Re: rpm's treatment of unversioned provides

2011-02-21 Thread Petr Pisar
On 2011-02-21, Paul Howarth p...@city-fan.org wrote:
 RPM traditionally treats unversioned provides as meaning any version. 
 Over on perl-devel list, it's been suggested that this is a bug in rpm.

 Googling around, I can't find any specific rationale for why rpm does 
 this as opposed to say providing version 0. Can anybody enlighten me?

The full story begins on
https://bugzilla.redhat.com/show_bug.cgi?id=672246#c7.

I'm really interrested why RPM dependecny solver behaves like Paul says
and what it is good for.

-- Petr

-- 
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel


RE: rpm's treatment of unversioned provides

2011-02-21 Thread pinto.e...@gmail.com
Perhaps, should be most useful to post question as this, interesting as they 
are, on the rpm mailing list. Just an opinion. Regards
-Original Message-
From: Petr Pisar
Sent:  21/02/2011, 16:43 
To: devel@lists.fedoraproject.org
Subject: Re: rpm's treatment of unversioned provides


On 2011-02-21, Paul Howarth p...@city-fan.org wrote:
 RPM traditionally treats unversioned provides as meaning any version. 
 Over on perl-devel list, it's been suggested that this is a bug in rpm.

 Googling around, I can't find any specific rationale for why rpm does 
 this as opposed to say providing version 0. Can anybody enlighten me?

The full story begins on
https://bugzilla.redhat.com/show_bug.cgi?id=672246#c7.

I'm really interrested why RPM dependecny solver behaves like Paul says
and what it is good for.

-- Petr

-- 
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel

-- 
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel


Re: rpm's treatment of unversioned provides

2011-02-21 Thread Michael Schroeder
On Mon, Feb 21, 2011 at 02:51:11PM +, Paul Howarth wrote:
 RPM traditionally treats unversioned provides as meaning any version. 
 Over on perl-devel list, it's been suggested that this is a bug in rpm.
 
 Googling around, I can't find any specific rationale for why rpm does 
 this as opposed to say providing version 0. Can anybody enlighten me?

I think it is for symmetry reasons:

  Requires: foo require any version/release of foo
  Requires: foo = 1 require version 1 of foo, any release
  Requires: foo = 1-1   require version 1 of foo, release 1

  Provides: foo provide any version/release of foo
  Provides: foo = 1 provide version 1 of foo, any release
  Provides: foo = 1-1   provide version 1 of foo, release 1
  
Also, if it always provides version 0 there would be no way to
tell it to provide all versions. So it's more flexible the
way that it is.

(Yes, Debian is different in that regard: versioned requires
never match unversioned provides for them. But they also don't
support an any release matcher.)

Cheers,
  Michael.

-- 
Michael Schroeder   m...@suse.de
SUSE LINUX Products GmbH, GF Markus Rex, HRB 16746 AG Nuernberg
main(_){while(_=~getchar())putchar(~_-1/(~(_|32)/13*2-11)*13);}
-- 
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel