On Mon, 2002-09-09 at 09:07, Ramesh Ananthakrishnan wrote: > > Er... it is a silly thing to ask, but is there any way to write C code that comes >out assembled in Parrot? > > Has C been targeted at parrot? Is it a logical thing to do? Does it make sense. (I >tried this in my brain for three days and am still confused over whether it is a >sensible thing to do). It seems to me that lots of stuff already written in C could >be brought to Parrot with the least of fuss.
There are many reasons why you would want to gain access to existing C code. Here are some examples: Access to existing libraries -- The POSIX/ANSI C libraries are probably the most obvious example here. Parrot will need some way to dynamically link to existing libraries (be they C, C++, FORTRAN or whatever) in the same way that Perl 5 did. The difference will be that any language compiled down into PBC will then have access to them, where Perl 5's XS was only available to Perl. Using small sections of code -- If you buy "Algorithms in C" and find something you want to use, it's probably best to re-write it as a module in Perl or some other high-level language and compile down to PBC. This will force you to re-think where Parrot and your high-level language might already provide some or all of what you want. Moving projects over to a Parrot-based language -- Here, I expect you will be combining the above strategies, depending on project size. Some of the existing project's libraries will stand alone well enough to be used externally (no translation required). The rest, you will probably want to translate to something like Perl. So, no I don't think translating C to PBC will make a whole lot of sense, and translating to a high-level language will almost have to be manual. I hope this helps. -- Aaron Sherman <[EMAIL PROTECTED]> http://www.ajs.com/~ajs
