Open Whissi opened 3 years ago
Thanks, I was going to message you and double-check what the issue remaining was.
Regarding dropping the repodir function, I did it because the tatt script is a monolith right now, but I also accept that I can't think of another use for it right now! So we can always return to it if we need to.
I run the tests in chroot, and do the commits outside or even on a different machine using a crude script inherited from @trofi and then heavily modified. I would welcome if someone comes up with a better solution, like writing the nattka commits to a script file so I can move them over or whatever.
You can easily get the same result if you overwrite your commit-header. I usually have a ~/tatt in my build chroots and symlink all files from /usr/share into that and overwrite those that need it.
Yeah, providing your own templates is another 'solution' for my problem. I don't like it or try to avoid in case templates will be updated: In this case I will probably not notice the changes. So I prefer using default templates but customize via option.
I call tatt on the x86 system where I do AT work but I do not commit from this system. Therefore I set different
repodir
in tatt config. But since commit 042010b840bd920cece9ef00d06db392f142a7b3 this doesn't work anymore becauserepodir
value is now also used for nattka and must point to a valid repository.This PR will partially revert commit 042010b840bd920cece9ef00d06db392f142a7b3 because it doesn't make sense to call a helper function multiple times. Instead tatt will now define
repodir
if not set and validate value at start after config loading.In addition an additional option
repodir_commits
was introduced which will get used in commit-header template and defaults torepodir
value.