Alad Wenter

PKGBUILD parsing will always be broken and adds a lot of complexity. If you really want to include user edits in the build process, run `makepkg --printsrcinfo` to regenerate `.SRCINFO`....

Not sure why you're using both a "compile once, distribute" local repo and `yay`. `paru` and `aurutils` (both mentioned in the `yay` README) better support this use case.

There's no reason `--noconfirm` should result in automatic replacement of packages, because this replacement might not be what you want or unexpected. `pacman` has `N` for these questions as well....

>This is a really bad idea, as sudo credentials are cached by the TTY and if you walk away from the computer while yay is running, another person can come...

That rule would allow to run `sudo yay` without password. Not sure there's much benefit there, since with the proper privilege dropping and `sudo yay` you only need one password...

> Could argue that setting sudo to nopassword could be more dangerous as it's systemwide It's probably fine if you use a separate user for that which does no other...

Only a comment thread was resolved in that link - the fix was never merged.

Doing this in `aur-search` was straightforward. The above scheme on solving `provides` could be implemented directly in `aur-sync` or in a helper like `aur-sync--provides`. That said, it might be preferable...

`v6` of the AUR RPC isn't coming anytime soon. Reopening. Test case from #1060: > For example, [satdump-git](https://aur.archlinux.org/packages/satdump-git) depends on airspyhf, but the only package satisfying this is [airspyhf-git](https://aur.archlinux.org/packages/airspyhf-git)'s provides:...

yay uses `provides` heuristics that severely increase the AUR load and have lead to rate limiting in certain cases. I do not recommend any other projects to implement them.