Albert Graef

I don't see anything suspicious in there. The config.log also looks fine -- g++ a.k.a. clang is properly recognized as capable of compiling Objective C++ source. And, as your logs...

It's possible in purr-data as long as HTML SVG supports it. Of which I'm almost certain, but I'm not sure how much work it would be. (There's stacked lines of...

This is not a bad idea at all, as it would arguably make some kinds of initialization much easier. The problem is that postinitialize needs to run before any other...

> i'm wondering if it might be an option to delay postinitialize() [...]. would this possibly allow to read the arguments via get_args() then? This specific problem might be somewhat...

Hi Claude, thanks for chiming in! > so it seems to me now that the rationale for postinitialize is to do pd-dependent things (like binding to receive names, create clocks,...

> if the new get_args/set_args are part of the Pd API [...] They aren't. They are just part of the Pd-Lua API. (Tim introduced this with his Pd-Lua graphics API...

This doesn't work for me. The basic multichannel.pd example (included in pdlua/examples/multichannel) now fails with this error message: ~~~ test~.pd_lua: 9: signal_setmultiout: invalid signal pointer. must be called from dsp...

Ok, thx, looks good to me now. Maybe @timothyschoen can have a look at well. Tim?

Yes, so you also need to keep track of the allocated size somewhere. AFAICT, that's for Pd's own internal tracking of allocated memory if it was compiled with `DEBUGMEM`. Or...

Works for me. Not sure what to test with the example in #63, though. Did you check that the clone stuff still works?