With a patch to compile libjava on a per-directory basis, I found a bug with
imports
Take file AWTEvent.java
package java.awt;
import java.util.EventObject;
public abstract class AWTEvent extends EventObject {}
and file B.java
package java.awt;
import java.awt.event.ComponentEve
void
sys_futex0 (int *addr, int op, int val)
{
register long int r0 __asm__ ("r0");
register long int r3 __asm__ ("r3");
register long int r4 __asm__ ("r4");
register long int r5 __asm__ ("r5");
register long int r6 __asm__ ("r6");
r0 = 221;
r3 = (long) addr;
r4 = op;
r5 = val;
GCC is supposed to suppress division by zero warnings when floating point.
However the warning is based on the lexical form, i.e. the unpromoted type. I
suggest it should depend on the promoted type.
float f[] = {
1.0f/0.0f, // No warning
1.0f/0 // Warning (and diagnostic is on the wrong
Comparison between
Wed Apr 27 04:58:50 UTC 2005
and
Wed Apr 27 14:13:15 UTC 2005
New tests that FAIL:
...
gcc.dg/tree-ssa/gen-vect-26.c scan-tree-dump-times Alignment of access forced
using peeling 1
This was when the test was introduced, and it has failed since then.
It still fails with "Fri Ma
The following test program works with gcc-2.95.3 and gcc-3.4.1,
but aborts with this error on gcc-4.0.0:
z.ii: In function 'void bar()':
z.ii:11: internal compiler error: in create_tmp_var, at gimplify.c:368
struct Foo {
~Foo() {}
int i;
operator bool() { return !i; }
};
void bar() {
--- Additional Comments From dank at kegel dot com 2005-05-07 08:31 ---
Forgot to mention: I have applied the fixes for
PR20973 and PR21173 as given in bugzilla.
--
What|Removed |Added
-
Comparison between
Fri Apr 29 20:34:58 CEST 2005
Fri Apr 29 18:34:58 UTC 2005
and
Sat Apr 30 23:09:53 CEST 2005
Sat Apr 30 21:09:53 UTC 2005
Tests that now fail, but worked before:
gcc.c-torture/execute/20040703-1.c compilation, -Os
What's even worse, the same bug (according to the message in
--
What|Removed |Added
AssignedTo|unassigned at gcc dot gnu |hp at gcc dot gnu dot org
|dot org |
Status|UNCONFIRMED
--- Additional Comments From bonzini at gcc dot gnu dot org 2005-05-07
09:22 ---
With this version of AWTEvent.java, this is an ice-on-valid:
package java.awt;
import java.util.EventObject;
public abstract class AWTEvent extends EventObject {
AWTEvent() { super(null); }
--
What|Removed |Added
OtherBugsDependingO||21442
nThis||
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21436
Another bug in multifile processing. With these two classes:
Pipe.java:
package java.nio.channels;
import java.nio.channels.spi.AbstractSelectableChannel;
public abstract class Pipe { }
SelectableChannel.java:
package java.nio.channels;
import java.nio.channels.spi.AbstractInterruptibl
It uses -O1 -fdump-tree-store_ccp and tests for lack of link_error,
but note that store_ccp is enabled with -O2 or higher.
So the testcase is testing something that happens in
some other optimizer like DOM.
--
Summary: ssa-ccp-9.c etc is broken
Product: gcc
Versi
--- Additional Comments From steven at gcc dot gnu dot org 2005-05-07
13:54 ---
Looks like random instruction selection weirdness.
Not much can be done about this.
--
What|Removed |Added
--
--- Additional Comments From ornati at fastwebnet dot it 2005-05-07 14:04
---
ok
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19133
--- Additional Comments From steven at gcc dot gnu dot org 2005-05-07
14:04 ---
This is fixed on the tree-profiling-branch.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17667
--- Additional Comments From pinskia at gcc dot gnu dot org 2005-05-07
15:02 ---
Confirmed. Related to PR 20103.
--
What|Removed |Added
BugsThisDependsOn|
--- Additional Comments From kazu at cs dot umass dot edu 2005-05-07 15:03
---
FWIW, here is the last SSA form I get with vops.
foo (p)
{
int D.1241;
int global.0;
int D.1239;
:
# VUSE ;
D.1239_2 = p_1->c;
if (D.1239_2 != 0) goto ; else goto ;
:;
# VUSE ;
global.0_
--- Additional Comments From danglin at gcc dot gnu dot org 2005-05-07
15:05 ---
Here's the story on this problem. Ada bootstrap was broken on this
target by this change:
2005-04-08 Diego Novillo <[EMAIL PROTECTED]>
Merge from tree-cleanup-branch: VRP, store CCP, store
Run the test case below, hold CTRL and try clicking multiple items.
Test case:
import javax.swing.*;
import java.awt.*;
public class testlist
{
public static void main(String[] args)
{
JFrame f = new JFrame();
DefaultListModel v = new DefaultListModel();
v.addElement("alsdkfj");
--
What|Removed |Added
Status|UNCONFIRMED |NEW
Ever Confirmed||1
Last reconfirmed|-00-00 00:00:00 |2005-05-
--- Additional Comments From steven at gcc dot gnu dot org 2005-05-07
15:27 ---
Looks like -D__NO_MATH_INLINES makes gcc produced code much better...
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=13712
--- Additional Comments From pinskia at gcc dot gnu dot org 2005-05-07
15:52 ---
(In reply to comment #0)
> We have the following coverage for the COMPONENT_REF case.
Actually I think at one point this was valid to do but that was before gimple
and generic and tree-ssa.
I think the code
--
What|Removed |Added
URL||http://gcc.gnu.org/ml/gcc-
||patches/2005-
|
--
What|Removed |Added
Status|UNCONFIRMED |NEW
Ever Confirmed||1
Last reconfirmed|-00-00 00:00:00 |2005-05-
--- Additional Comments From tobi at gcc dot gnu dot org 2005-05-07 15:56
---
*** This bug has been marked as a duplicate of 19425 ***
--
What|Removed |Added
S
--- Additional Comments From tobi at gcc dot gnu dot org 2005-05-07 15:56
---
*** Bug 21434 has been marked as a duplicate of this bug. ***
--
What|Removed |Added
--- Additional Comments From cvs-commit at gcc dot gnu dot org 2005-05-07
16:46 ---
Subject: Bug 21412
CVSROOT:/cvs/gcc
Module name:gcc
Changes by: [EMAIL PROTECTED] 2005-05-07 16:46:09
Modified files:
gcc: ChangeLog
gcc/config/rs6000:
--
What|Removed |Added
Component|fortran |libfortran
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21435
--
What|Removed |Added
Severity|normal |minor
Summary|gfortran fails to open |fails to open nonexisting
|nonexisting
--- Additional Comments From cvs-commit at gcc dot gnu dot org 2005-05-07
16:50 ---
Subject: Bug 21412
CVSROOT:/cvs/gcc
Module name:gcc
Branch: gcc-4_0-branch
Changes by: [EMAIL PROTECTED] 2005-05-07 16:49:59
Modified files:
gcc: Change
--- Additional Comments From ebotcazou at gcc dot gnu dot org 2005-05-07
16:52 ---
> I don't know that I'm in a position to make a good decision here, but I
> do agree with Jeff -- it would be a lot more compelling to make the
> change if there was actually a demonstrable problem -- es
--- Additional Comments From pinskia at gcc dot gnu dot org 2005-05-07
16:56 ---
Confirmed.
--
What|Removed |Added
Status|UNCONFIRMED |NEW
E
--- Additional Comments From pinskia at gcc dot gnu dot org 2005-05-07
16:56 ---
Confirmed.
--
What|Removed |Added
Status|UNCONFIRMED |NEW
E
--
What|Removed |Added
AssignedTo|unassigned at gcc dot gnu |kazu at cs dot umass dot edu
|dot org |
Status|NEW
--- Additional Comments From kazu at cs dot umass dot edu 2005-05-07 17:02
---
A quick grep for fdump-tree-store_ccp shows
20030731-2.c:/* { dg-options "-O1 -fdump-tree-store_ccp" } */
20030917-1.c:/* { dg-options "-O1 -fdump-tree-store_ccp" } */
20030917-3.c:/* { dg-options "-O1 -fno-t
--- Additional Comments From pinskia at gcc dot gnu dot org 2005-05-07
17:30 ---
Confirmed, this happens with any file. Some how the scheduler is broken.
Something changed between 20050502 and now.
--
What|Removed |Added
-
--- Additional Comments From pinskia at gcc dot gnu dot org 2005-05-07
17:43 ---
This is fall out from Aldy's changing to use .opt.
--
What|Removed |Added
CC
Hi folks,
The attached testcase dies on Debian (amd64) if
compiled with -O2.
Here is the usual data:
% gcc -v
Using built-in specs.
Target: x86_64-linux
Configured with: ../src/configure -v
--enable-languages=c,c++,java,f95,objc,ada,treelang --prefix=/usr
--libexecdir=/usr/lib --enable-shared
On May 7, 2005, at 2:04 PM, Harald Dunkel wrote:
Hi folks,
The testcase has been created by grabbing some code
fragments from libX11, XFree86 4.3 on Debian, amd64.
-Wall complains about dereferencing type-punned pointer.
Stop right there, you said it warns, well this is a case where
you are violati
--- Additional Comments From pinskia at gcc dot gnu dot org 2005-05-07
18:18 ---
Confirmed, it is obvious from looking at the code what was wrong. The
following patch should fix it
but I don't have time to test it fully:
Index: c-typeck.c
==
--- Additional Comments From pinskia at gcc dot gnu dot org 2005-05-07
18:20 ---
Hmm, I wonder why C++ warns though:
t.c:2: warning: division by zero in '1.0e+0f / 0.'
t.c:3: warning: division by zero in '1.0e+0f / 0'
--
What|Removed |Added
--
Just a meta-bug showing where a reassociate pass could help.
--
Summary: [meta-bug] we need a reassociate pass
Product: gcc
Version: 4.0.0
Status: UNCONFIRMED
Keywords: missed-optimization
Severity: enhancement
Priority: P
--- Additional Comments From pinskia at gcc dot gnu dot org 2005-05-07
18:37 ---
Fixed by the patch for PR 18316 and PR 19126.
--
What|Removed |Added
Status|NEW
--- Additional Comments From pinskia at gcc dot gnu dot org 2005-05-07
18:45 ---
CCing Danny as he is writting this pass right now.
--
What|Removed |Added
CC
--- Additional Comments From pinskia at gcc dot gnu dot org 2005-05-07
18:52 ---
The orginal testcase in comment #0 is fixed now but not the one in comment #2.
--
What|Removed |Added
--- Additional Comments From pinskia at gcc dot gnu dot org 2005-05-07
19:02 ---
Fixed by:
* config.gcc ({powerpc,rs6000}-ibm-aix*): Remove rs6000/aix.opt
from extra_options.
* config.in (HAVE_AS_POPCNTB): New.
* configure.ac (HAVE_AS_MFCRF): Add .machine
--- Additional Comments From pinskia at gcc dot gnu dot org 2005-05-07
19:09 ---
On the mainline I get:
void foo() ()
{
void * ivtmp.3;
struct NonPod x;
:
ivtmp.3 = &x.foo[2];
:;
ivtmp.3 = ivtmp.3 - 4B;
if (ivtmp.3 == &x.foo) goto ; else goto ;
:;
return;
}
Which is stil
Andrew Pinski wrote:
>
> Stop right there, you said it warns, well this is a case where
> you are violating C89/C99/C++ aliasing rules so it is a bug in your
> code and not in GCC.
>
This is not my code. It is XFree86 4.3. I am just trying to
help by investigating a problem and providing an easy t
--- Additional Comments From pinskia at gcc dot gnu dot org 2005-05-07
19:30 ---
Huh, I don't understand why VRP does not take care of this:
i_13: [-2147483648, ]
D.1236_4 = i_13 <= ;
i_13 = ASSERT_EXPR ;
D.1236_4 = i_13 <= ;
The following is the testcas
On May 7, 2005, at 3:25 PM, Harald Dunkel wrote:
Andrew Pinski wrote:
Stop right there, you said it warns, well this is a case where
you are violating C89/C99/C++ aliasing rules so it is a bug in your
code and not in GCC.
This is not my code. It is XFree86 4.3. I am just trying to
help by investiga
Trying to build gcc-3.4.3 for IBM AIX 5.3L on an IBM p570 machine. Fails, and a
message asks me to file a bug report.
I have generated the .i file using ./xgcc -v -save-temps ... ... but I don't
see a way to attach the .i file (and .s file if you need) in this Bugzilla
form. I can email it on r
--- Additional Comments From pinskia at gcc dot gnu dot org 2005-05-07
19:33 ---
*** This bug has been marked as a duplicate of 9382 ***
--
What|Removed |Added
--- Additional Comments From pinskia at gcc dot gnu dot org 2005-05-07
19:33 ---
*** Bug 21446 has been marked as a duplicate of this bug. ***
--
What|Removed |Added
--- Additional Comments From cvs-commit at gcc dot gnu dot org 2005-05-07
19:40 ---
Subject: Bug 21441
CVSROOT:/cvs/gcc
Module name:gcc
Changes by: [EMAIL PROTECTED] 2005-05-07 19:40:39
Modified files:
gcc: ChangeLog
Log message:
PR t
--- Additional Comments From hp at gcc dot gnu dot org 2005-05-07 19:43
---
See http://gcc.gnu.org/ml/gcc-patches/2005-05/msg00595.html>.
--
What|Removed |Added
--- Additional Comments From cvs-commit at gcc dot gnu dot org 2005-05-07
19:48 ---
Subject: Bug 18958
CVSROOT:/cvs/gcc
Module name:gcc
Branch: gcc-4_0-branch
Changes by: [EMAIL PROTECTED] 2005-05-07 19:48:19
Modified files:
libgfortran: Change
--- Additional Comments From tkoenig at gcc dot gnu dot org 2005-05-07
19:49 ---
Fixed in the 4.0 branch. Closing.
--
What|Removed |Added
Status|NEW
--- Additional Comments From pinskia at gcc dot gnu dot org 2005-05-07
19:56 ---
Fixed.
--
What|Removed |Added
Status|NEW |RESOLVED
--
Bug 11706 depends on bug 19401, which changed state.
Bug 19401 Summary: Trivial loop not unrolled
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19401
What|Old Value |New Value
--
What|Removed |Added
Target Milestone|--- |4.1.0
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21441
This problem occurs compiling wxWidgets 2.6.0 and also if I try to compile a
program using wxWidgets 2.5.x compiled with gcc 3.4.2.
It's the one when compiling wxWidgets 2.6.0:
$ sh ../configure CXXFLAGS='-O4 -s -DNO_GCC_PRAGMA' --with-msw --enable-stl
--enable-shared --with-odbc
$ make
../src/c
--- Additional Comments From pinskia at gcc dot gnu dot org 2005-05-07
20:20 ---
*** This bug has been marked as a duplicate of 21275 ***
--
What|Removed |Added
--- Additional Comments From pinskia at gcc dot gnu dot org 2005-05-07
20:20 ---
*** Bug 21447 has been marked as a duplicate of this bug. ***
--
What|Removed |Added
--- Additional Comments From hjl at lucon dot org 2005-05-07 20:24 ---
That is a binutils bug:
http://sources.redhat.com/bugzilla/show_bug.cgi?id=940
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21070
--- Additional Comments From dje at gcc dot gnu dot org 2005-05-07 20:24
---
i386 synthesizes a DImode compare for 32-bit mode in target-dependent code.
PowerPC probably needs to do the same.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21237
The following test failures have appeared on mainline between 20050506 and
20050507. All are "internal compiler error: in estimate_num_insns_1, at
tree-inline.c:1433".
On i686-pc-linux-gnu and ia64-hpux:
FAIL: gcc.dg/vect/vect-ifcvt-2.c (test for excess errors)
FAIL: gcc.dg/vect/vect
--- Additional Comments From pinskia at gcc dot gnu dot org 2005-05-07
20:38 ---
I think this was caused by:
2005-05-06 Zdenek Dvorak <[EMAIL PROTECTED]>
PR tree-optimization/19401
* tree-flow.h (tree_unroll_loops_completely): Declaration changed.
* tre
--- Additional Comments From pinskia at gcc dot gnu dot org 2005-05-07
20:47 ---
On the mainline, we get:
.L13:
fld %st(0)
incl%ecx
fmuls (%edx)
addl%ebx, %edx
fadds (%eax)
fstps (%eax)
addl%esi, %eax
c
--
Bug 18754 depends on bug 18755, which changed state.
Bug 18755 Summary: invariant/constant not recomputed when ADDR_EXPR changed
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18755
What|Old Value |New Value
-
--
Bug 9895 depends on bug 18755, which changed state.
Bug 18755 Summary: invariant/constant not recomputed when ADDR_EXPR changed
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18755
What|Old Value |New Value
--
--- Additional Comments From pinskia at gcc dot gnu dot org 2005-05-07
21:00 ---
This has now been fixed.
--
What|Removed |Added
Status|NEW
--- Additional Comments From pinskia at gcc dot gnu dot org 2005-05-07
21:22 ---
For this one, GCC is over estimating the size of the loop.
--
What|Removed |Added
Andrew Pinski <[EMAIL PROTECTED]> writes:
| On May 6, 2005, at 8:09 PM, Julian Cummings wrote:
|
| > People are reporting trouble compiling blitz with gcc-4.0.0, and the
| > compiler errors are resulting from the use of unnamed enums. A
| > simple code
| > illustrates the problem:
| >
| > stru
The loop unroller over estimates the unrolled sized of an empty loop:
int f(void)
{
int i;
for(i = 0 ;i < 10; i++);
}
We get:
Loop size: 5
Estimated size after unrolling: 10
--
Summary: Loop unroller is way over estimating the unroll size of
a loop
P
--
What|Removed |Added
Version|4.0.0 |4.1.0
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21449
"Julian Cummings" <[EMAIL PROTECTED]> writes:
| Hmmm... I just read through the bug reports you cited. Sounds to me like
| this is still somewhat of an open issue, as to whether the compiler should
| issue an error in these cases or simply silently discard any templated
| function as a possible m
On May 7, 2005, at 6:40 PM, Gabriel Dos Reis wrote:
"Julian Cummings" <[EMAIL PROTECTED]> writes:
| Hmmm... I just read through the bug reports you cited. Sounds to me
like
| this is still somewhat of an open issue, as to whether the compiler
should
| issue an error in these cases or simply sile
This C++ code snippet aborts on amd64 and i386 if compiled with -O[23]:
(Note: tree dumps are correct)
---
extern "C" void abort (void);
struct T {};
struct U : T {};
int main ()
{
int i;
U *p;
T **q;
for (i = 0
--- Additional Comments From jvdelisle at verizon dot net 2005-05-07 22:16
---
A patch to fix the problem in comment #7 has been submitted for review.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17423
--- Additional Comments From pinskia at gcc dot gnu dot org 2005-05-07
22:19 ---
Confirmed, here is a testcase which makes it fail with 3.4.0:
extern "C" void abort (void) throw();
struct T {};
struct U : T {};
int size = 2;
int main ()
{
struct T * * pretmp3;
struct U * p;
int i
--- Additional Comments From falk at debian dot org 2005-05-07 22:27
---
I guess I'm missing something in here. *q accesses an object of type
U* (namely p). The type of the lvalue *q is T*. Right? Which case in
3.10/15 covers this?
--
What|Removed |A
--- Additional Comments From cvs-commit at gcc dot gnu dot org 2005-05-07
22:28 ---
Subject: Bug 21403
CVSROOT:/cvs/gcc
Module name:gcc
Branch: gcc-4_0-branch
Changes by: [EMAIL PROTECTED] 2005-05-07 22:28:10
Modified files:
gcc: Change
--- Additional Comments From pinskia at gcc dot gnu dot org 2005-05-07
22:30 ---
*** Bug 21450 has been marked as a duplicate of this bug. ***
--
What|Removed |Added
--- Additional Comments From pinskia at gcc dot gnu dot org 2005-05-07
22:30 ---
Yes the code is invalid after I thought about it and looked for another bug
which has the same issue,
see PR 11376.
*** This bug has been marked as a duplicate of 11376 ***
--
What|Remove
--- Additional Comments From cvs-commit at gcc dot gnu dot org 2005-05-07
22:32 ---
Subject: Bug 21403
CVSROOT:/cvs/gcc
Module name:gcc
Changes by: [EMAIL PROTECTED] 2005-05-07 22:32:04
Modified files:
gcc: ChangeLog configure.ac configure
Lo
--- Additional Comments From green at redhat dot com 2005-05-07 22:33
---
Patch applied to HEAD and 4.0 branch.
--
What|Removed |Added
Status|ASSIGNED
--
What|Removed |Added
Target Milestone|--- |4.0.1
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21403
--- Additional Comments From pinskia at gcc dot gnu dot org 2005-05-07
22:42 ---
Confirmed.
--
What|Removed |Added
Status|UNCONFIRMED |NEW
E
--- Additional Comments From pinskia at gcc dot gnu dot org 2005-05-07
22:47 ---
Confirmed.
--
What|Removed |Added
Status|UNCONFIRMED |NEW
E
--
What|Removed |Added
Summary|mudflap reports errors |mudflap reports errors for
||external variables
Versi
Andrew Pinski wrote:
>
> Wrong, try again. Violating aliasing rules cause undefined behavior
> so seg faulting is an okay thing to do.
But producing a warning message and bad code is not OK. Either
using a "type-punned pointer" should be treated as a fatal
error, because gcc would create bad code
91 matches
Mail list logo