Le 07/02/2017 à 18:01, John D Jones III a écrit :
I'm here, I'll take a good at the firefox-nightly-fr pkgbuild and work to make the necessary adjustments. I've never particularly cared for how it just skips the sha512 sum on the tarball, but I also wasn't thrilled updating the sha512 every 1-180 minutes so I went with what was available to me at the time.
I too agree that auto-updating a PKGBUILD without some form of vetting is a bad idea. Given the number of pkgs I maintain, I'd love to let things update automatically, but my experience in these past several years suggests that there is always something that fails in the process and that nothing but good old fashioned vim/diff can properly ensure a working and clean pkgbuild.
My point was more about the fact that updating VCS-like PKGBUILD just for the sake of updating them has no purpose (whether it’s auto or not doesn’t matter here). The only case where a VCS-like PKGBUILD needs to be updated is when there is a change in build process or dependencies. It’s even more true in the case of Firefox channels, since you can quite reliably when the next build is going to be, so they are plenty of way to handle this on the user-end side. But even for -git packages for instance there is no point, and I even would like this to be against the rules for several reasons, which is something that should be discussed amongst TUs. I was particularly scared by this one for instance: https://aur.archlinux.org/cgit/aur.git/log/?h=krita-git, just take a look at the full commit message of any “Updated version” commit…
As for co-maintaining... I'm not really ready for that, but thanks anyways.
That’s your call. As long as the PKGBUILD is fine for everyone and you update it in a timely manner, that’s OK. ;) Regards, Bruno