Mathieu Bouchard wrote:
On Wed, 28 Nov 2007, IOhannes m zmoelnig wrote:
right, i found out that hexloader does a bad job when trying to load patches. this part of the hexloader is obviously still experimental. what it does is to register a dummy class for an abstraction which it thinks it should handle.
Why do you need dummy classes?
because i don't know how to do it otherwise. i would be glad for advice on how to do it in a better way. e.g. i still don't understand how to use pd_objectmaker.
what i am currently doing is: if hexloader is called for an object that needs hexmangling and it finds a relevant abstraction, it registers a class with the object-name. when this object is instantiated, a setup-function provided by hexloader is called, which then instantiates the abstraction, by calling " binbuf_evalfile()" (and all the rest i copied from m_class.c:542..552)
this works fine, as long as no abstractions are within abstractions (that is the problem patrick experienced). i think the problem is, that Pd has some recursion-prevention mechanism ("tryingalready"), that hinders the instantiation of abstractions with externals. probably i should have a look how [dyn] handles this.
fgm,asdr IOhannes