On Fri, Sep 4, 2009 at 3:49 AM, Xyne<[email protected]> wrote: > On Thu, 3 Sep 2009 15:20:03 +0200 > Xavier <[email protected]> wrote: > >> I have a preliminary patch for this. I am not really happy with the C >> code I wrote for doing the substitutions, but it works. >> >> Does it make any sense to have this for -R and -U operations ? I >> implemented that but I don't really see any utility so I wanted to >> drop it. > > I think it makes sense for -R at the moment. It would also make sense > for -U if -U could handle dependency resolution. If you have working > code then I think you should just leave it in for now as it will be > good to have later. > >> Final concern : do you think this feature is useful when pacman asks >> questions and prints a lot of stuff before the actual output ? >> For many operations, the output is quite crap, and I don't know how to fix >> this. >> >> $ pacman -Rc --print gtk2 >> >> Remove (14): tango-icon-theme-0.8.1-1 midori-0.1.9-1 firefox-3.5.2-1 >> xulrunner-1.9.1.2-1 wireshark-1.2.1-1 mtr-0.75-4 >> libwebkit-1.1.10-2 libsoup-2.26.3-1 libunique-1.0.8-1 >> gconf-2.26.2-2 emacs-23.1-3 librsvg-2.26.0-1 audacious-2.1-1 >> gtk2-2.16.5-1 >> >> Total Removed Size: 279.73 MB >> >> Do you want to remove these packages? [Y/n] >> tango-icon-theme 0.8.1-1 >> midori 0.1.9-1 >> firefox 3.5.2-1 >> xulrunner 1.9.1.2-1 >> >> $ pacman -S --print base >> base package not found, searching for group... >> :: group base (including ignored packages): >> acl attr bash binutils bzip2 coreutils cpio cracklib >> cryptsetup dash db dbus-core dcron device-mapper dhcpcd dialog >> diffutils e2fsprogs file filesystem findutils gawk gcc-libs >> gdbm gen-init-cpio gettext glibc grep groff grub gzip hdparm >> initscripts iputils jfsutils kbd kernel-headers kernel26 >> kernel26-firmware klibc klibc-extras klibc-kbd >> klibc-module-init-tools klibc-udev less libarchive libfetch >> libgcrypt libgpg-error libpcap libusb licenses logrotate lvm2 >> lzo2 mailx man-db man-pages mdadm mkinitcpio mlocate >> module-init-tools nano ncurses net-tools pacman pacman-mirrorlist >> pam pciutils pcmciautils pcre perl popt ppp procinfo procps >> psmisc readline reiserfsprogs rp-pppoe sdparm sed shadow >> sysfsutils syslog-ng sysvinit tar tcp_wrappers texinfo tzdata >> udev usbutils util-linux-ng vi wget which wpa_supplicant >> xfsprogs zlib >> :: Install whole content? [Y/n] >> kernel-headers 2.6.30.5-1 >> tzdata 2009l-1 >> glibc 2.10.1-4 >> attr 2.4.43-1 >> acl 2.2.47-2 >> ncurses 5.7-2 > > > Hmmmm... that's a good point. It might seem superfluous but I think > it's good to keep that. It would enable scripts to handle the dialogue > through a pipe and then easily parse the output so I see a utility in > it. It would probably be less useful for the user directly though, but > even then it would be interesting for some dialogues to see which > dependencies some of the group members will pull in (or cascadingly > remove). Consider the DE groups for example. They should pull in other > packages that are not part of the group dialogue. > >
That does not go well with my idea that print operations should imply noconfirm though (see attached patch and commit log). I guess this requires further thinking and discussion. I also forgot to mention that I don't like the group interactivity , but that is another problem.
0001-use-noconfirm-with-print-operations.patch
Description: application/mbox
