Am Freitag, den 13.03.2009, 19:54 +0100 schrieb Friedrich Weber: > Hi all, > > thanks for your thoughts :) > Nicolas, I think the approach is cool for immutable values, but if I use > mutable Neko values (like an array for example), my wrap function will > create a wrapper object each time I use this value (and that's not > optimal) - or am I missing something? > Felix, I think I got it, interesting! Do you plan to release the Python > compiler some day? > > I think I'm going to use the laziest way. I'll operate as much as > possible with Mylanguage's objects and add a neko block to the language, > so that the contents of the neko block is copied verbatim to the nxml > code file. > > Thank you, > > Friedrich >
Hi, I certainly intend to release that thing, once it's in an actually bearable form ;) (I hope I actually have enough patience to get it there). I didn't start working on the compiler, but the runtime stuff is coming along nicely. Of course, I'll post it here once I release the code. Using mainly your own objects is probably the simplest way to go. But one of my goals when I thought up that scheme was to get the cleanest possible integration between Python code and other Neko code, mostly because that simplifies the runtime code, of which Python needs quite some. Regards, Felix Krull -- Neko : One VM to run them all (http://nekovm.org)
