On Thu, Aug 11, 2011 at 7:29 PM, Dan McGee <dpmcgee@xxxxxxxxx> wrote: > On Thu, Aug 11, 2011 at 12:28 PM, Dan McGee <dpmcgee@xxxxxxxxx> wrote: > > For the daring, pick your poison (by architecture): > > > > * pacman -U > http://dev.archlinux.org/~dan/pacman-4.0.0rc1-1-i686.pkg.tar.gz > > * pacman -U > http://dev.archlinux.org/~dan/pacman-4.0.0rc1-1-x86_64.pkg.tar.gz > > > > Allan, Dave, and I (and probably a few others) run pacman-git on most > > of our systems with no problems, so their should be no real shockers > > or problems if you give this a spin. There are no database upgrades or > > changes this time so downgrading later should work fine if truly > > necessary. > > > > What we're looking for feedback on: > > > > * any build failures in makepkg you may see > > * if you manage a custom repo, how does repo-add work for you > > * does pacman behave as it did before > > * if you want to sign packages, does the functionality in makepkg and > > the documentation make sense > > * same for signing repos- does it work for you > > > > What we know isn't there yet: > > * translations > > * a developer keyring (or keyring package)- if you delve into this, > > you will need to look at pacman-key for now > > * great error messages on verification failure, or ability to import > > keys on the fly if it is unknown > > > > Note that we'd love testing even if you don't plan on touching any of > > the new signing stuff- there were 500+ commits worth of changes in > > this release, including a switch to curl as the download library, so > > anything out of the ordinary should be reported. Please choose -git as > > the version in the bugtracker if you do it that way, otherwise email > > pacman-dev. > > > > Happy testing! > > Darn it- forgot to link to this if you want to see the glorious changelog: > http://mailman.archlinux.org/pipermail/pacman-dev/2011-August/014039.html > Being daring is fun. I got some rather amusing results on my work machine with this release, please see FS #25527 :) I also observed that pacman failed to remove the lock file when killed by ^C, I'll file a bug-report on this if I manage to reproduce it. Regards, Bjørn Øivind
