Derek Homeier wrote:
> Hi,
>
> I only came upon this thread after having some trouble with matplotlib
> myself...
>
> On 18 Sep 2009, at 22:44, John K. Parejko wrote:
>
>>> On Wed, Sep 16, 2009 at 02:09:05PM -0400, John K. Parejko wrote:
>>>>
>>>> I'm trying to install the new version of matplotlib-py26, but I get
>>>> the following error about it trying to overwrite files that belong to
>>>> dateutil:
>>>
>>> I can't reproduce this on my 10.4/ppc machine with unstable activated
>>> (dateutil-py26-1.4-1 matplotlib-py26-0.99.0-1) and with all packages
>>> built locally by fink (vs downloaded precompiled .deb).
>>
>> After some digging, I remembered that I'd built a pre-release of
>> matplotlib 0.98.something so that I could use some of the new 0.98.x
>> features. When I did the build, I used the SVN version with "python
>> setupegg.py develop" to make following SVN changes easier. This sets
>> up some symlinks and references in /site-packages back to the build
>> directory.
>>
>> I removed fink's python2.6 and all *-py26 packages and matplotlib now
>> installs cleanly and seems to work fine. Let this be a warning to
>> those who try to run on the bleeding edge!
>
> Just wondering - by this do you mean you removed all the fink
> python2.6-libraries
> and reinstalled them and afterwards succeeded in building matplotlib-py26,
> or did you then compile matplotlib independently of fink? The latter
> does not
> really sound like a solution to the fink package's problem...
> What I found, also perhaps trying to be more on the bleeding edge, was that
> matplotlib 0.99 and in particular 0.99.1 introduced a number of new
> configuration
> settings, among others setting a default to try to install its own pytz
> and dateutil
> in setup.cfg.
> I managed to put a working package for 0.99.1.1 afterwards, attached below.
>
> On this occasion, I am wondering if sending updated package descriptions
> to the
> list, as I've done before, is the appropriate way, or whether I should
> rather send
> them directly to the maintainer or simply be content to wait until
> they've checked in
> their own updates or...
>
> Whatever the correct upgrade path, I'd also like to make a suggestion at
> this point:
> Is it possible to remove the numeric-py and numarray-py dependencies,
> perhaps at
> least make them "Suggests:" or "Enhances:" instead? They are clearly no
> longer required
> to build or use matplotlib (in fact I am not even sure if they are still
> officially supported as
> backends, since that configuration option has been removed now), and I
> find it rather
> painful to still being forced to keep two long outdated libraries
> installed.
> I'd think anyone who still needs either of them for legacy routines or
> data would know
> so on their own - just my 2d.
>
> Cheers,
> Derek
>
>
>
>
You should consider sending your package description to the maintainer
of the matplotlib package ("fink info matplotlib") because he would be
the one to make any changes.
--
Alexander Hansen
Fink User Liaison
------------------------------------------------------------------------------
Come build with us! The BlackBerry® Developer Conference in SF, CA
is the only developer event you need to attend this year. Jumpstart your
developing skills, take BlackBerry mobile applications to market and stay
ahead of the curve. Join us from November 9-12, 2009. Register now!
http://p.sf.net/sfu/devconf
_______________________________________________
Fink-users mailing list
Fink-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/fink-users
