[PyMOL] aqual tcltk Pmw gui

[William S. <wg...@ch...>]

Dear Murali:

Python2.3 distributed by apple in 10.4 does not appear to make use of  
their aqua tcltk framework.  I don't understand why they didn't do  
this, but that was my motivation to compile Python2.4 using Apple's  
tcltk framework.  In order to get the "hybrid" program that Warren  
distributes to use this, you would have to either recompile Python2.3  
in the same way we did for 2.4, or wait for Warren to distribute the  
hybrid program using Python2.4.  He could probably be persuaded to  
have it use Apple's TclTk as long as we all continue to faithfully  
support his efforts with MacPymol.  I just renewed my maintenance  
agreement for MacPyMol; it is definitely very much worth it.

All the best,

Bill

Begin forwarded message:

>
> Message: 1
> From: Murali Vadivelu <mk...@ca...>
> Subject: Re: [Fink-devel] Python, frameworks and Tkinter in 10.4
> Date: Sun, 15 May 2005 16:14:07 +0100
> To: William Scott <wg...@ch...>,
>  fin...@li...
>
> Dear Fink Developers,
>
> Pymol's install plugin menu gives the following error:
> AttributeError Exception in Tk callback
>    Function: <function <lambda> at 0x56d0cf0> (type: <type  
> 'function'>)
>    Args: ()
> Traceback (innermost last):
>    File "/Users/delwarl/pymol/products/PyMOLX11Hybrid.app/py23/lib/
> python2.3/Pmw/Pmw_1_2/lib/PmwBase.py", line 1747, in __call__
>      None
>    File "/Users/muralikv/Desktop/PyMOLX11Hybrid-0.98-rc5.app/pymol/
> modules/pmg_tk/skins/normal/__init__.py", line 2135, in <lambda>
>      command = lambda s=self: s.install_plugin())
> AttributeError: Normal instance has no attribute 'install_plugin'
>
> Then of interest to developers wanting to use Aqua-Tcl/Tk:
> I compiled 0.98 using the .info file from Prof. Bill Scott. as
> mentioned by him below.
> It works fine. I tested against MacPython version 2.4.1 on Tiger.
>
> I also took the PymolX11Hybrid (version 0.98rc5 downloaded from pymol
> site) and removed the X11 Tcl/Tk, Python files and replaced it with
> the Tiger provided equivalents. It did not work as the GUI didn't
> respond. Then I replaced all the Pymol files, with the fink compiled
> Pymol. It works perfectly fine now. However, the Pymol Viewer Window
> is still an X11 window! Its the opposite of the original
> PymolX11Hybrid.app where the menu/ command window is X11 and the
> Pymol Viewer window is native mac!! Does anyone know how to create a
> native Pymol Viewer Window so that it does not need X11 at all.
>
> Many thanks.
>
> Best regards,
> Murali.
>
> On 14 May 2005, at 10:01 pm, William Scott wrote:
>
>
>> Dear Muraili:
>>
>> For pymol, I just did this:
>>
>>
>> <pymol.info>
>>
>> This uses the original patch file.
>>
>> Then when it installed I changed the /sw/bin/pymol file to call /
>> usr/bin/pymol2.4
>>
>> HTH,
>>
>> Bill
>>
>>
>>
>>
>> On May 14, 2005, at 1:53 PM, Murali Vadivelu wrote:
>>
>>
>>
>>> Dear Professor Scott,
>>>
>>> Do you have these binaries and/ or specific instructions (if any/
>>> required) in your web site?
>>>
>>> Many thanks.
>>>
>>> Yours sincerely,
>>> Murali.
>>>
>>> On 14 May 2005, at 07:52 pm, William Scott wrote:
>>>
>>>
>>>
>>>
>>>>
>>>> Here's a screenshot of how it looks with the fink version of pymol.
>>>> The Pmw menu launches 5 times faster and looks nicer.
>>>>
>>>> http://babyurl.com/pDCJhv/pymol_screenshot.png
>>>>
>>>>
>>>>
>>>> On Sat, 14 May 2005, Murali Vadivelu wrote:
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>> I think its a fantastic idea. I built R against the Tiger
>>>>> version of
>>>>> aqua-tcl/tk and used a tcl/tk package from inside R (called
>>>>> RCommander). It works flawlessly! Its just fantastic. The output
>>>>> graph from the package was displayed in another quartz window  
>>>>> which
>>>>> has to be opened from inside the R command line manually. Its a
>>>>> good
>>>>> work by the R team. They carbonized the binary using a single line
>>>>> command from a command line tool included with the Developer  
>>>>> tools.
>>>>>
>>>>>
>>>>> On 14 May 2005, at 05:46 pm, William Scott wrote:
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>> Hi citizens:
>>>>>>
>>>>>> I noticed that 10.4 now comes with an aquafied TclTk framework.
>>>>>> For awhile now I had been manually installing this, and then
>>>>>> manually installing a framework build of Python.
>>>>>>
>>>>>> With 10.4, this just became much easier, as the only thing
>>>>>> required
>>>>>> for me to build Python 2.4 was to configure enabling frameworks
>>>>>> and
>>>>>> then to issue make frameworkinstall.  To get python to use the
>>>>>> non-
>>>>>> fink TclTk by default, I just put prepended /usr/bin to my path
>>>>>> temporarily.  After it installed, I created the symbolic link
>>>>>>  /Library/Frameworks/Python.framework/Versions/Current/lib/
>>>>>> python2.4/site-packages -> /sw/lib/python2.4/site-packages/   to
>>>>>> replace the (empty) framework site-packages directory with mine
>>>>>> from fink, that contains other modules like numeric, scientific,
>>>>>> Pmw, and so on.
>>>>>>
>>>>>> As a result, programs (such as pymol) that use Tkinter or Pmw
>>>>>> (which depends on Tkinter) launch FIVE TIMES faster than with
>>>>>> the X-
>>>>>> windows version in fink (6 seconds vs. 30 seconds on my 2x2.5 GHz
>>>>>> G5), and using IDLE is now faster and it is aquafied.
>>>>>>
>>>>>> I was wondering if it would be possible to have a version of
>>>>>> Python
>>>>>> within fink that builds against the aqua TclTk in fink?  I think
>>>>>> you can potentially put frameworks anywhere.
>>>>>>
>>>>>> The only drawback I can see currently is that there are some
>>>>>> functions within Pmw that depend on blt, which is not yet
>>>>>> available
>>>>>> in the aqua form.
>>>>>>
>>>>>> Just an idea ...
>>>>>>
>>>>>> Bill Scott
>>>>>>
>>>>>>