On Sunday 27 March 2005 04:45, Canqun Yang wrote:
Another question is why the new RTL loop-unroller does
not support giv splitting.
Apparently because for most people it is not a problem that it does
not do it, and while you have indicated earlier that it may be useful
for you, you have neither
%cat LAST_UPDATED
Sat Mar 26 21:31:28 EST 2005
Sun Mar 27 02:31:28 UTC 2005
stage1/xgcc -Bstage1/ -B/opt/gcc-head//powerpc-apple-darwin7.8.0/bin/
-c -g -O2 -mdynamic-no-pic -DIN_GCC -W -Wall -Wwrite-strings
-Wstrict-prototypes -Wmissing-prototypes -pedantic -Wno-long-long
Geert Bosch wrote:
%cat LAST_UPDATED
Sat Mar 26 21:31:28 EST 2005
Sun Mar 27 02:31:28 UTC 2005
stage1/xgcc -Bstage1/ -B/opt/gcc-head//powerpc-apple-darwin7.8.0/bin/
-c -g -O2 -mdynamic-no-pic -DIN_GCC -W -Wall -Wwrite-strings
-Wstrict-prototypes -Wmissing-prototypes -pedantic -Wno-long-long
Dear GNU,
Hello? I'm Anderson Shin([EMAIL PROTECTED]), director of
IPEAN(ipean.com) company in South Korea.
To begin with, please let me introduce shortly about our company.
We are developping EDA(Electronic Design Automation) tools that is CAD
system for digital/analog IC/circuit designing.
I guess its due to the following patch (
http://gcc.gnu.org/ml/gcc-cvs/2005-03/msg01320.html).
I get the following error when trying to build gcc on powerpc-apple-darwin:
gcc -c -DHAVE_CONFIG_H -g -O2 -I. -I../../../gcc/libiberty/../include -W
-Wall -Wtraditional -pedantic
From: Ian Lance Taylor ian@airs.com
Paul Schlie [EMAIL PROTECTED]
(is this not feasible if the target is accurately described in rtl?)
I don't know how to respond to this. I'm discussing a way to achieve
an incremental improvement in gcc. You seem to be discussing a
different compiler.
Snapshot gcc-4.1-20050327 is now available on
ftp://gcc.gnu.org/pub/gcc/snapshots/4.1-20050327/
and on various mirrors, see http://gcc.gnu.org/mirrors.html for details.
This snapshot has been generated from the GCC 4.1 CVS branch
with the following options: -D2005-03-27 17:43 UTC
You'll find
Hello,
On Sunday 27 March 2005 04:45, Canqun Yang wrote:
Another question is why the new RTL loop-unroller does
not support giv splitting.
Apparently because for most people it is not a problem that it does
not do it, and while you have indicated earlier that it may be useful
for you,
If you run 'make check' after --enable-mapped-location (even
just --enable-languages=c) you'll find some apparant regressions.
They aren't real regressions - it's just now we now get column numbers
in some of the diagnostic messages, and this confuses dejagnu.
Now I'm willing to fix those tests by
On Sunday, March 27, 2005, at 08:55 AM, anderson shin wrote:
However we always respect an opinion of the GNU. So we will follow
your decision and we hope our suggestion will be accepted.
Mostly, this is off-topic for this list. gnu.misc.discuss is the
canonical place for such discussions. I'll
On Sunday, March 27, 2005, at 11:58 AM, Per Bothner wrote:
If you run 'make check' after --enable-mapped-location (even
just --enable-languages=c) you'll find some apparant regressions.
They aren't real regressions - it's just now we now get column numbers
in some of the diagnostic messages, and
On Sunday, March 27, 2005, at 11:58 AM, Per Bothner wrote:
Now I'm willing to fix those tests by adding -fno-show-column where
necessary
Ick. I favor adding it unconditionally to compile lines over this.
See -fmessage-length code (gcc/testsuite/lib/g++.exp) for hints. And
even that, I'm not
On Sunday, March 27, 2005, at 09:31 AM, Gabriel Dos Reis wrote:
Fixed with this
http://gcc.gnu.org/ml/gcc-patches/2005-03/msg02450.html
Please try again and let me know.
A quick check of build's libiberty, seems to build for me now on
darwin8.
On 27/03/2005, at 4:00 AM, Geert Bosch wrote:
%cat LAST_UPDATED
Sat Mar 26 21:31:28 EST 2005
Sun Mar 27 02:31:28 UTC 2005
stage1/xgcc -Bstage1/ -B/opt/gcc-head//powerpc-apple-darwin7.8.0/bin/
-c -g -O2 -mdynamic-no-pic -DIN_GCC -W -Wall -Wwrite-strings
-Wstrict-prototypes
ÒýÑÔ Zdenek Dvorak [EMAIL PROTECTED]:
Hello,
On Sunday 27 March 2005 03:53, Canqun Yang wrote:
The last ChangeLog of rtlopt-branch was written
in
2003. After more than one year, many impovements
in
this branch haven't been put into the GCC HEAD.
Why?
Almost all of the rtlopt
ÒýÑÔ Zdenek Dvorak [EMAIL PROTECTED]:
Hello,
On Sunday 27 March 2005 04:45, Canqun Yang wrote:
Another question is why the new RTL loop-
unroller does
not support giv splitting.
Apparently because for most people it is not a
problem that it does
not do it, and while you have
Hi Diego,
By merging, do you mean *replacing* CCP with VRP? Yes, it's
doable. No, it's not a good idea.
Understood.
Also, if we are inserting ASSERT_EXPRs, it seems to be a good idea to
run copy-prop before VRP. Otherwise, we would end up with lots of
D.18001_101 = D.18001_198;
Hi Diego,
There is a copy-propagation pass before VRP. Or do you mean
right before? Sure, the ordering of these passes is in eternal
flux anyway.
Before, but doesn't have to be right before. The current ordering
is reasonable.
Currently, we still have these even after copy prop because
Hi Ian, (getting back to reality) upon reviewing things further, it appears
that if GCC could relax it's single-set restriction to enable a restricted
form of multi-set instructions to be included in optimizations; then ISA's
who's instructions either implicitly set or depend on global machine
i build a crosscompiler for gcc, abi=n32
gcc-4.1-20050327/configure -target=mips64el-linux
-prefix=/opt/gcc-4.1-20050327/ -enable-languages=c --disable-shared
make
it will error with config/mips/mips.c
--- Additional Comments From aj at gcc dot gnu dot org 2005-03-27 08:16
---
glibc needs to be changed for this, for details chech the thread starting at:
http://sourceware.org/ml/libc-hacker/2005-03/msg00061.html
--
What|Removed |Added
This error only seems to occur when -O2 -march=k6-2 -ftracer are used
together. Removing -ftracer causes the error to go away. This problem does not
seem to exist in the GCC4.0 branch.
i686-pc-linux-gnu-gcc -DHAVE_CONFIG_H -I. -I. -I.. -I../include -O20 -ffast-math
-D_REENTRANT -fsigned-char
--- Additional Comments From halcy0n at gentoo dot org 2005-03-27 08:50
---
Created an attachment (id=8461)
-- (http://gcc.gnu.org/bugzilla/attachment.cgi?id=8461action=view)
Preprocessed file for above failure
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=20653
--- Additional Comments From steven at gcc dot gnu dot org 2005-03-27
10:19 ---
Two things:
1) Test case? No test case, no way to reproduce it without re-doing the
investigating you have already done. Stop work duplication, provide
test cases to your fellow GCC hackers. I
--- Additional Comments From dorit at il dot ibm dot com 2005-03-27 12:36
---
patch: http://gcc.gnu.org/ml/gcc-patches/2005-03/msg02442.html
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=20626
binutils ar was recently changed to exclude path when comparing object
filenames, to agree with POSIX. This combines with Windows' case-insensitive
filesystem to cause java/lang/Exception.o to replace exception.o in the
following command while creating libgcj.a.
ar rc
--
What|Removed |Added
Status|UNCONFIRMED |NEW
Ever Confirmed||1
Last reconfirmed|-00-00 00:00:00
--- Additional Comments From andrewhutchinson at cox dot net 2005-03-27
14:33 ---
The problem here is that gcc is using a DImode register to handle 6 byte
(int+long) structure. Why I have no idea!
Since the target has no insn for DI move, gcc turns this into individual QImode
byte
/*** SNIP ***/
/*
Place this file into `bug.c' (the code between the SNIP comments).
Then try `gcc -c -Wall bug.c'
The compiler silently transforms the type of the structure component
below to some kind of a generic pointer or something.
However it should output a diagnostic on
--- Additional Comments From david dot nospam dot hopwood at blueyonder dot
co dot uk 2005-03-27 15:48 ---
This is particularly bad for some warnings, such as -Wpadded, that really only
make sense when applied to a subset of code. This came up for the Xen virtual
machine manager: we
--- Additional Comments From pinskia at gcc dot gnu dot org 2005-03-27
16:15 ---
This is still valid code because the struct could be defined below still.
For an example:
struct a
{
struct b *c;
};
struct b
{
int i;
struct a *c;
};
--
What|Removed
--- Additional Comments From pinskia at gcc dot gnu dot org 2005-03-27
16:21 ---
This is why POSIX ar is bad. Oh well.
--
What|Removed |Added
Component|java
--- Additional Comments From pinskia at gcc dot gnu dot org 2005-03-27
16:36 ---
Confirmed, there might be a testcase for 4.0.0 which can reproduce this too but
I don't know of any.
Anyways the problem looks like not taking into counting some alignment or
something into the length
so
--- Additional Comments From ian at airs dot com 2005-03-27 17:14 ---
This bug appears to still exists in mainline. When I compile the test case
without optimization, I get both these lines
.def_Test; .val_Test; .scl2; .type 044;.endef
.def
--
What|Removed |Added
Target Milestone|3.3.1 |---
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=9963
--- Additional Comments From schwab at suse dot de 2005-03-27 18:14 ---
Triggered by this patch from PR15242:
2005-02-01 Steven Bosscher [EMAIL PROTECTED]
PR optimization/15242
* params.def (PARAM_MAX_GOTO_DUPLICATION_INSNS): New param.
* basic-block.h
--
What|Removed |Added
CC||pinskia at gcc dot gnu dot
||org
--- Additional Comments From ian at airs dot com 2005-03-27 18:54 ---
Patch here:
http://gcc.gnu.org/ml/gcc-patches/2005-03/msg02460.html
--
What|Removed |Added
The following two function should be equivalent:
int f(int offset, int len, int num_bytes)
{
int i;
num_bytes = 0;
for (i = 0; i len; i++) num_bytes ++;
return num_bytes;
}
int f1(int offset, int len, int num_bytes)
{
if (len0)
return 0;
return len;
}
Why don't we strength reduce the
--- Additional Comments From hubicka at gcc dot gnu dot org 2005-03-27
19:35 ---
This regression should be solved by the patch so I guess I will close it and
move on the new regression :(
--
What|Removed |Added
--- Additional Comments From kazu at cs dot umass dot edu 2005-03-27 19:38
---
Fixed with Diego's recent updates to incremental SSA updates.
--
What|Removed |Added
Consider
int
foo (int a)
{
if (a == 0)
if (a == 0)
return 1;
return 0;
}
Note that the second if statement is redundant.
--
Summary: VRP does not get rid of a redundant if statement.
Product: gcc
Version: unknown
Status: UNCONFIRMED
--
What|Removed |Added
Summary|VRP does not get rid of a |[tcb] VRP does not get rid
|redundant if statement. |of a redundant if
--- Additional Comments From hubicka at gcc dot gnu dot org 2005-03-27
19:41 ---
How this is supposed to be failing? I do get undefined reference to baz, but
same I do get with my system compiler here and it seem to be right as baz is
extern inline function
--
--
What|Removed |Added
Status|NEW |WAITING
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=20635
--- Additional Comments From kazu at cs dot umass dot edu 2005-03-27 19:42
---
Created an attachment (id=8462)
-- (http://gcc.gnu.org/bugzilla/attachment.cgi?id=8462action=view)
Patch
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=20657
long long a=-9223372036854775808ll;
int b=-2147483648;
test2.cpp:1:14: warning: integer constant is so large that it is unsigned
it is not.
test2.cpp:1: warning: this decimal constant is unsigned only in ISO C90
test2.cpp:2: warning: this decimal constant is unsigned only in ISO C90
huh?
++ --no-create --no-recursion
Thread model: posix
gcc version 4.1.0 20050327 (experimental)
./cc1 -quiet -v -iprefix
/usr/src/gcc/obj/gcc/../lib/gcc/x86_64-unknown-linux-gnu/4.1.0/ -isystem
./include pr20635.c -quiet -dumpbase pr20635.c -mtune=k8 -auxbase pr20635 -O2
-version -o /tmp/ccmjI4n7.s
--- Additional Comments From jakub at gcc dot gnu dot org 2005-03-27 20:55
---
Forgot to add, this testcase was meant for gcc.c-torture/compile, i.e. an
compile test only. It of course won't link, because baz is extern inline.
But it should compile.
--
gcc version 4.0.0 20041213
javac Bla.java works.
gcj -C Bla.java (or -c)
Bla.java: In class 'Bla$B':
Bla.java: In constructor '(Bla)':
Bla.java:18: error: Can't reference 'this' before the superclass constructor
has been called.
super(doit());
^
1 error
--- Additional Comments From hubicka at gcc dot gnu dot org 2005-03-27
21:42 ---
I see I forgot my tree with checking enabled. This is obviously latent bug in
handling extern inline functions, I am looking into it.
--
What|Removed |Added
--- Additional Comments From pinskia at gcc dot gnu dot org 2005-03-27
21:46 ---
No, the warning is correct - is an operator and not part of the number,
--
What|Removed |Added
This is Walt Brainerd's fc001.f95
INQUIRE says only units 5, 6 exist,
but open(11... works. This is wrong.
Units 5 and 6 are preconnected, which
means that they are open, but all
units that can be used should exist.
--
Summary: INQUIRE incorrectly reports the existence of UNITS
--- Additional Comments From kargl at gcc dot gnu dot org 2005-03-27 21:56
---
Created an attachment (id=8463)
-- (http://gcc.gnu.org/bugzilla/attachment.cgi?id=8463action=view)
Walt's program
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=20660
--- Additional Comments From pinskia at gcc dot gnu dot org 2005-03-27
21:59 ---
Confirmed.
--
What|Removed |Added
Status|UNCONFIRMED |NEW
This fc002.f95 from Walt Brainerd.
! End of record is not detected
!on second READ
! iostats should be 0, 0, -2, -1
--
Summary: End of record not detected
Product: gcc
Version: 4.0.0
Status: UNCONFIRMED
Severity: normal
--- Additional Comments From kargl at gcc dot gnu dot org 2005-03-27 22:00
---
Created an attachment (id=8464)
-- (http://gcc.gnu.org/bugzilla/attachment.cgi?id=8464action=view)
Walt's program
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=20661
--- Additional Comments From pinskia at gcc dot gnu dot org 2005-03-27
22:02 ---
Confirmed.
--
What|Removed |Added
Status|UNCONFIRMED |NEW
This is Walt Brainerd's fc004.f95. A trace of f951 shows
(gdb) run fc004.f95
Starting program:
/home/kargl/work/41/libexec/gcc/i386-unknown-freebsd6.0/4.1.0/f951 fc004.f95
zero
MAIN__
to_sub
Program received signal SIGSEGV, Segmentation fault.
0x0809b832 in gfc_conv_intrinsic_bound
--- Additional Comments From kargl at gcc dot gnu dot org 2005-03-27 22:08
---
Created an attachment (id=8465)
-- (http://gcc.gnu.org/bugzilla/attachment.cgi?id=8465action=view)
Walt's program
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=20662
This is Walt Brainerd's fc005.f95.
kargl[227] gfc41 -static -o z fc005.f95
In file fc005.f95:23
if (.not. close(rx, rr)) then
1
Error: Symbol 'close' at (1) has no IMPLICIT type
--
Summary: Generic function is not resolved
Product: gcc
Version:
--- Additional Comments From kargl at gcc dot gnu dot org 2005-03-27 22:12
---
Created an attachment (id=8466)
-- (http://gcc.gnu.org/bugzilla/attachment.cgi?id=8466action=view)
Walt's program
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=20663
--- Additional Comments From pinskia at gcc dot gnu dot org 2005-03-27
22:13 ---
Confirmed.
--
What|Removed |Added
Status|UNCONFIRMED |NEW
This is Walt's fc006.f95. It uses lbound and ubound, so it may
be related to other PRs.
kargl[231] gfc41 -static -o z fc006.f95
fc006.f95: In function 'fcn':
fc006.f95:33: internal compiler error: in gfc_conv_descriptor_dtype, at
fortran/trans-array.c:177
--
Summary:
--- Additional Comments From pinskia at gcc dot gnu dot org 2005-03-27
22:15 ---
Confirmed, but I think this is a dup of bug 20482.
--
What|Removed |Added
--- Additional Comments From kargl at gcc dot gnu dot org 2005-03-27 22:15
---
Created an attachment (id=8467)
-- (http://gcc.gnu.org/bugzilla/attachment.cgi?id=8467action=view)
Walt's program
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=20664
--- Additional Comments From pinskia at gcc dot gnu dot org 2005-03-27
22:17 ---
*** This bug has been marked as a duplicate of 4695 ***
--
What|Removed |Added
--- Additional Comments From pinskia at gcc dot gnu dot org 2005-03-27
22:17 ---
*** Bug 20659 has been marked as a duplicate of this bug. ***
--
What|Removed |Added
--- Additional Comments From pinskia at gcc dot gnu dot org 2005-03-27
22:20 ---
*** This bug has been marked as a duplicate of 17202 ***
--
What|Removed |Added
--- Additional Comments From pinskia at gcc dot gnu dot org 2005-03-27
22:20 ---
*** Bug 20664 has been marked as a duplicate of this bug. ***
--
What|Removed |Added
--- Additional Comments From pinskia at gcc dot gnu dot org 2005-03-27
22:32 ---
Small testcase:
int f(double a, double b)
{
int a1 = a;
int b1 = b;
return a1+b1;
}
You can reproduce this on powerpc-darwin with -mcpu=601
--
What|Removed |Added
--- Additional Comments From pinskia at gcc dot gnu dot org 2005-03-27
22:34 ---
This was caused by:
2005-03-25 Geoffrey Keating [EMAIL PROTECTED]
* config/rs6000/rs6000.md (UNSPEC constants): Add UNSPEC_STFIWX.
(fix_truncdfsi2): Allow registers or memory as
--- Additional Comments From pinskia at gcc dot gnu dot org 2005-03-27
22:35 ---
Confirmed.
--
What|Removed |Added
Status|UNCONFIRMED |ASSIGNED
- -
: [bug fmodulo-sched/gcc]
: zouq [EMAIL PROTECTED]
: Mon, 28, 2005 8:09 am
: [EMAIL PROTECTED]
--
1.i build a cross-compiler for powerpc
the version of
On Mar 27, 2005, at 8:06 PM, zouq wrote:
- ~{T4SJ~~} -
~{VwLb~}: [bug fmodulo-sched/gcc]
~{7~HK~}: zouq [EMAIL PROTECTED]
~{HUFZ~}: Mon, ~{H}TB~} 28, 2005 8:09 am
~{JU~HK~}: [EMAIL PROTECTED]
In:
templatetypename T class foo {}
enum A{b, c};
gets you:
~/ootbc/members/bin$ g++ foo.cc
foo.cc:2: error: template declaration of `enum A'
foo.cc:2: confused by earlier errors, bailing out
The actual error is a missing semicolon.
Ivan
--
Summary: poor diagnostic
--- Additional Comments From pinskia at gcc dot gnu dot org 2005-03-28
02:07 ---
On the mainline I get:
t.cc:1: error: template declaration of 'enum'
t.cc:2: error: multiple types in one declaration
There might be a dup of this bug somewhere.
--
Right now now folding opportunities are taken for sparc builtin functions.
This should be fixed by implementing sparc_fold_builtin.
--
Summary: SPARC builtins should be folded if possible
Product: gcc
Version: 4.1.0
Status: UNCONFIRMED
--
What|Removed |Added
Severity|normal |enhancement
Keywords||missed-optimization
--- Additional Comments From phython at gcc dot gnu dot org 2005-03-28
05:57 ---
Created an attachment (id=8468)
-- (http://gcc.gnu.org/bugzilla/attachment.cgi?id=8468action=view)
ignored result testcase
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=20666
--- Additional Comments From phython at gcc dot gnu dot org 2005-03-28
06:00 ---
Created an attachment (id=8470)
-- (http://gcc.gnu.org/bugzilla/attachment.cgi?id=8470action=view)
initial folding of fexpand
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=20666
--- Additional Comments From aaronavay62 at aaronwl dot com 2005-03-28
06:09 ---
This also happens with gnu/java/security/OID.o and org/ietf/jgss/Oid.o, again
causing build to break in jv-collect. These files are less easily renamed,
though. I don't know how to fix this.
--
--- Additional Comments From aaronavay62 at aaronwl dot com 2005-03-28
07:40 ---
I just checked against Microsoft (R) Library Manager Version 7.10.3052 and
binutils's case-insensitivity replacement semantics are consistant with
Microsoft's implementation. LIB also has the same
83 matches
Mail list logo