On 08/26/09 at 01:35pm, David C. Rankin wrote: > On Wednesday 26 August 2009 03:19:30 am Nathan K. Bathory wrote: > > > > > > pacman -Q obviously not the answer, but whilst on that - is it > > > possible to do a pacman -Q, save the output somewhere, then on > > > another machine, just reverse the process? Would save me heaps of > > > time putting a system back together as happened last week when i had > > > to replace my HDD, taken me ages to 'remember' everything I had on it. > > > > > > > > > Richard > > > > To save the download time, if the machines are the same architecture: > > > `pacman -Qq > pkg.list` > > then on the other machine > > (as root) rsync -uav username@oldmachine:/var/cache/pacman/pkg > /var/cache/pacman > > > `sudo pacman -S $(cat pkg.list)` > > > > If you rsync (or scp or move with a usb stick) the package files from the > oldmachine to the newmachine before issuing the "sudo pacman -S $(cat > pkg.list)`, then the install uses the local files and avoid the time required > to download them. You can easily do a 10 minute install that way. > as a small aside; after being corrected a few times on bbs, i now know that pacman -Qqe | grep -vx "$(pacman -Qqm)" > pkg.list and cat pkg.list | xargs pacman -S --noconfirm --needed work the best for this. list out only explicitly installed minus any foreign packages; then only install those that are needed (pull in your own deps). -- patrick brisbin
