[PATCH] fix int vs enum issues with in stabs support
Olivier Hainque
hainque@adacore.com
Wed May 20 13:39:00 GMT 2009
Ian Lance Taylor wrote:
> Please add 2009 to the list of copyright years.
...
> Use stab_code_type instead.
Sure.
> This is OK with those changes.
Thanks for your feedback Ian.
FWIW, I'm having troubles with the ppc-aix port these days.
I'm attempting to resubmit
http://gcc.gnu.org/ml/gcc-patches/2008-08/msg00047.html
and hit a number of intermediate issues just trying to build a pristine
reference:
* After this enum thing, there is
.../src/gcc/toplev.c: In function 'process_options':
.../src/gcc/toplev.c:2043:
error: logical 'and' of mutually exclusive tests is always false
on "if (!FRAME_GROWS_DOWNWARD && flag_stack_protect)" because
rs6000.h has
#define FRAME_GROWS_DOWNWARD (flag_stack_protect != 0)
An extra 'true &&' circumvents this but is probably not the proper
way to go.
* Later on, a pleasant
.../ada/targparm.adb:641: error: insn does not satisfy its constraints:
(insn 587 585 586 65 .../src/gcc/ada/targparm.adb:313
(set (reg/f:SI 0 0 [721])
(mem/u/c:SI (plus:SI (mem/c:SI (plus:SI (reg/f:SI 31 31)
(const_int 196 [0xc4])) [41 %sfp+196 S4 A32])
(const:SI (unspec:SI [
(symbol_ref/u:SI ("*LC..18") [flags 0x2])
] 49))) [33 S4 A8])) 309 {*movsi_internal1}
(expr_list:REG_EQUIV (symbol_ref:SI ("constraint_error") [flags 0xc0]
<var_decl 30190120 constraint_error>)
(nil)))
| 4.5.0 20090519 (experimental) (powerpc-ibm-aix5.3.0.0) GCC error:
| in reload_cse_simplify_operands, at postreload.c:396
pops up in stage2,
* and the debug info for locals is apparently screwed up: for
instance ...
int main (void)
{
int x = 20;
return 0;
}
stage1 compiles into
.stabx "x:-1",28,129,0
**
li 0,20 # tmp121,
stw 0,24(31) # x, tmp121
**
so
(gdb) start
...
main () at t.c:4
4 int x = 20;
(gdb) n
6 return 0;
(gdb) p x
$1 = -559038737
Some of these might be related to my base compiler. I don't know yet.
More information about the Gcc-patches
mailing list