Re: [tree-ssa] COND_EXPR lowering.

[law at redhat dot com]

In message <1067009634.14175.2907.camel@p4>, Andrew MacLeod writes:
 >On Fri, 2003-10-24 at 09:22, Andrew MacLeod wrote:
 >
 >> 
 >> Oops, my x86 bootstrap with this patch just died in PRE:
 >> 
 >> insn-recog.c:17561: internal compiler error: virtual array
 >> basic_block_info[5349]: element 3537031808 out of bounds in
 >> insert_occ_in_preorder_dt_order_1, at tree-ssa-pre.c:903
 >> Please submit a full bug report,
 >> with preprocessed source if appropriate.
 >> See <URL:http://gcc.gnu.org/bugs.html> for instructions.
 >> make[1]: *** [insn-recog.o] Error 1
 >> make[1]: Leaving directory `/build/tree-ssa/2003-10-21/gcc'
 >> make: *** [stage2_build] Error 2
 >> 
 >> My source tree is a couple of days old, I'll try it with a fresh tree
 >> too.
 >> 
 >There is a change to the dominator version of thread_edge that is
 >incompatible. I hacked around a bit, (leaving 2 thread_edge routines,
 >and modifying the one in dom.c to handle COND_EXPR's), but I get a
 >bootstrap error:
 >
 >/src/tree-ssa/2003-10-24/gcc/gcc/combine.c: In function `simplify_shift_const
 >':
 >
 >/src/tree-ssa/2003-10-24/gcc/gcc/combine.c:9137: internal compiler error: Seg
 >mentation fault
 >Please submit a full bug report,
 >with preprocessed source if appropriate.
 >See <URL:http://gcc.gnu.org/bugs.html> for instructions.
 >make[2]: *** [combine.o] Error 1
 >make[2]: Leaving directory `/build/tree-ssa/2003-10-24/gcc'
 >make[1]: *** [stage2_build] Error 2
 >make[1]: Leaving directory `/build/tree-ssa/2003-10-24/gcc'
 >make: *** [bootstrap] Error 2
 >
 >
 >You might want to take a look at the threading bits and see if you can
 >get it to bootstrap. It bootstraps without the patch.
I don't mind updating the treading bits if that's what's holding up
COND_EXPR lowering.  I've already spent some time pondering how to 
merge the two :-)


jeff