Re: [gentoo-dev] Upcoming repoman check will fail on large files in the tree

[Robin H. Johnson]

On Fri, Jul 30, 2004 at 05:07:27AM +0100, Ciaran McCreesh wrote:
> On Fri, 30 Jul 2004 04:22:29 +0100 Ciaran McCreesh <ciaranm@g.o>
> wrote:
> | On Fri, 30 Jul 2004 12:15:49 +0900 Georgi Georgiev <chutz@×××.net>
> | wrote:
> | | What about not allowing binary files at all?
> |
> | That'd break ChangeLog.
No it won't.

> Ok, since I've been asked about this several times on IRC... Some people
> have names which include non-[A-Za-z] characters.
UTF8 != binary.
UTF8 is fine, it's specifically designed to co-exist with ASCII/ISO-8559-1.
As such, CVS handles UTF8 perfectly well (UTF16 is a different matter).

I'm in favour of removing all properly binary files from CVS, to get the
checkout size down. Brian's earlier list of files over 20k comprises
more than 9Mb of files, and that's a not even a complete list.

Could we please consider removing sys-apps/portage/files/portage-rescue*
from the tree?
These files are there, and comprise relatively large chunk of space.
portage-rescue-2.0.44-ppc.tbz2
portage-rescue-2.0.48-r1-sparc64.tbz2
portage-rescue-2.0.49-r3-alpha.tbz2
portage-rescue-2.0.48-r1-sparc.tbz2
portage-rescue-2.0.49-r15-amd64.tbz2

The x86 portage rescue has already gone, and the others should join it.
Leaving an README.RESCUE would be prudent.

> Incidentally, when are we going to officially standardise on UTF-8 for
> ChangeLog and ebuilds?
There is absolutely nothing I'm aware of stopping you from using UTF
already in ChangeLogs and ebuilds, barring some common sense
limitations, like bash not accepting non-ASCII for the names of
variables and functions.

-- 
Robin Hugh Johnson
E-Mail     : robbat2@××××××××××××××.net
Home Page  : http://www.orbis-terrarum.net/?l=people.robbat2
ICQ#       : 30269588 or 41961639
GnuPG FP   : 11AC BA4F 4778 E3F6 E4ED  F38E B27B 944E 3488 4E85