On 11/04/2016 10:20 AM, Levente Polyak wrote:
I don't think its worth trying to comfort all AUR wrappers in the way a PKGBUILD is handled. I have seen too often including horrible hacks here and there because one wrapper does or doesn't do X or Y and fails.Its not as harmful because at the end its just a different pkgver but the conclusion is the same: No gain other then to satisfy AUR wrappers. Therefor my personal opinion is to avoid that all together no matter if you think that it is overreacting or not.
That wasn't initially clear from what you said. :) I will stick with making sure the PKGBUILD is well-written then, and if I feel in the mood, commit the changes makepkg itself makes to the PKGBUILD. ;) (No real gain != loss) And I don't really see it as a "workaround for AUR wrappers". I see it as getting more useful information into the AUR rpc interface. It's similar to why I make sure to bump the pkgrel to trigger a rebuild for soname bumps -- and there are people who insist that AUR users should be required to check for that by hand. -- Eli Schwartz