On Wed, 1 Oct 2003 19:29:15 +0200 (Romance Daylight Time) Vadim Zeitlin <[EMAIL PROTECTED]> wrote:
VZ> Basically, SWIG parses special .i files containing something looking like VZ> C++ declarations and generates the glue necessary to call use the C++ VZ> classes/functions from Python (and/or other languages, there is really no So it is a two way process? M <--> Python? VZ> particular reason to limit ourselves to Python if we use SWIG, supporting VZ> Perl, TCL or Ruby is not more difficult than supporting any one of them). I actually figured that out too reading the documents too! If we continue making M to be the power users Emailer / News Reader, the SWIG interfaceing setup could be a good feature if we do it in a thoughtful / comprehensive manner given some of the wishlist we seem to have in Buglist :-) VZ> The idea I had (point (2) below) is that we could try to use our .h files VZ> directly instead of maintaining .i files in parallel as new SWIG (the VZ> official one, nothing to do with Robin's changes) is supposed to be able to VZ> parse real C++ just as well. This would be ideal as the burden of VZ> maintenance would be almost inexistent then. It wouldn't work without some VZ> effort, however... VZ> VZ> ML> VZ> Yes. Fixing Python is easy but not trivial. There are 2 posssibilities: VZ> ML> VZ> VZ> ML> VZ> 1. simple but not very interesting: update .i files manually, i.e. simply VZ> ML> VZ> integrate the changes from .h files into them VZ> ML> VZ> ML> Is this hard to do? I have not looked at it at all yet. VZ> VZ> No, it shouldn't be. VZ> VZ> ML> VZ> 2. try new SWIG (1.3.19?) which is supposed to be able to directly parse .h VZ> ML> VZ> files and get rid of .i completely or, maybe, keep them but just for VZ> ML> VZ> SWIG directives (i.e. we could tell it to ignore some methods or rename VZ> ML> VZ> some others) OK. Given (realistically) that it is mostly going to fall to you to implement (any other volunteers?) , how do you suggest we proceed? VZ> ML> Would we be better off first trying to use the SWIG that is used in wxPython as VZ> ML> I understand Robin has specially modified it to be used with wxWindows? It also VZ> ML> at least comes in the wxWindows CVS tree. VZ> VZ> No, we don't need Robin's changes and we don't expose wxWindows in our .i VZ> files, just functionality specific to Mahogany. OK. Thanks for the explaination. Dr. QA ------------------------------------------------------- This sf.net email is sponsored by:ThinkGeek Welcome to geek heaven. http://thinkgeek.com/sf _______________________________________________ Mahogany-Developers mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/mahogany-developers
