[pacman-dev] additional (small) pacman issues

Dan McGee dpmcgee at gmail.com
Sun Mar 18 22:43:27 EDT 2007

On 3/18/07, Scott Horowitz <stonecrest at gmail.com> wrote:
> I've been somewhat systematically going through the various less
> commonly used options, and there are some issues. Most of these are
> smaller than the ones James has previously described ;)
> 1. pacman -S --noprogressbar seems to add some extra "done"s  and is
> missing some other output like (1/2) next to the installed package.

Extra done text is fixed (locally) and will soon be in CVS. The (1/2)
stuff is done with the progress bar, so that isn't going to show up
now with --noprogressbar. That was an addition since pacman 2 anyway,
so no one should care.

> 2. pacman -S --noscriptlet -> should this be a permanent action? In
> other words, if I pacman -S --notscriptlet foo and then pacman -S foo
> (or pacman -Su when foo is updated), should the script run the second
> time? I can understand it working either way: 1) if you install with
> noscriptlet, you never want the script to run, or 2) if you pacman -S
> a second time without specifying noscriptlet, you'd want the script to
> run. The docs aren't explicit enough for me to understand. Maybe -S/U
> foo should be handled differently than pacman -Su.

This is something that needs further thought. I don't really like the
concept of --noscriptlet all that much, but that's a different story.

> 3. How does pacman -Qc work? I've tried it on numerous packages and
> have to find one with a changelog displayed by pacman.

We can't put changelogs in yet, because pacman 2 will install them in
the root directory as .CHANGELOG, which you probably don't want.
Perhaps we should do this now to solve this problem in the future- any
file in the root directory of the archive that begins with '.' should
not be installed? This would allow for extensibility once everyone is
on the pacman3 bandwagon.

> 4. The following are missing from the man page: --noscriptlet, --ask, -Qc, -Se

Added where necessary. --ask seems stupid, it looks half-baked and
half-implemented in the code. My new manpage entry says as much. -Se
isn't an option at all, I think...

> 5. "--noconfirm      do not ask for anything confirmation" -> that
> should probably read "do not ask for any confirmation"

Fixed, and I even updated it in the PO files.


More information about the pacman-dev mailing list