Package: gcc-4.1
Version: 4.1.1-5
Severity: important
Tags: patch, upstream
Forwarded: http://gcc.gnu.org/PR26957
The attached fixes an ICE when building packages like ace.
#! /bin/sh -e
# All lines beginning with `# DPATCH:' are a description of the patch.
# DP: Fix for c++/26957; committed ups
> For hppa, the glibc builds well with gcc 4.0, but create problem with
> python/perl. It still has to be investigated.
Can you explain what the problems are (or at least the symptoms) so that
we can look into them? Are there any bugs filed on this?
thanks
randolph
--
Randolph Chung
Debi
While investigating gstreamer0.10-ffmpeg's build failure under hppa[1],
I received the following error from gcc:
gcc -DHAVE_CONFIG_H -I. -I. -I.. -I../libavutil -DHAVE_AV_CONFIG_H=1 -Wall
-Wno-switch -g -O2 -MT mpegaudiodec.lo -MD -MP -MF .deps/mpegaudiodec.Tpo -c
mpegaudiodec.c -fPIC -DPIC
Randolph Chung wrote:
Could someone send the preprocessed source for mpegaudiodec.c?
http://www.parisc-linux.org/mpegaudiodec.i.gz
gcc bugzilla has a reduced testcase.
randolph
oops, that should be http://www.parisc-linux.org/~tausq/mpegaudiodec.i.gz
randolph
--
To UNSUBSCRIBE, email
Could someone send the preprocessed source for mpegaudiodec.c?
http://www.parisc-linux.org/mpegaudiodec.i.gz
gcc bugzilla has a reduced testcase.
randolph
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
sed paer.d.o, the developer
> machine, to diagnose and I couldn't install it myself.
>
> I'm putting debian-hppa@ in Cc:, perhaps they have a clue. Please Cc:
> me on replies!
>
>Bye,
>
> [1]
> http://buildd.debian.org/fetch.php?&pkg=gstreamer0
im to merge this upstream when he returns,
meanwhile perhaps somebody can roll a new glibc package for debian?
randolph
#! /bin/sh -e
# DP: Description: hppa floating point exception handling fix
# DP: Related bugs: Debian #342545
# DP: Dpatch author: Randolph Chung <[EMAIL PROTECTED]>
ry with -mlong-calls (instead of -ffunction-sections)
These are obviously all workarounds until we fix the real problem.
randolph
--
Randolph Chung
Debian GNU/Linux Developer, hppa/ia64 ports
http://www.tausq.org/
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "uns
ndolph
--
Randolph Chung
Debian GNU/Linux Developer, hppa/ia64 ports
http://www.tausq.org/
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Package: gcc-3.3-hppa64
Version: 3.3.3-2
Severity: important
When building a linux kernel with gcc-3.3-hppa64, some unaligned loads
are emitted which are rejected by gas.
Filed upstream as http://gcc.gnu.org/bugzilla/show_bug.cgi?id=14782
This prevents us from building working kernels with an of
; See http://gcc.gnu.org/bugs.html> for instructions.
filed upstream as #14838
thanks,
randolph
--
Randolph Chung
Debian GNU/Linux Developer, hppa/ia64 ports
http://www.tausq.org/
been used
successfully to port sablevm to hppa.
randolph
--
Randolph Chung
Debian GNU/Linux Developer, hppa/ia64 ports
http://www.tausq.org/
hppa-patches.tgz
Description: GNU Unix tar archive
In reference to a message from Daniel Jacobowitz, dated Oct 31:
> Are you sure he doesn't mean --enable-checking? --enable-debug should
> not affect compile time significantly.
yup, David confirmed he meant enable-checking. thanks for the quick
reply.
randolph
--
Randolph Chung
FYI -randolph
- Forwarded message from David Mosberger <[EMAIL PROTECTED]> -
From: David Mosberger <[EMAIL PROTECTED]>
Date: Fri, 31 Oct 2003 15:20:36 -0800
To: debian-ia64@lists.debian.org
Subject: gcc in debian/unstable
Reply-To: [EMAIL PROTECTED]
X-Mailing-List: archive/latest/1242
W
- Forwarded message from Borislav Petkov <[EMAIL PROTECTED]> -
From: Borislav Petkov <[EMAIL PROTECTED]>
Reply-To: [EMAIL PROTECTED]
To: [EMAIL PROTECTED]
Subject: Bug in libstdc++2.10-dev
Date: Wed, 15 Oct 2003 07:33:34 +0200
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
(.text+0x1c): In function `call_foo':
: undefined reference to `foo'
this worked fine on older binutils (e.g. 2.13.90.0.16 worked).
does anyone know what might be broken? this is causing a lot of failures
in the binutils test suite
thanks,
randolph
--
Randolph Chung
Debian GNU/Linux
nt binary
compatibility, they should follow what Debian is doing (on hppa), since
Debian was there first :-)
my 2 cents,
randolph
--
Randolph Chung
Debian GNU/Linux Developer, hppa/ia64 ports
http://www.tausq.org/
tag 193207 +patch
thanks
Please replace the hppa-fptr patch in gcc-3.2 with this one, which has a
fix (already in gcc-3.3 upstream) for the fptr comparision problem.
thanks
randolph
#! /bin/sh -e
# DP: Handle function pointer comparisions on hppa properly
# DP: Backported from gcc-3.3 branch
#
>Category: c++
>Synopsis: [3.3 regression] [hppa] Segfault when building jade
>Confidential: no
>Severity: serious
>Priority: medium
>Class: ice-on-legal-code
>Submitter-Id: net
>Originator: Randolph Chung <[EMAIL PROTECTED]>
will get retried. once i get confirmation from him that this is ok, i
will close this bug.
thanks
randolph
--
Randolph Chung
Debian GNU/Linux Developer, hppa/ia64 ports
http://www.tausq.org/
In reference to a message from Matthias Klose, dated Apr 05:
> Will do. What about applying this patch upstream for 3.2.3?
since this is not a regression, i'm not sure it will be a candidate for
3.2.3, but if Gabriel is ok with it, all the better :-)
randolph
causes some X programs to be miscompiled and not
work.
thanks
randolph
--
Randolph Chung
Debian GNU/Linux Developer, hppa/ia64 ports
http://www.tausq.org/
#! /bin/sh -e
# DP: Handle function pointer comparisions on hppa properly
# DP: Backported from gcc-3.3 branch
# DP: http://gcc.gnu.org/ml/gcc-cvs
forgot to list debian-gcc in cc.. oops. here's a copy.
- Forwarded message from Randolph Chung <[EMAIL PROTECTED]> -
To: [EMAIL PROTECTED]
Cc: [EMAIL PROTECTED]
From: Randolph Chung <[EMAIL PROTECTED]>
Reply-To: Randolph Chung <[EMAIL PROTECTED]>
Subject: [parisc-
>Category: target
>Synopsis: [3.0/3.2/3.3/3.4] Floating point args not correctly loaded for
>function calls
>Confidential: no
>Severity: serious
>Priority: medium
>Class: wrong-code
>Submitter-Id: net
>Originator: Rando
Package: gcc-3.2
Severity: important
[Debian note: this may be related to #185184 and #105816]
When compiling the bug.c code below with -O2, the floating point
argument to the fprintf does not get reloaded for the second fprintf()
call, so the second call prints junk. At -O1 the arguments (r23/
>Submitter-Id: net
>Originator:Randolph Chung <[EMAIL PROTECTED]>
>Organization: Debian
>Confidential: no
>Synopsis: [3.3 regression] [parisc-linux] ICE when building lesstif1 at
>-O1/O2
>Severity: serious
>Priority: medium
>Category:
>Category: c++
>Synopsis: [3.3 regression] internal compiler error: in cp_expr_size, at
>cp/cp-lang.c:307
>Confidential: no
>Severity: serious
>Priority: medium
>Class: ice-on-legal-code
>Submitter-Id: net
>Originator: Rando
.. no one has tested 3.3 hppa64-linux-gcc at all, so it will just
be introducing more unknowns into the problem.
randolph
--
Randolph Chung
Debian GNU/Linux Developer, hppa/ia64 ports
http://www.tausq.org/
ay, I have the debs installed in a chroot now. Will start building
random stuff and see what happens :)
stay tuned,
randolph
--
Randolph Chung
Debian GNU/Linux Developer, hppa/ia64 ports
http://www.tausq.org/
nly because it overwrites things in the current
3.2 package)
randolph
--
Randolph Chung
Debian GNU/Linux Developer, hppa/ia64 ports
http://www.tausq.org/
oad.
http://gcc.gnu.org/ml/gcc-cvs/2003-02/msg00993.html
For reference, the upstream PR is optimization/9768
randolph
--
Randolph Chung
Debian GNU/Linux Developer, hppa/ia64 ports
http://www.tausq.org/
* pa.c (output_millicode_call): Use $PIC_pcrel$0 for long PIC
millicode calls when !TARGET_SOM and TARGET_GAS is true.
* pa.md (jump): Likewise.
thanks
randolph
--
Randolph Chung
Debian GNU/Linux Developer, hppa/ia64 ports
http://www.tausq.org/
#! /bin/sh -e
# All lines beginning with `# DPATCH:&
tag 180266 +unreproducible +moreinfo
thanks
> You can't compile any C or C++ programs because the CPP is going
> to dump some memory (or whatnot) beginning at EOF.
>
> Suggestion: Maybe you guys should be doing a make check before packing
> those binaries up.
Suggestion: Maybe you should check h
tag 180102 +patch
thanks
> This build failed due to assembler errors, lots of "Error: can't
> resolve" messages.
>
> http://buildd.debian.org/fetch.php?&pkg=arts&ver=1.1.0-2&arch=hppa&stamp=1044489665&file=log&as=raw
the fix is here: http://gcc.gnu.org/ml/gcc-cvs/2003-02/msg00361.html
randolph
c hppa-linux
Thread model: posix
gcc version 3.2.2 20030131 (Debian prerelease)
we already enable that flag...
randolph
--
Randolph Chung
Debian GNU/Linux Developer, hppa/ia64 ports
http://www.tausq.org/
reassign 171411 glibc
severity 171411 important
merge 171411 16
thanks
this is already reported. -randolph
Package: gcc-3.2
Version: 1:3.2.1ds6-1
Severity: serious
I guess this is probably known already, but I don't see a bug on this
yet. gcc-3.2 fails to build on sparc. Dies with an ICE at:
../../src/gcc/p/script/mkdir-p p/rts
src=`cd ../../src/gcc && pwd` && \
cd p/rts && \
CC="`echo ./xgcc -B./ -B/
reassign 169497 g++-3.2
severity 169497 important
merge 169497 16
thanks
already reported, thanks -randolph
> Package: gcc-3.2
> Version: 1:3.2.1-0pre4
> Severity: important
> Justification: fails to build from source
>
> apt-get build-dep gcc-3.2
> Reading Package Lists... Done Building
> Dependency Tree... Done
> E: Build-Depends dependency on gcc-3.2 cannot be satisfied because the
> package libc6
> > -8<---
> > /* compile with gcc -c foo.c */
> >
> > typedef struct {
> > volatile unsigned int lock;
> > } spinlock_t;
> >
> > typedef struct {
> > spinlock_t lock;
> > volatile int counter;
> > } rwlock_t;
> >
> > void foo(void)
ernel/fork.c has something like this:
struct *fs = /* allocate memory for fs */;
fs->lock = (rwlock_t) { ... };
thanks,
randolph
--
Randolph Chung
Debian GNU/Linux Developer, hppa/ia64 ports
http://www.tausq.org/
> 1) When compiling with -g, the final link fails, and explicitly asks
>for recompilation with -ffunction-sections. This works, and
>produces a maxima passing all tests.
known problem with gcc-3.0.x
> 2) Any optimization at all breaks the build. I have a gdb
>session/backtrace showi
> I have no idea what I modified to my system for this; I also purged and
> re-installed all gcc packages at no avail.
M_PI should be defined in /usr/include/math.h try reinstalling
libc6-dev?
randolph
forwarded to bts for the record...
- Forwarded message from John David Anglin <[EMAIL PROTECTED]> -
Subject: Re: [parisc-linux] FWD: Bug#162690: [parisc]gcc 3.x produces openssl
test code that SEGVs with -march=2.0 and optimization >0
To: [EMAIL PROTECTED]
From: "John David Anglin" <[EM
In reference to a message from Aaron Lehmann, dated Sep 07:
> Maybe this is because gcc 2.96 is not an official release of gcc and
> instead is a buggy snapshot?
ia64 uses 2.96 as default as a compromise for many issues. we knew from
the beginning that it's not an official release of gcc; but call
In reference to a message from LaMont Jones, dated Sep 03:
> Bug #158290 talks about an issue with gcc 3.0 on hppa. Specifically,
> building perl 5.8 with -O2 fails tests, while -O1 works.
>
> Any thoughts on the subject? mawk also suffers from the same ailment.
bod tells me that with gcc-3.2 t
> a) 3.1 as to be released (without dwarf2 support)
> b) 3.1 + dwarf2 support
> c) 3.2 CVS 20020429 plus/minus patches
for woody+1, is it an option to skip 3.1 altogether and go directly to
3.2? from what i understand gcc-3.2 is supposed to be officially
released in the fall... (i.e. likely befo
> maybe if linking with static libraries is an option?
i guess that's always possible, but kinda ugly.
randolph
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
> Is that one available somewhere on an ia64 box, preferably one accessible
> to John?
gcc-snapshot package, but we cannot use that to build binaries to go into the
archive (uses different library versions)
I'm not sure which box John has access to, but mail
[EMAIL PROTECTED] and they can install
> There appear to (still) be some issues with g++-3.0 on ia64. Under some
> optimisation switches, the executable is built, but segfaults. Under others,
> g++-3.-0 dies.
fwiw with gcc version 3.1 20020331 (prerelease) the problem appears to
be fixed. I think this might be the same/related to pr/53
Package: gcc-3.0
Version: 3.0.4ds3-1
Severity: important
Tags: patch
Attached is an updated hppa-build.dpatch that fixes two bugs:
1. large switch statements were not being compiled properly
2. fp arguments were not being passed correctly to indirect calls
Both of these patches came from J David
Package: gcc-snapshot
Version: 20020224-1
Severity: serious
legolas[21:58] ~% /usr/lib/gcc-snapshot/bin/gcc -o test test.c
/usr/bin/ld: unrecognized option '--eh-frame-hdr'
/usr/bin/ld: use the --help option for usage information
collect2: ld returned 1 exit status
randolph
--
Debian Developer <
Works great, thanks Gary!
Ben, FYI - i'll file a bug against glibc for this.
randolph
In reference to a message from Gary Hade, dated Feb 25:
> Randolph,
> I believe this problem is due to a bug in /usr/lib/gcrt1.o
> that was fixed by recent glibc changes to csu/gmon-start.c
> (revs 1.13 and 1
FYI -randolph
- Forwarded message from Randolph Chung <[EMAIL PROTECTED]> -
Date: Sat, 23 Feb 2002 19:10:30 -0800
From: Randolph Chung <[EMAIL PROTECTED]>
To: [EMAIL PROTECTED]
Cc: gcc-bugs@gcc.gnu.org
Subject: gcc profiling broken on Linux/ia64?
Reply-To: Randolph C
>Submitter-Id: net
>Originator:Randolph Chung
>Organization: Debian
>Confidential: no
>Synopsis: ICE in reload_cse_simplify on hppa-linux
>Severity: serious
>Priority: medium
>Category: optimization
>Class: ice-on-legal-code
>Releas
# bcc [EMAIL PROTECTED]
tag 123685 + patch
thanks
For the record, this updated NaT patch will allow ia64 to build... but we
are waiting for confirmation from upstream.
randolph
--
Debian Developer <[EMAIL PROTECTED]>
http://www.TauSq.org/
#! /bin/sh -e
# DP: ia64 NaT bug fix
# DP:
https://e
Package: gcc-3.0
Version: 3.0.3ds0-1
Severity: serious
/xgcc -B./ -B/usr/ia64-linux/bin/ -isystem /usr/ia64-linux/include -isystem /us
r/ia64-linux/sys-include -O2 -DIN_GCC-W -Wall -Wwrite-strings -Wstrict-pro
totypes -Wmissing-prototypes -isystem ./include -fPIC -g1 -DHAVE_GTHR_DEFAULT -
> I want leave the final decision to Randolph Tausq, which does (?) the
> ia64 parts in gcc.
It's always refreshing to get a new name ;-)
Matthias, please update gcc-defaults so that for ia64:
g77 stays the way it is (points to 2.96)
gobjc points to gobjc points to 3.0
Thanks!
randolph
--
Debia
Package: gcc-defaults
Severity: serious
There is no gobjc-2.96 on ia64, so gcc-defaults needs to be updated to
point gobjc to gobjc-3.0
randolph
--
Debian Developer <[EMAIL PROTECTED]>
http://www.TauSq.org/
Here's an updated patch from upstream to fix the strength reduction ICE.
randolph
--
Debian Developer <[EMAIL PROTECTED]>
http://www.TauSq.org/
#! /bin/sh -e
# DP: Fix for ICE caused by strength-reduction optimization
# DP: http://gcc.gnu.org/ml/gcc-cvs/2001-12/msg4.html
# DP: from Olivier
Package: gcc-3.0
Version: 3.0.2ds4-4
Severity: important
Tags: patch
The attached patch fixes a strength reduction ICE seen when compiling
groff on hppa-linux. It was contributed by Alan Modra. I'm still in the
process of verifying this with the gcc hppa maintainer (Dave Anglin) but
wanted to send
patch attached
randolph
--
Debian Developer <[EMAIL PROTECTED]>
http://www.TauSq.org/
#! /bin/sh -e
# DP: ia64 NaT bug fix
# DP:
https://external-lists.valinux.com/archives//linux-ia64/2001-November/002495.html
dir=
if [ $# -eq 3 -a "$2" = '-d' ]; then
pdir="-d $3"
dir="$3/"
elif [ $#
Package: gcc-3.0
Severity: serious
Details at:
http://external-lists.valinux.com/archives//linux-ia64/2001-November/002495.html
The bug can cause miscompiled userapps to crash the kernel.
randolph
--
Debian Developer <[EMAIL PROTECTED]>
http://www.TauSq.org/
Just a heads up -- there may be a ia64 patch needed for gcc-3.0. When I
have some time I'll look at pulling the patches and testing them for
ia64..
randolph
- Forwarded message from David Mosberger <[EMAIL PROTECTED]> -
From: David Mosberger <[EMAIL PROTECTED]>
To: [EMAIL PROTECTED]
Repl
Looks like there's a strength-reduce optimization bug in g++-3.0.2 on
hppa
Trying to compile qt-x11 gives a segmentation fault when compiling
src/kernel/qurl.cpp. (willy and lamont also saw this I believe...). Turning
off strength-reduction (-fno-strength-reduce) allows the file to
compile to
John had mentioned this bug to me on irc... after looking at this a
bit more, it looks like someone else had reported this to the gcc folks
and they determined it's not a bug.
See PR#3747 at http://gcc.gnu.org/cgi-bin/gnatsweb.pl
The fix is to not use unsigned int as a type in an overridden []
op
I'm not the maintainer for libstdc++2.10 Matthias, can you comment
on this? I guess the maintainer of record got messed up with one of my
previous ia64 uploads.. not sure how to fix :-(
randolph
In reference to a message from justin, dated Oct 25:
> Hi,
>
> I was wondering if you'd looked
Package: gcc-3.0
Version: 1:3.0.2ds1-0pre010908
Severity: serious
Looks like 3.0.2 has bootstrapping problems... I tried building it with
gcc-2.96, gcc-3.0.1 and gcc-3.0.2 (*) and they all die with internal
compiler errors.
(*) With gcc-2.96/gcc-3.0.1 it dies compiling libstdc++; removing -O2 fro
Hi,
On Debian/hppa, we have 32 or 64 bit kernels, but userspace is always 32
bit (sorta like sparc aiui). As such, there's a need for a
hppa32->hppa64 cross-compiler.
Right now there is no such package in Debian yet; the hppa porters use a
tarball that is distributed separately.
Are the gcc main
> hmm, it's easy to build a libg2c-pic.a library. So I will do this
> instead of building the standard library with pic. Building a shared
> library would introduce a incompatibility to other distros.
Matthias,
This might be an issue -- because a generic invocation of g77 won't know
library to li
> for which architectures should the patch be enabled (besides ia64?)
my understanding is that this will be needed for any architectures that
cannot link non-PIC code into a PIC library, so that's iirc:
ia64, hppa, mips, mipsel, s390, sh, powerpc, m68k
thanks,
randolph
--
Debian Developer <[EM
> > The fortrain library libg2c is built as a non-PIC archive in the default
> > gcc package (at least it was on ia64). On some architectures, linking
> > non-PIC code into PIC shared objects is not allowed (ia64, hppa,
> > mips{,el}, etc). This causes applications that try to build fortrain
>
Package: gcc-3.0
Version: 1:3.0.1-0pre010801
The fortrain library libg2c is built as a non-PIC archive in the default
gcc package (at least it was on ia64). On some architectures, linking
non-PIC code into PIC shared objects is not allowed (ia64, hppa,
mips{,el}, etc). This causes applications tha
I've uploaded experimental gcc-3.0 and supporting binutils packages to
http://people.debian.org/~tausq/ia64/
They appear to fix some of the issues we've seen with gcc-2.96. However,
at the present time there is no plan to move to gcc-3.0 as the default
compiler toolchain for IA64 for the w
Package: gcc-3.0
Version: 3.0.ds9
Severity: serious
Tags: patch
X-Debbugs-Cc: debian-ia64@lists.debian.org
Here's a patch for the ia64 packaging bits.
I get a lot of regression test errors, but they seem to be comparable to
what we get on hppa and mips.
Enjoy :)
randolph
diff -uNr gcc-3.0-3.0.
(re ajt's comments to bug #101878)
gcc-defaults doesn't build packages with epochs by default, so if we can
keep it that way we'll be more consistent with the other architectures.
But I think we just want to get something into the archive that works,
so whatever we can agree to .
cc'ing the m
76 matches
Mail list logo