From: Ullrich von Bassewitz (uz_at_musoftware.de)
Date: 2000-11-30 10:24:25
On Thu, Nov 30, 2000 at 01:35:26AM -0500, Todd Vierling wrote: > Gcc can be targeted to an 8-bit target quite easily; it's the guts of > implementation that is difficult, but only because it is tedious to flesh > out all the various RTL instructions for 16-bit and 32-bit quantities > (particularly to reduce code size with known constant values that fit into > smaller int sizes). The compiler is so flexible that you could conceivably > have compiler options to provide several different machine modes; one to use > cc65 register and stack conventions, one to use a bunch of "virtual > registers", one to use a virtual stack, one to use completely vanilla 6502 > everything, and so forth. My assumption was based on some statements from the porting manual, on the fact that there is no known gcc port to any 8 or 16 bit machine, and on the fact that the existing gcc6502 port emulates some sort of 24 bit CPU. I have to admit that I have never tried porting gcc myself, and I did not have a look at the backend myself, so I may be wrong. > Context: I actually started writing such a port many moons ago, but didn't > get very far because personal time shrunk to 0. Who knows, I might try > again for 65xx (particularly 65816) as a hobby when I have more time. This > would require, actually, using as65 and ld65.... There seems to be a proliferation of 6502 C compilers lately:-) In a not too distant future we may have the same situation as with assemblers: Anyone capable of doing so will write his own compiler. A lot of these projects will never make it to become actually usable, because most people underestimate the work needed to complete such a project. You don't need just a compiler, you need several more tools, and if you want the compiler to be useful, you have to write a complete C library for it. I'm of course the last one to tell other people how to spent their time. I know, that fiddling with a compiler is fun for itself (I'm working on cc65 for fun myself:-). However, it is my opinion that it would be much better for the community if people would work together more closely and use/improve what is there instead of recreating existing things over and over. There are not many 6502 developers left, and splitting resources means that a lot less will get accomplished. This is of course just my humble opinion. I know that there are more reasons to start such a project than just the goal to create a compiler for the old machines. So, I wish anyone luck who is working on another compiler implementation for 6502 machines. I will be the first one to download your stuff once it is available. May the fun be with you:-) Regards Uz -- Ullrich von Bassewitz uz_at_musoftware.de
This archive was generated by hypermail 2.1.3 : 2001-12-14 22:05:37 CET