--- Additional Comments From lerdsuwa at gcc dot gnu dot org 2004-10-24 08:44
---
Will look at it.
--
What|Removed |Added
AssignedTo|unassigned at gcc dot gnu
CVS HEAD 20041024, works ok for i486-linux
configure:3264: /build/packages/gcc/snap/gcc-snapshot-20041024/build/gcc/xgcc -s
hared-libgcc -B/build/packages/gcc/snap/gcc-snapshot-20041024/build/gcc/ -nostdi
nc++ -L/build/packages/gcc/snap/gcc-snapshot-20041024/build/hppa-linux-gnu/libst
dc++-v3/src
The C++ parser fails to parse sizeof compound-literal correctly. For example,
int x = sizeof (int) { 1 };
receives the error
t.C:1: error: expected %,% or %;% before '{' token
As shown in the grammar for GNU extensions in comments in cp/parser.c,
compound literals are considered
--- Additional Comments From pinskia at gcc dot gnu dot org 2004-10-24 12:58
---
I think this is related to PR 17919 (or is a dup of that bug). I think this is GC
related.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18125
--
What|Removed |Added
Severity|normal |enhancement
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=16464
../../xgcc -B../../ -c -g -O2 -W -Wall -gnatpg s-fore.adb -o s-fore.o
+===GNAT BUG
DETECTED==+
| 4.0.0 20041023 (experimental) (powerpc-apple-darwin7.5.0) GCC error: |
| in rs6000_emit_minmax, at config/rs6000/rs6000.c:11805
--
What|Removed |Added
Keywords||ice-on-valid-code
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18127
--- Additional Comments From pinskia at gcc dot gnu dot org 2004-10-24 16:23
---
*** Bug 18127 has been marked as a duplicate of this bug. ***
--
What|Removed |Added
--- Additional Comments From pinskia at gcc dot gnu dot org 2004-10-24 16:23
---
*** This bug has been marked as a duplicate of 17956 ***
--
What|Removed |Added
--- Additional Comments From pinskia at gcc dot gnu dot org 2004-10-24 16:23
---
Confirmed via the dup bug.
--
What|Removed |Added
Status|UNCONFIRMED
--- Additional Comments From bangerth at dealii dot org 2004-10-24 18:48 ---
Confirmed. As an additional data point, icc accepts the code without any
problem.
W.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18126
--
What|Removed |Added
Status|UNCONFIRMED |NEW
Ever Confirmed||1
Keywords|
Produces wrong code on both 3.3.4 and 3.4.2 (in Debian Sid). It works perfectly
with gcc 2.95.
The problem is that,
-A+B == B-A
-A()+B == B-A()
-A+B() == B()-A
but,
-A()+B() ?= B()-A()
if A or B depend on each other, the last condition is !=. Otherwise, it is ==.
This breaks any code that relies
--- Additional Comments From pinskia at gcc dot gnu dot org 2004-10-24 19:01
---
Invalid:
printf( -4.0 is %f\n, -A() + A());
either A() can be evaluated first which is the misunderstanding in your part.
--
What|Removed |Added
struct S { char *a, *b; };
extern void foo (struct S *);
int
main ()
{
struct S s[] = {
{ABCDEFGH0123, T},
{ABCDEFGH4567, T},
{ABCDEFGH89ZYX, T},
{IJK012, T},
{IJK345, T},
{IJK678, T},
{IJKLMN, T},
{IJKOPQ, T},
{0, 0}
};
foo (s);
return 0;
}
at -O2
--
What|Removed |Added
Keywords||wrong-code
Summary|-fwritable-strings doesn't |[3.4 Regression] -fwritable-
--- Additional Comments From pinskia at gcc dot gnu dot org 2004-10-24 19:56
---
Confirmed.
20030422 (236) between 20030504 (237)
--
What|Removed |Added
--- Additional Comments From pinskia at gcc dot gnu dot org 2004-10-24 20:31
---
Patch here: http://gcc.gnu.org/ml/gcc-patches/2004-10/msg02020.html.
--
What|Removed |Added
I doubt this code is valid, but anyway, GCC should display a proper
error message rather than just crash.
I am using a GCC checked out from CVS HEAD yesterday.
I also get an ICE with GCC 3.3.4-11 and 3.4.1-17 from Debian,
but i haven't tried the corresponding FSF releases.
--- Source file ---
--- Additional Comments From pinskia at gcc dot gnu dot org 2004-10-24 21:46
---
Confirmed.
--
What|Removed |Added
Status|UNCONFIRMED |NEW
--
Summary: [meta-bug] inner class problems in java front-end
Product: gcc
Version: 4.0.0
Status: UNCONFIRMED
Keywords: ice-on-valid-code, ice-on-invalid-code, rejects-valid,
accepts-invalid, meta-bug
--
What|Removed |Added
OtherBugsDependingO||18131
nThis||
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18130
--
What|Removed |Added
BugsThisDependsOn||12788
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18131
--
What|Removed |Added
BugsThisDependsOn||15714
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18131
--
What|Removed |Added
OtherBugsDependingO||18131
nThis||
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17747
--- Additional Comments From karl at grebyn dot com 2004-10-24 22:39 ---
Problem is that the documentation has an incorrect specification of the interface.
By changing ex7.h to the following, the program works.
//ex7.h
class Origin {
public:
int
# make
make[1]: Entering directory `/tmp/gcc-3.4.2/libiberty'
if [ x-fpic != x ]; then \
gcc -c -DHAVE_CONFIG_H -s -O3 -march=i686 -I. -I./../include -W -Wall
-Wtraditional -pedantic -fpic regex.c -o pic/regex.o; \
else true; fi
regex.c:132: warning: conflicting types for built-in function
--- Additional Comments From dje at gcc dot gnu dot org 2004-10-24 22:52 ---
This problem appears to be a confluence of issues between fix_truncdfdi2 and
limitations of reload, exacerbated by Apple's LL64 mode.
The problematic code is
typedef unsigned long clock_t;
clock_t clock(void);
--- Additional Comments From pinskia at gcc dot gnu dot org 2004-10-24 22:53
---
How did you configure GCC and what are your CFLAGS?
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18132
--
What|Removed |Added
Status|UNCONFIRMED |WAITING
Keywords||build
void foo (void)
{
void *p;
p = L0;
goto *p;
L0:
return;
}
This is only needed for tree level.
--
Summary: computed gotos are not folded back to regual gotos when
it is found it is not a computed goto at all
Product: gcc
Version: 4.0.0
--- Additional Comments From pinskia at gcc dot gnu dot org 2004-10-24 23:36
---
Confirmed via Steven on IRC (really he was the one who gave the example).
--
What|Removed |Added
int foo (int a)
{
void *p;
if (a!=0)
p = L0;
else
p = L1;
goto *p;
L0:
return 1;
L1:
return 0;
}
this should be equivant to:
int foo(int a) { return a !=0;}
--
Summary: computed goto and if statement
Product: gcc
Version: 4.0.0
Status:
--
What|Removed |Added
BugsThisDependsOn||18133
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18134
--- Additional Comments From steven at gcc dot gnu dot org 2004-10-24 23:52
---
Fixing this is easy
for all bb
if bb ends in a computed goto
(which on tree means there is *only* a computed goto)
find the computed goto target
if the target is an SSA_NAME and the
--- Additional Comments From pinskia at gcc dot gnu dot org 2004-10-24 23:55
---
Confirmed via Steven.
--
What|Removed |Added
Status|UNCONFIRMED
I found this when I was running the testsuite on a new compiler with a new prefix that
was empty.
creating gctest
make AR_FLAGS=rc CC_FOR_BUILD=gcc CFLAGS=-O2 -g -O2 CXXFLAGS=-g -O2
CFLAGS_FOR_BUILD= CFLAGS_FOR_TARGET=-O2 -g -O2 INSTALL=/usr/bin/install -c
INSTALL_DATA=/usr/bin/install -c -m
--- Additional Comments From pinskia at gcc dot gnu dot org 2004-10-25 00:48
---
Can you tell us if this patch works?
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=14939
--
What|Removed |Added
Attachment #6828|application/octet-stream|text/plain
mime type||
Since the arguments are gimple registers, gimple optimizers are happy to create many
references to it.
While lowering them to RTL these however compile to memory loads causing number of
redundant loads.
The attached quicksort loop on when compiled with -O2 -fno-loop-optimize (the other
being
--- Additional Comments From hubicka at gcc dot gnu dot org 2004-10-25 00:50
---
Created an attachment (id=7406)
-- (http://gcc.gnu.org/bugzilla/attachment.cgi?id=7406action=view)
the quicksort loop
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18136
Since the arguments are gimple registers, gimple optimizers are happy to create many
references to it.
While lowering them to RTL these however compile to memory loads causing number of
redundant loads.
The attached quicksort loop on when compiled with -O2 -fno-loop-optimize (the other
being
--- Additional Comments From hubicka at gcc dot gnu dot org 2004-10-25 00:51
---
Created an attachment (id=7407)
-- (http://gcc.gnu.org/bugzilla/attachment.cgi?id=7407action=view)
testcase
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18137
--- Additional Comments From pinskia at gcc dot gnu dot org 2004-10-25 00:59
---
Confirmed, here is a reduced testcase:
unsigned char a, b, c, d, e, f, g;
void safer_ecb_encrypt(unsigned char *block_out, unsigned char *key, unsigned int
round)
{
while(round-- 0)
{
a ^=
--- Additional Comments From pinskia at gcc dot gnu dot org 2004-10-25 01:03
---
*** This bug has been marked as a duplicate of 18136 ***
--
What|Removed |Added
--- Additional Comments From pinskia at gcc dot gnu dot org 2004-10-25 01:03
---
*** Bug 18137 has been marked as a duplicate of this bug. ***
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18136
--- Additional Comments From pinskia at gcc dot gnu dot org 2004-10-25 01:08
---
Lets reopen this one as this is the one with the testcase.
--
What|Removed |Added
--- Additional Comments From pinskia at gcc dot gnu dot org 2004-10-25 01:09
---
*** This bug has been marked as a duplicate of 18137 ***
--
What|Removed |Added
--- Additional Comments From pinskia at gcc dot gnu dot org 2004-10-25 01:09
---
*** Bug 18136 has been marked as a duplicate of this bug. ***
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18137
--- Additional Comments From geoffk at gcc dot gnu dot org 2004-10-25 02:21
---
GCC is now more consistent about when this bug triggers, so some cases which passed in
3.4 fail in
4.0.
There are testcases for this bug, gcc.dg/c90-const-expr-3.c and
gcc.dg/c99-const-expr-3.c.
--
--
What|Removed |Added
Target Milestone|--- |4.0.0
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=5675
--- Additional Comments From pinskia at gcc dot gnu dot org 2004-10-25 02:35
---
Fixed by:
2004-10-24 Geoffrey Keating [EMAIL PROTECTED]
* config/rs6000/darwin.h (BLOCK_REG_PADDING): Define.
--
What|Removed |Added
--- Additional Comments From pinskia at gcc dot gnu dot org 2004-10-25 02:35
---
The darwin problem has now been fixed by:
2004-10-24 Geoffrey Keating [EMAIL PROTECTED]
* config/rs6000/darwin.h (BLOCK_REG_PADDING): Define.
--
What|Removed
--
Bug 15923 depends on bug 14262, which changed state.
Bug 14262 Summary: Structure size computed wrong
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=14262
What|Old Value |New Value
--
What|Removed |Added
CC||pinskia at gcc dot gnu dot
||org
--
What|Removed |Added
Keywords||missed-optimization
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18137
--- Additional Comments From pinskia at gcc dot gnu dot org 2004-10-25 03:33
---
This is really a rtl problem, the problem comes from greg. before that we have:
(insn:HI 7 11 8 0 (set (reg/v:SI 68 [ start ])
(mem/i:SI (reg/f:SI 16 argp) [3 start+0 S4 A32])) 44 {*movsi_1} (nil)
The following patch breaks the 64-bit testsuite for mips-sgi-irix6.5. I have
tracked it down in HEAD but the same patch causes a similar problem in the 3.4
branch.
2004-11-18 H.J. Lu [EMAIL PROTECTED]
PR bootstrap/17684
* Makefile.in (clean): Remove
--- Additional Comments From pinskia at gcc dot gnu dot org 2004-10-25 03:41
---
Confirmed, Eric Botcazou posted a patch here:
http://gcc.gnu.org/ml/gcc-patches/2004-10/
msg02011.html to fix the problem.
--
What|Removed |Added
--- Additional Comments From david dot billinghurst at comalco dot riotinto dot
com dot au 2004-10-25 03:56 ---
Subject: RE: [3.4/4.0 regression] libgcc_s.so.1 not found by 64-bit testsuite
The gfortran.exp part works for me on mips-sgi-irix6.5. Will try the rest overnight.
60 matches
Mail list logo