matplotlib-devel Mailing List for matplotlib

On Feb 23, 2007, at 8:00 PM, Andrew Straw wrote:
>
> 2) make our own distutils monkeypatch a la setuptools. Looking at
> setuptools/dist.py, this doesn't look trivial -- certainly beyond my
> free bandwidth capacity.
I've written a script that attempts to simplify writing setup.py's 
that includes automagic support for package_data in Python 2.3:
	http://agni.phys.iit.edu/~kmcivor/downloads/metasetup.py
You can see a simple example of it in the WxMpl source:
	http://svn.csrri.iit.edu/mr-software/wxmpl/trunk/setup.py
Ken

On 2/24/07, Andrew Straw <str...@as...> wrote:
> Robert Kern wrote:
> > IPython does something similar and possibly better.
> >
> > http://ipython.scipy.org/svn/ipython/ipython/trunk/setupext/install_data_ext.py
> >
> From a quick look at the code, it's hard to determine whether this new
> distutils command (install_data_ext) can handle installation to a nested
> directory structure. Can it? If it can, there's still the question of
> whether we want to continue rolling our own solution or simply using
> Python >= 2.4's standard "package_data".
I think it does, but you should actually ask someone who knows about ipython :)
I say 'I think' from reading setup.py, where this code is invoked:
datafiles = [('data', docdirbase, docfiles),
 ('data', os.path.join(docdirbase, 'examples'),
 examfiles),
 ('data', os.path.join(docdirbase, 'manual'),
 manfiles),
 ('data', manpagebase, manpages),
 ('lib', 'IPython/UserConfig', cfgfiles)]
[...]
and then the setup() call contains:
 cmdclass = {'install_data': install_data_ext},
 data_files = datafiles,
So it certainly looks like it works fine for copying things like the
example files and the html manual (which is nested, below doc/manual).
This code was actually given to me years ago, by some kind soul who
cringed at the horrid hacks I had in place to achieve the goal. It
has never been modified and has served us well for years, so feel free
to grab it if it happens to be useful to mpl. It's a tiny bit of
code, so if it gets you out of a bind, I'd say just use it.
It does work fine with python 2.3, if that's one of your goals.
Cheers,
f

Robert Kern wrote:
> Andrew Straw wrote:
>
>> 1) revert to the old way. The primary issues with this are a) 
>> "package_data" is supported as standard Python from 2.4 on, and the old 
>> way required carrying our own distutils command and b) we switched the 
>> data directory to have a nested structure, which required code changes 
>> and repository layout changes that would have to be undone.
>> 2) make our own distutils monkeypatch a la setuptools. Looking at 
>> setuptools/dist.py, this doesn't look trivial -- certainly beyond my 
>> free bandwidth capacity.
>> 
>
> Actually, it ought to be pretty trivial without setuptools (but compatible with
> setuptools, AFAICT). Here is a Cookbook recipe that ought to work:
>
> http://wiki.python.org/moin/DistutilsInstallDataScattered
>
> 
That's exactly "the old way", referred to in point #1.
> IPython does something similar and possibly better.
>
> http://ipython.scipy.org/svn/ipython/ipython/trunk/setupext/install_data_ext.py
> 
 From a quick look at the code, it's hard to determine whether this new 
distutils command (install_data_ext) can handle installation to a nested 
directory structure. Can it? If it can, there's still the question of 
whether we want to continue rolling our own solution or simply using 
Python >= 2.4's standard "package_data".
AFAICT, the monkeypatching setuptools does for 2.3 to support 
"package_data" goes beyond adding a new distutils command. (I don't 
consider adding a distutils command to be monkeypatching -- that's just 
extending distutils is a pre-designed way.) Instead, setuptools actually 
replaces the distutils.dist.Distribution class with 
setuptools.dist.Distribution.

Andrew Straw wrote:
> 2) make our own distutils monkeypatch a la setuptools. Looking at 
> setuptools/dist.py, this doesn't look trivial -- certainly beyond my 
> free bandwidth capacity.
Actually, it ought to be pretty trivial without setuptools (but compatible with
setuptools, AFAICT). Here is a Cookbook recipe that ought to work:
 http://wiki.python.org/moin/DistutilsInstallDataScattered
IPython does something similar and possibly better.
 http://ipython.scipy.org/svn/ipython/ipython/trunk/setupext/install_data_ext.py
-- 
Robert Kern
"I have come to believe that the whole world is an enigma, a harmless enigma
 that is made terrible by our own mad attempt to interpret it as though it had
 an underlying truth."
 -- Umberto Eco

Andrew,
I agree with your proposal; I think it makes more sense than either 
alternative. Let's see what John says when he gets back from his vacation.
Eric
Andrew Straw wrote:
> (Picking up this thread a bit late... And I just wrote a longer email 
> which got munched due to email configuration issues...)
> 
> I'm responsible for the "package_data" keyword being added to setup.py. 
> The bottom line is Python 2.3 is still supported. I simply didn't 
> realize that it would screw things up. I propose that setuptools be a 
> requirement for matplotlib with Python 2.3 and have committed a change 
> that does this. So the issue is now closed unless we want to implement 
> an alternative solution. These, as I see them, are:
> 
> 1) revert to the old way. The primary issues with this are a) 
> "package_data" is supported as standard Python from 2.4 on, and the old 
> way required carrying our own distutils command and b) we switched the 
> data directory to have a nested structure, which required code changes 
> and repository layout changes that would have to be undone.
> 
> 2) make our own distutils monkeypatch a la setuptools. Looking at 
> setuptools/dist.py, this doesn't look trivial -- certainly beyond my 
> free bandwidth capacity.
> 
> -Andrew
> 
> Eric Firing wrote:
>> Darren Dale wrote:
>> 
>>> We support setuptools, but we do not require it.
>>> 
>>
>> So, it sounds like setuptools is required now if one has Python 2.3. 
>> If so, is that acceptable--is the gain worth the pain? Is there any 
>> significant pain associated with requiring setuptools, at least for 
>> people with Python 2.3?
>>
>> Eric
>>
>> 
>>> On Friday 23 February 2007 5:46:58 am Edin Salkovic wrote:
>>> 
>>>> I'm learning a bit about setuptools and distutils, so sorry if this is
>>>> a no brainer: Are we using only distutils for matplotlib? I.e. - no
>>>> setuptools?
>>>>
>>>> This is because I stumbled across this at the setuptools page:
>>>> http://peak.telecommunity.com/DevCenter/setuptools
>>>> ====
>>>> Feature Highlights:
>>>>
>>>> ....
>>>> * Include data files inside your package directories, where your
>>>> code can actually use them. (Python 2.4 distutils also supports this
>>>> feature, but setuptools provides the feature for Python 2.3 packages
>>>> also, and supports accessing data files in zipped packages too.)
>>>> ....
>>>>
>>>> Cheers,
>>>> Edin
>>>>
>>>> On 2/22/07, Darren Dale <dd...@co...> wrote:
>>>> 
>>>>> I noticed today that setup.py is using package_data. Is this 
>>>>> absolutely
>>>>> necessary? The most recent version of Red Hat Enterprise Linux 
>>>>> includes
>>>>> python-2.3, which does not support package_data. We are still 
>>>>> supporting
>>>>> python-2.3, aren't we?
>>>>>
>>>>> Darren
>>>>> 
> 

(Picking up this thread a bit late... And I just wrote a longer email 
which got munched due to email configuration issues...)
I'm responsible for the "package_data" keyword being added to setup.py. 
The bottom line is Python 2.3 is still supported. I simply didn't 
realize that it would screw things up. I propose that setuptools be a 
requirement for matplotlib with Python 2.3 and have committed a change 
that does this. So the issue is now closed unless we want to implement 
an alternative solution. These, as I see them, are:
1) revert to the old way. The primary issues with this are a) 
"package_data" is supported as standard Python from 2.4 on, and the old 
way required carrying our own distutils command and b) we switched the 
data directory to have a nested structure, which required code changes 
and repository layout changes that would have to be undone.
2) make our own distutils monkeypatch a la setuptools. Looking at 
setuptools/dist.py, this doesn't look trivial -- certainly beyond my 
free bandwidth capacity.
-Andrew
Eric Firing wrote:
> Darren Dale wrote:
> 
>> We support setuptools, but we do not require it.
>> 
>
> So, it sounds like setuptools is required now if one has Python 2.3. If 
> so, is that acceptable--is the gain worth the pain? Is there any 
> significant pain associated with requiring setuptools, at least for 
> people with Python 2.3?
>
> Eric
>
> 
>> On Friday 23 February 2007 5:46:58 am Edin Salkovic wrote:
>> 
>>> I'm learning a bit about setuptools and distutils, so sorry if this is
>>> a no brainer: Are we using only distutils for matplotlib? I.e. - no
>>> setuptools?
>>>
>>> This is because I stumbled across this at the setuptools page:
>>> http://peak.telecommunity.com/DevCenter/setuptools
>>> ====
>>> Feature Highlights:
>>>
>>> ....
>>> * Include data files inside your package directories, where your
>>> code can actually use them. (Python 2.4 distutils also supports this
>>> feature, but setuptools provides the feature for Python 2.3 packages
>>> also, and supports accessing data files in zipped packages too.)
>>> ....
>>>
>>> Cheers,
>>> Edin
>>>
>>> On 2/22/07, Darren Dale <dd...@co...> wrote:
>>> 
>>>> I noticed today that setup.py is using package_data. Is this absolutely
>>>> necessary? The most recent version of Red Hat Enterprise Linux includes
>>>> python-2.3, which does not support package_data. We are still supporting
>>>> python-2.3, aren't we?
>>>>
>>>> Darren
>>>>