It's the same idea I toyed with (including using boost, and had a brief and disgusting venture into SWIG territory too), while a bit above my par, I might have pulled together the education and learning needed to do it over some time, but the initial effort to get anything going was well above what I was willing to commit.
Much like pyMEL and the success it's been (despite the fact it couldn't support a few things initially, for obvious reasons), I'm sure people would more than gladly pay the trade-off in comfort and compatibility and missing features for the sheer performance and the possibilty to couple it with other useful bindings and arbitrary python versions. I appreciate the standard scripting interface, the consistency across the platform, and all other things that COM and what it enforced brought with it, don't get me wrong, but in some regards the trade offs (static versioning and dependencies, mainwin etc.) aren't worth paying, and a set of good bindings to the API and a real, well implemented pyQt Host that goes with it would open doors that would surclass the ones that'd get closed :) On Tue, Sep 4, 2012 at 10:42 AM, Luc-Eric Rousseau <luceri...@gmail.com>wrote: > I didn't mention that yes, indeed, you could just make python bindings > to the c++ sdk, because the ideal scenario is to have it works through > the standard scripting interface. That way the history log would work, > the ppg logic, and the script remain compatible. > > Marc Andre did toy around with the idea of making bindings to the c++ > (with boost I think) and that could run a lot faster too but it would > mean two incompatible ways to write scripts so we hesitated. > > On Sun, Sep 2, 2012 at 9:58 PM, Raffaele Fragapane > <raffsxsil...@googlemail.com> wrote: > > "All" it would take is some proper python bindings of the cpp API. It > could > > easily be done outside of mainwin as the cpp API was done a while ago > > diverging from the silly COM business which mainwin helps taking care of. > > It'd probably turn out better than stff like pyMEL (which has a similar > > approach) given that the cpp API is a lot close to the scripting one than > > MEL ever was to the proper Maya API. > > > > I had toyed with the idea a while ago, and had started looking into some > > assistend binding building, but it's a fair chunk of work, and just > getting > > part of App to even register and LogMessage working required going pretty > > far up the chain (just to give you an idea). > > > > If you fancy a nightly project... :p > -- Our users will know fear and cower before our software! Ship it! Ship it and let them flee like the dogs they are!