[GRASS-dev] Modules

[Glynn Clements]

J-Báchym Čepický wrote:-A

> > I may or may not be able to see the whole picture here, but what I took
> > from the recent discussions was that v.edit is not usable as a backend for
> > a v.digit replacement because of its requirement to commit all changes and
> > re-build topology after every edit. v.digit functionality needs to include
> > undo/redo, incremental topology building etc. which means it needs to be
> > programmed in C with direct access to the vector library functions.
> 
> We could access the vector files using python-swig bindings...

The point is that you need a monolithic process which calls the Vect_*
functions directly. If you're using an interpreted language, you need
language bindings to the C functions; you can't delegate this to an
external program.

IOW, a hybrid Tcl/Tk+C (or Python+C) program will work, but a
combination of a pure Tcl/Tk (or Python) program with a pure C helper
program won't work.

-- 
Glynn Clements <glynn at gclements.plus.com>