AstraLuma#SwissArmyUpload update:
Recursive copy for Teahouse both up and down.
Just gotta implement GitHub oidc stuff, and it'll be ready for dogfood.
And then I can write more docs. And tests. And CI integrations. And...
today in #SwissArmyUpload , got recursive download working for both teahouse and git-pages
After I get recursive upload working, I can round out some features and start dogfooding this?
(The goal is to have a usable Action in time for the Forgejo v15 release March 26, so you can upload to Teahouse from Forgejo as soon as your instance turns on actions oidc.)
Yay! New version of #SwissArmyUpload out the door.
https://codeberg.org/teahouse/swiss-army-upload/releases/tag/v0.2.0
This adds git-pages support, so that I have known-good multi-backend handling.
It's still not really installable.
Next step is docs, packaging cleanup, and other things to make it maybe kinda usable.
git-pages - teahouse/swiss-army-upload
This adds preliminary support for [git-pages](https://git-pages.org/) and [grebedoc](https://grebedoc.dev/). This allows you to push and pull individual files. Note that this only works with the root project--if you've published a sub-site (ie, `https://{username}.grebedoc.dev/{repository}/...
milestone! swiss-army-upload v0.1 is out!
It's got none of the features I want, and it's not uploadable to PyPI because I'm using git dependencies in a few places, and I haven't tested the RPM.
But you can download it at https://codeberg.org/teahouse/swiss-army-upload/releases/tag/v0.1.0
Lunchtime check in
Integrated keyring into #SwissArmyUpload, so now it can read your username/password. Unfortunately, the 1password and BitWarden plugins for keyring both implement get_credential() badly, so I can't use them.
I'm thinking of applying for a job after lunch.
Oh, #SwissArmyUpload update from yesterday:
Got all the logic for Teahouse authentication done. It's not complicated, but it did force me to straighten out my dependency injection and sort out how Teahouse's ephemeral credentials interact with Handtruck's AWS-orientated credential tools.
(Oh, and I implemented GitHub OIDC handling while I was at it, because hyperfocus.)
Next step is credentials store, because this is an user tool and I don't want to do envvars.