On Wed, 25 Sep 2002, Leopold Toetsch wrote:
> This patch obsoletes all previous imcc 0.0.9 patches and contains all
> current fixes and improvements.
> I suppose, problems Andy has, to be related with the parser.
Actually, the core dump I got in reg_sort_f turned out to be a simple bug
in reg_sort_f! That's probably one of the few times I've ever encountered
where a core dump occurred near the bug that caused it :-).
The problem is that the qsort() function wasn't being called with a
correct comparison function. The supplied comparison function only ever
returned 0 or 1, instead of the required -1, 0, 1. Why your qsort
was happy anyway, I don't know.
What Solaris's qsort ended up doing was walking off the front of the
reglist[] array, effectively trying to sort reglist[-1], which, of
course, didn't exist. (*Why* it did that is a bit of a mystery; I'd have
to look at the detailed implementation of Sun's qsort to say for sure.)
Here's the patch that fixes it.
--- parrot-orig/languages/imcc/imc.c Mon Sep 23 14:05:20 2002
+++ parrot-andy/languages/imcc/imc.c Wed Sep 25 13:18:34 2002
@@ -153,7 +153,15 @@
static int reg_sort_f(const void *a, const void *b) {
SymReg *ra = *(SymReg**) a;
SymReg *rb = *(SymReg**) b;
- return ra->first_ins->index > rb->first_ins->index ? 1 : 0;
+ if (ra->first_ins->index < rb->first_ins->index) {
+ return -1;
+ }
+ else if (ra->first_ins->index == rb->first_ins->index) {
+ return 0;
+ }
+ else {
+ return 1;
+ }
}
static void sort_reglist()
--
Andy Dougherty [EMAIL PROTECTED]
