On Mon, 22 Dec 2008 17:12:39 +0000
Bradley Smith <bradsm...@debian.org> wrote:

> On Mon, 22 Dec 2008 16:26:52 +0000
> Neil Williams <codeh...@debian.org> wrote:
> 
> > I'm sure you're aware of the release freeze - do you have a
> > particular need for 3.0.0 in experimental?
> 
> Sure, I'm just curious as to whether a version will hit experimental
> or whether it'll just be after lenny, since I'd quite like to use
> some of the features in the newer version.

OK, I've had a look at 3.0.0 and the news is truly appalling - this
has to be the worst upstream release I have ever had the misfortune to
encounter.

I don't think 3.0.0 is going to hit Debian any time soon - either in
experimental or unstable. The package has changed radically in this
release, large parts of the build system have become brain-dead and
non-configurable. Entire sections of the package have disappeared
(notably sqlite support) and the build system is simply not sane any
more.

The little bit of the package I've been able to build so far results
in a broken libsoci_core.so anyway - no SONAME.

I'm going to have to spend large amounts of time persuading upstream
that removing 75% of the upstream tarball - mostly all the build
environment scripts (like configure.in) is not a sane way to release a
package. Currently, I'm beginning to regret packaging soci 2.2.0 for
Debian as 3.0.0 is severely crippled by design. (I'm not kidding, the
old .orig.tar.gz is 600Kb and the new one less than 200Kb.)

I'm actually debating whether to seek removal of 2.2.0 rather than
introduce the mess that is 3.0.0 - at least until 3.0.1 becomes a
possibility.

I'll upload 2.2.0-5 to fix the gcc-4.4 but that can easily wait until
after Lenny - unless I decide to remove soci before that time.

-- 


Neil Williams
=============
http://www.data-freedom.org/
http://www.linux.codehelp.co.uk/
http://e-mail.is-not-s.ms/

Attachment: pgpNJApGz7Zv9.pgp
Description: PGP signature

Reply via email to