--- Additional Comments From charles at kde dot org 2005-07-06 01:24
---
Oh, and to add, if you change the double to an int, it works.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=22318
--- Additional Comments From charles at kde dot org 2005-07-06 01:26
---
I should point out that fixing the actual bug in question will fix the ICE
itself.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=16829
--- Additional Comments From pinskia at gcc dot gnu dot org 2005-07-06
01:30 ---
Fixed in 3.4.0 and above. 3.3.x is only gcc releases which are effected so
closing as fixed as 3.3.6 was
the last release of 3.3.
--
What|Removed |Added
../../xgcc -B../../ -c -g -O2 -fPIC -DELF=1 -DLINUX=1 -W -Wall -gnatpg g-
socket.adb -o g-socket.o
+===GNAT BUG DETECTED==+
| 4.1.0 20050705 (experimental) (hppa-unknown-linux-gnu) GCC error:|
| tree check: expected integer_cst
--- Additional Comments From pinskia at gcc dot gnu dot org 2005-07-06
01:59 ---
And I just ran into this too.
--
What|Removed |Added
CC|
--
What|Removed |Added
CC||dberlin at gcc dot gnu dot
||org
--- Additional Comments From pinskia at gcc dot gnu dot org 2005-07-06
02:05 ---
This looks like PR 22140.
--
What|Removed |Added
BugsThisDependsOn|
--- Additional Comments From mathieu at malaterre dot com 2005-07-06 02:32
---
Using gcc --version:
g++ (GCC) 4.1.0 20050607 (experimental)
I still cannot get anything using the -ftrapv:
Ex is:
#include iostream
int main()
{
const unsigned int a = 11;
const unsigned int b = 10;
--- Additional Comments From pinskia at gcc dot gnu dot org 2005-07-06
02:33 ---
(In reply to comment #6)
It looks like -ftrapv only works on signed overflow.
Yes because signed overflow is undefined. unsigned overflow wraps by both the
C and C++ standards.
--
What
--- Additional Comments From mathieu at malaterre dot com 2005-07-06 02:59
---
Using gcc --version:
g++ (GCC) 4.1.0 20050607 (experimental)
I can still reproduce the same unsafe operation:
#include iostream
int main()
{
unsigned int l = (unsigned int)-1;
const int a = l/2;
--- Additional Comments From mathieu at malaterre dot com 2005-07-06 03:00
---
Using gcc --version:
g++ (GCC) 4.1.0 20050607 (experimental)
I can still reproduce the same unsafe operation:
#include iostream
int main()
{
unsigned int l = (unsigned int)-1;
const int a = l/2;
The GCC 4.0.0 documentation, page 71 in the PostScript version,
http://gcc.gnu.org/onlinedocs/gcc-4.0.0/gcc/Optimize-Options.html#Optimize-Options
in the HTML version, says the following:
-ftree-dce
Perform dead code elimination (DCE) on trees. This flag is enabled by default
at -O and higher.
--- Additional Comments From pinskia at gcc dot gnu dot org 2005-07-06
04:12 ---
Confirmed, still wrong on the mainline.
--
What|Removed |Added
--- Additional Comments From pinskia at gcc dot gnu dot org 2005-07-06
04:45 ---
Confirmed. I hate static linking.
--
What|Removed |Added
Severity|normal
--
What|Removed |Added
GCC build triplet|i686-pc-linux-gnu, hppa-|
|unknown-linux-gnu |
GCC host triplet|i686-pc-linux-gnu, hppa-|
--- Additional Comments From pinskia at gcc dot gnu dot org 2005-07-06
04:58 ---
How much memory do you have?
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=20462
--- Additional Comments From albertm at uphs dot upenn dot edu 2005-07-06
05:09 ---
Created an attachment (id=9212)
-- (http://gcc.gnu.org/bugzilla/attachment.cgi?id=9212action=view)
minimal test case
The bug persists in gcc version 4.1.0 20050702 (experimental)
If I replace:
--- Additional Comments From dank at kegel dot com 2005-07-06 05:11 ---
Still seeing this on gcc-4.0.1-20050702.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=20583
101 - 118 of 118 matches
Mail list logo