[Pkg-mailman-hackers] Re: mailman_2.1.8-1_sparc.changes REJECTED

Lionel Elie Mamane lmamane at debian.org
Wed May 10 11:44:04 UTC 2006


On Wed, May 10, 2006 at 12:33:11PM +0200, Joost van Baal wrote:
> Op wo 10 mei 2006 om 07:44:24 +0200 schreef Lionel Elie Mamane:
>> On Sun, Apr 16, 2006 at 05:43:51PM +0200, Lionel Elie Mamane wrote:
>>> On Sun, Apr 16, 2006 at 08:17:09AM -0700, Debian Installer wrote:

>>>> Rejected: mailman_2.1.8-1_sparc.deb: old version (0:2.1.7-2.1.8rc1-1) in unstable >= new version (2.1.8-1) targeted at unstable.
>>>> Rejected: mailman_2.1.8-1_sparc.deb: old version (0:2.1.7-2.1.8rc1-1) in testing >= new version (2.1.8-1) targeted at unstable.
>>>> Rejected: mailman_2.1.8-1.dsc: old version (0:2.1.7-2.1.8rc1-1) in unstable >= new version (2.1.8-1) targeted at unstable.
>>>> Rejected: mailman_2.1.8-1.dsc: old version (0:2.1.7-2.1.8rc1-1) in testing >= new version (2.1.8-1) targeted at unstable.

>>> That's all completely wrong:

>>> dmaster at bagnat:~$ dpkg --compare-versions 0:2.1.7-2.1.8rc1-1 lt 2.1.8-1 && echo yes || echo no
>>> yes

>>> According to policy, an explicit zero epoch is exactly the same as no
>>> epoch.

>> I downloaded katie and traced the bug back to libapt-pkg (via
>> python-apt), and submitted a patch, which got applied. Version 0.6.44
>> solves this error.

> http://ftp-master.debian.org/reject.html says:

> ----

> Rejected: foobar_0.8.1-5_i386.deb: Old version `0.8.1-6' >= new version
> `0.8.1-5'.

> Problem
>     Every new upload to the archive must have a greater version number
>  than existing versions in the target suite (stable, unstable).
> Solution
>     Increase the version number (with an epoch if necessary).

> ----

> So perhaps there's no easy way to prevent the mailman package to go
> carry 1:2.1.8-1: I suggest to upload your 2.1.8 package as
> 1:2.1.8-1.

The point is that I _did_ increase the version number, but katie
_wrongfully_ thinks I did not.

Anyway, I uploaded 2.1.8-1 as 0:2.1.8-1 on 2 May.

-- 
Lionel



More information about the Pkg-mailman-hackers mailing list