i kind of bumbled through this by copying the joplin-desktop and standardnotes derivations

the gsettings-desktop-schemas seems to be needed to prevent crashes when uploading profile pics

seems to run ok but there are warnings.

any help debugging would be appreciated!

this on boot:

(process:20493): Gtk-WARNING **: 22:25:30.050: Locale not supported by C library.
        Using the fallback 'C' locale.
{ path: '/home/thedavidmeister/.ssb',
  party: true,
  timeout: 0,
  pub: true,
  local: true,
  friends: { dunbar: 150, hops: 2 },
  gossip: { connections: 3 },
  connections: 
   { outgoing: { net: [Array] },
     incoming: { net: [Array], ws: [Array] } },
  timers: { connection: 0, reconnect: 5000, ping: 300000, handshake: 5000 },
  caps: 
   { shs: '1KHLiKZvAvjbY1ziZEHMXawbCEIM6qwjCDm3VYRan/s=',
     sign: null },
  master: [],
  logging: { level: 'notice' },
  port: 8008,
  blobsPort: 8989,
  server: true,
  _: [],
  host: '::',
  ws: { host: '::1', port: 8989, scope: [ 'device' ], transform: 'shs' },
  keys: 
   { curve: 'ed25519',
     public: '8VYZW+U0zbewBQSaKpi1fCTD2cPEbkKy8B6amcy06IY=.ed25519',
     private: null,
     id: '@8VYZW+U0zbewBQSaKpi1fCTD2cPEbkKy8B6amcy06IY=.ed25519' } }
ssb-friends: createFriendStream legacy api used
git-ssb is not installed, or not available in path
Error: listen EADDRINUSE :::8989
    at Object._errnoException (util.js:1024:11)
    at _exceptionWithHostPort (util.js:1046:20)
    at Server.setupListenHandle [as _listen2] (net.js:1351:14)
    at listenInCluster (net.js:1392:12)
    at Server.listen (net.js:1476:7)
    at EventEmitter.emitter.listen (/home/thedavidmeister/.cache/appimage-run/9811e63b0395fab7bf604f9dcd63c60ad91457d1378881365e26ae2aa8c159b4/squashfs-root/app/resources/app.asar/node_modules/pull-ws/server.js:50:14)
    at Object.server (/home/thedavidmeister/.cache/appimage-run/9811e63b0395fab7bf604f9dcd63c60ad91457d1378881365e26ae2aa8c159b4/squashfs-root/app/resources/app.asar/node_modules/scuttlebot/node_modules/multiserver/plugins/ws.js:20:31)
    at Object.server (/home/thedavidmeister/.cache/appimage-run/9811e63b0395fab7bf604f9dcd63c60ad91457d1378881365e26ae2aa8c159b4/squashfs-root/app/resources/app.asar/node_modules/multiserver/compose.js:96:29)
    at /home/thedavidmeister/.cache/appimage-run/9811e63b0395fab7bf604f9dcd63c60ad91457d1378881365e26ae2aa8c159b4/squashfs-root/app/resources/app.asar/node_modules/multiserver/index.js:45:21
    at Array.map (<anonymous>)
ssb-friends: stream legacy api used
ssb-friends: get legacy api used

then a lot of this:

(ssb-patchwork:20493): dconf-WARNING **: 22:27:31.414: failed to commit changes to dconf: GDBus.Error:org.freedesktop.DBus.Error.ServiceUnknown: The name ca.desrt.dconf was not provided by any .service files

(ssb-patchwork:20493): dconf-WARNING **: 22:27:31.431: failed to commit changes to dconf: GDBus.Error:org.freedesktop.DBus.Error.ServiceUnknown: The name ca.desrt.dconf was not provided by any .service files

(ssb-patchwork:20493): dconf-WARNING **: 22:27:31.445: failed to commit changes to dconf: GDBus.Error:org.freedesktop.DBus.Error.ServiceUnknown: The name ca.desrt.dconf was not provided by any .service files

(ssb-patchwork:20493): dconf-WARNING **: 22:27:31.446: failed to commit changes to dconf: GDBus.Error:org.freedesktop.DBus.Error.ServiceUnknown: The name ca.desrt.dconf was not provided by any .service files

@thedavidmeister I think the error you're seeing is coming from ssbc/patchwork#972

Is this "patchwork" or "patchwork-classic"? Because if it is the latter, don't mind continuing... patchwork-classic is afaik deprecated and only patchwork should be used.

If it is patchwork, I'd welcome to have this in nixpkgs!

@matthiasbeyer patchwork, i got the latest appimage

i tested 3.11.4 with nix-env -f . -iA patchwork

running this version has a lot fewer errors than newer version(s): ssbc/patchwork#972

tried adding git-ssb but it didn't help (build time vs. run time path differences?), i don't think it is needed as it is debug info (check comment thread)

moved the derivation to networking/ssb/patchwork

fixed maintainer-list.nix

don't think appimageTools is correct here as we are consuming not building an app image

thanks for the review @asymmetric i implemented what i could, how do you feel about it now?

Added a comment.

@asymmetric i don't see it 🤔

Because you defined version outside of the package's expression, you need to inherit version;, otherwise the full package name cannot be composed.

@asymmetric great, seems to work with the version on the derivation rather than the let

how you feeling about it now?

Nice!

You should clean up the commit history and make it fit this.

In short, there should probably be just one commit (maybe two, to add yourself to the maintainers), and it should look like patchwork: init at 3.11.4.

@asymmetric doesn't github provide the option to squash commits when merging the PR?

Yes, but it does not provide options for cleaning up the branch. You made several changes which should end up in different commits, but not several changes which should end up in one commit.

Also I think it's nice not to ask the maintainers to clean up your branch when merging, even if in theory they could :)

But I agree that there's a lot of things to check manually when contributing to nixpkgs.

This is the real history of the code development, rewriting it to fit someone(s) else’s aesthetics is beyond me I’m afraid

IMO “cleaning up” is a git smell, I don’t expect anyone to ever do it! but I’ll roll with this idea to get a merge approved

Would 1 commit for maintainers and 1 for everything else as suggested earlier provide sufficient dopamine for everyone? 😉

Yes two commits is fine. One for adding yourself to the maintainers list, and the other one for the package init is what would be expected. And thanks for the contribution!

The Making patches section of the manual has the exact format for packages and modules.

The manual explains it as “Make commits of logical units”. The main idea being that yes, while developing things are messy, and it's natural, they're not necessarily what's published here. It is “a smell”, but I'm not sure it's a bad smell, considering the scope of Nixpkgs, and how git, this way, becomes the history of the main changes, not the development history of separate features. This is part of CONTRIBUTING.md.

@samueldr what is the commit message for adding myself to maintainers?

When in doubt, I check the history and try to reason with the general rules.

Here maintainers: Add thedavidmeister could be fine, as an example.

Package inits and upgrades have rules that allows automation.

@samueldr kk, here it is #61530