Allow different repodir in scripts #69
Conversation
Fixes 042010b ("tool: Refactor repodir handling") Signed-off-by: Thomas Deutschmann <whissi@gentoo.org>
This commit is basically reverting commit 042010b. Instead calling a helper function all the time we need repodir value, define/validate value once at start. Signed-off-by: Thomas Deutschmann <whissi@gentoo.org>
This will allow for running commit scripts on a different host. Signed-off-by: Thomas Deutschmann <whissi@gentoo.org>
|
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. |
| # Now that we use nattka, we need a working repodir | ||
| if config['repodir'] == "": | ||
| if os.path.isdir("/var/db/repos/gentoo"): | ||
| config['repodir']="/var/db/repos/gentoo" |
There was a problem hiding this comment.
I adopted existing coding style, see assignment in line 92 and 94 above.
|
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. |
thesamesam
left a comment
thesamesam
left a comment
There was a problem hiding this comment.
Note that get_repo_dir was a function to isolate the logic (to make it easier to find and modify) and cleanup the monolithic script file, so I don't think it was worth reverting it.
Looks fine otherwise, I'll merge it if you can fix the merge conflict. Ideally, undo the revert for aforementioned reasons.
I agree this is a good use for templates, but given I hit the same use case, I guess it's common enough to justify a config option.
3 participants
I call tatt on the x86 system where I do AT work but I do not commit from this system. Therefore I set different
repodirin tatt config. But since commit 042010b this doesn't work anymore becauserepodirvalue is now also used for nattka and must point to a valid repository.This PR will partially revert commit 042010b because it doesn't make sense to call a helper function multiple times. Instead tatt will now define
repodirif not set and validate value at start after config loading.In addition an additional option
repodir_commitswas introduced which will get used in commit-header template and defaults torepodirvalue.