Differences

This shows you the differences between two versions of the page.

--- sd:sdcc [2026/04/13 17:33] – created appledog
+++ sd:sdcc [2026/04/13 17:44] (current) – appledog
@@ Line 1: / Line 1: @@
 = SD/CC
+* SD/CC is TinyC for the SD-8516.
-SD/CC TinyC for the SD-8516
 == Introduction
-A self-hosted C compiler has been on the table for quite some time. A few weeks ago I was able to get something set up where it could compile:
+A self-hosted C compiler has been my goal for quite some time. I did BASIC and Forth first. BASIC was useful, Forth -- not so much (from a practical standpoint).
+At the end of March I was able to get something set up where it could compile:
     int main(void) { return 0; }
-The problem was that isn't very much. So I looked at the BASIC expression parser and was trying to figure out how to get expressions. It's a recursive descent parser so the fist thing I added was additive and multiplicative expressions:
+The problem was that isn't very much. Weeks went by and I couldn't figure it out. It was supposed to be just like the BASIC expression parser. A recursive descent parser. So the first thing I added was additive and multiplicative expressions. Then suddenly it started working. I honestly don't know what changed because it wasn't working then suddenly it was and I couldn't remember changing anything in the code.
 <codify armasm>
@@ Line 31: / Line 32: @@
 </codify>
-This kind of works, we can return 3+5 now; but I need more. I've been trying to get a symbol table up and to define int; but it isn't working.
+This works for + and -, we can return 3+5 now. Next I added negative numbers. I had to add a check for TOK_MINUS first and then emit a SUB A, B to get it to negate.
+== Integers
+* April 14th, 2026.
+I've been trying to get a symbol table up and to define int; but it isn't working. I think there is an off by one bug on the stack. Storing variables on the stack is something I don't fully understand or trust yet.
+What is really biting me in the ass is the system I came up with to pair 16-bit registers into 24-bit pointers. It's a charmingly retro, period-authentic thing. But it makes coding a bit hard when you need multiple pointers. I have a 'secret weapon' -- the ability to add more registers any time I want, but, I want to avoid going down that route if possible. Having so many registers already is supposed to be good enough.
 More to come.