Re: [gentoo-portage-dev] handling config stuff in portage (for package compression, etc)

2005-11-06 Thread Marius Mauch
On Sat, 5 Nov 2005 20:58:57 -0600
Jason Pepas [EMAIL PROTECTED] wrote:

 So, I have been going over how class config works in portage, but I
 am still unsure of where to fit in the changes I would need.
 
 I suppose I'll lay out the structure of what I had in mind and ask
 y'all for advice.
 
 Compression would be supported in a modular fashion.  The following
 config options would be supported for each type of compression:
 
 ZEXT - the compression filename extension
 TZEXT - the binary package filename extension
 ZCOMPRESS - the command used to compress a file
 ZDECOMPRESS - the command used to uncompress a file
 ZFILTER - command used to compress in a pipeline
 ZUNFILTER - command used to uncompress in a pipeline
 ZRECOMPRESS - should files already compressed using another method be
 uncompressed and then recompressed using the preferred method? (ie, if
 manpages are shipped as foo.1.gz and you want foo.1.bz2)

I really don't like this Z prefix, should rather be
PORTAGE_COMPRESS_{EXT,BINEXT,COMPRESSCOMMAND,...}
(avoid env namespace pollution a IMO a lot cleaner)

 For example, if Z=bzip2, a file would be sourced (bzip2.sh), which
 would contain the these settings:
 
 ZEXT=bz2
 TZEXT=tbz2
 ZCOMPRESS=bzip2
 ZDECOMPRESS=bunzip2
 ZFILTER=bzip2
 ZUNFILTER=bunzip2
 ZRECOMPRESS=no

Why this indirection? Just for convienience or are there technical
reasons? If it's just convienience then you don't need this, just
utilize the source command in make.conf. Anyway, there is no place for
one-letter vars in make.*

 The following subsystems would source one of these files to get
 their settings:
 
 PKG_
 DOC_
 MAN_
 INFO_
 
 possibly others.
 
 So, if PKG_Z=bzip2, bzip2.sh would be read to set PKG_ZEXT,
 PKG_TZEXT, PKG_ZCOMPRESS, etc.

Why are those vars needed? Can't they be directly derived from the
global ones?

 As these module files are sourced, individual config options could be
 overridden via values in the environment.  For example, if I wanted
 gzip used across the board, but I wanted different levels of
 compression for man pages and packages, I could do this in make.conf:
 
 Z=gzip
 PKG_ZCOMPRESS=gzip --best
 MAN_ZCOMPRESS=gzip --fast

Somehow this part looks like overengineering to me. Doesn't seem worth
to introduce 7*4=28 new vars just for this.

Marius

-- 
Public Key at http://www.genone.de/info/gpg-key.pub

In the beginning, there was nothing. And God said, 'Let there be
Light.' And there was still nothing, but you could see a bit better.


pgpw5V5iOCAyQ.pgp
Description: PGP signature


Re: [gentoo-portage-dev] branches/2.0 moved to trunk

2005-11-06 Thread Mike Frysinger
On Sun, Nov 06, 2005 at 05:40:26PM +0900, Jason Stubbs wrote:
 I've moved trunk/ to branches/2.1-experimental/ and branches/2.0 to trunk/.
 Make sure to update before doing any changes...

i thought 2.1 / trunk was dead ?  why not just punt it and be done ?
-mike
-- 
gentoo-portage-dev@gentoo.org mailing list



Re: [gentoo-portage-dev] branches/2.0 moved to trunk

2005-11-06 Thread Alec Warner
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1

Mike Frysinger wrote:
 On Sun, Nov 06, 2005 at 05:40:26PM +0900, Jason Stubbs wrote:
 
I've moved trunk/ to branches/2.1-experimental/ and branches/2.0 to trunk/.
Make sure to update before doing any changes...
 
 
 i thought 2.1 / trunk was dead ?  why not just punt it and be done ?
 -mike

IIRC, people are still backporting things... I was going to look at
confcache and porting it to 2.0... if I ever find the time and
subsequent motivation ;)
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.1 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org
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=TWpq
-END PGP SIGNATURE-
-- 
gentoo-portage-dev@gentoo.org mailing list



Re: [gentoo-portage-dev] branches/2.0 moved to trunk

2005-11-06 Thread Mike Frysinger
On Sun, Nov 06, 2005 at 09:24:58PM -0500, Alec Warner wrote:
 -BEGIN PGP SIGNED MESSAGE-
 Hash: SHA1
 
 Mike Frysinger wrote:
  On Sun, Nov 06, 2005 at 05:40:26PM +0900, Jason Stubbs wrote:
  
 I've moved trunk/ to branches/2.1-experimental/ and branches/2.0 to trunk/.
 Make sure to update before doing any changes...
  
  
  i thought 2.1 / trunk was dead ?  why not just punt it and be done ?
 
 IIRC, people are still backporting things... I was going to look at
 confcache and porting it to 2.0... if I ever find the time and
 subsequent motivation ;)

i'm backporting things to 2.0 from trunk only because i didnt know trunk was 
dead ... otherwise, i think people are backporting from savior/3.0, not from 
trunk/2.1
-mike
-- 
gentoo-portage-dev@gentoo.org mailing list