Shouldn't the AURweb git hooks deny commits where only the PKGBUILD but not the .SRCINFO got updated? Am Sonntag, 1. August 2021 18:24:53 CEST schrieb Peter Nabbefeld via aur-general:
Am 01.08.21 um 17:47 schrieb Ralf Mardorf via aur-general:
On Sun, 1 Aug 2021 17:35:54 +0200, Peter Nabbefeld wrote: ... It's only been the first time for me I noticed a conflict between PKGBUILD and .SRCINFO - next time I will look into both files, when I notice such suspicious behaviour. ;-) On Sat, 31 Jul 2021 12:39:59 +0200, Ralf Mardorf wrote:
the very first thing to do by yourself [...] is trying to build without an AUR helper. https://lists.archlinux.org/pipermail/arch-general/2021-July/049129.html
Don't replace one AUR helper by another, if an AUR helper fails. You can build it using "makepkg". I thought, pikaur fails because of a problem with some caching; then it makes sense to check by using some other AUR helper, and aura failing finally led me to the correct point, too - but it was the last one I checked, and it took me quite some time.
BTW, I've now installed meshroom-git using the following command: pikaur -Sa --assume-installed alembic-qfix meshroom-git
After building, this results in the following strange pacman command for installation: 'sudo pacman --color=always --assume-installed alembic-qfix --upgrade --assume_installed /home/peter/.cache/pikaur/pkg/meshroom-git-2021.1.0.r134.gf1331c6a-1-x86_64.pkg.tar.zst'
That causes pacman to abort, but I could continue installing after removing "--assume-installed", i.e. using: sudo pacman --color=always --upgrade /home/peter/.cache/pikaur/pkg/meshroom-git-2021.1.0.r134.gf1331c6a-1-x86_64.pkg.tar.zst
However, that should probably be fixed ...
Regards, Peter
Regards, Ralf