Am Montag, 17. September 2007 schrieb Jason Chu:

On Mon, Sep 17, 2007 at 01:11:45AM +0200, Thomas Bächler wrote:

...

I am CC'ing this to the tur-users mailing list for reference, as the following applies to official repositories, as well as the community repository:

I am planning to add all packages in core/base to the 'base' group and all packages in core/devel to the 'base-devel' group. The first allows quick installations in different roots with 'pacman -S base', the second will install a complete build environment for makepkg. This has been partially done already, but should be completed. In the next few days, all packages in core/base and core/devel will be rebuilt.

This rebuild comes with a rule: No package outside of the 'core' repository may have group membership in 'base' or 'base-devel'.

Thank you and good night Thomas Hi i think i forgot to add csup to core/devel it is needed for abs or am i wrong here?

greetings tpowa -- Tobias Powalowski Archlinux Developer & Package Maintainer (tpowa) http://www.archlinux.org tpowa@archlinux.org

On 9/17/07, Roman Kyrylych wrote:

2007/9/17, Tobias Powalowski :

...

Hi i think i forgot to add csup to core/devel it is needed for abs or am i wrong here?

You are correct.

This seems like a good candidate for devel/, since you don't really need it to run a system... -Dan

On 9/17/07, Roman Kyrylych wrote:

2007/9/17, Dan McGee :

...

On 9/17/07, Roman Kyrylych wrote:

...

2007/9/17, Tobias Powalowski :

...

Hi i think i forgot to add csup to core/devel it is needed for abs or am i wrong here?

You are correct.

This seems like a good candidate for devel/, since you don't really need it to run a system...

fakeroot too ;-) it's not a dependency of pacman and for makepkg to be useful you'll need base-devel anyway.

That kind of defeats the purpose of a split then, doesn't it. I understand your point, but it doesn't justify it not living in devel/. fakeroot should probably go there as well. And although I'd support it, did we finally dump cvsup? -Dan

Am Mon, 17 Sep 2007 07:00:00 -0500 schrieb "Dan McGee" :

And although I'd support it, did we finally dump cvsup?

-Dan

csup does everything we need for abs. and it's much faster than cvsup. everything csup can't do can be done with cvs. so we indead can dump cvsup. if someeone still finds it usefull we can put it into AUR. otherwise to trash. Andy

I'd vote +1 for putting cvsup into unsupported.

My only complaint, and it may be easily fixable- csup shows "updating from <ip addr>" instead of "updating from <hostname>" messages. Any way to change it back to the hostname (cvsup) behavior?

Dan, Try putting `options attempts:4` into your resolv.conf file, and see if that helps at all.

On 9/17/07, eliott wrote:

...

I'd vote +1 for putting cvsup into unsupported.

My only complaint, and it may be easily fixable- csup shows "updating from <ip addr>" instead of "updating from <hostname>" messages. Any way to change it back to the hostname (cvsup) behavior?

Dan,

Try putting `options attempts:4` into your resolv.conf file, and see if that helps at all.

Nope, and that wouldn't make a whole lot of sense since we already did a hostname -> IP translation: $ cat /etc/abs/supfile.core # # /etc/abs/supfile.core # # this is the host containing the core PKGBUILD files *default host=cvs.archlinux.org ... $ abs Connected to 66.211.213.17 Updating collection core/cvs Edit core/base/ncurses/PKGBUILD ... -Dan

On 9/17/07, eliott wrote:

...

Nope, and that wouldn't make a whole lot of sense since we already did a hostname -> IP translation:

I was thinking maybe a reverse lookup was timing out... but it appears that the IP you listed reverses to velocity.net anyway...

Interesting..I ratched up the verbosity output on csup, and this is what I get..

Parsing supfile "/etc/abs/supfile.core" Connecting to cvs.archlinux.org Connected to 66.211.213.17 Server software version: SNAP_16_1h Negotiating file attribute support Exchanging collection information Establishing multiplexed-mode data connection Running Updating collection core/cvs Shutting down connection to server Finished successfully

Some gaolergle'ing turned up this: http://lists.freebsd.org/pipermail/cvs-projects/2006-October/000602.html http://www.freebsd.org/cgi/cvsweb.cgi/projects/csup/proto.c.diff?r1=1.92;r2=... so...I pulled a more recent snapshot of csup..and now I get this..with L 1 verbosity (default).

Connected to cvs.archlinux.org (66.211.213.17) Updating collection core/cvs Finished successfully

Seems pretty win. PKGBUILD for my change can be found here... http://archlinux.org/~eliott/zomgstuff/csup/PKGBUILD the snapshot pull from freebsd is here: http://archlinux.org/~eliott/zomgstuff/csup/csup-snap-20061014.tgz

Roman Kyrylych schrieb:

...

This seems like a good candidate for devel/, since you don't really need it to run a system...

fakeroot too ;-) it's not a dependency of pacman and for makepkg to be useful you'll need base-devel anyway.

I am rather shocked that neither fakeroot nor make are in devel, but rather in base. diffutils are in base, but patch is in devel, which doesn't make sense either.

Thomas Bächler schrieb:

I am shuffling some stuff around in the core cvs now, I will move some stuff from base to devel for a better distinction. I am doing this in the repository directly (hail CVS), so you may have to delete some directories and check them out again. I will post again when I am finished.

Okay, this is what I did: removed base/diffutils (tpowa had already commited it to devel) moved ed, fakeroot, libtool and make from base to devel removed base/unifdef and base/devfsd Delete those directories from your working copy and check them out again. We should also remove fakeroot from pacman as a dependency, as pacman -S base-devel will install a full build environment.

Thomas Bächler wrote:

I did some of the rebuilds already, will continue later. Some of core/base and all of core/devel needs to be rebuilt.

I will also move the new util-linux (which is actually util-linux-ng) and nfs-utils from testing to core later, unless there are objections.

Can this be renamed to util-linux-ng as that's it's official name. Andrew

Thomas Bächler wrote:

Andrew Fyfe schrieb:

...

...

I will also move the new util-linux (which is actually util-linux-ng) and nfs-utils from testing to core later, unless there are objections. Can this be renamed to util-linux-ng as that's it's official name.

It's easier to not change the name for now. If util-linux ever releases a new version, we can rename the package to avoid confusion, but currently it looks like this is not happening.

+1 for using the official name, as that's what we do in Arch, and that's also what we tell users to do when creating packages. I'd prefer the right way to the easy way. T.

Thomas Bächler wrote:

Tom K schrieb:

...

+1 for using the official name, as that's what we do in Arch, and that's also what we tell users to do when creating packages.

I'd prefer the right way to the easy way.

T.

Okay, we'll rename when moving it to core (which I will do when I finished the rest of the base and devel rebuilds).

Now that it's in core, the expected confusion is arising on IRC etc, so I've put an announcement about it in the news.

On 9/20/07, Pierre Schmitz wrote:

Am Donnerstag, 20. September 2007 01:22:20 schrieb Thomas Bächler:

...

I will remove the fakeroot dependency from pacman. A post-install message will tell the user to install 'base-devel' to be able to build packages with makepkg.

Doesn't makepkg itself warn about missing fakeroot when trying to build as user?

Yes, it does. I had to ask Dan. IIRC we slated this for removal, making fakeroot and non-root builds mandatory, instead of just suggested, but it is still in in the current incarnation.