[flext] Flext and jMax
Thomas Grill
gr at grrrr.org
Wed Jun 9 17:37:10 CEST 2010
Hi Maurizio,
to be honest i would only support the first strategy.
It's already hard for me to track down problems in the current codebase because i only look at it every few months. Therefore i would not advocate changes that go beyond a more or less superficial level.
gr~~~
Am 09.06.2010 um 10:43 schrieb Maurizio De Cecco:
> OK, i took a quick look.
>
> Apparently, things are not so easy: the flext code use the basic Max/Pd data types and constants, and have a lot of #ifdefs around.
>
> jMax do not use the Max/Pd data types: jMax was redesigned are rewritten from scratch.
>
> OK, i actually added a compatibility layers supporting some of the data type and constants, but the source, not binary level.
>
> What i means is that data types and constants are mapped of different
> implementation types; for example, constants like A_FLOAT are mapped
> to symbols, not to integers.
>
> So, a possible strategy would be: use the pd compatibility layer for jMax (a specific version of the m_pd.h file, essentially), pretend that
> jMax is PD, compile and see what happens.
> Add the proper #ifdefs in the flext and the proper extensions to the jMax m_pd.h file. Reiterate until it compiles.
> Try to compile/run simple objects and reiterate.
>
> Another possible strategy is to work at a different level, and
> see how flext can be readapted or reimplemented on top the jMax
> architecture.
>
> Both strategies are possible and not cheap. I will insert this
> work on my jMax roadmap, but i do not have a big slot available
> in the very short term; we will rediscuss the situation
> after the summer ...
>
> Maurizio
> _______________________________________________
> http://grrrr.org/ext/flext
>
> flext mailing list
> flext at grrrr.org
> http://grrrr.org/cgi-bin/mailman/listinfo/flext
More information about the flext
mailing list