[GRASS-dev] how to start wx vdigit?

William Kyngesburye woklist at kyngchaos.com
Sun Mar 9 17:16:24 EDT 2008

Oops, forgot the -shared flag in the vdigit makefile.  It ended up  
linking as a program.  See: http://trac.osgeo.org/grass/ticket/61

Changed the shared flag to something appropriate for OSX, and I didn't  
need to add the "import wx" to grass6_wxvdigit.py.

I successfully digitized a boundary.  Lots of messages in the Terminal  
- failed assertions and context errors.

I'll add to http://trac.osgeo.org/grass/ticket/58 with my results.

On Mar 9, 2008, at 3:18 PM, William Kyngesburye wrote:

>> important is the last part of the message which comes when you try to
>> import grass6_wxvdigit Python module .
>> Try:
>> cd gui/wxpython/vdigit
>> python
>> import grass6_wxvdigit
>> you will get this error
> So is there a bug in grass6_wxvdigit.py?
>>> I'm not sure if this is from the way I setup my test build or
>>> something else.
>>> Here is what I did to get it compiled: since _gdi_.so is technically
>>> an extension to [wx]Python, I figure that it should get loaded  
>>> (maybe
>>> automatically?) by wxPython.  So instead of directly linking it into
>>> vdigit, I used an OSX linker flag "-undefined dynamic_lookup" to  
>>> tell
>>> the linker that the gdi.so symbols will be found at runtime.
>> Hm, the last days I have not so much time to work on this, the way  
>> you
>> suggest maybe could solve it.
> I'm guessing "import wx" should import _gdi and load _gdi_.so, since  
> _core imports _gdi.  Didn't help with the init_grass6_wxvdigit error.

William Kyngesburye <kyngchaos*at*kyngchaos*dot*com>

"History is an illusion caused by the passage of time, and time is an  
illusion caused by the passage of history."

- Hitchhiker's Guide to the Galaxy

More information about the grass-dev mailing list