This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
internal error with -fbranch-probabilities
- To: gcc-bugs at gcc dot gnu dot org
- Subject: internal error with -fbranch-probabilities
- From: Dominik Strasser <dominik dot strasser at mchp dot siemens dot de>
- Date: Thu, 19 Aug 1999 16:35:33 +0200
- Organization: Siemens AG, ZT SE 4
To follow up myself, the complete error messages reads like this:
verify_1.C: In method `char *
basic_string<char,string_char_traits<char>,__default_alloc_template<true,0>
>::Rep::clone()':
verify_1.C:6487: instantiated from here
verify_1.C:6985: internal error--unrecognizable insn:
(jump_insn 297 296 302 (set (pc)
(if_then_else (ltu (cc0)
(const_int 10000 [0x2710]))
(label_ref 26)
(pc))) -1 (nil)
(expr_list:REG_BR_PROB (const_int -1 [0xffffffff])
(nil)))
As I wrote yesterday, I can provide the preprocessed source + the
associated .da-file.
Today's experiments were with gcc 2.95.1. I was able to reduce the input
a little bit, but it is still 18k lines, 95K gzipped.
Please ask for more dtails.
Dominik
PS: gcc 2.95.1 seems to choke on a different source line compared to gcc
2.95
--
Dominik Strasser | Phone: +49 89 636-43691
SIEMENS AG | Fax: +49 89 636-42284
ZT SE 4 | E-Mail:Dominik.Strasser@mchp.siemens.de
Otto-Hahn-Ring 6 |
D-81739 Muenchen |