time this came up in #lad I was like "another name is sensible!" ("how about osm? like, One Session Manager!" (that being, at least, British humour)), but, by the second time, I was resigned to the fact that renaming what would be the same (but more robust protocol) to another acronym would be pointlessly confusing to users (and devs). NSM was a hopelessly overloaded term already.

It's interesting to see the instance of HN groupthink speculation (and lack of google fu) there has been in this comment section. Folk should come idle/chat in #lad. It's available via Matrix.


Because people here look dimly on those taking other's hard work with underhanded tactics. The name NSM and its binary files are the BRAND of the original project and the original developer spent lots of time to build the project and build the brand. All the integration with the names required work and time built over time that the original developer had put in. Don't tell me it's just a name and has no value. If so, you guys won't pick the same name and same binary names.

Fork the project and pick another name. Spend the effort to market and build the integration with the new name. No one would have a problem.


Underhanded tactics? JML was abusive and has exaggerated.

I think it should be clear to anyone the relationship; the README mentioned JML from the get go, it's not framed as NON related, there is reference client and daemon, and it's just the same acronym because there's no reason to change.

The new projects name is New Session Manager.


Picking the same acronym based on a similar name is VERY underhanded, especially when the acronym is used throughout as the name for the project. The term NSM is used more than a dozen times while Non Session Manager is used 3 times in their about page. NSM is basically the identity of project. http://non.tuxfamily.org/wiki/Non%20Session%20Manager

Using the same acronym and the same binary file names to push out Non from the distribution is very underhanded.

This looks like a deliberate aim to confuse the users and taking over the project code and identity. It's not cool.

I don't know JML and don't know his development style. Since he has been gracious enough to open source his project, I assume he's generous enough to share his hard work with others. If people don't like his style, fork the project with a different identity and compete on the merits, rather than stealing his project's identity with confusion.


Push out non nsm from a distribution? Mate, it was the individual distro packagers that decided enough was enough on that front.

You are trying to argue a course of action that would just have confused both users and devs.


For those interested, the linux audio consortium mailing list has archives available at: https://lists.linuxaudio.org/archives/consortium/ I see mentions of LAD mailing list archives elsewhere in the comments section, but for anyone looking for the public consortium emails they're here.


(citation needed)


He literally stated where you can find the information.


He hasn't provided a citation. He's told someone they can do their own research an a mailing list. A parallel might be me claiming that we have a highly algorithm for factoring large primes and saying you can look it up on arxiv.org - I've pointed at a repository of knowledge, but not actually cited why I believe the fact.

Providing a citation is pointing at a specific place where the information came from at the smallest granular level. In this case, a citation would be something like links to specific threads.


Your analogy is way off in terms of scale.

Edit: also, I've no obligation to provide a citation, I'm one of those that's bored by this particular drama, folk can go search for posts of they want, I've given enough to go by.


Hyperlinks were invented for a reason.


For people at CERN to make bold claims, with citations


>(citation needed)

>The full story is on the archives of the Linux Audio Development and Consortium mailing lists.


How can you trust the mailing lost when the accused are the heavy-handed moderators of that list?


It is one sided, by definition.

However creating a fork with the same acronym and name for the daemon or binaries or whatever seems pretty under-handed.

Was a release announcement for the original software moderated, on a list where release announcements had been posted without moderation before the fork? Was the author banned from lists?


The Linux-Audio-Announce mailing list is a moderated mailing list and has been for quite a few years (10+ IIRC). It had been several years since the original non-* projects had a release announcement there and within those years the standard format of release postings had shifted from a more casual "tell us about your software" tone into "just provide changelogs and links to resources" type emails. The change over the years had not been documented in written list policies, so the initial LAA email was rejected. To the best of my knowledge no LAA ban was implemented, just a rejection of the initial submission after it entered the moderation queue.

A ban did occur on the LAD mailing list, but to the best of my knowledge the ban occurred due to some off list emails with one of the list moderators.


The same acronym is... a questionable choice. But the matching binary names are standard for projects which are compatible replacements for the other project. You're basically releasing something with a different name, but implementing "previous-software-name" interface. This helps with drop-in replacement distro packages to.

That's why Ubuntu has the alternatives system. And why mawk / gawk / etc. are often linked to /bin/awk.


You don't actually need to name your binary after your project you know.


The same acronym is underhanded.

And it's entirely possible to use different software names for projects that implement a compatible API or service, the claim that it's for compatibility is pretty bogus. It's not "standard", especially not for an adversarial fork. And that's not why the alternatives system exists, the alternatives system exists so differently named packages can implement a particular service (i.e., exactly the opposite, for the equivalent in shell commands). So that's underhanded.


Alternatives works that way, because conflicting binary names are messy to handle, but at the same time it works that way, because that's the behaviour you normally want in your system. You want "a grep" and "an awk" and many others.


I know how alternatives work, and they are not used to make packages with the same named binaries drop in replacements.

They are used to make packages with differently named binaries able to implement standard shell commands.

There is no need to match binary names. It doesn't really help with anything.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: