| 26 Apr 2025 |
 K900 | Sphinx the tool can also parse numpy style comments so it's really whatever | 06:29:32 |
 Rutile (Commentator2.0) | In reply to @jade_:matrix.org
re frustrations with pytest, those still exist, but i don't think it is necessarily that big of a hassle to port the test suite to a different one when we have the time to write a better one.
the fixtures in pytest are almost good, the only thing i want different from them is that they should be driven directly by type annotations
the big bad thing about pytest is xdist related jankiness, but i also doubt it is going to wind up ultimately being That Bad. the ideal testsuite would have threaded or asyncio running instead of the xdist multiprocessing monster, but it's kinda like whatever. Iirc django provides that, but i wouldnt recommend using django here just for tests | 06:50:40 |
| Rutile (Commentator2.0) | In reply to @k900:0upti.me
Sphinx the tool can also parse numpy style comments so it's really whatever Sphinx is the style most ides default to afaik, and imo it is the easiest to write from hand correctly | 06:54:21 |
| K900 | In reply to @commentator2.0:elia.garden
Iirc django provides that, but i wouldnt recommend using django here just for tests Django tests are also effectively xdist | 06:54:50 |
| K900 | Except even more cursed because it spins up multiple database instances | 06:54:59 |
| K900 | There's https://github.com/willemt/pytest-asyncio-cooperative for true async tests but it's giga cursed because pytest is REALLY not designed for this | 06:55:39 |
| K900 | And if your tests ever race, good luck figuring that out | 06:55:52 |
| Rutile (Commentator2.0) | Yeahhhh, and all the other stuff it does to make it work with its model api | 06:55:54 |
| K900 | IMO (sadly) xdist is as uncursed as we're going to get until someone writes an actually async test runner that doesn't sick | 06:57:33 |
| K900 | *suck | 06:57:36 |
| Rutile (Commentator2.0) | while we are almost on topic already;
helle (just a stray cat girl) can you reply/clarify this? https://git.lix.systems/lix-project/lix/issues/601#issuecomment-10352 | 07:11:29 |
| K900 | @helle (just a stray cat girl) do you want a co-authored-by on the fucked up ssh thing | 07:47:09 |
 helle (just a stray cat girl) | In reply to @commentator2.0:elia.garden
while we are almost on topic already;
helle (just a stray cat girl) can you reply/clarify this? https://git.lix.systems/lix-project/lix/issues/601#issuecomment-10352 I will get to a pc soon and reply, sorr(y | 09:45:40 |
| helle (just a stray cat girl) | the mails that a reply was made must have slipped through probably my own mental filter | 09:46:13 |
 griff | When was the whole SSH "echo started" introduced since that would probably break my setups but hasn't yet? | 10:49:10 |
| K900 | A few days ago | 10:50:56 |
| helle (just a stray cat girl) | In reply to @k900:0upti.me
@helle (just a stray cat girl) do you want a co-authored-by on the fucked up ssh thing ah, what is the current state on that? | 10:51:30 |
| K900 | We have it | 10:51:38 |
| K900 | I think | 10:51:39 |
| griff | OK. How does it interact with authorized_keys command like NixOS nix.sshServe uses? | 10:52:57 |
 KFears (burning out) | In reply to @jade_:matrix.org
it's a multi-language python project, really Nitpicking, but it's a pain to work with multi-language stuff :( | 10:56:08 |
| KFears (burning out) | I still don't quite understand what value xonsh brings over like a function wrapper in Python to run commands | 10:56:38 |
 raitobezarius | In reply to @griff79:matrix.org
OK. How does it interact with authorized_keys command like NixOS nix.sshServe uses? That's a good question | 10:57:01 |
| griff | Also why is it needed? Both daemon protocol and legacy worker protocol start by writing a magic number and reading a magic number. Can't SSH code just handle this magic and then fake the magic for what connection is handed to. | 11:12:59 |
| helle (just a stray cat girl) | so uuuh, RemoteCommand runs before the main command, so uh, let me get this straight, the intent is that we get a bash shell, guaranteed and then just pipe over stdin to it the echo that we want and then pipe into it exec command | 11:14:09 |
| helle (just a stray cat girl) | Redacted or Malformed Event | 11:15:51 |
| helle (just a stray cat girl) | glares at matrix | 11:23:22 |
| helle (just a stray cat girl) | I am still curious what makes -oRemoteCommand preferable over just ssh with the command being bash tacked on the end, because iirc it is the same, just going for a less well known syntax | 11:25:48 |
| helle (just a stray cat girl) | In reply to @raitobezarius:matrix.org
That's a good question authorized_keys command btw just overrides RemoteCommand just as it does for a command tacked at the end, but now the command passed in ${SSH_ORIGINAL_COMMAND} will just contain "bash" instead of the actual stuff we wanted to execute | 11:31:40 |
| K900 | It fucks with pipes somehow | 11:42:06 |
