- http://www.cornebise.com/julien
-
Joined on
2020-01-14
Awesome, thanks :)
Woohoo, thanks @witten for your reactivity :)
I suggest unsafe_skip_path_validation_before_create. It's a mouthful and a few characters longer than your suggested unsafe_skip_create_path_validation, but it avoids giving the impression…
Thanks! For full transparency: I don't have much merit, I've mostly prompted Claude Code precisely and verified its work. It came up with the tests based on those already in the code, so: "nice…
Make sense. I had missed that documentation page, so it had taken me a while to wrap my head around the runtime directory. I'll have to look better in the doc next time ! Done.
Make sense, done !
Makes sense ! Done.
Thanks a lot for the review @witten ! It all makes a lot of sense.
I've incorporated all of it. I suggest unsafe_skip_path_validation_before_create. It's a mouthful and a few characters longer…
@witten could you please approve the CI worfklow for the tests? They pass on my machine :)
Thanks @witten for your reactivity! I've created PR #1222 2 with a name and description that hopefully is clear enough. Let me know there if you need any change to it.
I needed to delete a range of backups that contained a set of files.
But I realise this is tricky, because each repository has different archive name with using {utcnow}.
So alternatively, and instead of having to reimplement every single borg command, how about a borgmatic command that just outputs a pre-filled borg command-line with the right setup (environment variables, repos, maybe repo passphrase in a securefile sourced with "PASSCOMMAND", etc). Then the user would just have to add the final borg commands they need.
Would that make more sense?