Hi Konrad, > I was just wondering what the current status of 64 bit picolisp is.
The good news: It is under work since more than one year now :-) The bad new: I'm afraid it will still take quite a while :-( I do not want to go public with it before I'm sure that most details are settled down. But I already have a skeleton running, a proof of concept. The basic subsystems are ready so far (eval, apply, symbol tables, exception handling, garbage collection), but the huge amount of I/O functionalities is still missing. Currently I'm working on the bignum arithmetics. It is a complete rewrite. Even the implementation language changed. Instead of C it is written in a generic assembler (which in turn is written in PicoLisp :) that generates GNU assembler code (currently there is only a x86-64 generator, but other CPUs are possible). The problem is that I can work on it only at my spare time, and this is usually scarce. Things would be different if I'd get paid for it ;-) I hope I can show a protype during the next year. > I;m on a 64 bit system which for the moment has prevented me from > trying out Thomas's Async read and write code. Is it a non-x86-64 system, preventing you from running 32bit binaries? > Does the larger cell size mean that there is an extra unneeded bit on > every pointer, which could be used as an additional flag bit. Exact. > For say floating point numbers? No, but short and big numbers will be separate types internally. The structure model is cnt xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxS010 big xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxS100 sym xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx1000 cell xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx0000 Floating point numbers would have been an option, but are not intended. > I ask the latter as this is the one of the two points I could see > Picolisp being flamed for, when it gets sufficently popular. The other > is use of dynamic scope, though i suspect we agree on this one, its a > design decision live with it. Yep. As you know, I'm convinced that dynamic binding is far superior to lexical binding ;-) Cheers, - Alex -- UNSUBSCRIBE: mailto:[EMAIL PROTECTED]
