Release: New Session Manager Version 1.4.0

[nils]

Linuxaudio.org presents: New Session Manager Version 1.4.0

## Summary

The theme for this release is to fix known bugs and issues with both the original Non-Session-Manager ("nsmd") server as well as its GUI (our "nsm-legacy-gui"). Problems and inconsistencies that have piled up for years were finally cleaned up (full changelog below).

Users do not need to take any action (besides updating), the session save-format remains unchanged. Client programs, or rather their developers, do not need to change or adjust anything; all changes are "under the hood".

We have no self-hosted website yet, but plan to establish one in the future. Until then the project can be found on Github:

https://github.com/linuxaudio/new-session-manager/
https://github.com/linuxaudio/new-sessi ... tag/v1.4.0

## Full Release Announcement:

New Session Manager (NSM) is a tool to assist music production by grouping standalone programs into sessions. Your workflow becomes easy to manage, robust and fast by leveraging the full potential of cooperative applications.

It is a community version of the "NON Session Manager" and free in every sense of the word: free of cost, free to share and use, free of spyware or ads, free-and-open-source.

You can create a session, or project, add programs to it and then use commands to save, start/stop, hide/show all programs at once, or individually. At a later date you can then re-open the session and continue where you left off.

All files belonging to the session will be saved in the same directory.

Check your distributions in a few days for New-Session-Manager 1.4.0

You can find the source release on Github:

https://github.com/linuxaudio/new-sessi ... tag/v1.4.0


Bullet Points
* Drop-In replacement for the non-session-manager daemon nsmd and tools (e.g. jackpatch)
* Simple and hassle-free build system to make packaging easy
* Possibility to react to sensible bug fixes that would not have been integrated into original nsmd
* Stay upwards and downwards compatible with original nsmd
* Conservative and hesitant in regards to new features and behaviour-changes, but possible in principle
* Keep the session-manager separate from the other NON* tools Mixer, Sequencer and Timeline.
* Protect nsmd from vanishing from the internet one day.
* The goal is to become the de-facto standard music session manager for Linux distributions

Changes since new-session-manager v1.3.2 (2020-06-20)

Code: Select all

Add documentation and manpages.
Highlight: Provide updated API-Document (core documentation) on
https://linuxaudio.github.io/new-session-manager/api/index.html

Legacy-GUI:
 Overhaul look and feel.
 Rewrite labels and buttons with unambiguous descriptions.
 Protect text-input dialog windows from empty strings, like "Add New Client" or "New Session"
 Scale icons, support more icon formats.
 Show all icons and buttons when attaching to a running nsmd session
 Various small fixes.
 Always show correct session name, no matter how the session was loaded or how the GUI was started

nsmd:
 NSM_API_VERSION_MINOR from 0 to 1 (1.0 -> 1.1)
 Repair nsmd to correctly send client data when running headless and a GUI announces later.
 ClientId generation now prevent collision with existing IDs.
 nsmd command line option --load-session to directly load one (Berkelder, Rik)
 Better detection of clients that failed to launch leads to faster session startup (by 5 seconds)
 Users get informed by client-label if an executable is not present on the system or permission denied
 Fixed reply for listing sessions from a plain "Done." to proper reply path with empty string as terminal symbol  "/reply", "/nsm/server/list", ""
 Fix operation reply to last treated client instead to reply to sender (Picot, Mathieu  / houston)
 /nsm/gui/session/name send consistent session name/relative-path pair to the annouced GUI, no matter how the session was loaded.

nsm.h
 :optional-gui: support to nsm.h, for other applications to include and use. (Meyer, Hermann / brummer )

This is a joint release from multiple people under the linuxaudio.org "brand".

https://github.com/linuxaudio/new-session-manager


Greetings,

dvzrv, falktx and nils

[studio32]

Something about NSM

[raboof]

Why the name changes in the menus of the FLTK GUI? Makes very little sense to me, to change the original ones (duplicate, abort). You've 4 buttons now starting with save?

Are you trying to start a discussion with the New Session Manager authors, or a general forum discussion?

If you want to have a discussion with the New Session Manager authors, the project issue tracker at https://github.com/linuxaudio/new-sessi ... ger/issues might be a better place.

If you want to have a general forum discussion, it would be good to give some more context: what were the old names, what are the new names, why do you think the new names are less good, and what would you like to discuss about this here?

As an aside: I understand that you are frustrated by some of the changes in this forked version of a piece of software you are obviously passionate about. Please, let's work to keep this discussion friendly, even when disagreeing.

[studio32]

Something about NSM

[tramp]

Original non-session-manager vs nsm-legacy-gui:

quit / refresh / open / close / abort / save / new / duplicate

quit / refresh / new session / save / save & close / save & dupl. / save & open / close without saving

For me this is a clear example of solving a non-issue by making the situation more complicated.

It would be better for the user to go back to the original naming.

No, it wouldn't. For example, "abort", this is usually used for stopping a running process, like a download, not for close a document (here the session) of a application.
"close" in the orginal GUI do save and close, again, that isn't conform with the usual Desktop application guidelines, were close means exactly that, just close. It's okay to ask if the user want to save changes, even if I found that annoying.
So, you don't see a issue because you are used to use the ordinal GUI, but others may be offended by the naming scheme of it. (like me)

Debian could ship just Argodejo with new-session-manager nsmd, without the 'NSM-legacy-GUI'.

Wasn't it you who do the request on https://github.com/original-male/non/issues/280 to do exactly what is done now, just that you never receive a answer from Jonathan.
Also, as far I know you've not so much experience with what debian will ship and what not, and why. So, it may be better to leave that decision by those who do the work.
Still, no-one hinder you to use the original non-session-manager with all it's unfixed mem-leaks, no-one hinder you to promote the use of the original version, but, I wonder why you are so furious about this evolution. You've worked yourself on bringing this tool into debian, and now, it's on the best way to make that true.

[studio32]

Something about NSM

[sysrqer]

If my 2 cents are worth anything the new options make sense to me and seem to be clearer.

[nils]

All changes to the legacy GUI have been made because it was 1) obviously a bug (showing the wrong session title, not showing program icons, accepting invalid input for sessions and clients, such as empty strings) 2) after testing with real world users, some of them never used a session manager, some never audio programs, some never a computer (only phones). Testgroup size was around 40, additionally to people that have something to say on their own (like tramp here)

Exactly these tests prompted me to write Agordejo in the first place (the program launcher as the centerpiece), but falktx and others voiced the wish to use the old GUI, so we fixed and improved some issues here as well. Some of them were pure side effects of wrong nsmd code (from the above: wrong session titles, no icons) some were actual GUI-code problems.

I previously said (in a deleted thread) something like "original NSM gui is basically untested". studio32 protested.
Explanation: The code and UI design clearly shows that one particular workflow is well establish and basically bug free. My guess is that this is the one the original developer Jonathan most likely used. If you diverge from that particular way, by alternatives described by the original NON-manual, you will encounter more and more bugs and problems. Biggest culprit so far has been attaching the GUI to a running server or a new server to a running GUI (same local machine), which are usecases that are treated by the NON website as "unique selling points". But also mentioned issues with the icons can be traced down to that narrow testing behaviour (doing the same test for 10 years is not real testing). The original author Jonathan mentioned himself, over the years, that he uses NSM to just run the same setup of programs all the time (sorry, no source for that now).

At this point the code of both the nsmd-server as well as the legacy-gui are faster, more robust and with less bugs while keeping 100% compatibility. If one would look at the code changes one can see that this has been made through surgical and minimal changes .
These are exactly the fixes and changes that have been submitted for years but were either ignored or met with verbal abuse by the original author. (see non bug tracker)

No amount of shit-talking (to put it mildly(!)) and claiming the opposite will change that. Stating the opposite such as here (website change from 2020-07-14, day of the recent new-session-manager release) are concious, plain lies:

They have contributed nothing to this project.

I personally have worked with and for NSM at least since 2015. I submitted (accepted) code patches, documentation patches, many bug reports. I kept the list of supported applications up to date. I wrote a library that makes NSM support for Python-programs easier (that is linked on the official NSM page) . Additionaly the non-quantifiable work of educating people online and offline (OSAMC, Sonoj Convention) which lead not only to user support but also client developers and other resources. The same is true to some extend for falktx and dvzrv. Saying we have "contributed nothing" is a lie and an insult.

They have offered nothing (not so much a good idea).

Since the text above is already so long and I already linked so much here is just one example of a "good idea". Even studio/grammoboy2 said so himself (see link)

And in case "offered nothing" was meant as "How can be improve the situation in general?": I suggested to NON around February to take my ideas that are unwanted in upstream NSM and create a new GUI (which turned out to be Agordejo) instead which works around problems but will run on the original nsmd-server without conflicting packages or files. This idea was met with strong opposition by the original author (no sources because the comments have been deleted on github or were in private e-mails) and I was called names and was insulted for wanting to develop my own GUI in a way that was suggested by the Non Website itself: "Session Management Daemon Controllable via OSC. Don't like the NSM UI? You can build your own and communicate with the server via a simple OSC protocol."

The quality of their implementations are likely as poor as the morality of their "authors".

As written above that is not true. Read the code and you will see that most bugfixes are very small changes, sometimes just even switching the position of two existing lines of code so that NSM-messages arrive in the correct order ("correct" is defined here as described by the NON-API document)

They didn't even ask to take over maintainership.

We did not, so much is true. But would you do that, dear reader, if for years your suggestions had been met regularly with being called (moderation note: removed some nasty words)... I thought so...

Their projects are not made in the spirit of free software

New Session Manager and Agordejo are released under a GPL license. All fixes and improvements are documented through the git history and can be copied over to the original Non Session Manager in bulk or cherry picked. We encourage others to work on the code, report problems or give general feedback, such as people offering to proof-read the API document, which happened for our version of the document, which corrected typos and grammatical errors. That sounds pretty much like "free software" by FSF-definition and in spirit to me.

[studio32]

something about NSM

[tramp]

deleted because not completed

[tramp]

edit] Actually, this recent comments about the FLTK GUI here are because someone asked me what my plans where for non-daw in Debian and how I saw the situation. I told him I don't care anymore, but that I think I would add Argodejo with the nsmd of new-session-manager without the nsm-legacy-gui (especially now they changed the menu naming) to Debian and then making it possible for Debian users to both install Argodejo from the Debian repos with nsmd of new-session-manager and build and install non-daw with non-session-manager including the nsmd of non-session-manager, so that the non-daw developer doesn't get bugreports from new-session-manager. This way it's still possible to package the comlete non-daw software suite for Debian later. Does that make sense? I don't know. I've always been a proponent of a less dramatic fork (if needed), where nsmd of non-session-manager was in it's own repository (maintained by Nils? still closely related to the original project and the original author). This whole new situation, (which looks to me at least partly, like a act of revenge towards the non author, presented with a nice flavor of 'community sauce') adds complexity to the ecosystem I guess. But that's not up to me to solve.[/edit]

See, who is it who don't listen? You've yourself contacted debian maintainers to try to get NSM into debian, you fail. But still you believe how to handle that and what is the best way. I#st that silly? Your way fail, so now let's try a other way.

Of course some bugs will be fixed, but don't pretend that the non-session-manager is a heavily misbehaving application sucking system resources, it's hard to find lighter applications then the non-daw.

It's easier then you guess, but have nothing to do with the current discussion. Point is, when, a developer, presented to me, as a developer, a implementation point, to support his work, and, this point is already spiced with mem-leaks, I, as a developer loss very far the interest to implement support for that feature. Even more when I realise that the main application is spiced even more with mem-leaks. More harder will it come, when the developer of the "to supported feature" refuse bugfixes. No one say, that the idea behind non-session-manager is bad, the opposite is the case. But it really needs a responsible maintainer in order to make it a standard across distributions.

I'm repeating myself, but NSM support in applications is the only real thing that helps me. In fact, the recent mamba virtual midi keyboard with NSM support helped me more then all that work on the fork (which helped me nothing so far, seriously). (Yes, now you tell me that it's the effect of the fork, so keep forking I guess... but I find it somewhat childish to not support NSM in your application when you don't like the designer of it, but that's a other discussion.

see my stance before.

As a side note, I've nothing to do with this fork, but, I anticipate from it, and have to say thanks to nils for getting the eggs and just do it.