Doesn't use setup.py

On Sun, 2007-05-27 at 14:33 +0000, Jelmer Vernooij wrote:
> Public bug reported:
>
> affects /products/pqm
>
> The makefile only installs the binary but not the required 'pqm' python
> package.

I've always run PQM in-place.

> Is there any reason for using the auto* tools rather than distutils ?

Hysterical raisons.

-Rob
--
GPG key available at: <http://www.robertcollins.net/keys.txt>.

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Robert Collins wrote:
> On Sun, 2007-05-27 at 14:33 +0000, Jelmer Vernooij wrote:
>> Is there any reason for using the auto* tools rather than distutils ?
> Hysterical raisons.
Hmm.. so patches to change that situation would be welcome ?

Cheers,

Jelmer

P.S. Did you see my followup to your review of my other pqm-bzr patch?

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iQCVAwUBRl6Wfgy0JeEGD2blAQLBowP9E0foXdq34rbpPjrtCIZ5YM91X9pMk0u5
Gy7lyLZWBnc1A7Br4DYhlff0BroLCyNYnD58Fo8/Xdc4gSsjoZCP7k+BSbE7IdgN
q71fH9tutknWrXmPwkIRSySQNEMyr/jBG8XDOEAoCuRAzsPQ0jVZxEb9MsxZB73V
SsRn3GCoAt0=
=plVu
-----END PGP SIGNATURE-----

On Thu, 2007-05-31 at 09:33 +0000, Jelmer Vernooij wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> Robert Collins wrote:
> > On Sun, 2007-05-27 at 14:33 +0000, Jelmer Vernooij wrote:
> >> Is there any reason for using the auto* tools rather than distutils ?
> > Hysterical raisons.
> Hmm.. so patches to change that situation would be welcome ?

Hmmm, if they are a net win, sure. Theres a bunch of stuff that
setuptools is no better at managing - see bzr's use of a plain Makefile
for instance.

-Rob
--
GPG key available at: <http://www.robertcollins.net/keys.txt>.

Rather important to make it packagable.

  status fixcommitted

Fixed in the attached bundle.
--
Jelmer Vernooij <email address hidden> - http://samba.org/~jelmer/
Jabber: <email address hidden>

So I've looked at the patch and it does not seem to be a net win to me. On reflection I'd really rather keep just fix the failure to install the package and not convert everything to setup.py - setup.py doesn't seem simpler to me.

automake is a very heavy dependency for a simple python package such as pqm.

We could still ditch autoconf and automake and use a simple Makefile for building the docbook stuff?

Given that PQM is primarily written in python, I'd have to argue that a setup.py is much easier to understand for the majority of people that would be looking to work with the code. By keeping autoconf and automake you are scaring away potential contributors.

I'd vote for moving to setup.py.

I'd prefer a setup.py also. I'm simply not familiar with autotools.

Updated version of my patch against current trunk.

Cheers,

Jelmer

--

Resetting the status - this isn't in the trunk yet.

The net win includes familiarity to Python developers, ability to distribute via egg on the Python package index, platform independence, Python version independence, and that nobody will need to write the INSTALL document.

It took a few minutes for me to work out the undocumented procedure on how to generate and run the Makefile (autogen.sh?) that gives me a big autogenerated Makefile that I certainly can't be arsed decoding to work out what the rules are and what they do.

 status fix-committed

It is committed in a branch; pqm uses the same status values as bzr;
committed -> fix-committed
on trunk -> fix-released

 status fixcommitted

It is committed in a branch; pqm uses the same status values as bzr;
committed -> fix-committed
on trunk -> fix-released