In case folks haven't noticed, there is a major
problem with the perl script to fix the shared lib
names for the MacOS X X11 beta files. If you look
at the log from update_prebinding on a machine
that has had the perl script run on it, you will
notice it breaks prebinding on all of the files
Ben,
I see the following type of errors for all the binaries in X11 when
update_prebinding gets run after your perl script has been used.
2003-01-29 22:45:58.751 update_prebinding[9384] redo_prebinding on
/usr/X11R6/bin/xcmsdb: executable: /usr/X11R6/bin/xcmsdb (architecture ppc)
Has Apple indicated any time frame for when the next
X11 beta might be released?
Jack
---
This SF.NET email is sponsored by:
SourceForge Enterprise Edition + IBM + LinuxWorld = Something 2 See!
I just upgraded my v0.1 installation of the X11 beta to v0.2
and was annoyed to see that now a window manager starts up by default
and all of the windows are using that for frames instead of the aqua
frames from before. While I can understand why this is happening
does anyone know what the
Hello,
Have any other fink users been using the molmol package from
http://dryden.biol.yorku.ca/darwin.html? On my machine with either
the 0.1 or 0.2.1 MacOS X X11 beta I find this package doesn't do
3D images (like the ball-and-stick or ribbon molecular representations)
correctly. Instead I
One question for the fink developers...is there a particular reason
why we don't use the debhelper package while we are using dpkg and
apt-get? Just curious.
Jack
---
This SF.NET email is sponsored by: FREE SSL Guide
I had to reinstall Jaguar and Fink this weekend and it seems
now that fink now longer can find the molmol package. This is
after doing a 'fink selfupdate-cvs' and then trying 'fink install molmol'.
Any idea what is going on?
Jack
Is there some trick besides using ./inject.pl and fink selupdate-rsync
to keep fink current to the cvs? I am finding that the last couple of days
worth of changes that appear in the fink-commits mailing list aren't pulled
down to my machine. I have tried switching back and forth from rsync to
Well after struggling with building the most current g77 3.4 package
on my G5 with fink 10.2-gcc3.3 rsync installed I have defined the problem
slightly. It must be a flaw in the current fink distribution itself.
If I avoid sourcing fink to be used and repeat the same exact build
steps using
I have run into a particular problem several times this weekend
while trying (unsuccessfully) to get g77 to build under 10.2-gcc3.3.
I find that if I move a copy ofg77-3.4-20031015.info into the
my finkinfo directory that it will be used if I do 'fink install g77'.
However since that build
Is it possible to suppress the removal of the source build tree in
fink? I have run into a number of instances where I would have liked to
had the fully populated build tree left behind after a 'fink install'
so I could run the make checks. Is there a hidden flag in fink to do
that? As far as
Chris,
Thanks for the pointing out the option in fink.conf. However I think
fink should have a command line flag to allow KeepBuildDir to be set
for a single fink build without changing the global behavior in
/etc/fink.conf. That would greatly limit the bloat from using the
option since it
Darian,
Thanks. Are -Kk and --keep-build-dir --keep-root-dir documented
in the fink man page? The fink maintainer should definitely add
something about those options to the fink man page.
Jack
---
This SF.net email is
Martin,
I don't think the newer g77 was proven to build on a G5.
On my G5, I find that building the package manually with
the commands contained in the .info works but if I use
fink instead, the build fails.
Jack
---
This
In asking some questions about the G5 optimizations
in gcc on the gcc mailing list, someone suggested that
we might want to use the Apple gcc tree to build our
g77. Has anyone considered doing that...using the
gcc3 tarball from Darwin to build a f77 that could
be merged more cleanly into
David,
Under my copy of 10.2-gcc3.3, fink building g77 fails the
same way as reported earlier...
http://article.gmane.org/gmane.os.apple.fink.devel/5660/match=g77
I am tending to suspect this is somehow do to a mismatch with
the libiconv that been installed and used to build against.
Dave,
Hmmm. I could be confused but it looks like the
dpkg in fink must not call dpkg-shlibdeps when packaging.
If I do...
dpkg-deb --info readline-shlibs_4.3-5_darwin-powerpc.deb
the Depends: line appears empty. I would have thought it
would contain entries obtained from calls like...
Well I just finished a bootstrap of 10.2-gcc3.3 from fink 1.5.1-beta
and the fink install of g77 still results in the same compile failure.
Well hopefully enough of 10.3 will be present next week to try another
bootstrap once I upgrade my G5 to Panther and its devtools.
To see if it helps building g77 on my G5, I am interested in
doing a bootstrap build of 10.2-gcc3.3. I have am bit unclear on
a couple of points however. I have downloaded fink 1.5.1beta
and have read the INSTALL file. However it is unclear to me if
I want or need to download the fink 0.5.3
I currently have fink 1.5.1-beta and 10.2-gcc3.3 installed
on my G5 running 10.2.8. Does anyone know what the recommended
approach will be to convert a fink 10.2-gcc3.3 to 10.3 once
the machine is upgraded to Panther? Thanks in advance for any
clarifications on this.
Jack
Unlike g77 3.4, which still fails to build under fink
on my G5 with 10.2-gcc3.3, I was able to modify the same
.info file to build gcc 3.3.2's g77 and it seems to
package up fine. I uncommented the make check-g77 for the
build and got...
=== g77 Summary ===
# of expected
Martin,
I downloaded the release gcc 3.3.2.tar.bz2 and modified the
g77 info script for gcc 3.4-20031015 to build gcc 3.3.2 instead.
The machine is a dual G5 with Darwin 6.8.5 installed and the
Aug 2003 devtool update applied over the Dec 2002 devtools and
Dec 2002 G5/gcc 3.3 update. I don't
Martin,
The gcc 3.3.2 package I built can compiler a program with
only a line containing...
end
The compile works fine and the resulting a.out runs without
actually doing anything (as expected).
Jack
---
This
Does anyone know what the cleanest approach is to
changing the shell that xterm uses under Panther? I want
to set it back to csh or tcsh.
Jack
---
This SF.net email is sponsored by: SF.net Giveback Program.
Does
Has anyone made any progress towards getting
pymol 0.90, 0.92 or 0.93 to build under Fink 0.6.1
and MacOS X 10.3?
Jack
---
This SF. Net email is sponsored by: GoToMyPC
GoToMyPC is the fast, easy and secure way to access
I am trying to recompile some fortran code with shared libraries
under IBM's XL Fortran. However since XL doesn't have a
-dynamiclib flag I need to change my link line to use -Wl,
instead. However for both g77 and XL this doesn't seem to
work the same as...
g77 -dynamiclib -flat_namespace
Matt,
The 10.2 raster3d package could be moved immediately into 10.3
stable if you just change the patch from...
-INCDIRS = -I/usr/include -I/usr/local/include
-LIBDIRS = -L/usr/local/lib
+INCDIRS = -I/usr/include [EMAIL PROTECTED]@/include
+LIBDIRS = [EMAIL
Matt,
I asked on darwin-development about the compile failures when building
pymol 0.93 on current fink 10.3. One of the darwin developers suggested
we try removing -ansi from the CCOPT1 flag in Rules.make. I did that
change and it does indeed eliminate the compilation problems.
Gary,
Have you tested your iMac 1.8 GHz powerpc with TechTool Deluxe
for memory problems and bad disk blocks? I saw a problem like that
awhile back on my dual G5 when the drive had developed some
bad blocks.
Jack
Alexander,
We would need to modify the section of sparky-py.patch with...
CXX= g++
-CXXFLAGS =
+CXXFLAGS = -O3 -ffast-math -fPIC -fno-common -Wno-long-double
CC = gcc
-CFLAGS =
+CFLAGS = -O3 -ffast-math -fPIC -fno-common
PYFLAGS
Is there any information available on
the proper use of the fink-prebinding package?
The paraview package in fink 10.4 unstable is
taking an abnormally long time to launch the
program. The paraview developers believe this
is due to the absence of prebinding for the
paraview shared libraries.
Dave,
Actually I was running paraview-mpi-openmpi. I'll try a build
of the non-mpi paraview package to make sure I'm not seeing a
delay related to mpi issues.
Jack
On Tue, Jun 26, 2007 at 07:06:26AM -0700, David R. Morrison wrote:
Prebinding was officially deprecated by Apple
Robert,
If your drive has developed bad blocks, you will get
all sorts of oddities in building gcc42. I ran into that
with one of the two drives in my AppleRAID mirror earlier
in the year. You should try erasing the drive in Disk
Utility with zeroing of the data. If the drive is SATA
or IDE,
Benoît,
Unfortunately, the cctools package doesn't make clear that it
is a placekeeper for a particular version of Xcode. The requirement
of cctools (= 622-1) for gcc42 to build or install is meant to force
the user to install Xcode 2.4.1 or newer. Hopefully at some point
the cctools virtual
Benjamin,
This is a bug in kmail and not Leopard. The Tcl/Tk 8.4.15
release suffers from this bug. Apple's programmers describe it
in that case as...
tcl/tk applications do a fork(2) without an execve(2) (or equivalent).
The application is attempting to use CF or Foundation on the child
side
Benjamin,
Actually, the code changes made to tcltk 8.4.16 to resolve this
same issue can be found at...
http://tcl.cvs.sourceforge.net/tcl/tcl/macosx/tclMacOSXNotify.c?r1=1.1.2.12r2=1.1.2.13pathrev=core-8-4-16
Jack
it is working - I can't say for sure. But the programs,
I already tried to compile against this libraries did work. Thanks.
Cheers,
Gernot
On Nov 7, 2007, at 5:51 PM, Jack Howarth wrote:
The g77 package for MacIntel on fink tracking is only
there because the submitter did not realize
The g77 package for MacIntel on fink tracking is only
there because the submitter did not realize that a current
gfortran was available in the gcc42-4.2.2-1000 package.
I haven't heard back from him since this was pointed out.
Jack
Robert,
How about identifying the link in the atlas Makefile.in's which
does that 'rm -f *.o x* config?.out *core*' and just changing it
to 'rm -fr *.o x* config?.out *core*'?
Jack
-
This
Paul,
Apparently you haven't recently done a 'fink update-all'
after a 'fink selfupdate'. The gcc42-4.2.3-1000 package with
a BuildConflicts on gcc43 has been in fink 10.4/10.5 unstable
for sometime now. Each time a new gcc4X package is created
it will be necessary for users to update to the
, Mar 12, 2008 at 06:48:39PM +0100, Jean-Francois Mertens wrote:
Jack Howarth wrote:
Each time a new gcc4X package is created
it will be necessary for users to update to the previous ones
to revisions which have BuildConflicts on the newer ones.
Adding a buildconflict should not require a rev
Alexander,
It also would be a good idea if someone would update
fftw and fftw3 to the versions that are built with gcc43...
http://sourceforge.net/tracker/index.php?func=detailaid=1910497group_id=17203atid=414256
There are at least two bugs in the linker from Xcode 3.1
which will prevent gcc43 from building. The first one manifests
itself as a duplicate symbol in both tree-inline.o and libbackend.a
(which contains tree-inline.o already) when linking cc1plus-dummy.
This can be hacked around by removing
Michelle and Michael,
Try the new pymol-py-1.0r2-1002 packaging in fink 10.4/10.5
unstable. I've hacked around the problem of the python from
the python2.5 package incorrectly generating the Makefile in
contrib/modules. The package should build fine now although
we still need to figure exactly
/dpkg-lockwait -r fink-buildlock-pymol-py25-1.0r2-1001
(Reading database ... 15577 files and directories currently installed.)
Removing fink-buildlock-pymol-py25-1.0r2-1001 ...
Failed: phase compiling: pymol-py25-1.0r2-1001 failed
On May 9, 2008, at 3:57 PM, Jack Howarth wrote:
Michelle
Victor,
The orte_config.h header would be internal to the build
of openmpi itself. If you compare other openmpi packaging
like that from Fedora 9 linux, you will see that this header
is not packaged. It would be broken behavior for software
using openmpi to expect that header.
Jack
Guy,
I have no idea where that is coming from. I just built test
gcc43-4.3.2-rc1 packages against the current fink unstable tree
last night without issues. What do you get for...
whereis jar
I see...
/usr/bin/jar
Jack
On Wed, Aug 20, 2008 at 11:48:26AM +0200, Guy
Martin,
I don't see how this problem could be related to which
libtool fink has installed. The failing step...
/bin/sh ../../libtool --mode=link gcc -DXNO_MTSAFE_STRINGAPI
-DXNO_MTSAFE_PWDAPI -DXNO_MTSAFE_TIMEAPI -Wall -g -Wno-unused
-Wno-comment -version-info 3:0:0 -o libXm.la -rpath
Alexander,
I been complaining about fink warning that
packages can't Depend on xcode for ages. It is
simply wrong-headed to do that. There have been
several times that FSF gcc has required fixes
from a newer Xcode's cctools to properly
assemble or link code generated by FSF gcc.
It is essential
Alexander,
I just rebuilt gcc43 and octave on a dual G5 running
10.5.6 with Xcode 3.1.2. Everything builds normally here.
Are you sure you don't have some non-fink software installed
that is confusing the fink builds? I've had no other reports
of this problem.
Jack
On Tue,
Using a clean bootstrap, I can recreate this
problem with a simple...
fink -k rebuild glib2
Information about 8404 packages read in 1 seconds.
The following package will be rebuilt:
glib2
The following 4 additional packages will be installed:
glib2-dev glib2-shlibs gnome-doc-utils gtk-doc
The change...
Index: glib2.info
===
RCS file: /cvsroot/fink/dists/10.4/unstable/main/finkinfo/gnome/glib2.info,v
retrieving revision 1.43
diff -r1.43 glib2.info
9d8
gtk-doc (= 1.8-1),
clears the way for 'fink rebuild glib2'
On Fri, Apr 17, 2009 at 12:35:40PM +0200, Martin Costabel wrote:
Trying to build gcc43-4.3.3-1000 (needed for openmpi) on 10.5.6/intel
with xcode-3.0 crashes for me (reproducibly) with a bus error:
T=`${PWDCMD-pwd}`/ \ cd ../.././gcc \
make
Thomas,
You need to delete those files that are marked as
being in the way. Just prefix /sw/fink to the marked
path to delete them. Afterwards a 'fink selfupdate'
should complete without errors.
Jack
--
I've considered adding ada to the compilers built in the
gcc4X packages and have decided against it. The ada developers
refuse to attempt to create a version of ada that can be boot
strapped from a c compiler. This means that gcc4X would develop
a BuildDepends on a prebuilt ada compiler from an
On Thu, Apr 30, 2009 at 12:09:54AM -0500, Peter O'Gorman wrote:
Jack Howarth wrote:
I've considered adding ada to the compilers built in the
gcc4X packages and have decided against it. The ada developers
refuse to attempt to create a version of ada that can be boot
strapped from a c
Rocco,
What happens if you add...
--build=%m-apple-darwin`uname -r|cut -f1 -d.` --host=%m-apple-darwin`uname
-r|cut -f1 -d.`
to ConfigureParams? In some instances, packages rely on configure to
properly detect the architecture that is being compiled for. In those
instances, you may have to
If you don't have /usr/X11R6/lib/libGL.dylib on your drive,
this means you never installed the X11 SDK in Xcode.
Jack
--
Enter the BlackBerry Developer Challenge
This is your chance to win up to $100,000 in
More precisely, try http://connect.apple.com and create a
free Apple Developer account. The Xcode you want is under
the Developer Tools downloads section and is Xcode 2.5
which is the last release for Tiger. When you install
Xcode in Tiger, you have to manually select the X11 SDK.
Jack
The packaging for povray needs the change...
-ConfigureParams: --mandir=%i/share/man COMPILED_BY=Andrew Rohl as part of
Fink
+ConfigureParams: --mandir=%i/share/man --build=%m-apple-darwin`uname -r|cut
-f1 -d.` --host=%m-apple-darwin`uname -r|cut -f1 -d.` COMPILED_BY=Andrew Rohl
as part of
to have issues, please check to see if the
FAQ on fink's website solves the problem. If not, ask on the fink-users
or fink-beginners mailing lists, with a carbon copy to the maintainer:
Jack Howarth howa...@bromo.med.uc.edu
Note that this is preferable to emailing the maintainer directly
able to reproduce this build error. I'm CCing the maintainer
of gcc42, Jack Howarth.
Cheers,
--
monipol
http://finkers.wordpress.com
Submitting a Fink bug report? Read this:
http://www.finkproject.org/doc/netiquette/index.php
http://finkers.wordpress.com/2009/06/03/bug-reports/
Are both
able to reproduce this build error. I'm CCing the maintainer
of gcc42, Jack Howarth.
Cheers,
--
monipol
http://finkers.wordpress.com
Submitting a Fink bug report? Read this:
http://www.finkproject.org/doc/netiquette/index.php
http://finkers.wordpress.com/2009/06/03/bug-reports/
Monipol
On Sat, Aug 01, 2009 at 08:45:49PM -0300, monipol wrote:
On 01/08/2009, at 18:36, monipol wrote:
On 01/08/2009, at 18:08, Jack Howarth wrote:
On Sat, Aug 01, 2009 at 05:52:38PM -0300, monipol wrote:
On 01/08/2009, at 16:24, Mike Zanker wrote:
This is what I'm getting:
bzip2 -dc /sw/src/gcc
On Sat, Aug 01, 2009 at 09:00:07PM -0500, Peter O'Gorman wrote:
Martin Costabel wrote:
I guess pogma needs to add another patch for autoconf-2.64.
Or Fink needs to go back to autoconf-2.63.
Well, gcc requires the use of specific, non-vendor patched versions of
the autotools, the
I can't reproduce these problems here. Someone will have to post
their config.log so we can see why exactly what configure was looking
at that it decided was too old. Alternatively, users could try removing
the autoconf2.5 and autoconf2.13 to see if that helps solve the issue.
Also what happens if
Fixed in current unstable cvs.
On Mon, Aug 10, 2009 at 01:58:07PM +0100, Alan wrote:
Hi there,
Updating my Fink (x86_64) and I got that:
How do you want to proceed? [2]
curl --connect-timeout 30 -f -L -A 'fink/0.29.99.cvs' -O
http://www.bernstein-plus-sons.com/software/RasMol_2.7.5.tar.gz
Dave,
Modifying the various transfig packages as below solves
the x86_64 fink build problems on Leopard.
Jack
--- unstable/main/finkinfo/graphics/transfig-3.2.4-6.info 2008-05-01
13:32:43.0 -0400
+++ local/main/finkinfo/transfig-3.2.4-7.info 2009-08-11
The failure to build the Xgrid support appears to be related to
its incompatibility with the -enable-static configure option...
http://www.open-mpi.org/community/lists/users/2007/08/3900.php
I would complain to upstream.
Alan,
I have uploaded a first pass of new apbs-mpi
packaging which builds apbs properly against openmpi.
It probably still needs work to make compatible with
the lammpi variant. The packaging is on fink tracking
at...
The ability of Apple's gcc to generate different code architectures from
the same compiler is unique to Apple's gcc branch. This feature has never been
ported over to FSF gcc and, with the GPLv3 conflict, now it will likely never be
either. It questionable when and if even the newer objective c
Daniel,
I added static library support to openmpi because of complaints from
folks trying to build static binaries via fink for distributing their
programs. I would rather leave that in as my philosophy is that fink
should do whatever possible to discourage users from installing their
own
It looks like somehow the removal of the gcc43 package didn't
occur when you tried to build gcc44. The following...
Conflicts: gcc4, gcc42, gcc44, gcc45
Replaces: gcc4, gcc42, gcc44, gcc45
has always been sufficient to cause the other conflicting
gcc4X packages to be removed during the build of
on gcc43-shlibs.
/sw/bin/dpkg: error processing gcc43-shlibs (--remove):
I could to try 'remove --recursive', if you think this is a wise.
On Aug 23, 2009, at 11:02 AM, Jack Howarth wrote:
It looks like somehow the removal of the gcc43 package didn't
occur when you tried to build gcc44
Boaz,
I would seriously consider doing an archive and install
of 10.4 to purge out these modifications you have made to
your system. Fink doesn't tolerate unexpected development
packages being installed. Usually folks do that in /usr/local
or /opt so it is easy to disable.
Jack
On
I'll look at this tonight. Here on a G4 running 10.5.8 and
Xquartz 2.4.0 I see...
Program received signal EXC_BAD_ACCESS, Could not access memory.
Reason: KERN_PROTECTION_FAILURE at address: 0x
0x00196104 in _XmClipboardReplaceItem ()
(gdb) bt
#0 0x00196104 in _XmClipboardReplaceItem
Hmmm...I wonder if this is a latent bug in either nedit
or lesstif. See...
https://bugzilla.redhat.com/show_bug.cgi?id=493002
--
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
trial.
Does anyone have a machine with MacPorts installed? They have lesstif 0.95.2
in current MacPorts and the same 5.5 release of nedit. Does that exhibit the
same problem? We are in a pickle here because the 0.95.2 release is supposed to
have 64-bit fixes in it. I'll check tonight if the problem is
Interestingly under x86_64 fink, nedit's paste works fine but bus errors
under i386 fink with darwin10. I see the same bus errors under ppc darwin9
fink. We could create a x86_64 variant for lesstif 0.95.2 and revert the
other architectures back to the last 0.95.0 packaging.
I have updated lesstif to fix error upstream when
the debian 64-bit patches were applied (missing braces).
However this is only part of the fix. It eliminates the
bus errors in nedit but the copy/paste seems non-functional.
I have created updated nedit packaging using the same
Is anyone using the older lesstif-0.95.0-1 and nedit-5.5-1 packages
from fink unstable? Do you have any issues with pasting? In my hands,
these packages seem to often silently refuse to paste under X11 2.4.0
without bus errors. If I disable Update Pasteboard when
CLIPBOARD changes, pasting
I believe if the indexing is choking on a corrupted file
you can identify that while the indexing is hung with...
lsof | grep mdimport
from a Terminal window.
Jack
--
Let Crystal Reports handle the
, 2009 at 5:15 PM, Jack Howarth howa...@bromo.med.uc.eduwrote:
I believe if the indexing is choking on a corrupted file
you can identify that while the indexing is hung with...
lsof | grep mdimport
from a Terminal window.
Jack
Martin,
Look in the i386 fink build log for ocaml and see if there
are there are any instances where the gcc compiler is called
with a path (ie /usr/bin/gcc rather than just gcc). This is
the major cause for breakage in i386 fink builds on SL as
it shortcircuits the use of the compiler
The current SDL sources use the version 1 of the
API for CoreAudio which is depreciated in Snow Leopard.
The maintainer is aware that this code has to be
rewritten to use the newer version 2 API for CoreAudio
for the package to build on Snow Leopard. I asked him
to set Distribution: 10.4, 10.5
Alan,
See the message molmol bug -- minor on the fink-beginners
list. Just because some of these programs superficially appear
to work after updating the openmotif release doesn't mean there
aren't minor bugs being introduced. FYI.
Jack
ps Molmol is unmaintained upstream and I don't
I am puzzled by this sudden focus on resuscitating the g77
package in fink. You do realize that even debian has
dropped g77 from their current stable and unstable releases.
It only exists in Debian's oldstable. What exactly is preventing
gfortran from being used? Unless I am mistaken, all of
Keep in mind that not all source packages support
a parallel make. If I recall correctly, openmotif for
example has issues. It would be better to encourage
package maintainers to simply enable the parallel
make in their info files by using something like...
num_cpu=$(echo `sysctl -n hw.ncpu`)
So far I am seeing the following results with gcc 4.4.2 when trying
to compile the testme.java sample code...
public class testme {
public static void main(String args[]){
System.out.println(Hello);
}
}
with gcj --main=testme -O testme.java
darwin9 darwin10
Has anyone looked at how MacPorts is building screen?
http://trac.macports.org/browser/trunk/dports/sysutils/screen/Portfile?rev=62355
They seem to have addressed both the build issues on Leopard and Snow Leopard
with their patch set...
On Fri, Apr 16, 2010 at 03:49:57AM +0200, Jean-François Mertens wrote:
When instaling gdb, I get:
# dpkg --force-overwrite -i /sw/fink/dists/unstable/main/binary-
darwin-i386/devel/gdb_7.1-1_darwin-i386.deb
(Reading database ... 768642 files and directories currently
installed.)
On Fri, Apr 16, 2010 at 04:50:34PM +0200, Jean-François Mertens wrote:
On 16 Apr 2010, at 15:14, Jack Howarth wrote:
JF,
When you say you had to remove binutils to build gcc45, do you
mean that FSF gcc tries to use binutils instead of cctools if present?
If so, we definitely need either
On Fri, Apr 16, 2010 at 11:21:52AM -0500, Peter O'Gorman wrote:
On 04/16/2010 09:50 AM, Jean-François Mertens wrote:
checking for exported symbols... unable to read unknown load command
0x1b
/sw/bin/objdump: conftest: Invalid operation
checking for -rdynamic... unable to read unknown
On Fri, Apr 16, 2010 at 12:35:16PM -0500, Peter O'Gorman wrote:
On 04/16/2010 12:30 PM, Jean-François Mertens wrote:
since conftest was just compiled, not linked, the -T operation may
not have been considered meaningful by objdump ?
Ie _ is it really an error ?
Yes, because it
On Fri, Apr 16, 2010 at 12:35:16PM -0500, Peter O'Gorman wrote:
On 04/16/2010 12:30 PM, Jean-François Mertens wrote:
since conftest was just compiled, not linked, the -T operation may
not have been considered meaningful by objdump ?
Ie _ is it really an error ?
Yes, because it
On Fri, Apr 16, 2010 at 12:35:16PM -0500, Peter O'Gorman wrote:
On 04/16/2010 12:30 PM, Jean-François Mertens wrote:
since conftest was just compiled, not linked, the -T operation may
not have been considered meaningful by objdump ?
Ie _ is it really an error ?
Yes, because it
I believe the openmpi packages were left that
way because in general most package maintainers didn't
want to be forced to use the FSF gcc/g++ just because
they needed gfortran. This trend was extended to the
openmpi package.
Jack
I have posted new packaging for openmpi-1.4.1-1000 on
fink tracking
https://sourceforge.net/tracker/?func=detailaid=2994449group_id=17203atid=414256
This package requires the new gcc45-4.5.0-1001 package
also on fink tracking...
On Wed, Jun 02, 2010 at 07:54:31AM -0400, Alexander Hansen wrote:
It errored out with:
cp
/sw/src/fink.build/gcc45-4.5.0-1000/gcc-4.5.0/gcc/config/darwin-sections.def
/sw/src/fink.build/root-gcc45-4.5.0-1000/sw/lib/gcc4.5/lib/gcc/powerpc-apple-darwin${darwinvers}/4.5.0/plugin/include/config
On Mon, Jun 07, 2010 at 10:25:51PM +0200, Dominique Dhumieres wrote:
I won't disable gcj in fink's libtool yet, I'd like to see if I get more
reports of failures.
What is working is to use gcc44.
Is this the same issue of gcj in gcc 4.5.0 not compiling java under
darwin9? If so, gcc 4.4.x
1 - 100 of 284 matches
Mail list logo