| 26 Apr 2025 |
 Rutile (Commentator2.0) | i might genuinly tend towards having black be in the pre commit hooks/ passing conditions for commits (black is a py style formatter/linter) | 21:31:41 |
 jade_ | yes, that's the idea | 21:31:55 |
| jade_ | but i believe ruff is the better supported one these days | 21:32:04 |
| jade_ | i have an autocmd that sets syntax but not filetype to python | 21:32:33 |
 KFears (burning out) | I'm not much of a Python person at all, but... working with releng mixed between xonsh and Python felt like major pain | 21:32:33 |
| jade_ | fixes the problem | 21:32:37 |
| Rutile (Commentator2.0) | i don't have any experience with ruff and only little with black, so i can't say which is better or worse | 21:32:42 |
| KFears (burning out) | Isn't Xonsh syntax literally different from Python? | 21:33:09 |
| jade_ | it's like, Fine. the only problem with it is that you don't have IDE support in your straight line code | 21:33:13 |
| jade_ | it's a superset that only blows up language servers but the syntax highlighter doesn't care that much | 21:33:33 |
| KFears (burning out) | Not having an LSP feels very bad, though | 21:34:17 |
| jade_ | i guess, the main thing that sucks about working on the xonsh is, yes, the lack of language server support, which is why you put the complicated code in a .py file | 21:34:19 |
| KFears (burning out) | I in general don't like understand the value proposition of xonsh | 21:34:42 |
| jade_ | maybe i am too used to writing nix and large scale haskell but it's really not that bad. also ctags can help | 21:34:50 |
| KFears (burning out) | To me it feels like an annoying side thing that bites me in the ass while not bringing much useful | 21:35:02 |
| jade_ | the value proposition is it's not a shell script and it's more legible than python for stuff that is a shell script | 21:35:15 |
| jade_ | * the value proposition is it's not an actual shell script and it's more legible and writable than python for stuff that is a shell script | 21:35:32 |
| KFears (burning out) | Can't we write helper function to like run exec and pipes? | 21:35:45 |
| jade_ | also it manages environment variables better | 21:35:53 |
| jade_ | perhaps, i just don't see any value in rewriting working code to become harder to read when we have bigger fish to fry | 21:36:53 |
| KFears (burning out) | Well, it was hard to work with for me, that is all | 21:37:54 |
| KFears (burning out) | I'll try to get the local releng to the finish line, once I have it in me to continue | 21:39:13 |
| KFears (burning out) | Hope it helps to get people to work on releng | 21:40:20 |
| KFears (burning out) | Because writing code blind without any way to test it feels very bad | 21:40:48 |
| jade_ | i was testing it in staging but it is indeed quite hard to do for external contributors.
the garage ephemeral key is a remote script on the other end; you might either be able to not use it and just scatter keys all over the temporary box cuz whatever, or you can integrate it. probably easier to not use it. | 21:41:54 |
| KFears (burning out) | I've made some decent progress | 21:42:22 |
| KFears (burning out) | Like, garage ephemeral stuff is sorted out | 21:42:36 |
| KFears (burning out) | But the code hasn't turned out any good, because I basically insert if/else everywhere to check if we're running local or not and code in very different behavior | 21:43:08 |
 K900 | I was going to make it talk to a VM for this | 21:43:29 |
| K900 | We have good tools for those | 21:43:42 |
