--- Comment #7 from rguenth at gcc dot gnu dot org 2005-10-21 12:11 ---
The problem is we didn't fold
arg 0
readonly
arg 0 >
which originates from build_binary_op() where default_conversion converts
unit size
align 32 symt
--- Comment #6 from ebotcazou at gcc dot gnu dot org 2005-10-21 11:06
---
> I have a patch that papers over the problem.
The best ones. ;-)
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=24470
--- Comment #5 from rguenth at gcc dot gnu dot org 2005-10-21 11:00 ---
I have a patch that papers over the problem.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=24470
--- Comment #4 from rguenth at gcc dot gnu dot org 2005-10-21 10:54 ---
The problem is the condition of the COND_EXPR is not folded:
arg 0
arg 0
readonly
arg 0 >
arg 1 >
so the transformation
/* If this is A
--- Comment #3 from rguenth at gcc dot gnu dot org 2005-10-21 10:41 ---
We are endlessly folding
(gdb) call debug_tree(expr)
unit size
align 32 symtab 0 alias set -1 precision 32 min max >
arg 0
unit size
align 8 symtab 0 alias set -1 p
--- Comment #2 from rguenth at gcc dot gnu dot org 2005-10-21 10:28 ---
With checking enabled and 3.4 branch HEAD we go loop infinitely allocating
_lots_ of ram.
Investigating.
--
rguenth at gcc dot gnu dot org changed:
What|Removed |Added
---
--- Comment #1 from raasch at mathematik dot uni-marburg dot de 2005-10-21
10:10 ---
Created an attachment (id=10038)
--> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=10038&action=view)
gzipped .ii file
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=24470
