Re: [Fink-devel] opengl-py broken?

[Martin Costabel]

Daniel Macks wrote:
> On Tue, Mar 24, 2009 at 04:53:38PM +0100, Jean-Fran?ois Mertens wrote:
>> The new pkg wxgtk2.8-py26 has a dep on opengl-py26.
>> Checking whether the latter did build, I got:
[]
>>> /usr/libexec/gcc/powerpc-apple-darwin8/4.0.1/ld: Undefined symbols:
>>> _glCurrentPaletteMatrixARB
[]
>> Indeed, those symbols are to be found nowhere..
>> The difference with py25 is that the latter uses "-undefined  
>> dynamic_lookup"
>> (I guess caused by the python used)
>> in this link command _ thus creating a (or _ many ?) broken bundle _ cf
>> "nm -m /sw/lib/python2.5/site-packages/OpenGL/GL/ARB/ 
>> matrix_palette.so| fgrep _glMatrixIndex"
>>
>> So I presume all variants of opengl-py are broken..
>> (don't know sufficiently of python and its buildsystem to fix this  
>> myself)
>>
>> Is this diiference in flags due to python ? and 
> 
> I elided the -undefined dynamic_lookup in py26 specifically to catch
> this type of problem. A missing symbol is a missing symbol, at least
> now it's caught at build-time instead of runtime.

I have now updated opengl-py to the latest version 3.0.1, and there this 
error seems to have gone. opengl-py26 builds now for me.

-- 
Martin



--
This SF.net email is sponsored by Sprint
What will you do first with EVO, the first 4G phone?
Visit sprint.com/first -- http://p.sf.net/sfu/sprint-com-first
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] mayavi2 fink

[Martin Costabel]

Alexander Hansen wrote:
> -BEGIN PGP SIGNED MESSAGE-
> Hash: SHA1
> 
> On 6/29/10 11:40 AM, Jed Frechette wrote:
>> On Tue, 29 Jun 2010 09:20:54 -0600, RaDaniel Christian 성준  
>>  wrote:
>>
>>> I have found your fink package for MayaVi2-py25, but I can't install
>>> it. I was wondering if you could help me with this:
>>>
>>> choel...@bcm-17-82:~/ > fink install mayavi2-py25
>>> Password:
>>> Scanning package description files..
>>> Information about 9034 packages read in 2 seconds.
>>> Can't resolve dependency "vtk-py25 (>= 5.0)" for package
>>> "mayavi2-py25-3.2.0-2" (no matching packages/versions found)
>>> Exiting with failure.
>> ...
>>> I have the 64 bit version of fink installed, and I have also installed
>>> http://stronginference.com/scipy-superpack/ with the .sh script
>>> provided.
>>> However, I have not been able to find or run any of the commands to
>>> start up the program.
>>
> 
> Fink isn't guaranteed to work with any arbitrary third-party stuff.
> 
> vtk-py25 has been pulled from the 10.6 distribution because it doesn't
> work, and mayavi2 should be pulled as well, until somebody can modify it
> to build.

I am not trying to fix the old vtk-py25 for 10.6/64bit (it would work on 
10.6/32bit),but I made a new package for vtk56 the latest release of 
vtk, which uses py26. There is now also a new version of mayavi2 that 
uses vtk56. Hopefully this will work on all 10.6 architectures.

-- 
Martin




--
This SF.net email is sponsored by Sprint
What will you do first with EVO, the first 4G phone?
Visit sprint.com/first -- http://p.sf.net/sfu/sprint-com-first
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

[Fink-devel] wx-config missing libraries

[Martin Costabel]

With wxgtk2.8-2.8.10-23, `wx-config --libs` gives the following output:

-L/sw/lib  -L/sw/lib/fontconfig2/lib 
-Wl,-dylib_file,/System/Library/Frameworks/OpenGL.framework/Versions/A/Libraries/libGL.dylib:/System/Library/Frameworks/OpenGL.framework/Versions/A/Libraries/libGL.dylib
 
-L/usr/X11R6/lib -L/sw/lib   -L/sw/lib -framework IOKit -framework 
CoreServices -framework System -framework ApplicationServices 
-lwx_gtk2u_richtext-2.8 -lwx_gtk2u_aui-2.8 -lwx_gtk2u_xrc-2.8 
-lwx_gtk2u_qa-2.8 -lwx_gtk2u_html-2.8 -lwx_gtk2u_adv-2.8 
-lwx_gtk2u_core-2.8 -lwx_baseu_xml-2.8 -lwx_baseu_net-2.8 -lwx_baseu-2.8

This makes the build of wxgtk2.8-py26 crash with undefined symbols. For 
wxgtk2.8-py25, there were also undefined symbols, but the build didn't 
crash, which was incorrect. There are indeed quite a few libs that otool 
-L shows for the above list and which are not listed. I think they 
should. For wxget2.8-py26, the missing libs are

-lgdk-x11-2.0 -lgobject-2.0 -lX11

I have added a corresponding PatchFile line to wxgtk2.8.info, and I am 
checking this now in, so wxgtk2.8-py26 and consequently mayavi2-py26 can 
build. Pierre-Henri, I hope you don't mind.

-- 
Martin



--
This SF.net email is sponsored by Sprint
What will you do first with EVO, the first 4G phone?
Visit sprint.com/first -- http://p.sf.net/sfu/sprint-com-first
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] Orphaning my packages

[Martin Costabel]

Jed Frechette wrote:
> Due to ever decreasing use of OS X for work and a general lack of time I  
> have decided to orphan my Fink packages. Hopefully, someone will be  
> interested enough to adopt a few of the packages listed below, if not, can  
> someone change the Maintainer field to fink-devel in svn.
> 
> Best wishes,
> 
> setupdocs-py
> wxgtk2.8-py.info
> app-tools-py.info
> configobj-py.info
> enthought-base-py.info
> envisage-core-py.info
> envisage-plugins-py.info
> traits-backend-qt-py.info
> traits-backend-wx-py.info
> traits-gui-py.info
> traits-py.info
> mayavi2-py.info

I have updated mayavi2 to the latest version and let it use vtk56 now.
Along the way, I had to update most of the other packages in your list 
and am taking them over for now.

-- 
Martin



--
This SF.net email is sponsored by Sprint
What will you do first with EVO, the first 4G phone?
Visit sprint.com/first -- http://p.sf.net/sfu/sprint-com-first
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] libjpeg-shlibs and libjpeg dependencies in your packages!

[Martin Costabel]

Koen van der Drift wrote:
[]
> make[262]: Makefile:11: fork: Resource temporarily unavailable
> make[262]: vfork: Resource temporarily unavailable
> make[261]: *** [dirs] Error 2

The [262] tells you that there are 262 levels of "make" running. Either 
there is an infinite loop (but then this should happen to others, too), 
or it is a question of parallel make running wild. Do you somehow have 
your fink tweaked so that it runs "make -j" without a finite (small) 
number of instances given, that is, not "make -j4" or something? This 
has been seen on google as causing such a runaway vforking, because make 
then spawns as many instances of itself as it can.

I haven't followed the recent discussion about parallel makes, but I 
seem to dimly remember that some code was added to fink that runs "make 
-jn" with a value for "n" that is determined from the machine itself. If 
such code is in your version of fink, then this could be a source of the 
error if it doesn't work as intended.

-- 
Martin



--
ThinkGeek and WIRED's GeekDad team up for the Ultimate 
GeekDad Father's Day Giveaway. ONE MASSIVE PRIZE to the 
lucky parental unit.  See the prize list and enter to win: 
http://p.sf.net/sfu/thinkgeek-promo
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] SDL or libiconv problem?

[Martin Costabel]

David Lowe wrote:
>   This error just showed up during an 'update-all'.  
> 
>>  #include 
>> configure:23968: result: no
>> configure:23968: checking SDL_mixer.h presence
>> configure:23968: gcc -E -g -O2 -I/sw/include/SDL -D_GNU_SOURCE=1 
>> -D_THREAD_SAFE conftest.c
>> In file included from /sw/include/SDL/SDL_types.h:28,
>>  from /sw/include/SDL/SDL_mixer.h:28,
>>  from conftest.c:62:
>> /sw/include/SDL/SDL_stdinc.h:74:20: error: iconv.h: No such file or directory
>> configure:23968: $? = 1
> 
>   I do have headers for SDL_types, SDL_mixer, and SDL_stdinc in 
> /sw/include/SDL/.  Is this by any chance related to recent changes in 
> libiconv?

It is probably related to your recent test of the libiconv package. 
Normally, the above would find /usr/include/iconv.h. I suspect you 
removed that for the test and forgot to move it back. It is, however, a 
good thing that /usr/include/iconv.h was not found in this situation, 
because this could lead to incompatibilities.

-- 
Martin



--
ThinkGeek and WIRED's GeekDad team up for the Ultimate 
GeekDad Father's Day Giveaway. ONE MASSIVE PRIZE to the 
lucky parental unit.  See the prize list and enter to win: 
http://p.sf.net/sfu/thinkgeek-promo
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] Fixing libiconv (build-test needed)

[Martin Costabel]

Daniel Macks wrote:
> Could someone(s) test the libiconv in my cvs exp? Goal: compare build
> transcript:
> 
>   1) with fink libiconv-dev installed,
>   2) without libiconv-dev,
>   3) without libiconv-dev *and with /usr/include/iconv.h and
>  /usr/lib/libiconv.dylib (and .la if present) moved away.
> 
> Need to make sure we aren't getting circular dependency and aren't
> mixing system with internal lib/headers of same name. Clean on 10.4,
> need checked on 10.6/32 and 10.6/64
> 
> http://fink.cvs.sf.net/viewvc/*checkout*/fink/experimental/dmacks/finkinfo/libiconv.info?revision=1.2
> http://fink.cvs.sf.net/viewvc/*checkout*/fink/experimental/dmacks/finkinfo/libiconv.patch?revision=1.1

No problem here on 10.6/64. The only difference in the build logs is a line

libtool: install: warning: `../lib/libiconv.la' has not been installed 
in `/sw/lib'

that appears when libiconv is built while libiconv-dev is not installed. 
This seems to indicate that when libiconv-dev is installed, the build 
process sees at this point the installed /sw/lib/libiconv.la and not the 
newly-built ../lib/libiconv.la. But why it looks at libconv.la at this 
point at all is unclear. This is between the two lines

/bin/sh ../libtool --mode=install /usr/bin/install -c iconv 
/sw/src/fink.build/root-libiconv-1.12-4/sw/bin/iconv

and

/usr/bin/install -c .libs/iconv 
/sw/src/fink.build/root-libiconv-1.12-4/sw/bin/iconv

just after the linking of the iconv executable is done, which for some 
strange reason happens during the install phase.

-- 
Martin





--
ThinkGeek and WIRED's GeekDad team up for the Ultimate 
GeekDad Father's Day Giveaway. ONE MASSIVE PRIZE to the 
lucky parental unit.  See the prize list and enter to win: 
http://p.sf.net/sfu/thinkgeek-promo
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] sudo apt-get remove question

[Martin Costabel]

Koen van der Drift wrote:
[]
> and from dpkg -l
> 
> ri  xml-parser-pm588  2.36-2Perl ext interface to  
> XML parser/expat

"ri" says that it is installed, but marked for removal. I don't quite 
know what this means, probably that a removal had been attempted, but 
was unsuccessful. Too unclean for apt-get.

-- 
Martin



--
ThinkGeek and WIRED's GeekDad team up for the Ultimate 
GeekDad Father's Day Giveaway. ONE MASSIVE PRIZE to the 
lucky parental unit.  See the prize list and enter to win: 
http://p.sf.net/sfu/thinkgeek-promo
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] sudo apt-get remove question

[Martin Costabel]

Koen van der Drift wrote:
> I'm trying to force remove a package, but get this output:
> 
> $ sudo apt-get remove xml-parser-pm588
> Reading Package Lists... Done
> Building Dependency Tree... Done
> You might want to run `apt-get -f install' to correct these:
> Sorry, but the following packages have unmet dependencies:
>intltool40: Depends: xml-parser-pm588 but it is not going to be  
> installed
>software-license-pm588: Depends: data-section-pm588 but it is not  
> going to be installed
>Depends: mro-compat-pm588 but it is not  
> going to be installed
>Depends: sub-exporter-pm588 but it is not  
> going to be installed
>Depends: sub-install-pm588 but it is not  
> going to be installed
>Depends: text-template-pm but it is not  
> going to be installed
> E: Unmet dependencies. Try 'apt-get -f install' with no packages (or  
> specify a solution).
> 
> 
> I don't understand this, why would it complain about unmet  
> dependencies when I want to *remove* a package??

You will probably get the same complaint from apt-get whenever you ask 
it to do *anything*. It just refuses to work as long as it feels 
something is unclean in its database. dpkg is much less inhibited, you 
can use it to --remove --force-all whatever you wish.

OTOH, the complaint seems to indicate among other things that 
xml-parser-pm588 is not installed, at least not according to apt-get. I 
don't know if apt-get or apt-cache can give you more detailed 
information, but "dpkg -l" will.

-- 
Martin



--
ThinkGeek and WIRED's GeekDad team up for the Ultimate 
GeekDad Father's Day Giveaway. ONE MASSIVE PRIZE to the 
lucky parental unit.  See the prize list and enter to win: 
http://p.sf.net/sfu/thinkgeek-promo
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] How to remove a certain group of packages

[Martin Costabel]

Koen van der Drift wrote:
> Hi,
> 
> Is there a way to remove a certain group of packages in fink? For  
> instance, all packages that I maintain, or all installed perlmodules?
> 
> I can list them:
> 
> fink list --installed -m 'driftkop'
> 
> or
> 
> fink list --installed | grep pm588
> 
> But then how to proceed?

For example (1 line)

sudo dpkg -r --force-all `fink list -i -m 'driftkop' | awk '{print $2}'`

-- 
Martin



--
ThinkGeek and WIRED's GeekDad team up for the Ultimate 
GeekDad Father's Day Giveaway. ONE MASSIVE PRIZE to the 
lucky parental unit.  See the prize list and enter to win: 
http://p.sf.net/sfu/thinkgeek-promo
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] linker error: ld: library not found for -lcrt1.10.5.o

[Martin Costabel]

Koen van der Drift wrote:
[]
> I tried that, and now get the following error instead at the same location:
> 
> Linking CXX executable massXpert.app/Contents/MacOS/massXpert
> cd /sw/src/fink.build/massxpert-x11-2.1.0-1001/massxpert-2.1.0/gui && 
> /sw/bin/cmake -E cmake_link_script CMakeFiles/massXpert.dir/link.txt 
> --verbose=1
> /sw/var/lib/fink/path-prefix-g++-4.0/c++-mmacosx-version-min=10.3 
> -fvisibility-inlines-hidden -isysroot /Developer/SDKs/MacOSX10.4u.sdk 

OK, the -mmacosx-version-min=10.5 is gone, but -mmacosx-version-min=10.3 
is still there, and -isysroot /Developer/SDKs/MacOSX10.4u.sdk too.
I don't know if these have anything to do with the error you are seeing 
now, though.

[]
> QtXml -framework QtSvg -Wl,-all_load ../lib/libmasslib.a -Wl,-noall_load 
> -framework QtGui -framework QtCore ../lib/libmasslib.a
> 
> ld warning: option -noall_load is obsolete and being ignored
> Undefined symbols:
>   "_CFRelease", referenced from:
>   massXpert::ConfigSettings::initializeSystemConfig()  in 
[]

Looks like a missing -framework CoreFoundation. No idea who should have 
brought this in and didn't.

> And, indeed, I don't have  
> /Developer/SDKs/MacOSX10.4u.sdk/usr/lib/crt1.10.5.o but have 
> /Developer/SDKs/MacOSX10.4u.sdk/usr/lib/crt1.o .

That's why removing -mmacosx-version-min=10.5 worked (for your first 
error, at least).

-- 
Martin

--

___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] linker error: ld: library not found for -lcrt1.10.5.o

[Martin Costabel]

Koen van der Drift wrote:
> Hi,
> 
> As posted earlier this week on IRC, I am getting the following linker  
> error:
> 
> Linking CXX executable massXpert.app/Contents/MacOS/massXpert
> cd /sw/src/fink.build/massxpert-x11-2.1.0-1001/massxpert-2.1.0/gui && / 
> sw/bin/cmake -E cmake_link_script CMakeFiles/massXpert.dir/link.txt -- 
> verbose=1
> /sw/var/lib/fink/path-prefix-g++-4.0/c++-mmacosx-version-min=10.3 - 
> fvisibility-inlines-hidden -isysroot /Developer/SDKs/MacOSX10.4u.sdk - 
> mmacosx-version-min=10.5 -Wl,-search_paths_first - 
[]
> ld: library not found for -lcrt1.10.5.o
>  
> This is on:
> 
> Package manager version: 0.29.99.cvs
> Distribution version: selfupdate-cvs Fri May  7 18:18:20 2010, 10.5,  
> powerpc
> 
> And I have on my system:
> 
> /usr/lib/crt1.10.5.o  
> /usr/lib/crt1.o
> 
> 
> Any ideas what could cause this?

The -isysroot flag. You probably don't have 
/Developer/SDKs/MacOSX10.4u.sdk/usr/lib/crt1.10.5.o

Unfortunately, cmake is trying successfully to out-bloat the 
autotools-based build system. In 2.8 they added this -isysroot garbage 
as default on MacOSX.

If one doesn't want to build a universal binary, one now needs to get 
rid of the -isysroot flag and to remove the -mmacosx-version-min crap, 
too. With cmake, you can do this by adding the two options

-DCMAKE_OSX_DEPLOYMENT_TARGET:STRING="" \
-DCMAKE_OSX_SYSROOT:PATH="/" \

to the cmake command line. I haven't tested this for massXpert on OSX 
10.5, but I had to do it also for scribus135-aqua in order to make it 
build on OSX 10.4.

-- 
Martin




--

___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] distfiles mirrors and License: Restrictive (and SSL linking as Restrictive)

[Martin Costabel]

David R. Morrison wrote:
[]
> However, I can't think of a case other than the openssl nonsense  
> where this would apply.  And many openssl packages have been  
> converted to use the openssl which ships with os x (rather than  
> fink's) which makes it ok to distribute.  So I'm thinking that it  
> might be better to spend energy in revising those packages, rather  
> than in writing fink code (and distfiles code) to handle a new  
> special case.

A healthier option would be to simply ignore that silly pissing contest 
between gnu and openssl and declare gpl packages as gpl, full stop. To 
any non-American at least it is clear that we are splitting hairs here 
that don't even exist, since Fink has stopped distributing binaries long 
ago.

-- 
Martin



--
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] gcc4x/gcc4x-compiler packages

[Martin Costabel]

Jack Howarth wrote:
> JF,
>   Okay. Untested gcc45-x86_64.info with a gcc45-compiler
> splitoff. Still needs the libffi update-alternatives.

To me, this one looks OK (untested, of course), contrary to the 
description in one of dmacks' latest messages (the one with *1, *2 etc), 
which even after reading it 3 times still looks really backwards.

-- 
Martin





--
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] co-existing gcc4x packages

[Martin Costabel]

Jack Howarth wrote:
> On Sun, Apr 25, 2010 at 09:46:51AM +0200, Martin Costabel wrote:
>> Jack Howarth wrote:
>> []
>>>What I am considering is to add a gcc4X-bin
>>> split-off to all of the gcc4X packages which
>>> will contain all of the %/bin symlinks currently
>>> provided by the main gcc4X package.
>> Why do we neeed *any* of these executables directly in %p/bin?
>> They could as well live in some subdirectory of %p/lib, and if you want  
>> to use them, you modify PATH correspondingly.
> 
> Ease of use? Users already have to adapt to the using
> a different compiler name in some cases and now they
> have to invoke a new path as well? 

In many cases, you don't even have to change PATH. Just use the full 
name of the executable.

No sane person will remember for longer than 24 hours that 'gcc-4' is 
Fink's gcc version 4.2.4 if gcc42 is installed or version 4.4.2 if gcc44 
is installed, whereas 'gcc-4.2' is Apple's gcc version 4.2.1 and 
'gcc-4.0' is Apple's gcc version 4.0.1, even if it is invoked as 
'/sw/bin/gcc-4.0'.

So in any case you will have to look it up before using it either 
manually or in a script or package. And I would claim that when you need 
to look it up anyway before using it, it is only marginally more 
complicated to use '%p/lib/gcc-4.4/bin/gcc' than 'gcc-4'.

-- 
Martin



--
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] co-existing gcc4x packages

[Martin Costabel]

Jack Howarth wrote:
[]
>What I am considering is to add a gcc4X-bin
> split-off to all of the gcc4X packages which
> will contain all of the %/bin symlinks currently
> provided by the main gcc4X package.

Why do we neeed *any* of these executables directly in %p/bin?
They could as well live in some subdirectory of %p/lib, and if you want 
to use them, you modify PATH correspondingly.

-- 
Martin



--
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] Atlas library for stable macos X 10.4.11

[Martin Costabel]

Panagiotis Maniadis wrote:
[]
> I change the "cp -P" to "cp -d" on the info file

Why do you think this is a good idea? /bin/cp does not understand the -d 
flsg.

-- 
Martin




--
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] gcc45-4.5.0-1000 release packaging

[Martin Costabel]

Jean-François Mertens wrote:
[]
> [ Under melina, "options_machine FC_name" or "options_machine LINKER"
> currently yield just the plain "gfortran", and nothing in done in the 
> info file
> to force this. ]

That's right, melina simply needs a working gfortran and a working C 
compiler. The ideal would be to have a gfortran from the same gcc 
version as the system gcc. I (and melina) don't care about the latest 
and greatest gcc, so any of gcc4[2-5] would be OK, but I don't like 
adding arbitrary preventive dependencies on packages that I haven't 
tested and won't test for a while.

-- 
Martin



--
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] update-alternatives usage

[Martin Costabel]

Jack Howarth wrote:
> Martin,
>So what is the recommended method for resolving conflicts over
> manpages if update-alternatives is the wrong approach?

Personally, I would rename one of them. I would rather not to find a man 
page than be shown one with the right name that is not the one I am 
looking for.

-- 
Martin





--
Download Intel® Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] update-alternatives usage

[Martin Costabel]

Jack Howarth wrote:
>In order to work around the conflicts over ffi.3,
> ffi_call.3 and ffi_prep_cif.3 between the libffi and
> gcc45 package, I am looking at adding the calls to
> update-alternatives to gcc45. Looking at the openmotif4.info
> as an example, I see...
> 
> PostInstScript: <<
> update-alternatives --install %p/share/man/man3/Object.3 Object.3 
> %p/share/man/man3/Object.3.openmotif 65
> <<
> PreRmScript: <<
> if [ $1 != "upgrade" ]; then
>   update-alternatives --remove Object.3 %p/share/man/man3/Object.3.openmotif
> fi
> <<
> 
> However, this doesn't seem to be functional as my installed copy of openmotif4
> has...
> 
> ls -l /sw/share/man/man3/Object.3
> -rw-r--r--  1 root  admin  2452 Sep 12  2009 /sw/share/man/man3/Object.3
> cd /sw/etc/alternatives
> ls -l Object.3
> lrwxr-xr-x  1 root  admin  33 Jul 24  2009 Object.3 -> 
> /sw/share/man/man3/Object.3.tcltk
> 
> but... 
> 
> ls -l /sw/share/man/man3/Object.3.tcltk
> ls: /sw/share/man/man3/Object.3.tcltk: No such file or directory

Although it is a blatant misuse of the update-alternatives system, this 
example is actually an excellent illustration for the reasons why I hate 
this system:

The system was certainly not meant for avoiding random file name 
conflicts between arbitrary packages, but for allowing users to choose 
variants of certain generic utilities. A good example would be 
/sw/bin/sed, which can point to various different versions of sed. But 
even in this legitimate case I would dispute its utility, because its 
effect is that in practice you *lose* control over what you have installed.

In your example, update-alternatives was originally used to hide the 
fact that 2 packages used the same name for two quite different man 
pages. The update-alternatives command was put into the Fink info files 
many years ago and then carried on, although it had long since lost its 
usefulness, because nobody (you included, when you were maintainer of 
openmotif) understood what it was doing, and therefore nobody dared 
remove it.

> In this case don't we need to explicitly move aside the Object.3 as
> Object.3.tcltk in the install directory before creating the deb files?

tcltk hasn't had an Object.3 man page for many years, and the package 
doesn't use update-alternatives --install, only --remove. But you are 
right, the openmotif package, if it really was serious about using the 
update-alternatives system, would have had to rename its Object.3 file 
first, which it appears to have never done. So even when it was 
introduced into the openmotif info file in 2003 (or earlier; cvs log 
shows that it was there in April 2003), the update-alternatives command 
was probably not doing what it was supposed to do.

In short: Do not use update-alternatives, unless you absolutely need to 
and you know what you are doing.

-- 
Martin




--
Download Intel® Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] Fink cmake.info patch

[Martin Costabel]

James Bunton wrote:
[]
> The problem is not a build failure. ccmake (which uses ncurses to
> display a menu-like configuration interface) no longer works with the
> arrow keys on the keyboard to edit the configuration.

Wunderful. This breaks also the ccmake that comes with the precompiled 
CMake.app. And probably lots of other software that uses ncurses. They 
changed neither the compatibility version nor even the current version 
of libncurses.5.4.dylib, although the symbols in the two versions are 
quite different. I am going to file a bug with Apple as soon as I have 
time. I'll also fix the cmake package. I don't know, though, why anyone 
would ever want to run ccmake. Feels like dselect for installing Fink 
packages :-)

-- 
Martin

--
Download Intel® Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] Unpacking failure during package build

[Martin Costabel]

Ather Qadri wrote:
[]
> sh-3.2# fink -m --build-as-nobody rebuild maxwell
[]
> sudo -u fink-bld [ENV] sh -c/var/tmp/tmp.1.6CT8nu
> Can't locate Fink/Services.pm in @INC (@INC contains: 
> /Library/Perl/Updates/5.8.8 
> /System/Library/Perl/5.8.8/darwin-thread-multi-2level 
> /System/Library/Perl/5.8.8 
> /Library/Perl/5.8.8/darwin-thread-multi-2level /Library/Perl/5.8.8 
> /Library/Perl /Network/Library/Perl/5.8.8/darwin-thread-multi-2level 
> /Network/Library/Perl/5.8.8 /Network/Library/Perl 
> /System/Library/Perl/Extras/5.8.8/darwin-thread-multi-2level 
> /System/Library/Perl/Extras/5.8.8 /Library/Perl/5.8.6 
> /Library/Perl/5.8.1 .) at /var/tmp/tmp.1.6CT8nu line 2.
> BEGIN failed--compilation aborted at /var/tmp/tmp.1.6CT8nu line 2.

You are running fink as root.
Does it work if you run it as yourself, without first becoming root?

-- 
Martin

--
Download Intel® Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] Error while build as nobody (CILK)

[Martin Costabel]

Pranay Airan wrote:
> i removed all %i from file field but got following error
> 
> Validating .deb dir /sw/src/fink.build/root-cilk-5.4.6-1...
> Package looks good!
> dpkg-deb -b root-cilk-5.4.6-1 /sw/fink/10.5/local/main/binary-darwin-i386
> dpkg-deb: building package `cilk' in 
> `/sw/fink/10.5/local/main/binary-darwin-i386/cilk_5.4.6-1_darwin-i386.deb'.
> Reverting ownership of install dir to root
> - Depends line is: darwin (>= 9-1)
> Writing control file...
> Creating shlibs files...
> Validating .deb dir /sw/src/fink.build/root-cilk-shlibs-5.4.6-1...
> Error: Files with names less specifically versioned than ones in public 
> Shlibs entries do not belong in this package
> Offending file: /sw/lib/libcilk.g.dylib
> Offending file: /sw/lib/libcilk.g.p.dylib
> Offending file: /sw/lib/libcilk.p.dylib

This is explained in the Packaging Manual, section 3.4 "Shared 
Libraries" 

-- 
Martin

--
Download Intel® Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] Issue with numpy-py26 / scipy-core-py26

[Martin Costabel]

ピエールランリ・ラヴィン wrote:
> Hello all,
> 
> I have an issue with update-all. Same problem on Tiger and Snow Leopard,
> unstable tree.
> 
> I don't know but one package seems to require some python stuff, and an
> issue with dependencies will force me to remove gnome.

 From the fink-users and -beginners lists:

sudo dpkg -r --force-all scipy-core-py26
fink install scipy-core-py26

-- 
Martin Costabel

--
Download Intel® Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] pyobjc-py26-1.4-3

[Martin Costabel]

Alexander Hansen wrote:
[]
> In file included from Modules/AppKit/_AppKit.m:1042:
> build/codegen/_App_Enum.inc: At top level:
> build/codegen/_App_Enum.inc:702: error: 'NSOneByteGlyphPacking'
> undeclared here (not in a function)

Alex,

it seems we had an exchange about this same problem already a couple of 
months ago: 

The situation has not evolved since.

-- 
Martin

--
Download Intel® Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] PostRmScript

[Martin Costabel]

Daniel Macks wrote:
[]
> According to the Debian package-script docs, postrm does run during an
> upgrade/reinstall (because the old/previous one is being rm'ed). 

According to man dpkg, what should work is to use PreRmScript instead of 
PostRmScript.

-- 
Martin


--
Download Intel® Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] ncmpcpp build error with 10.6. Apparently due to ncurses library.

[Martin Costabel]

Alexander Hansen wrote:
[]
> I get the same error on 10.4, as well, so we appear not to be building
> on any supported platform.

On 10.4 I suspect it cannot be built (the system curses.h is as old as 
Fink's there). On 10.5/10.6, it can be built when libncurses5 is not 
installed. A BuildConflicts: libncurses5 is sufficient for this. In my 
case, I first had to remove the yafc package, which illegally Depends on 
libncurses5. OTOH, the ncmpcpp executable links with 
/sw/lib/libncursesw.5.dylib, which is not nice when the corresponding 
header files are not used.

I wonder how Jack was ever able to build this package, because the 
libncurses5 package, while not essential, is a build dependency of 
hundreds of packages, so the chances are very high that it is installed.

-- 
Martin



--
The Planet: dedicated and managed hosting, cloud storage, colocation
Stay online with enterprise data centers and the best network in the business
Choose flexible plans and management services without long-term contracts
Personal 24x7 support from experience hosting pros just a phone call away.
http://p.sf.net/sfu/theplanet-com
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] ncmpcpp build error with 10.6. Apparently due to ncurses library.

[Martin Costabel]

Alexander Hansen wrote:
[]
>> screen.h:251: error: 'NCURSES_BUTTON_PRESSED' was not declared in this scope
[]
> This is about an error in a maintained package, so discussion would be
> more appropriate for fink-users.  fink-devel is more for packaging
> questions and unmaintained packages.
> 
> We'll leave it here, though, and cc the maintainer.
> 
> It doesn't build on 10.5, either (same error).

This might even be suitable for fink-core :-)

I suspect this is not a problem of ncmpcpp, but a sign of Fink's ncurses 
package finally showing its age (6 years against 4 years for Leopard's, 
whereas Snow Leopard's seems to be up to date).

Does anyone remember why Fink needs its own ncurses package?

-- 
Martin

--
The Planet: dedicated and managed hosting, cloud storage, colocation
Stay online with enterprise data centers and the best network in the business
Choose flexible plans and management services without long-term contracts
Personal 24x7 support from experience hosting pros just a phone call away.
http://p.sf.net/sfu/theplanet-com
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] Problem building vtk-py25

[Martin Costabel]

Dr. Wolfram Schroers wrote:
> No, I do not need this old version. However, I wanted to install mayavi2 (or, 
> at least, the older mayavi) and this is a required dependency. If one could 
> also use the more recent vtk, I would be very happy to use that one, instead.

Ah yes, forgot about mayavi. The problem with more recent versions of 
vtk is that in python "import vtk" does not work, because of linking 
problems of the python modules.

There are several different ways of resolving the python-supplied 
symbols in python modules, and so far I have not had success with any of 
them. This is a very time-consuming process, not only because vtk takes 
a long time to compile, but mainly because of the cmake build system, 
which is a royal PITA in this situation. Trying to convince it to use a 
certain linker flag and to link or not to link to a certain library is a 
major hacking challenge.

If anyone on the list wants to give it a try, I'd be happy. The 
possibilities are:

1. Let cmake do what it wants (after setting a lot of cmake flags, of 
course). This resolves undefined symbols from libpython2.5.dylib and 
links the modules with that dylib. For many python modules (and this 
happens with vtk52 and vtk54, but strangely not with vtk50), this 
situation leads to the famous "interpreter version mismatch" crash when 
the module is imported into python. I think this problem might disappear 
with python2.6, but I haven't had time to check.

2. Add the "-bundle_loader /sw/bin/python2.5" linker flag. This uses the 
python executable to resolve symbols. It is the recommended method for 
linking python modules, but most of the time leads to linker errors.

3. Try to get rid of libpython.dylib on the linker line and add either 
"-undefined dynamic_lookup" or "-flat_namespace -undefined suppress". 
This works for many python modules, but with vtk I got linker errors, too.

I'll continue trying this, but I cannot estimate when I will succeed. In 
any case, I will never try to fix anything that involves funny java 
compiler errors.

-- 
Martin









--
Throughout its 18-year history, RSA Conference consistently attracts the
world's best and brightest in the field, creating opportunities for Conference
attendees to learn about information security's most important issues through
interactions with peers, luminaries and emerging and established companies.
http://p.sf.net/sfu/rsaconf-dev2dev
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] Problem building vtk-py25

[Martin Costabel]

Wolfram Schroers wrote:
> Hi,
> 
> I have a problem building the package vtk-py25. At 79% I get the following 
> error message:
> 
> [ 79%] Generating ../../java/vtk/vtkBuildAllDriver.class
> cd /sw/src/fink.build/vtk-py25-5.0.4-3/VTKBuild/Wrapping/Java && 
> /usr/bin/javac -classpath 
> /sw/src/fink.build/vtk-py25-5.0.4-3/VTKBuild/java/vtk/.. -d 
> /sw/src/fink.build/vtk-py25-5.0.4-3/VTKBuild/java/vtk/.. 
> /sw/src/fink.build/vtk-py25-5.0.4-3/VTKBuild/java/vtk/vtkBuildAllDriver.java
> Exception in thread "main" java.lang.OutOfMemoryError: Java heap space
>   at java.util.ResourceBundle$CacheKey.clone(ResourceBundle.java:600)

Do you need this old version? There is vtk54, which builds on 10.6.
I will probably just restrict vtk-py25 to 10.4 and 10.5, because I have 
no idea (and don't want to spend time on) how to fix this error.

-- 
Martin



--
Throughout its 18-year history, RSA Conference consistently attracts the
world's best and brightest in the field, creating opportunities for Conference
attendees to learn about information security's most important issues through
interactions with peers, luminaries and emerging and established companies.
http://p.sf.net/sfu/rsaconf-dev2dev
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] [Fink-users] Please remove screen from fink

[Martin Costabel]

William G. Scott wrote:
> 
> On Jan 14, 2010, at 11:08 PM, Michal Suchanek wrote:
> 
>> 2010/1/15 William G. Scott :
>>> In my case, fink's screen works flawlessly (including sudo), and Apple's is 
>>> problematic (hangs).  I am not using bash.  Maybe bash is not playing well 
>>> with fink's screen...
>> I don't think the shell could affect how sudo works but it could hang
>> the screen. Perhaps your shell does no play well with it?
> 
> Quite possibly.
> 
>> What version of OS X do you have?
> 
> 10.6.2

This is strange. On 10.6.2, I get the "sudo: unknown uid: 502" error 
with Fink's screen (both bash or tcsh as login shell). Apple's screen 
works correctly with both login shells.

To see the error, you have to start screen remotely, then quit it (C-a 
C-d), log out, then ssh in again and reconnect to the screen. Then sudo 
or chsh etc will give the error.

-- 
Martin


--
Throughout its 18-year history, RSA Conference consistently attracts the
world's best and brightest in the field, creating opportunities for Conference
attendees to learn about information security's most important issues through
interactions with peers, luminaries and emerging and established companies.
http://p.sf.net/sfu/rsaconf-dev2dev
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] [Fink-users] Please remove screen from fink

[Martin Costabel]

William G. Scott wrote:
> In my case, fink's screen works flawlessly (including sudo), and Apple's is 
> problematic (hangs).  I am not using bash.  Maybe bash is not playing well 
> with fink's screen...

For me, Fink's screen works on 10.4, but gives the "uid xxx does not 
exist in the passwd file!" error on 10.5 and 10.6.

Some more information on the history of screen:

Apple's screen is built from the same sources as Fink's screen. On 10.4, 
Apple used screen-4.0.2, and after ca 10.4.8 with a patch to bring it up 
to 4.0.3. On 10.5 and 10.6, they use 4.0.3 with some additional patches. 
Fink's package uses some of the patches that Apple introduced on 10.5 
, and this 
was a good idea while on 10.4. The patch for screen.c, which takes care 
of the bug we are talking about is not (and, on 10.4, cannot and does 
not need to be) applied, because it uses some private header that exists 
only for 10.5+.

After digging all this up, I think now that the right thing to do is to 
restrict screen's Distribution to 10.4.

-- 
Martin









--
Throughout its 18-year history, RSA Conference consistently attracts the
world's best and brightest in the field, creating opportunities for Conference
attendees to learn about information security's most important issues through
interactions with peers, luminaries and emerging and established companies.
http://p.sf.net/sfu/rsaconf-dev2dev
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] Please remove screen from fink

[Martin Costabel]

Michal Suchanek wrote:
[]
> Apple's version of screen works flawlessly.

In addition, the version that comes with MacOSX since 10.4 is the same 
as the one in Fink, minus the bug.

Therefore I second Michal's suggestion. This package has no reason to exist.

-- 
Martin




--
Throughout its 18-year history, RSA Conference consistently attracts the
world's best and brightest in the field, creating opportunities for Conference
attendees to learn about information security's most important issues through
interactions with peers, luminaries and emerging and established companies.
http://p.sf.net/sfu/rsaconf-dev2dev
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] boost1.35.systempython-1.35.0-5 failed to build

[Martin Costabel]

Hanspeter Niederstrasser wrote:

> I get the same error on 10.5/i386.  Previous revisions of 
> boost1.35.systempython (-2 and -3) built fine, but -5 crashes with the 
> "error: 'PyVarObject_HEAD_INIT' was not declared in this scope" errors 
> that Sjors also gets.

Like Alex, I didn't get the error on 10.5/i386, but I see now that that 
I could have got it, too. In fact, the "systempython" variant does not 
mean that the package uses necessarily the python from /System/Library. 
It means rather "use whatever python you can find", and this can be 
different on different machines.

This used to be OK, because boost does not really _use_ python, it only 
creates an interface. Unfortunately, this does not seem to be OK any 
longer, because the python developers have taken lessons from the 
freetype2 or gnome guys; they now routinely introduce API 
incompatibilities, even in micro-version updates. The boost package 
version 1.35.0-5 has a patch against such an incompatibility between 
python 2.6.2 and 2.6.3.

Anyway, I have now put an updated version 1.35.0-6 on the server which 
should solve the problem for the moment. It doesn't apply the patch for 
the systempython variant, so the build error will go away.

-- 
Martin


--
Throughout its 18-year history, RSA Conference consistently attracts the
world's best and brightest in the field, creating opportunities for Conference
attendees to learn about information security's most important issues through
interactions with peers, luminaries and emerging and established companies.
http://p.sf.net/sfu/rsaconf-dev2dev
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] atlas-3.9.11-1

[Martin Costabel]

Edoardo Milotti wrote:
> Hello, 
> 
> I'm having trouble installing ATLAS on a new MacPro with quad-core 
> Nehalem processor and Mac OS X 10.6.2. I attach the final installation 
> messages. I recently installed ATLAS on another Intel Mac (double core 
> MacBook Pro, same system version) and I had no problems. What can it 
> possibly be? 

[]
> make[7]: *** [dinstall_cpsc] Bus error

The crash log of this bus error might give some more information. Look 
for it with Console.app.

> ERROR 577 DURING CACHESIZE SEARCH!!.  CHECK INSTALL_LOG/Stage1.log FOR 
> DETAILS.

This file may contain some more information, too.

> mv error_Corei732SSE3.tar.gz error_Corei732SSE3.tgz
> Error report error_.tgz has been created in your top-level ATLAS
> directory. 

And this one.

-- 
Martin



--
This SF.Net email is sponsored by the Verizon Developer Community
Take advantage of Verizon's best-in-class app development support
A streamlined, 14 day to market process makes app distribution fast and easy
Join now and get one step closer to millions of Verizon customers
http://p.sf.net/sfu/verizon-dev2dev 
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] Bug in in mc install (10.5) and some others

[Martin Costabel]

Daniel Macks wrote:
[]nd is

> We obviously can't fix the original binary installer

Read this as: "We have an old tradition of not fixing binary 
installers." Nothing obvious there.

-- 
Martin




--
This SF.Net email is sponsored by the Verizon Developer Community
Take advantage of Verizon's best-in-class app development support
A streamlined, 14 day to market process makes app distribution fast and easy
Join now and get one step closer to millions of Verizon customers
http://p.sf.net/sfu/verizon-dev2dev 
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] Bug in in mc install (10.5) and some others

[Martin Costabel]

monipol wrote:

> On 07/01/2010, at 00:07, DJamé Seddah wrote:
[]
>> $ sudo fink cleanup
>> Information about 2516 packages read in 0 seconds.
>> Collecting active source filenames...
>> Can't locate Config.pm in @INC (@INC contains: /usr/local/lib/
>> perl5/5.8.8/darwin-2level /usr/local/lib/perl5/5.8.8 /usr/local/lib/
>> perl5/site_perl/5.8.8/darwin-2level /usr/local/lib/perl5/site_perl/
>> 5.8.8 /usr/local/lib/perl5/site_perl .).
> (...)
> That's odd, Djamé. Are you sourcing (as in . or source) /sw/bin/ 
> init.sh in your shell startup script? It looks like the PERL5LIB  
> variable is not being properly set, which /sw/bin/init.sh should've  
> done for you.

Maybe if you just run fink without sudo, it will work. "sudo" removes a 
couple of environment variables, among them PERL5LIB.

-- 
Martin

--
This SF.Net email is sponsored by the Verizon Developer Community
Take advantage of Verizon's best-in-class app development support
A streamlined, 14 day to market process makes app distribution fast and easy
Join now and get one step closer to millions of Verizon customers
http://p.sf.net/sfu/verizon-dev2dev 
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] Problem installing sbcl [Bug in dpkg?]

[Martin Costabel]

Daniel Macks wrote:
[]
> main/archives.c:tarobject() uses a static buffer to store a filename
> during installation. Wanna guess the size of fnamebuf[]? Try bumping
> it to something "much larger". Latest dpkg upstream still has this
> same hardcoded size.

This may even be legal, seeing that /usr/include/sys/syslimits.h has

#define NAME_MAX   255   /* max bytes in a file name */

I don't know if this number includes the path component or only the 
basename, though. The basename itself, even extended by "-dpkg-new", is 
only 244 characters.

As buffer overflows do not always lead to crashes, this would explain 
the sort of non-deterministic behavior of the package.

Anayway, I guess this should be treated as a bug in sbcl. It doesn't 
really need that outrageously named file, which only contains a URL 
redirection, anyway, and does not seem to be linked to by any of the 
other documentation files.

-- 
Martin



--
This SF.Net email is sponsored by the Verizon Developer Community
Take advantage of Verizon's best-in-class app development support
A streamlined, 14 day to market process makes app distribution fast and easy
Join now and get one step closer to millions of Verizon customers
http://p.sf.net/sfu/verizon-dev2dev 
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] Framework linker flag in Compile script

[Martin Costabel]

Ebrahim Mayat wrote:
[]
> CompileScript: <<
>   ./configure %c "LDFLAGS=-framework CoreFoundation" && make
> <<
> 
>   I still get the following complaint:
> 
> Validating .deb dir /sw/src/fink.build/root-fluidsynth-dev-1.1.1-280...
> Error: The -framework flag may get munged by libtool. See the gcc  
> manpage for information about passing multi-word options to flags for  
> specific compiler passes.
>  Offending file: /sw/lib/libfluidsynth.la
>  Offending line: inherited_linker_flags='  -framework  
> CoreFoundation'

I think what the validator wants is

   LDFLAGS=-Wl,framework,CoreFoundation

Like this, a too eagerly alphabetically-sorting libtool cannot destroy 
the flag.

-- 
Martin

--
This SF.Net email is sponsored by the Verizon Developer Community
Take advantage of Verizon's best-in-class app development support
A streamlined, 14 day to market process makes app distribution fast and easy
Join now and get one step closer to millions of Verizon customers
http://p.sf.net/sfu/verizon-dev2dev 
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] Problem installing sbcl [Bug in dpkg?]

[Martin Costabel]

Daniel Macks wrote:
[]
>> This looks like dpkg is not able to handle files with a file name longer 
>> than 256 characters. On the other hand, I don't understand why this bug 
>> only shows up on 10.6/32bit and not on 10.6/64bit nor on 10.5/32bit.
> 
> Also doesn't seem to choke on 10.4/ppc. Wonder if there is some static
> buffer used to hold a pathname? I have:
> 
> /usr/include/limits.h:#define   _POSIX_PATH_MAX 256
> /usr/include/sys/param.h:#defineMAXPATHLEN  PATH_MAX
> /usr/include/sys/syslimits.h:#definePATH_MAX 1024   /* 
> max bytes in pathname */
> 
> The "paste" subprocess is in dpkg-deb/extract.c:extracthalf(), where I
> can't decipher the forking and stream processing that is used to read
> the .deb. Try a simpler dpkg-deb modes to make sure this is where it's
> choking: "dpkg-deb -c foo.deb" vs "dpkg-deb -X foo.deb /tmp". I think
> the actual archive reading is done with a system(tar) call, so that
> command (and its platform variations, and different fink-installed
> versions) could be another place to look for the root cause:
> 
>   dpkg-deb --fsys-tarfile foo.deb > archive.tar
> 
> to extract the .deb filesystem archive and then try 'tar -xf
> archive.tar' with various different tar commands.

All these command-line things work correctly for me. No crash and no 
missing files.

Two more data points: When running "fink install sbcl" or dpkg -k, after 
the failure the file with the long name is actually installed. It is the 
last one that is installed. The next one would be 
Methods-common-to-all-streams.html, but it is not there. If one rune 
install again, then the long file exists twice, once with a .dpkg-tmp 
extension and once without. This should narrow down the point where the 
crash happens.

There is also a crash report:

> Exception Type:  EXC_CRASH (SIGABRT)
> Exception Codes: 0x, 0x
> Crashed Thread:  0  Dispatch queue: com.apple.main-thread
> 
> Application Specific Information:
> __abort() called
> 
> Thread 0 Crashed:  Dispatch queue: com.apple.main-thread
> 0   libSystem.B.dylib   0x90064732 __kill + 10
> 1   libSystem.B.dylib   0x90064724 kill$UNIX2003 + 32
> 2   libSystem.B.dylib   0x900f798d raise + 26
> 3   libSystem.B.dylib   0x9010d9d9 __abort + 124
> 4   libSystem.B.dylib   0x900f071c 
> release_file_streams_for_task + 0
> 5   dpkg0x8a6a tarobject + 6277
> 6   dpkg0x00020a71 TarExtractor + 759
> 7   dpkg0xbc4c process_archive + 7772
> 8   dpkg0x997a archivefiles + 1849
> 9   dpkg0x3748 main + 323

-- 
Martin


--
This SF.Net email is sponsored by the Verizon Developer Community
Take advantage of Verizon's best-in-class app development support
A streamlined, 14 day to market process makes app distribution fast and easy
Join now and get one step closer to millions of Verizon customers
http://p.sf.net/sfu/verizon-dev2dev 
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] Problem installing sbcl [Bug in dpkg?]

[Martin Costabel]

Martin Costabel wrote:
[]
> On 10.6/32bit: Builds but doesn't install
> 
>> Unpacking sbcl (from .../sbcl_1.0.34-1_darwin-i386.deb) ...
>> ### execution of /sw32/bin/dpkg-lockwait failed, exit code 6
>> /sw32/bin/dpkg-deb: subprocess paste killed by signal (Broken pipe)
>> Updating the list of locally available binary packages.
>> Scanning dists/local/main/binary-darwin-i386
>> New package: 
>> dists/local/main/binary-darwin-i386/sbcl_1.0.34-1_darwin-i386.deb
>> Failed: can't install package sbcl-1.0.34-1

I have now found where this chokes: It happens when dpkg tries to 
install the file

/sw/share/doc/sbcl/html/sbcl/Method-sb_002dbsd_002dsockets_003asocket_002dmake_002dstream-_0028_0028socket-socket_0029-_0026key-input-output-_0028element_002dtype-_0027character_0029-_0028buffering-full_0029-_0028external_002dformat-default_0029-timeout_0029.html

If I remove this file in sbcl's InstallScript, then sbcl (1.0.33 and 
1.0.34) installs OK on 10.6/32bit (at least sometimes, at other times it 
produces the sb-posix module build failure I have seen also on 
10.6/64bit occasionally).

This looks like dpkg is not able to handle files with a file name longer 
than 256 characters. On the other hand, I don't understand why this bug 
only shows up on 10.6/32bit and not on 10.6/64bit nor on 10.5/32bit.

-- 
Martin


--
This SF.Net email is sponsored by the Verizon Developer Community
Take advantage of Verizon's best-in-class app development support
A streamlined, 14 day to market process makes app distribution fast and easy
Join now and get one step closer to millions of Verizon customers
http://p.sf.net/sfu/verizon-dev2dev 
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] Problem installing sbcl

[Martin Costabel]

Jesse Alama wrote:
[]
> I just committed an updated SBCL (version 1.0.34) to my experimental 
> repository:
> 
>   http://fink.cvs.sourceforge.net/viewvc/fink/experimental/jessealama/
> 
> Can you please give this a try?  I didn't do anything new apart from bumping 
> the version, but perhaps this will make a difference.  I don't expect it 
> will, but it's worth a shot.

The result is the same as I got with the previuous version:

On 10.6/32bit: Builds but doesn't install

> Unpacking sbcl (from .../sbcl_1.0.34-1_darwin-i386.deb) ...
> ### execution of /sw32/bin/dpkg-lockwait failed, exit code 6
> /sw32/bin/dpkg-deb: subprocess paste killed by signal (Broken pipe)
> Updating the list of locally available binary packages.
> Scanning dists/local/main/binary-darwin-i386
> New package: dists/local/main/binary-darwin-i386/sbcl_1.0.34-1_darwin-i386.deb
> Failed: can't install package sbcl-1.0.34-1

On 10.6/64bit: It now builds and installs OK, but so did the previous 
version last time I tried. I am not yet trusting it.

-- 
Martin

--
This SF.Net email is sponsored by the Verizon Developer Community
Take advantage of Verizon's best-in-class app development support
A streamlined, 14 day to market process makes app distribution fast and easy
Join now and get one step closer to millions of Verizon customers
http://p.sf.net/sfu/verizon-dev2dev 
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] Problem installing sbcl

[Martin Costabel]

Martin Costabel wrote:
> Alexander Hansen wrote:
> []
>> I was told that the package functioned on 10.6/64-bit.  I didn't have
>> time actually to verify that on a clean 10.6 system at that time, but
>> verified that it works on 10.5; I applied an update to maxima concurrently.
> 
> There is something weird going on: I have two build logs from yesterday 
> on 10.6/64bit that end with failure, with the error message I mentioned. 
> Now I tried it one more time, and the building and installation went 
> through without error. Nothing has changed in the meantime, except that 
> I switched a couple of times between 32bit and 64bit.

Another weird observation:

Now I am trying "fink rebuild sbcl" on 10.6/32bit, where the previous 
build worked and only the installation of the resulting *.deb file did 
not go through (several times repeated).

It does not build any more: I am now getting the same build error as 
previously on 10.6/64bit:

> WARNING! Some of the contrib modules did not build successfully or pass
> their self-tests. Failed contribs:"
>   sb-posix
> ### execution of SBCL_ARCH="x86" failed, exit code 1

This thing does not seem to build deterministically.

-- 
Martin


--
This SF.Net email is sponsored by the Verizon Developer Community
Take advantage of Verizon's best-in-class app development support
A streamlined, 14 day to market process makes app distribution fast and easy
Join now and get one step closer to millions of Verizon customers
http://p.sf.net/sfu/verizon-dev2dev 
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] Problem installing sbcl

[Martin Costabel]

Alexander Hansen wrote:
[]
> I was told that the package functioned on 10.6/64-bit.  I didn't have
> time actually to verify that on a clean 10.6 system at that time, but
> verified that it works on 10.5; I applied an update to maxima concurrently.

There is something weird going on: I have two build logs from yesterday 
on 10.6/64bit that end with failure, with the error message I mentioned. 
Now I tried it one more time, and the building and installation went 
through without error. Nothing has changed in the meantime, except that 
I switched a couple of times between 32bit and 64bit.

-- 
Martin



--
This SF.Net email is sponsored by the Verizon Developer Community
Take advantage of Verizon's best-in-class app development support
A streamlined, 14 day to market process makes app distribution fast and easy
Join now and get one step closer to millions of Verizon customers
http://p.sf.net/sfu/verizon-dev2dev 
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] Problem installing sbcl

[Martin Costabel]

Dr. Wolfram Schroers wrote:
> Martin,
> 
> I have followed the instructions and removed the entries from 
> /sw/var/lib/dpkg/status. There were no files with sbcl in their names in the 
> /sw/var/lib/dpkg/info subdirectory. Finally, I ran "fink scanpackages" and 
> "sudo apt-get update".
> 
> When trying to install sbcl again, I got an error message again. 

I can now confirm all the bad things recently said about sbcl on Snow 
Leopard. This package is really badly broken. Unfortunately the error 
messages it produces were not sufficient to make me see the origin of 
the errors. I am CCing the maintainer.

On 10.6/64bit, the build does not complete. The error message is:

> WARNING! Some of the contrib modules did not build successfully or pass
> their self-tests. Failed contribs:"
>   sb-posix
> ### execution of sh failed, exit code 1
> Removing runtime build-lock...
> Removing build-lock package...
> /sw64/bin/dpkg-lockwait -r fink-buildlock-sbcl-1.0.33-1
> (Reading database ... 131593 files and directories currently installed.)
> Removing fink-buildlock-sbcl-1.0.33-1 ...
> Failed: phase compiling: sbcl-1.0.33-1 failed

I don't know where the sh script starts that is reportedly failing. The 
code before the error message looks like a very long list of lisp 
selftests to me.

On 10.6/32bit, the building of the package succeeds, but the *.deb file 
does not install, neither via fink nor via dpkg or apt-get. This is the 
same error Wolfram has been reporting:

> Selecting previously deselected package sbcl.
> (Reading database ... 126383 files and directories currently installed.)
> Unpacking sbcl (from .../sbcl_1.0.33-1_darwin-i386.deb) ...
> /sw32/bin/dpkg-deb: subprocess paste killed by signal (Broken pipe)
> ### execution of /sw32/bin/dpkg-lockwait failed, exit code 6

The package is then "half-installed":

> % dpkg -l sbcl
> Desired=Unknown/Install/Remove/Purge/Hold
> | Status=Not/Installed/Config-files/Unpacked/Failed-config/Half-installed
> |/ Err?=(none)/Hold/Reinst-required/X=both-problems (Status,Err: 
> uppercase=bad)
> ||/ NameVersion Description
> +++-===-===-==
> iHR sbcl1.0.33-1

I have never seen this before. The files of the package are, in reality, 
  almost all installed:

>  ls -R /sw/lib/sbcl/ |wc
>  481 4608361

Likewise the files in /sw/share/doc/sbcl, but not the files 
/sw/share/info/{asdl,sbcl*}.info
Also, /sw/var/lib/dpkg/info does not contain any sbcl* files, and 
/sw/var/lib/dpkg/status contains

> Package: sbcl
> Status: install reinstreq half-installed
> Priority: optional
> Section: languages
> Version: 1.0.33-1

On can purge it:

>  sudo dpkg --purge --force-all sbcl
> dpkg - warning, overriding problem because --force enabled:
>  Package is in a very bad inconsistent state - you should
>  reinstall it before attempting a removal.
> (Reading database ... 
> dpkg: serious warning: files list file for package `sbcl' missing, assuming 
> package has no files currently installed.
> 126383 files and directories currently installed.)
> Removing sbcl ...

But this does actually not remove any files. It changes the status to 
"purged", but the installed files are still there, now orphaned. Not nice.

-- 
Martin






--
This SF.Net email is sponsored by the Verizon Developer Community
Take advantage of Verizon's best-in-class app development support
A streamlined, 14 day to market process makes app distribution fast and easy
Join now and get one step closer to millions of Verizon customers
http://p.sf.net/sfu/verizon-dev2dev 
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] Problem installing sbcl

[Martin Costabel]

Dr. Wolfram Schroers wrote:
[]
> Error message from "sudo apt-get remove sbcl":
[]
> /sw/bin/dpkg: error processing sbcl (--remove):
>  Package is in a very bad inconsistent state - you should
>  reinstall it before attempting a removal.

In such a situation, I found that the only solution is to manually erase 
the package from dpkg's memory:

1. Go to the directory /sw/var/lib/dpkg

2. Then copy the file "status" to "status.bak" to have a backup copy if 
anything goes wrong. There is already a backup copy called "status-old" 
which may be a little older.

3. Then edit "status" using a text editor that does not destroy text 
files (vi or pico etc., not emacs or TextEdit). The file contains little 
paragraphs for each package. Go to the paragraph starting with the line

Package: sbcl

and erase it (the whole paragraph, around 23 lines).

Repeat this for any other paragraph that has a "Package:" line with 
"sbcl" in it. Do not touch paragraphs that have "sbcl" in some other 
line. Save the edited file "status".

4. Finally go to the "info" subdirectory and remove any files with 
"sbcl" in their names, like "sbcl.list", "sbcl.postinst", "sbcl.prerm".

This should do it. Run "fink scanpackages" and "sudo apt-get update", 
and then resume whatever you wanted to do.

-- 
Martin

--
This SF.Net email is sponsored by the Verizon Developer Community
Take advantage of Verizon's best-in-class app development support
A streamlined, 14 day to market process makes app distribution fast and easy
Join now and get one step closer to millions of Verizon customers
http://p.sf.net/sfu/verizon-dev2dev 
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] fink update-all error

[Martin Costabel]

Eunjung Ko wrote:
> Dear Martin,
> 
> What I just did is "fink update-all". I did not intend to do 
> install transfig-3.2.4-7
> at all. However, "fink update-all" command made it. Please help me. 

Run

  fink remove transfig

-- 
Martin

--
This SF.Net email is sponsored by the Verizon Developer Community
Take advantage of Verizon's best-in-class app development support
A streamlined, 14 day to market process makes app distribution fast and easy
Join now and get one step closer to millions of Verizon customers
http://p.sf.net/sfu/verizon-dev2dev 
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] fink update-all error

[Martin Costabel]

Eunjung Ko wrote:
> Dear maintainer,
> 
> According to the FAQ entry, I tried to do as follows before I build.
> 
> sudo mv /usr/local /usr/local.moved
> 
> But I got another error messages as belows.
> 
[]
> /sw/bin/dpkg-lockwait -i 
> /sw/fink/dists/stable/main/binary-darwin-powerpc/graphics/transfig_3.2.4-7_darwin-powerpc.deb
>  
> (Reading database ... 92661 files and directories currently installed.) 
> Preparing to replace transfig 3.2.4-6 (using 
> .../transfig_3.2.4-7_darwin-powerpc.deb) ... Unpacking replacement 
> transfig ... /sw/bin/dpkg: error processing 
> /sw/fink/dists/stable/main/binary-darwin-powerpc/graphics/transfig_3.2.4-7_darwin-powerpc.deb
>  
> (--install): trying to overwrite `/sw/bin/fig2dev', which is also in 
> package transfig-graphicx 

Why do you want to install transfig-3.2.4-7 when you already have 
transfig-graphicx installed which is newer?

-- 
Martin

--
This SF.Net email is sponsored by the Verizon Developer Community
Take advantage of Verizon's best-in-class app development support
A streamlined, 14 day to market process makes app distribution fast and easy
Join now and get one step closer to millions of Verizon customers
http://p.sf.net/sfu/verizon-dev2dev 
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] tk-pm588-804.026-14

[Martin Costabel]

(Bouncing this back to the devel list. I have no idea ATM. But this
looks indeed like a fundamental flaw in Fink's tk perl module for
perl-5.8.8 on 10.6/32bit).

--
Martin

Thomas Proffen wrote:
> Here it is:
> 
> Process: perl5.8.8 [323]
> Path:/sw/bin/perl5.8.8
> Identifier:  perl5.8.8
> Version: ??? (???)
> Code Type:   X86 (Native)
> Parent Process:  bash [292]
> 
> Date/Time:   2009-12-21 02:46:20.767 -0700
> OS Version:  Mac OS X 10.6.2 (10C540)
> Report Version:  6
> 
> Exception Type:  EXC_BAD_ACCESS (SIGBUS)
> Exception Codes: KERN_PROTECTION_FAILURE at 0x000a
> Crashed Thread:  0  Dispatch queue: com.apple.main-thread
> 
> Thread 0 Crashed:  Dispatch queue: com.apple.main-thread
> 0   Tk.bundle   0x0056929a InitVtabs + 90
> 1   Tk.bundle   0x0056a729 Boot_Glue + 137
> 2   Tk.bundle   0x00560029 boot_Tk + 4649
> 3   perl5.8.8   0x00082121 Perl_pp_entersub + 929
> 4   perl5.8.8   0x00078d99 Perl_runops_standard + 25
> 5   perl5.8.8   0x5701 Perl_call_sv + 1601
> 6   perl5.8.8   0x5ad4 Perl_call_list + 628
> 7   perl5.8.8   0x000444dc Perl_newATTRSUB + 3868
> 8   perl5.8.8   0x00044f47 Perl_utilize + 535
> 9   perl5.8.8   0x0003095e Perl_yyparse + 2734
> 10  perl5.8.8   0xa005 S_parse_body + 3301
> 11  perl5.8.8   0xb945 perl_parse + 1125
> 12  perl5.8.8   0x266d main + 189
> 13  perl5.8.8   0x2565 start + 53
> 
> Thread 0 crashed with X86 Thread State (32-bit):
>   eax: 0x  ebx: 0x0056924b  ecx: 0x00610db8  edx: 0x
>   edi: 0xa0713500  esi: 0x00610db8  ebp: 0xbfffeb88  esp: 0xbfffeb60
>ss: 0x001f  efl: 0x00010246  eip: 0x0056929a   cs: 0x0017
>ds: 0x001f   es: 0x001f   fs: 0x   gs: 0x0037
>   cr2: 0x000a
> 
> Binary Images:
> 0x1000 -   0x11bff7 +perl5.8.8 ??? (???) 
> <88086649-BB3B-83C7-4278-CE5F1ECA9CA0> /sw/bin/perl5.8.8
>   0x143000 -   0x147ff7  libutil.dylib ??? (???) 
> <61FEDBC2-F98A-F343-BFB6-1F9C151F70DD> /usr/lib/libutil.dylib
>   0x16 -   0x161ff7 +Cwd.bundle ??? (???) 
>  
> /sw/lib/perl5-core/5.8.8/darwin-thread-multi-2level/auto/Cwd/Cwd.bundle
>   0x165000 -   0x16bfe7 +Encode.bundle ??? (???) 
> <6AB523C0-775D-C208-7E6E-06169F60EAA8> 
> /sw/lib/perl5-core/5.8.8/darwin-thread-multi-2level/auto/Encode/Encode.bundle
>   0x17 -   0x171ff7 +libXau.6.dylib ??? (???) 
>  /usr/X11/lib/libXau.6.dylib
>   0x175000 -   0x177fe7 +libXdmcp.6.dylib ??? (???) 
> <64BACC61-FFEA-FC73-5017-6F7483D8F129> /usr/X11/lib/libXdmcp.6.dylib
>   0x30 -   0x423fe7 +Event.bundle ??? (???) 
> <8532AC2C-1283-5B75-0225-9D1601B35093> 
> /sw/lib/perl5/5.8.8/darwin-thread-multi-2level/auto/Tk/Event/Event.bundle
>   0x448000 -   0x61aff7 +Tk.bundle ??? (???) 
> <42C2CF0F-E1DF-6F93-5BE3-1D499AE09D13> 
> /sw/lib/perl5/5.8.8/darwin-thread-multi-2level/auto/Tk/Tk.bundle
>   0x65b000 -   0x74fff7 +libX11.6.dylib ??? (???) 
> <99D5825E-7F59-2826-B730-3AAB596056D7> /usr/X11/lib/libX11.6.dylib
> 0x8fe0 - 0x8fe4162b  dyld 132.1 (???) 
> <211AF0DD-42D9-79C8-BB6A-1F4BEEF4B4AB> /usr/lib/dyld
> 0x91c3f000 - 0x91c42fe7  libmathCommon.A.dylib ??? (???) 
> <1622A54F-1A98-2CBE-B6A4-2122981A500E> /usr/lib/system/libmathCommon.A.dylib
> 0x93f5c000 - 0x93fc6fe7  libstdc++.6.dylib ??? (???) 
> <411D87F4-B7E1-44EB-F201-F8B4F9227213> /usr/lib/libstdc++.6.dylib
> 0x95bfb000 - 0x95d9ffeb  libSystem.B.dylib ??? (???) 
>  /usr/lib/libSystem.B.dylib
> 0x - 0x1fff  libSystem.B.dylib ??? (???) 
>  /usr/lib/libSystem.B.dylib
> 
> On Dec 21, 2009, at 2:43 AM, Martin Costabel wrote:
> 
>> Thomas Proffen wrote:
>>> Feedback: not good
>>> I lost the prior emails. I still have trouble to get perl-Tk to work. I 
>>> reinstalled and made sure it is running with perl 5.8.8, e.g.
>>> cheetah249:PDFgetN> /sw/bin/perl5.8.8 PDFgetN.pl Bus error
>> What does the crash log say? Look for it with Console.app (in 
>> /Applications/Utilities).
>>
>> -- 
>> Martin
---
Orange vous informe que cet  e-mail a ete controle par l'anti-virus mail.
Aucun virus connu a ce jour par nos services n'a ete detecte.




--
This SF.Net emai

Re: [Fink-devel] tk-pm588-804.026-14

[Martin Costabel]

Thomas Proffen wrote:
> Feedback: not good
> 
> I lost the prior emails. I still have trouble to get perl-Tk to work. I 
> reinstalled and made sure it is running with perl 5.8.8, e.g.
> 
> cheetah249:PDFgetN> /sw/bin/perl5.8.8 PDFgetN.pl 
> Bus error

What does the crash log say? Look for it with Console.app (in 
/Applications/Utilities).

-- 
Martin

--
This SF.Net email is sponsored by the Verizon Developer Community
Take advantage of Verizon's best-in-class app development support
A streamlined, 14 day to market process makes app distribution fast and easy
Join now and get one step closer to millions of Verizon customers
http://p.sf.net/sfu/verizon-dev2dev 
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] Apt cache limit

[Martin Costabel]

Martin Costabel wrote:
> Alexander Hansen wrote:
> []
>> Ah.  I got the same thing, myself.  Fortunately in my case the offending
>> package was a buildlock, so I just deleted it.
> 
> I am curious how long it will take you to see it again. I don't think 
> this can be solved by removing one package. It is like when your 
> harddisk is full: You can delete a file to buy a little time, but the 
> problem will soon come back.

I think the files to watch are /sw/var/cache/apt/(src,}pkgcache.bin. If 
they get bigger than 6MB (= 6291456 Bytes), you are in trouble. If the 
size is just below, you will be in trouble soon.

Adding the unofficial tokyo bindist (for 10.5) to 
/sw/etc/apt/sources.list adds a little over 4MB to the cache files.

-- 
Martin



--
Join us December 9, 2009 for the Red Hat Virtual Experience,
a free event focused on virtualization and cloud computing. 
Attend in-depth sessions from your desk. Your couch. Anywhere.
http://p.sf.net/sfu/redhat-sfdev2dev
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] Apt cache limit

[Martin Costabel]

Alexander Hansen wrote:
[]
> Ah.  I got the same thing, myself.  Fortunately in my case the offending
> package was a buildlock, so I just deleted it.

I am curious how long it will take you to see it again. I don't think 
this can be solved by removing one package. It is like when your 
harddisk is full: You can delete a file to buy a little time, but the 
problem will soon come back.

-- 
Martin

--
Join us December 9, 2009 for the Red Hat Virtual Experience,
a free event focused on virtualization and cloud computing. 
Attend in-depth sessions from your desk. Your couch. Anywhere.
http://p.sf.net/sfu/redhat-sfdev2dev
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

[Fink-devel] Apt cache limit

[Martin Costabel]

Recently, apt-get started crashing on me with the following error message:

E: Dynamic MMap ran out of room
E: Error occured while processing system-pkgconfig-xcb-property (NewPackage)
E: Problem with MergeList /tmp/finkaptstatus

The incriminated package does not matter, during my tests it was not 
always the same, but the error came up persistently, whenever anything 
apt-get related was run, apt-get or apt-cache directly, or fink 
scanpackages or selfupdate etc.

After some googling, it is clear that this is a real problem related to 
the increasing size of the package data bases and perhaps to the number 
of repositories mentioned in sources.list. On google I also found a 
workaround which might be interesting for others, as I am sure others 
will be hit by this, too:

Create a file /sw/etc/apt/apt.conf.d/30cache containing one line

   APT::Cache-Limit "125829120";

This sets the limit to 120MB instead of the ridiculously low default 
6MB. Even 12MB would suffice.

A permanent fix would involve a patch to the file apt-pkg/pkgcachegen.cc 
in the apt sources as follows:

--- apt-0.5.4/apt-pkg/pkgcachegen.cc~   2009-12-02 14:16:11.0 +0100
+++ apt-0.5.4/apt-pkg/pkgcachegen.cc2009-12-02 14:18:17.0 +0100
@@ -599,3 +599,3 @@
  {
-   unsigned long MapSize = _config->FindI("APT::Cache-Limit",6*1024*1024);
+   unsigned long MapSize = _config->FindI("APT::Cache-Limit",60*1024*1024);


There is another line in that file where MapSize is even lower at 4MB, 
but that does not seem to matter as much.

-- 
Martin

--
Join us December 9, 2009 for the Red Hat Virtual Experience,
a free event focused on virtualization and cloud computing. 
Attend in-depth sessions from your desk. Your couch. Anywhere.
http://p.sf.net/sfu/redhat-sfdev2dev
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] New libgettext8 package

[Martin Costabel]

Koen van der Drift wrote:
[]
> I didn't see a Distribution flag in the info file.

Not only is there no Distribution field, it also is built and installed 
automatically by selfupdate, no questions asked (well, it does actually 
ask permission, but declining would mean stopping selfupdate).

The reason is that libgettext3-shlibs, which is essential and got a new 
Revision, is automatically updated; it Depends on "gettext-bin (>= 
%v-%r)", hence gettext-bin will be updated, too, and in this case this 
means to the latest version gettext-bin-0.17-17, which builds and 
installs libgettext8-shlibs.

You could as well have made it essential right away.

-- 
Martin





--
Join us December 9, 2009 for the Red Hat Virtual Experience,
a free event focused on virtualization and cloud computing. 
Attend in-depth sessions from your desk. Your couch. Anywhere.
http://p.sf.net/sfu/redhat-sfdev2dev
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] ncbitools build failure

[Martin Costabel]

Koen van der Drift wrote:
> Hi Richard,
> 
> Ncbitools package fails as follows on 10.5.8, unstable. 

Looking at the cvs log, it semes to me that the maintainer has last 
touched this sometime in 2003. You are probably the first who has ever 
tried to build it on 10.5 or newer. If you need it, it would probably be 
a good idea to upgrade it to a more recent version. They have source 
tarballs from 2009, as opposed to the one from 2003 that Fink is using.

-- 
Martin


--
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day 
trial. Simplify your report design, integration and deployment - and focus on 
what you do best, core application coding. Discover what's new with
Crystal Reports now.  http://p.sf.net/sfu/bobj-july
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] WXGTK Flag Issue

[Martin Costabel]

ピエールランリ・ラヴィン wrote:
> Good day,
> 
> Back to devel list :-) I'm working on wxgtk version 2.8.10. After
> talking on IRC, it seems i need to patch it to remove the
> "-single_module -Wl,-x -dead_strip" flags. I don't know where it does
> come from in the source code. 

Perhaps from the line

  perl -pi -e 's/single_module/$& -Wl,-x -dead_strip/'
build/aclocal/bakefile.m4 configure

in the PatchScript?

-- 
Martin

--
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day 
trial. Simplify your report design, integration and deployment - and focus on 
what you do best, core application coding. Discover what's new with
Crystal Reports now.  http://p.sf.net/sfu/bobj-july
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] opensp4-1.5.1-1007

[Martin Costabel]

Edoardo Milotti wrote:
[]
> grep: /usr/lib/libiconv.la: No such file or directory

The recipe in this case is the following:

1. Find the file that contains the reference to the non-existing 
/usr/lib/libiconv.la

2. Destroy it. Or, at least, replace its "dependency_libs" line by

   dependency_libs=''

Candidates for 1. (in order of probability) are:

- /sw/lib/fontconfig2/lib/libfontconfig.fink.la
- Any file that answers to /usr/X11/lib/lib*.la
- Any other file named *.la in /sw/lib or in /usr/local/lib or any of 
their subdirectories.

-- 
Martin





--
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day 
trial. Simplify your report design, integration and deployment - and focus on 
what you do best, core application coding. Discover what's new with
Crystal Reports now.  http://p.sf.net/sfu/bobj-july
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] [Fink-beginners] Missing library, libXpm, for swi-prolog in 64-bit Fink

[Martin Costabel]

Brent E. Edwards wrote:
[]
> checking for X... no
> checking for main in -lSM... no
> checking for main in -lXext... no
> checking for main in -lXinerama... no
> checking X11/extensions/Xinerama.h usability... yes
> checking X11/extensions/Xinerama.h presence... yes
> checking for X11/extensions/Xinerama.h... yes
>   XLIBS=-lXt -lX11 
> checking for main in -ljpeg... no
> checking for main in -lXpm... no

It doesn't find X11 without help, as is the case for all autotools-based 
software on recent Apple (or xquartz) X11 versions. I have now applied 
the usual fix, so that your next selfupdate should bring you a version 
that builds.

(Since this is cross-posted to fink-devel, I will not swallow my anger 
about having to do this for the hundredth time: Hiding Fink's xmkmf in a 
subdirectory instead of placing it in %p/bin where it can be found was 
really a bad decision.)

-- 
Martin

--
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day 
trial. Simplify your report design, integration and deployment - and focus on 
what you do best, core application coding. Discover what's new with
Crystal Reports now.  http://p.sf.net/sfu/bobj-july
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] tk file dialogs broken?

[Martin Costabel]

Jack Howarth wrote:
[]
> ImportError: dlopen(/sw/lib/pymol-py25/modules/pymol/_cmd.so, 2): Symbol not 
> found: __cg_jpeg_CreateCompress
>   Referenced from: 
> /System/Library/Frameworks/ApplicationServices.framework/Versions/A/Frameworks/ImageIO.framework/Versions/A/ImageIO
>   Expected in: /sw/lib/libJPEG.dylib
> 
> I find this last error rather strange. 

It is a classical error that appears whenever DYLD_LIBRARY_PATH is set 
to /sw/lib and the ApplicationServices framework is invoked. Instead of 
/System/Library/Frameworks/ApplicationServices.framework/Versions/A/Frameworks/ImageIO.framework/Versions/A/Resources/libJPEG.dylib,
 
the linker then loads /sw/lib/libJPEG.dylib (which doea not exist, but 
on a case-insensitive file system, /sw/lib/libjpeg.dylib will do) and 
finds to its surprise that it does not contain the right symbols.

If things are compiled correctly, DYLD_LIBRARY_PATH and LD_LIBRARY_PATH 
are *never* necessary.

-- 
Martin





--
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day 
trial. Simplify your report design, integration and deployment - and focus on 
what you do best, core application coding. Discover what's new with
Crystal Reports now.  http://p.sf.net/sfu/bobj-july
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] [Fink-beginners] kdegraphics3-base installation fails

[Martin Costabel]

Alexander Hansen wrote:

> Umut Yildiz wrote:
>> Hi,
>>
>> I am trying to install kile with fink but I could not succeed at the
>> last step. It is giving me the following error. What can I do to fix this?
[]
>> 
>> checking for rpath... yes
>> checking for KDE... configure: error:
>> in the prefix, you've chosen, are no KDE libraries installed. This will
>> fail.
[]
> Check
> 
> /sw/src/fink.build/kdegraphics3-3.5.10-1/kdegraphics-3.5.10/config.log
> 
> and look for where it says "checking for KDE".  I have the following:
> 
> configure:31714: checking for KDE
> configure: 31767: /sw/include/ksharedptr.h
> taking that
> configure: 31797: /sw/lib/libkio.la
> taking that
> configure: 31815: /sw/lib/kde3/plugins/designer/kdewidgets.la
> taking that
> configure:31888: result: libraries /sw/lib, headers /sw/include

This is worth keeping for the record, as the first documented case where 
*.la files are necessary and where the advice to remove all *.la files 
breaks things.

Now why that configure script is doing this, and if it could be asked to 
cease and desist, is another question. It takes its cue from 
admin/acinclude.m4.in, a general kde thing, it seems.

-- 
Martin



--
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day 
trial. Simplify your report design, integration and deployment - and focus on 
what you do best, core application coding. Discover what's new with
Crystal Reports now.  http://p.sf.net/sfu/bobj-july
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] Old cruft in /sw/fink/update/ not 64bit aware

[Martin Costabel]

Max Horn wrote:
[]
> (1) Get rid of these UpdateFOO fields completely. Instead, require  
> package which have to update config.guess etc. to insert these updates  
> some other way: By rerunning auto-tools; by adding the correct  
> config.guess etc. version as a SourceN (care required to avoid name  
> clashes, though), as part of the patch (would lead to pretty big  
> patches, though) etc.
[]

> Personally, I prefer approach (1).

Ther is a practical problem with this approach: It requires real 
maintainer work, not just cosmetics. But if you look at the packages 
that use Update fields (a quick grep shows more than a hundred with 
UpdateConfigGuess and more than 50 with UpdateLibtool and even a dozen 
with UpdateLibtoolInDirs), you see that most of them are very old and 
haven't really been maintained for a long time. It will be difficult to 
do more than cosmetic changes in all of these.

-- 
Martin

--
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day 
trial. Simplify your report design, integration and deployment - and focus on 
what you do best, core application coding. Discover what's new with
Crystal Reports now.  http://p.sf.net/sfu/bobj-july
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] Old cruft in /sw/fink/update/ not 64bit aware

[Martin Costabel]

Peter O'Gorman wrote:
[]
> Feel free to change it in fink's ltconfig, etc., and the problem should 
> be "fixed" with the next fink release, but really I think that the 
> 'Update*' fields should be deprecated, in favor of rerunning some 
> version of autotools.

It is true that both packages gtk+ and recode are ancient and have had 
the 'Update' lines since the days of MacOSX 10.2 (maybe earlier, I 
didn't look further back). Apparently this didn't hurt until now (and 
even on 10.6/64bit you don't notice the breakage in the packages 
themselves, only when something wants to dlopen a binary linked to one 
of their dylibs.)

On 10.6/64bit, the 'Update' fields could have a renewed interest, 
because many packages have config.guess versions that guess wrong. In 
fact, does there even exist a config.guess that gives the right answer 
"x84_64-apple-darwin10"? Apple's own version in 
/usr/share/libtool/config/ doesn't.

-- 
Martin

--
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day 
trial. Simplify your report design, integration and deployment - and focus on 
what you do best, core application coding. Discover what's new with
Crystal Reports now.  http://p.sf.net/sfu/bobj-july
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

[Fink-devel] Old cruft in /sw/fink/update/ not 64bit aware

[Martin Costabel]

In the 2 threads "lablgtk 1.2.7-1002 on 10.6 64bit fails to build" and 
"python-bibtex-py26-1.2.4-1 build failed on 10.6 64bit" the error was 
caused by dylibs not linking with libintl.dylib, 
/sw/lib/libgtk-1.2.0.dylib in the first case and 
/sw/lib/librecode.0.dylib in the second case.

Both dylibs do link with libintl on 10.5/32bit (probably on 10.6/32bit, 
too).

The packages gtk+ and recode have in common that they use

UpdateConfigGuess: true
UpdateLibtool: true

This brings in particular the extremely old /sw/lib/fink/update/ltconfig 
which defines

   deplibs_check_method='file_magic Mach-O dynamically linked shared 
library'

In reality, the 64bit dylibs have "Mach-O 64-bit dynamically linked 
shared library".

The result is that libtool does not recognize that /sw/lib/libintl.dylib 
is a dynamic library, and it removes it from the linker list, 
accompanying this deed with the lie

*** Warning: This library needs some functionality provided by -lintl.
*** I have the capability to make that library automatically link in when
*** you link to this library.  But I can only do this if you have a
*** shared version of the library, which you do not appear to have.

Is there a reason why we have these old versions of ltconfig, ltmain.sh 
etc in fink, versions that date from before the release of MacOSX 10.0?

-- 
Martin

--
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day 
trial. Simplify your report design, integration and deployment - and focus on 
what you do best, core application coding. Discover what's new with
Crystal Reports now.  http://p.sf.net/sfu/bobj-july
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

[Fink-devel] validator request

[Martin Costabel]

"fink validate" has the following error message:

> Error: use of PatchFile requires declaring a BuildDepends on fink (>= 
> 0.24.12) or higher. (foo.info)

Could this please be removed from the validator? I can't see a situation 
where this would be useful. No contemporary Fink installation will work 
with fink versions as old as 0.24.x anyway.

-- 
Martin


--
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day 
trial. Simplify your report design, integration and deployment - and focus on 
what you do best, core application coding. Discover what's new with
Crystal Reports now.  http://p.sf.net/sfu/bobj-july
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] FinkCommander 0.5.5 (huh?)

[Martin Costabel]

monipol wrote:
> On 04/11/2009, at 20:09, Martin Costabel wrote:
>> monipol wrote:
>> []
>>> Here's the download URL:
>>> https://sourceforge.net/projects/finkcommander/files/FinkCommander/0.5.5/FinkCommander_0.5.5.dmg/download
>> Where are the sources for this release?
> 
> I committed the patches about half an hour ago, Martin. They're on  
> FinkCommander's CVS repository at SourceForge.net.

I know FinkCommander hasn't done this recently, but it would be nice to 
have a source tarball like most other open source software projects have.

-- 
Martin


--
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day 
trial. Simplify your report design, integration and deployment - and focus on 
what you do best, core application coding. Discover what's new with
Crystal Reports now.  http://p.sf.net/sfu/bobj-july
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] FinkCommander 0.5.5 (huh?)

[Martin Costabel]

monipol wrote:
[]
> Here's the download URL:
> 
> https://sourceforge.net/projects/finkcommander/files/FinkCommander/0.5.5/FinkCommander_0.5.5.dmg/download

Where are the sources for this release?

-- 
Martin

--
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day 
trial. Simplify your report design, integration and deployment - and focus on 
what you do best, core application coding. Discover what's new with
Crystal Reports now.  http://p.sf.net/sfu/bobj-july
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] libgegl0.0.0-shlibs-0.0.22-1 fail: 10.6.1 64-bit

[Martin Costabel]

William G. Scott wrote:
> 
> On Oct 30, 2009, at 11:32 PM, Martin Costabel wrote:
[]
>> "Apparently the build system still thinks it's building for i386."
>> (Maintainer CCed).
>> The standard autotools-made configure scripts guess the build host  
>> system wrong when building for x86_64. Anything cpu-specific will  
>> then fail.
>>
>> -- 
>> Martin
> 
> Thanks, Martin. (I tried the usual things like -m64 and even forcing  
> use of the fink 64-bit compilers, but no joy.)

I am away from from my 10.6 system, so I cannot check, but you could 
look what config.log says about "checking build system type" and 
"checking host system type", and what config.h says about "ARCH_X86_64".

Perhaps there need to be configure parameters like

  --build=%m-apple-darwin`uname -r|cut -f1 -d.` 
--host=%m-apple-darwin`uname -r|cut -f1 -d.`

as was successfully used in other packages.

-- 
Martin




--
Come build with us! The BlackBerry(R) Developer Conference in SF, CA
is the only developer event you need to attend this year. Jumpstart your
developing skills, take BlackBerry mobile applications to market and stay 
ahead of the curve. Join us from November 9 - 12, 2009. Register now!
http://p.sf.net/sfu/devconference
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] libgegl0.0.0-shlibs-0.0.22-1 fail: 10.6.1 64-bit

[Martin Costabel]

William G. Scott wrote:
> Hi Folks:
> 
> The only thing left standing between me and gimp2 on 10.6 is  
> libgegl0.0.0-shlibs-0.0.22-1.  It fails thusly:
> 
> 
> libtool: compile:  gcc -DHAVE_CONFIG_H -I. -I.. -I.. -I../gegl/buffer - 
> I../gegl/property-types -I../gegl/operation -I/sw/include/glib-2.0 -I/ 
> sw/lib/glib-2.0/include -I/sw/include/babl-0.0 -DLIBDIR=\"/sw/lib\" - 
> DGEGL_LOCALEDIR=\"/sw/share/locale\" -I/sw/lib/fontconfig2/include -I/ 
> sw/include -Os -Wall -Wdeclaration-after-statement -Wmissing- 
> prototypes -Wmissing-declarations -Winit-self -Wpointer-arith -Wold- 
> style-definition -mmmx -msse -c gegl-cpuaccel.c  -fno-common -DPIC - 
> o .libs/gegl-cpuaccel.o
> /var/tmp//ccIVYf6E.s:76:suffix or operands invalid for `pushf'
> /var/tmp//ccIVYf6E.s:77:suffix or operands invalid for `pushf'
> /var/tmp//ccIVYf6E.s:78:suffix or operands invalid for `pop'
> /var/tmp//ccIVYf6E.s:81:suffix or operands invalid for `push'
> /var/tmp//ccIVYf6E.s:82:suffix or operands invalid for `popf'
> /var/tmp//ccIVYf6E.s:83:suffix or operands invalid for `pushf'
> /var/tmp//ccIVYf6E.s:84:suffix or operands invalid for `pop'
> /var/tmp//ccIVYf6E.s:85:suffix or operands invalid for `popf'
> make[3]: *** [gegl-cpuaccel.lo] Error 1
> 
> Anyone know of a fix?

 From a message 2 weeks ago on the same subject, on the fink-beginners list:
"Apparently the build system still thinks it's building for i386."
(Maintainer CCed).
The standard autotools-made configure scripts guess the build host 
system wrong when building for x86_64. Anything cpu-specific will then fail.

-- 
Martin




--
Come build with us! The BlackBerry(R) Developer Conference in SF, CA
is the only developer event you need to attend this year. Jumpstart your
developing skills, take BlackBerry mobile applications to market and stay 
ahead of the curve. Join us from November 9 - 12, 2009. Register now!
http://p.sf.net/sfu/devconference
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] tk-pm588-804.026-14

[Martin Costabel]

Thomas Proffen wrote:
> Results in a bus error when running a perl script using Tk ..

Are you trying to use Tk-pm588 with perl-5.10? This won't work.

--
Martin

--
Come build with us! The BlackBerry(R) Developer Conference in SF, CA
is the only developer event you need to attend this year. Jumpstart your
developing skills, take BlackBerry mobile applications to market and stay 
ahead of the curve. Join us from November 9 - 12, 2009. Register now!
http://p.sf.net/sfu/devconference
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] Server issues: finkproject.org aliases and Subversion are down

[Martin Costabel]

monipol wrote:
[]
>  From a user perspective the situation should have normalised by now:  
> PDB and www.finkproject.org are running on the new machine,  
> finkmirrors.net name servers are working so rsync selfupdate,  
> distfiles (source files) and bindist (official binaries) mirror names  
> are being properly resolved unless the DNS servers being used by some  
> users haven't updated yet. Many thanks to Andreas Gockel for his  
> invaluable help with these services.

The following rsync mirrors are working, but stuck on 23 oct 2009:

rsync://distfiles.master.finkmirrors.net/finkinfo/ :
-rw-r--r--  11 2009/10/23 06:15:14 TIMESTAMP
rsync://sjc.ca.us.finkmirrors.net/finkinfo/ :
-rw-r--r--  11 2009/10/23 06:15:14 TIMESTAMP
rsync://dub.ie.eu.finkmirrors.net/finkinfo/ :
-rw-r--r--  11 2009/10/23 06:15:14 TIMESTAMP

And there are the 3 known corpses:
rsync://msn.wi.us.finkmirrors.net/finkinfo/ :
@ERROR: chroot failed
rsync://nte.fr.eu.finkmirrors.net/finkinfo/ :
rsync: failed to connect to nte.fr.eu.finkmirrors.net: Connection 
refused (61)
rsync error: error in socket IO (code 10) at 
/SourceCache/rsync/rsync-35.2/rsync/clientserver.c(105) [receiver=2.6.9]
rsync://trd.no.eu.finkmirrors.net/finkinfo/ :
-rw-r--r--  11 2009/07/03 00:48:45 TIMESTAMP

There are currently only 3 working rsync mirrors: the Dutch, German and 
Japanese.

-- 
Martin



--
Come build with us! The BlackBerry(R) Developer Conference in SF, CA
is the only developer event you need to attend this year. Jumpstart your
developing skills, take BlackBerry mobile applications to market and stay 
ahead of the curve. Join us from November 9 - 12, 2009. Register now!
http://p.sf.net/sfu/devconference
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] Differences in Directory Structure for texmf on PowerPC vs. Intel

[Martin Costabel]

rbambery wrote:
[]
>  Other Tex related tools installed on both systems
>  latex-beamer  3.07-1
>  latex-figbib 2004-08-12-2
>  libkpathsea4 3.5.7-1
>  ptexenc 1.06-1
>  rt2latex2e 1.0fc2-1
>  texi2html 1.64-16
>  texinfo 4.13-1001
>  texlive-texmf 0.20080822-1
>  texpower 0.2-4
>  unicode-tex 20041017-2
> 
>  Both systems have
>  dblatex,jadetex, tetex-texmf classified as archived
>  The intel system (only) has
>  texlive-base and texlive classified as archived

Although you don't say it, I suspect that you have mixtures of tetex and 
texlive installed, that is, you have tetex-base or ptex-base and maybe 
some other tetex or ptex packages installed in addition to 
texlive-texmf. Installing texlive-base should clean this out.

-- 
Martin




--
Come build with us! The BlackBerry(R) Developer Conference in SF, CA
is the only developer event you need to attend this year. Jumpstart your
developing skills, take BlackBerry mobile applications to market and stay 
ahead of the curve. Join us from November 9 - 12, 2009. Register now!
http://p.sf.net/sfu/devconference
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] freeglut-2.4.0-5 failed (Was: lablgl-togl-)

[Martin Costabel]

Tom Caulfield wrote:
> tried it twice.  updated fink too. 
> not working.

> Package manager version: 0.29.9 
> Distribution version: selfupdate-rsync Sun Oct 18 20:13:39 2009, 10.4, i386
> Mac OS X version: 10.4.11
> Xcode version: 2.5
> gcc version: 4.0.1 (Apple Computer, Inc. build 5370)
[]
> /bin/sh ../../../libtool --mode=link gcc  -O3  
> -Wl,-dylib_file,/System/Library/Frameworks/OpenGL.framework/Versions/A/Libraries/libGL.dylib:/System/Library/Frameworks/OpenGL.framework/Versions/A/Libraries/libGL.dylib
>  
> -L/sw/lib -o CallbackMaker -export-dynamic ../../../src/libfreeglut.la 
> CallbackMaker-CallbackMaker.o  
> mkdir .libs
> gcc -O3 -Wl,-dylib_file 
> -Wl,/System/Library/Frameworks/OpenGL.framework/Versions/A/Libraries/libGL.dylib:/System/Library/Frameworks/OpenGL.framework/Versions/A/Libraries/libGL.dylib
>  
> -o .libs/CallbackMaker CallbackMaker-CallbackMaker.o  -L/sw/lib 
> ../../../src/.libs/libfreeglut.dylib -lm -L/usr/X11R6/lib -lGL -lGLU 
> -lXext -lX11 -lXxf86vm
> /usr/libexec/gcc/i686-apple-darwin8/4.0.1/ld: Undefined symbols:
> _glClear
> _glClearColor

These symbols should be provided by /usr/X11R6/lib/libGL.dylib 
(indirectly from the OpenGL framework). What do you get from the command

   nm -m /usr/X11R6/lib/libGL.dylib | grep glClearColor

? I get the following:

9dfc (__TEXT,__text) external ___indirect_glClearColor
  (undefined) external _glClearColor (from libGL)
  (undefined) external ___indirect_glClearColor

If you get this, too, then I suspect that you have another, deficient, 
libGL.dylib, probably /usr/local/lib/libGL.dylib, that is interfering. 
If you see this file, remove it.

-- 
Martin



--
Come build with us! The BlackBerry(R) Developer Conference in SF, CA
is the only developer event you need to attend this year. Jumpstart your
developing skills, take BlackBerry mobile applications to market and stay 
ahead of the curve. Join us from November 9 - 12, 2009. Register now!
http://p.sf.net/sfu/devconference
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] fink update-all

[Martin Costabel]

Néstor Espino Briones wrote:
> I am having problem with an applications that uses cftisio in an imac 
> PPC G4. I think that is a library version problem. Then I run fink 
> selfupdate and fink update-all, but the latter does not work.

It would have been nice to know what version of Fink and of MacOSX you 
are running. Without this information, I will be guessing.

> I have problems with db-4.2.52.tar.gz:
> 
> The checksum of the file db-4.2.52.tar.gz of package db42-ssl-4.2.52-17 
> is incorrect. The most

This version db42-ssl-4.2.52-17 is not in any of the currently supported 
Fink trees. It exists in the distribution for 10.3 and in 
10.4-transitional, both of them unsupported for many years.

> likely cause for this is a corrupted or incomplete download
> Expected: cbc77517c9278cdb47613ce8cb55779f

This checksum hints at 10.4-transitional, unsupported for 3 years.

[]
> Make your choice:  [2] 3
> curl -f -L -O -C - http://distfiles.master.finkmirrors.net/db-4.2.52.tar.gz
> ** Resuming transfer from byte position 24439
>   % Total% Received % Xferd  Average Speed   TimeTime Time  
> Current
>  Dload  Upload   Total   SpentLeft  
> Speed
> 100 3803k  100 3803k0 0   101k  0  0:00:37  0:00:37 
> --:--:--  124k
> The checksum of the file is incorrect. The most likely cause for this is 
> a corrupted or
> incomplete download
> Expected: cbc77517c9278cdb47613ce8cb55779f
> Actual: 065954600e53b1845ce52f8cbeac0cef

I don't know what's wrong with your download, but this "Actual" md5sum 
is wrong, too. The right checksum of that file is 
8b5cff6eb83972afdd8e0b821703c33c (I just downloaded it again from this 
same server.)

[]
> curl -f -L -O http://distfiles.bru.be.eu.finkmirrors.net/db-4.2.52.tar.gz

This mirror hasn't worked for a long time. The fact that your fink tries 
  to use it is one of the side effects of being on an unsupported 
version of Fink.

[]
> Then, how can I tell fink that it has to ignore checksum? How can I 
> install it from source using fink? How can I know where to put the 
> package in order that db42-ssl-4.2.52-17 can find it, if I installed 
> db-4.2.52.tar.gz manually?

You can download db-4.2.52.tar.gz manually and place it into /sw/src/. 
Then, provided the checksums match, fink will use it (and there is then 
even an option to ignore checksums).

> I am using Fink 0.24.26.

As I said, this version is a couple of years out of date. Even if you 
are running MacOSX 10.4, you should be using fink-0.29.9 or later nowadays.

-- 
Martin





--
Come build with us! The BlackBerry(R) Developer Conference in SF, CA
is the only developer event you need to attend this year. Jumpstart your
developing skills, take BlackBerry mobile applications to market and stay 
ahead of the curve. Join us from November 9 - 12, 2009. Register now!
http://p.sf.net/sfu/devconference
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

[Fink-devel] path_separator test in older configure scripts breaks on 10.6

[Martin Costabel]

Building db3-3.3.11-1031 on 10.6 breaks with the error

   configure: error: No ar utility found.

Building mysql-shlibs-3.23.58-1023 on 10.6 breaks with the error

   configure: error: no acceptable C compiler found in $PATH

(see the corresponding threads on fink-devel and fink-beginners.)

Both failures have the same cause which I have now hunted down. It is 
likely to show up for other packages whose configure scripts were 
generated by autoconf version 2.53 or older. A workaround is given at 
the end of this message, and I have fixed db3 correspondingly.

The story is the following:

These old autoconf versions have a test for deciding whether 
PATH_SEPARATOR should be ";" or ":". (In reality they have even two 
independent pieces of code that do this, which is one of the many 
weirdnesses in this history.) The test looks whether the following line 
of code

   PATH=".;."; conftest.sh

executes without error, that is, finds the conftest.sh executable which 
is in the current directory. If yes, PATH_SEPARATOR is set to ";" which, 
of course, wreaks havoc on MacOSX.

Now this line of code is in fact executed by /bin/ksh, because 
configure, in its quest to find an acceptable variant of the "echo" 
command, re-executes itself several times in different shells, and it is 
/bin/ksh that it is trying out when the above test is being performed.

The two versions of /bin/ksh in MacOSX 10.5 and 10.6 behave differently 
with respect to that line of code (both of them are buggy IMHO). You can 
see this by executing the following line in /bin/ksh:

   cd /usr/bin; PATH=".;."; type -a machine; machine

On 10.5 (similar on 10.4), you get the output

> machine is a tracked alias for '/usr/bin/.;./machine'
> /bin/ksh: machine: not found [No such file or directory]

On 10.6 you get

> machine is a tracked alias for '/usr/bin/.;./machine'
> i486

Thus on 10.5, it "finds" the command, but then when asked to execute it, 
doesn't really find it. On 10.6, it actually finds it, although that 
weirdo path does not exist.

In newer versions of autoconf (2.57-2.59, ca 2004), the test is changed: 
It uses the line

   PATH="/nonexistent;."; conf$$.sh

which correctly fails in /bin/ksh on 10.6, too. Even newer versions of 
autoconf use some other variant of the test.

The workaround is to run configure with the environment variable

   PATH_SEPARATOR=":"

set, so that this whole crazy test is not performed at all.

-- 
Martin








--
Come build with us! The BlackBerry(R) Developer Conference in SF, CA
is the only developer event you need to attend this year. Jumpstart your
developing skills, take BlackBerry mobile applications to market and stay 
ahead of the curve. Join us from November 9 - 12, 2009. Register now!
http://p.sf.net/sfu/devconference
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] db3-3.3.11-1031 on 10.6 compile error

[Martin Costabel]

Adrian Prantl wrote:
> Hi,
> 
> on 10.6, fink 32-Bit, I am getting this rather confusing error when  
> updating db3:

This is now fixed in version 3.3.11-1032.

-- 
Martin

--
Come build with us! The BlackBerry(R) Developer Conference in SF, CA
is the only developer event you need to attend this year. Jumpstart your
developing skills, take BlackBerry mobile applications to market and stay 
ahead of the curve. Join us from November 9 - 12, 2009. Register now!
http://p.sf.net/sfu/devconference
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] tk-tablematrix-pm586-1.2-2

[Martin Costabel]

monipol wrote:
> On 15/10/2009, at 18:42, khlm lutrija wrote:
>> dont work
>>
>> --
>> Package manager version: 0.29.10
>> Distribution version: selfupdate-rsync Thu Oct 15 23:28:36 2009, 10.5,
>> i386
>> Mac OS X version: 10.5.7
>> Xcode version: 3.1.3
>> gcc version: 4.0.1 (Apple Inc. build 5493)
>> make version: 3.81
>> Feedback Courtesy of FinkCommander
> 
> Providing further (useful) information:
> 
> $ fink -m --build-as-nobody rebuild tk-tablematrix-pm586
> (...)
> Using -L/usr/X11R6/lib to find /usr/X11R6/lib/libX11.la
> Using -I/usr/X11R6/include to find /usr/X11R6/include/X11/Xlib.h
> ### execution of perl5.8.6 failed, exit code 10
> (...)
> Failed: phase compiling: tk-tablematrix-pm586-1.2-2 failed
> 
> This happens because tk-pm586 (unmaintained) doesn't work (my box is  
> similar to $user's):

OTOH, tk-tm588 does work, and building tk-tablematrix-pm588 (after 
adding the type 5.8.8 to the info file) does work, too. Is there any 
reason why this old perl5.8.6 cruft (which dates from the days of MacOSX 
10.3) should be conserved on MacOSX 10.5?

-- 
Martin

--
Come build with us! The BlackBerry(R) Developer Conference in SF, CA
is the only developer event you need to attend this year. Jumpstart your
developing skills, take BlackBerry mobile applications to market and stay 
ahead of the curve. Join us from November 9 - 12, 2009. Register now!
http://p.sf.net/sfu/devconference
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] x86_64 build of ffmpeg failing (still)

[Martin Costabel]

Max Horn wrote:
> 
> Am 14.10.2009 um 10:34 schrieb Jon Connell:
> 
>> William G. Scott wrote:
>>> In file included from h264.h:32,
>>>  from h264.c:31:
>>> cabac.h: In function 'decode_significance_x86':
>>> cabac.h:693: warning: cast from pointer to integer of different size
>>> cabac.h:694: warning: cast from pointer to integer of different size
>>> cabac.h: In function 'decode_significance_8x8_x86':
>>> cabac.h:742: warning: cast from pointer to integer of different size
[]
>>> {standard input}:9509:suffix or operands invalid for `add'
>>> {standard input}:9521:suffix or operands invalid for `add'
>> The configure script configures for x86_32 in spite of being told you
>> want to build for x86_64. This worked for me:
>>
>> diff -r1.12 ffmpeg.info
>> 21a22,23
>>>  ### Fix for selecting x86_32 in spite of --arch=x86_64
>>>  perl -pi -e '$c++ if m,arch="x86_32",;
>> s,arch="x86_32",arch="x86_64", if $c==2' configure
>>
>> I also had to configure with:
>>
>> --cc=gcc-4.0 --arch=x86_64 --extra-cflags="-I/sw64/include -m64"
>> --extra-ldflags=-m64
> 
> I haven't tried the following, so it might not work for ffmpeg; but  
> for other packages (e.g. my dosbox package), adding the following to  
> the configure params solves the issue, without any patching and  
> without needing to make multiple .info files.
> 
>--build=%m-apple-darwin`uname -r|cut -f1 -d.`

Unfortunately, ffmpeg's hand-crafted configure script does not like the 
--build flag. It wants --arch, but then it stupidly replaces x86_64 by 
x86_32. Jon's perl -pie fix does the trick here.

OTOH, the -m64 doesn't seem to be necessary if compiled with the default 
"gcc" (that is, gcc-4.2 with Fink's arch trickery). For gcc-4.2 to work, 
the "warning: cast from pointer to integer of different size" has to be 
patched away, too.

Since the maintainer seems to be busy, I have now committed a patched 
version of ffmpeg.info. Please test.

-- 
Martin

--
Come build with us! The BlackBerry(R) Developer Conference in SF, CA
is the only developer event you need to attend this year. Jumpstart your
developing skills, take BlackBerry mobile applications to market and stay 
ahead of the curve. Join us from November 9 - 12, 2009. Register now!
http://p.sf.net/sfu/devconference
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] "lipo: can't open input file" error

[Martin Costabel]

Koen van der Drift wrote:
> Yeah, Iwas thinking along those lines as well. So I found that if I
> comment out the line:
> 
> set( CMAKE_OSX_ARCHITECTURES ppc;i386 )
> 
> it will not create a universal binary. But then it will create
> different packages for ppc and i386, correct? How do I work around
> that, should I make variants, eg?

Packages for ppc and i386 are always different, that's why the *.deb 
files have different names and are stored in different directories.

-- 
Martin


--
Come build with us! The BlackBerry(R) Developer Conference in SF, CA
is the only developer event you need to attend this year. Jumpstart your
developing skills, take BlackBerry mobile applications to market and stay 
ahead of the curve. Join us from November 9 - 12, 2009. Register now!
http://p.sf.net/sfu/devconference
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] when builds fail (e.g., gcc44)

[Martin Costabel]

Dave Yost wrote:
> I suggest that fink be enhanced to do what some other systems have been 
> doing lately: package up all the info for a bug report and send it to 
> the right place (after the user has a chance to look it over and say OK).
> 
> The above functionality would replace the stdout dump of do-it-yourself 
> "Please submit a full bug report" instructions exemplified below, which 
> occurred as the first thing I tried to do after a clean install of fink 
> 029.9 on my 10.6.1 system with Xcode 3.2.1.

The problem is that bugs are rarely adhering to a standard that would 
allow an automatic treatment (otherwise they would probably not be 
there.) You case is a good example for this.

[]
> ...
> /sw/src/fink.build/gcc44-4.4.1-1000/darwin_objdir/./prev-gcc/xgcc 
> -B/sw/src/fink.build/gcc44-4.4.1-1000/darwin_objdir/./prev-gcc/ 
> -B/sw/lib/gcc4.4/i686-apple-darwin10/bin/ -c  -g -O2 
> -fomit-frame-pointer -DIN_GCC   -W -Wall -Wwrite-strings 
> -Wstrict-prototypes -Wmissing-prototypes -Wcast-qual 
> -Wold-style-definition -Wc++-compat -Wmissing-format-attribute -pedantic 
> -Wno-long-long -Wno-variadic-macros -Wno-overlength-strings   
> -DHAVE_CONFIG_H -I. -I. -I../../gcc-4.4.1/gcc -I../../gcc-4.4.1/gcc/. 
> -I../../gcc-4.4.1/gcc/../include -I../../gcc-4.4.1/gcc/../libcpp/include 
> -I/sw/include  -I../../gcc-4.4.1/gcc/../libdecnumber 
> -I../../gcc-4.4.1/gcc/../libdecnumber/dpd -I../libdecnumber 
> -I/sw/include  -I/sw/include -DCLOOG_PPL_BACKEND  -I/sw/include 
> ../../gcc-4.4.1/gcc/dbxout.c -o dbxout.o
> ../../gcc-4.4.1/gcc/dbxout.c: In function 'dbxout_begin_simple_stabs':
> ../../gcc-4.4.1/gcc/dbxout.c:584: internal compiler error: in 
> ggc_set_mark, at ggc-page.c:1319
> Please submit a full bug report,
> with preprocessed source if appropriate.
> See  for instructions. 

It is not fink that is speaking here, but gcc. So what would you think 
is "the right place" this should be sent to, Fink or gnu? And what would 
"all the info for a bug report" be?

[]
> ../../gcc-4.4.1/gcc/config/i386/sync.md: In function 'recog_35':
> ../../gcc-4.4.1/gcc/config/i386/sync.md:30: internal compiler error: in 
> df_reg_chain_mark, at df-scan.c:4277
> Please submit a full bug report,
> with preprocessed source if appropriate.
> See  for instructions.

Seeing that the bug seems to appear randomly, (a fact that an automated 
bug reporting system would probably have hidden), my wild guess is that 
you either have a shell startup script that sets a too small stacksize 
limit, or you have a hardware (perhaps memory) problem.

-- 
Martin


--
Come build with us! The BlackBerry(R) Developer Conference in SF, CA
is the only developer event you need to attend this year. Jumpstart your
developing skills, take BlackBerry mobile applications to market and stay 
ahead of the curve. Join us from November 9 - 12, 2009. Register now!
http://p.sf.net/sfu/devconference
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] openjade-1.3.2-1034 fails to install in os x 10.6 under fink?

[Martin Costabel]

I have that version of openjade installed on both 10.6/32bit and 
10.6/64bit, so there is no problem with the package itself.

Thomas Bewley wrote:
[]
>/sw/include/OpenSP/InternalInputSource.h:37: error: extra
> qualification 'OpenSP::InternalInputSource::' on member
> 'asInternalInputSource'

It seems you have an old version of opensp4-dev installed, which can 
happen if you upgraded your Fink from 10.5 or if you installed 
opensp4-dev from a (necessarily unofficial) binary distribution. The 
opensp4-dev package was fixed in December 2008, but unfortunately 
without changing its revision. Thus opensp4-dev-1.5.1-1006 can either be 
the old or the new version, depending on when it was built.

So, run "fink rebuild opensp4-dev" or wait a little until selfupdate 
shows you opensp4-1.5.1-1007, and then run "fink update opensp4-dev".
Then openjade should build OK.

-- 
Martin

--
Come build with us! The BlackBerry(R) Developer Conference in SF, CA
is the only developer event you need to attend this year. Jumpstart your
developing skills, take BlackBerry mobile applications to market and stay 
ahead of the curve. Join us from November 9 - 12, 2009. Register now!
http://p.sf.net/sfu/devconference
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] TeX Live in unstable

[Martin Costabel]

Tomoaki Okayama wrote:
> What will happen by applying the attached patch to lcoco.c?

With the patch, it builds OK.

-- 
Martin

--
Come build with us! The BlackBerry(R) Developer Conference in SF, CA
is the only developer event you need to attend this year. Jumpstart your
developing skills, take BlackBerry mobile applications to market and stay 
ahead of the curve. Join us from November 9 - 12, 2009. Register now!
http://p.sf.net/sfu/devconference
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] TeX Live in unstable

[Martin Costabel]

Tomoaki Okayama wrote:
> Dear Fink users/developers,
> 
> I am pleased to announce that TeX Live is now available on
> 10.4/10.5/10.6 in unstable. Enjoy it!

Building texlive-0.20080816-1 fails on 10.6/64bit with the error message

gcc -g -O2  -I/sw_x86_64/lib/freetype219/include/freetype2 
-I/sw_x86_64/lib/freetype219/include -I/sw_x86_64/include 
-I../../libs/obsdcompat -I../../libs/obsdcompat/.. 
-I../../../texk/web2c/../../libs/obsdcompat 
-I../../../texk/web2c/../../libs/obsdcompat/.. 
-I../../../texk/web2c/../../libs/zlib 
-I../../../texk/web2c/../../libs/libpng 
-I../../../texk/web2c/../../libs/xpdf -Dextra_version_info=`date 
+-%Y%m%d%H`  -c -o lcoco.o lcoco.c
In file included from lcoco.c:334:
/usr/include/ucontext.h:42:2: error: #error ucontext routines are 
deprecated, and require _XOPEN_SOURCE to be defined
[...]
make[4]: *** [lcoco.o] Error 1
make[3]: *** [posix] Error 2
make[2]: *** [../../libs/lua51/liblua.a] Error 2
make[1]: *** [all] Error 1
make: *** [all] Error 1

On the same machine, I have texlive-0.20080816-0test5 installed, which 
built without error.

-- 
Martin


--
Come build with us! The BlackBerry(R) Developer Conference in SF, CA
is the only developer event you need to attend this year. Jumpstart your
developing skills, take BlackBerry mobile applications to market and stay 
ahead of the curve. Join us from November 9 - 12, 2009. Register now!
http://p.sf.net/sfu/devconference
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] db3-3.3.11-1031 on 10.6 compile error

[Martin Costabel]

Adrian Prantl wrote:
[]
> + IFS=';'

This is the reason why it breaks. On 10.6, the configure script gets the 
weird idea that the path separator is ';' instead of ':'. It has a funny 
test for this, and this test behaves differently on 10.5 and on 10.6. It 
is the line

   PATH=".;."; conftest.sh

Here conftest.sh is a trivial script in the current directory. On 10.5, 
it gives the error message

../dist/configure[1123]: conftest.sh: not found [No such file or directory]

whereas on 10.6, it passes without error, inducing configure to believe 
that ';' is a valid path separator.

This is how far I got. I don't understand why this behaves differently 
on 10.5 and 10.6, but I don't even understand why this should be a valid 
test (which it is not, evidently). Weird.

-- 
Martin

--
Come build with us! The BlackBerry(R) Developer Conference in SF, CA
is the only developer event you need to attend this year. Jumpstart your
developing skills, take BlackBerry mobile applications to market and stay 
ahead of the curve. Join us from November 9 - 12, 2009. Register now!
http://p.sf.net/sfu/devconference
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] TeX Live in unstable

[Martin Costabel]

Tomoaki Okayama wrote:
> Dear Fink users/developers,
> 
> I am pleased to announce that TeX Live is now available on
> 10.4/10.5/10.6 in unstable. Enjoy it!

A big Thank You! I have been using your pre-release versions for many 
months now on a daily basis, without any problems, neither with my 
rather tex-intensive work nor with building other tex-dependent Fink 
packages.

> 
> - libkpathsea4.info (3.5.7-1, update)
> - texlive-texmf.info (0.20080822-1, new)
> - texlive.info (0.20080816-1, new)

If libkpathsea4 and texlive use the same patch file, they should perhaps 
also use the same MD5-sum? (The one in libkpathsea4.info does not work.)

-- 
Martin


--
Come build with us! The BlackBerry(R) Developer Conference in SF, CA
is the only developer event you need to attend this year. Jumpstart your
developing skills, take BlackBerry mobile applications to market and stay 
ahead of the curve. Join us from November 9 - 12, 2009. Register now!
http://p.sf.net/sfu/devconference
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] gnucash2

[Martin Costabel]

Alexander Hansen wrote:
[]
> My understanding is that this would primarily be the case for builds
> against static-only library packages.  Unfortunately, since we've got
> packages that install both static and dynamic libraries, scripting
> rebuilds of every package that has a ".a" file would be a pretty big
> rebuild on a typical user's system.

I must admit that I never understood the usefulness of the *.la files. 
As far as I can see, all libtool uses them for is to replace "-lfoo" on 
a "libtool --mode=link" line by the explicit filename of the library on 
the linker line, adding some dependencies at the same time - an exploit 
that is totally unnecessary. The linker understands perfectly well 
"-lfoo", whether it means a static or a dynamic library, and 
dependencies are usually also handled by other utilities.

Can someone give an example where the existence of an *.la file is 
useful or perhaps even necessary?

-- 
Martin



--
Come build with us! The BlackBerry® Developer Conference in SF, CA
is the only developer event you need to attend this year. Jumpstart your
developing skills, take BlackBerry mobile applications to market and stay 
ahead of the curve. Join us from November 9-12, 2009. Register now!
http://p.sf.net/sfu/devconf
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] gnucash2

[Martin Costabel]

David Fang wrote:
[]
> Martin,
>   What ever happened to that script you wrote to regenerate Apple's 
> X11 .la files?  I think I have a copy lying around somewhere.  Would that 
> work?

No, it wouldn't. As it turned out, the situation is worse than just an 
X11 problem, because also a couple of basic *.la files in /usr/lib have 
gone missing.

It would, of course, be possible to enhance that script so that it 
regenerates *all* the *.la files that were present in Leopard, but 
somehow I don't think this is a good idea. Killing all fink-generated 
*.la files gives the best solution so far, although there is a certain 
feeling that this might some day have bad side effects.

-- 
Martin



--
Come build with us! The BlackBerry® Developer Conference in SF, CA
is the only developer event you need to attend this year. Jumpstart your
developing skills, take BlackBerry mobile applications to market and stay 
ahead of the curve. Join us from November 9-12, 2009. Register now!
http://p.sf.net/sfu/devconf
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] gnucash2

[Martin Costabel]

Alexander Hansen wrote:
[]
> Yes.  You're missing libXrender.la and a bunch of other .la files :-)
> You didn't specify, but I'm assuming you upgraded from 10.5, and
> probably had Xquartz X11 installed.  Apple got rid of the .la files for
> X11, and for some of the other built-in libraries, too.
> 
> A workaround is to use (assuming your Fink tree is in the default path;
> modify as appropriate)
> 
> sudo -s
> find /sw/lib -name *.la | xargs rm
> 
> That will clear out all of your old .la files, and any new ones you
> generate from fink builds won't refer to the now-nonexistent ones.

Alex, wouldn't it be time for a FAQ entry? Or are the other Fink chiefs 
still hoping for a divine inspiration on how to save the 10.5->10.6 
upgrade path in a dignified way?

-- 
Martin

--
Come build with us! The BlackBerry® Developer Conference in SF, CA
is the only developer event you need to attend this year. Jumpstart your
developing skills, take BlackBerry mobile applications to market and stay 
ahead of the curve. Join us from November 9-12, 2009. Register now!
http://p.sf.net/sfu/devconf
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] pil-py25-1.1.6-6 selftest fails

[Martin Costabel]

monipol wrote:
> Intel OS 10.5.8, 32- and 64-bit Fink, Xcode 3.1.4, case-sensitive  
> filesystem.
> 
> $ fink -m --build-as-nobody rebuild pil-py25
> (...)
> building '_imagingmath' extension
> gcc -fno-strict-aliasing -mno-fused-madd -DNDEBUG -g -fwrapv -O3 -Wall  
> -Wstrict-prototypes -IlibImaging -I/sw/include/freetype2 -I/sw/include  
> -I/usr/include -I/sw/include/python2.5 -c _imagingmath.c -o build/ 
> temp.macosx-10.5-i386-2.5/_imagingmath.o
> gcc -L/sw/lib -bundle -undefined dynamic_lookup -L/sw/lib build/ 
> temp.macosx-10.5-i386-2.5/_imagingmath.o -L/sw/lib -L/usr/lib -o build/ 
> lib.macosx-10.5-i386-2.5/_imagingmath.so
> (...)
> changing mode of build/scripts-2.5/pilprint.py from 644 to 755
> sudo -u fink-bld [ENV] sh -c/var/tmp/tmp.5.ucAQri
> /sw/bin/python2.5 selftest.py || exit 2
> Traceback (most recent call last):
>File "selftest.py", line 11, in 
>  import ImageMath
>File "PIL/ImageMath.py", line 19, in 
>  import _imagingmath
> ImportError: No module named _imagingmath

The PIL guys seem to assume that everyone has the PIL directory in their 
PYTHONHOME path or that this test is run with the PIL directory as $CWD. 
I think they should use "from PIL import _imagingmath" instead of 
"import _imagingmath" or an appropriate cd command beforehand.

-- 
Martin



--
Come build with us! The BlackBerry® Developer Conference in SF, CA
is the only developer event you need to attend this year. Jumpstart your
developing skills, take BlackBerry mobile applications to market and stay 
ahead of the curve. Join us from November 9-12, 2009. Register now!
http://p.sf.net/sfu/devconf
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] TeX Live

[Martin Costabel]

Tomoaki Okayama wrote:
> Dear developers,
> 
> I'm planning to commit finkinfos related to TeX Live in the
> repository: experimental/todai/ecc-10.4/main/finkinfo/text/
> to 10.4/unstable tree. See also the discussions in Fink-devel:
>  http://thread.gmane.org/gmane.os.macosx.fink.devel/17670
> 
> My only worry is that there is no positive feedback from
> 10.6 users. Could anyone tell me it works or not?

This works OK for me on 10.6/64bit and on a 10.5->10.6 upgraded Fink tree.

> Any other feedbacks are welcome. To save time, EXPLICIT patches
> with your explanation are very appreciated.

No patches needed, as far as I am concerned. I can only repeat for the 
n-th time that I think it is high time that the old tetex be scrapped 
and replaced by texlive.

-- 
Martin

--
Come build with us! The BlackBerry® Developer Conference in SF, CA
is the only developer event you need to attend this year. Jumpstart your
developing skills, take BlackBerry mobile applications to market and stay 
ahead of the curve. Join us from November 9-12, 2009. Register now!
http://p.sf.net/sfu/devconf
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] XQuartz alpha1 for 10.6

[Martin Costabel]

Michal Suchanek wrote:
> 2009/9/18 Martin Costabel :
[]
>> completely over to /opt/X11, which means changing every single package
>> and forcing everyone to abandon the system X11.
>>
> 
> Well, perhaps you can introduce a variable which specifies the X
> location so most packages can be updated to compile against either
> location without further changes.

Without "further" changes, perhaps. But first every package has to be 
changed. And then this will be another variant to manage, which would 
make 8 (10.5/10.6, 32/64bit, X11-opt/usr), without Tiger. A nightmare.

-- 
Martin

--
Come build with us! The BlackBerry® Developer Conference in SF, CA
is the only developer event you need to attend this year. Jumpstart your
developing skills, take BlackBerry mobile applications to market and stay 
ahead of the curve. Join us from November 9-12, 2009. Register now!
http://p.sf.net/sfu/devconf
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] XQuartz alpha1 for 10.6

[Martin Costabel]

David R. Morrison wrote:
> 
> On Sep 18, 2009, at 7:23 AM, Jack Howarth wrote:
> 
>> Martin,
>>I don't follow your last statement. My understanding is that X11
>> in Snow Leopard is already preconfigured to hand off to an X11 in /opt
>> if present (remember that Jeremy Huddleston is both X11 maintainer at
>> Apple as well as the packager of the X11 Xquartz updates as well as
>> being the maintainer of all of MacPorts x11 packages). If you install
>> X11 Xquartz under Snow Leopard, it will automatically use any  
>> improvements
>> to the user interface since the OS will consider it the new default
>> X11 location (in terms of the runnning X11).
>>Jack
>>
> 
> That's interesting.  How does it work, do you know?  If you have  
> another program dynamically linked to an X11 library (in the standard  
> way), is it going to use the library in /opt instead of the library  
> in /usr/X11 ?  What would trigger that?

As I understood the announcement, there will be a preference pane 
allowing you to choose which X server will be started by launchd. What 
will happen if you run binaries compiled against the X11 in /usr/X11 on 
the X server in /opt/X11, well, I guess this will be interesting...

In any case, any improvements in xquartz living in /opt/X11 will perhaps 
be available at runtime, but not at compile time unless we switch Fink 
completely over to /opt/X11, which means changing every single package 
and forcing everyone to abandon the system X11.

-- 
Martin






--
Come build with us! The BlackBerry® Developer Conference in SF, CA
is the only developer event you need to attend this year. Jumpstart your
developing skills, take BlackBerry mobile applications to market and stay 
ahead of the curve. Join us from November 9-12, 2009. Register now!
http://p.sf.net/sfu/devconf
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] XQuartz alpha1 for 10.6

[Martin Costabel]

Alexander Hansen wrote:
[]
> There was an official update of 10.4's X11 via Software Update.  And on
> Leopard a number of the system updates included files in the X11
> distribution.

I agree with Jack on the fact that the xquartz development improved the
"official" X11. The latest softwareupdated X11 on Leopard has several
bug fixes for the user interface from early xquartz releases. And the
official X11 on Snow Leopard has all the fixes from xquartz before 2.4.0.

This concerns, for example, the new copy-paste system, or the 
interaction with Spaces. And there is the wrapper script in 
/Applications/Utilities/X11.app that runs X11 in a correct login 
environment for all known login shells. This was the result of a 
collaborative effort on the x11-users and xquartz-dev lists, put into 
xquartz by Jeremy. And it is now in the system X11 on Snow Leopard.

I doubt if such improvments of the user interface will still happen when 
  xquartz lives in /opt/X11.

-- 
Martin





--
Come build with us! The BlackBerry® Developer Conference in SF, CA
is the only developer event you need to attend this year. Jumpstart your
developing skills, take BlackBerry mobile applications to market and stay 
ahead of the curve. Join us from November 9-12, 2009. Register now!
http://p.sf.net/sfu/devconf
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] XQuartz alpha1 for 10.6

[Martin Costabel]

Jack Howarth wrote:
> Martin,
>It is a bit extreme to say the Jeremy has gone
> off the deep end with bleeding edge. Completely

I didn't say "off the deep end". What he does is still very valuable, 
but in a different way from before. In the long run, it will probably 
benefit all users of X11 on Macs, but the equation "xquartz = bugfix for 
Apple's X11" does no longer hold. And that's what it was for most X11 
users and most Fink users.

> untrue. Even if the X11 Xquartz was based entirely
> on the same code as that in Snow Leopard X11, this
> issue of potential incompatibilities would exist.
> The change causing the problem is the movement of
> X11 Xquartz into /opt. The reason he did that was
> to prevent system upgrades from constantly nuking
> the currently installed X11 Xquartz installation.
> Also note that while X11 Xquartz 2.4.0 now contains
> an automatic update mechanism, it isn't triggered
> automatically after a system update has been 
> applied on Leopard to restore the previous X11
> Xquartz installation's components. So short of such 
> a feature being added to X11 Xquartz, movement of
> the X11 Xquartz installation to /opt is the only
> way to prevent the system updates from overwriting
> the X11 Xquartz components.

True, but this goes also the other way: Installation to /opt means no 
more bug fix for the system X11. At runtime, there will be 
interferences. At compile-time, I don't know; I think /opt/X11 is not 
automatically found by configure scripts, which would mean no 
interference for most packages.

-- 
Martin

--
Come build with us! The BlackBerry® Developer Conference in SF, CA
is the only developer event you need to attend this year. Jumpstart your
developing skills, take BlackBerry mobile applications to market and stay 
ahead of the curve. Join us from November 9-12, 2009. Register now!
http://p.sf.net/sfu/devconf
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] XQuartz alpha1 for 10.6

[Martin Costabel]

Alexander Hansen wrote:
> -BEGIN PGP SIGNED MESSAGE-
> Hash: SHA1
> 
> Robert Wyatt wrote:
>> Jeremy H has just released XQuartz 2.4.1_alpha1 for Snow Leopard. I'm 
>> going to continue to work through building the bundle- packages with the 
>> X11 that shipped with 10.6 for now unless you folks think that it 
>> shouldn't matter (in which case I would install the alpha).
>> --Robert
>>
> 
> It may well matter.  Supposedly XQuartz for Snow Leopard is going to
> install under /opt, so Fink packages may need to be more explicit about
> their X11 detection to force the use of the official X11.

I rather think Fink should warn *not* to install this xquartz update.
Xquartz itself should give some warnings, too, like it should have 
warned not to install 2.4.0 on 10.5 if an update to 10.6 was ever planned.

There will certainly be some discussion on the x11-users list, once 
people become aware of the fact that Jeremy is now on a completely 
different trip and no longer mainly fixes bugs of Apple's X11, but 
offers bleeding-edge x.org X11 that may interfere with the working and 
use of Apple's X11. I am looking forward to this discussion...

-- 
Martin



--
Come build with us! The BlackBerry® Developer Conference in SF, CA
is the only developer event you need to attend this year. Jumpstart your
developing skills, take BlackBerry mobile applications to market and stay 
ahead of the curve. Join us from November 9-12, 2009. Register now!
http://p.sf.net/sfu/devconf
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] leaving fink

[Martin Costabel]

Jack Howarth wrote:
> Daniel and Benjamin,
>  I've been pondering switching my efforts over to
> MacPorts for awhile now since it is obvious that the
> X11 Xquartz situation is going to be very problematic
> for fink in the absence of manpower to maintain a proper
> set of X11 packages in fink. 

I shall not comment on the personal animosity part of this discussion (I 
actually understand both sides, with slightly more sympathy for Jack's 
points).

But as to X11 on Snow Leopard: Once the transition is really over (and 
this is far from done yet!), I think we can safely forget about xquartz 
and possible Fink X11 packages, at least until Apple starts doing crazy 
things again. I won't bet on how long this will take, but we should be 
safe at least for a couple of months.

Xquartz had two purposes that for a while went hand in hand:

First it was simply a bug fix release for Apple's rather botched X11 
release on Leopard.

Second, it was interesting for people who wanted the bleeding edge Xorg 
X11 on their Mac.

As of the latest declarations about the role and structure of X11 on SL, 
the second role is the only one that will remain.

On the other hand, for the large majority of people, it was the first 
role that was interesting. And the same thing holds for Fink users. On 
SL, most of the bugs of Leopard X11 are fixed, thanks to the xquartz 
project.

So we should concentrate on a future with Apple's SL X11. Let Macports 
spend their energy on bleeding edge X11 packages.

-- 
Martin


--
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day 
trial. Simplify your report design, integration and deployment - and focus on 
what you do best, core application coding. Discover what's new with 
Crystal Reports now.  http://p.sf.net/sfu/bobj-july
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] bash 4.0 to stable?

[Martin Costabel]

Robert Wyatt wrote:
[]
> gcc -L./builtins -L./lib/readline -L./lib/readline -L./lib/glob 
> -L./lib/tilde  -L./lib/sh -L/sw/lib-g -O2 -o bash shell.o eval.o 
> y.tab.o general.o make_cmd.o print_cmd.o  dispose_cmd.o execute_cmd.o 
> variables.o copy_cmd.o error.o expr.o flags.o jobs.o subst.o hashcmd.o 
> hashlib.o mailcheck.o trap.o input.o unwind_prot.o pathexp.o sig.o 
> test.o version.o alias.o array.o arrayfunc.o assoc.o braces.o 
> bracecomp.o bashhist.o bashline.o  list.o stringlib.o locale.o findcmd.o 
> redir.o pcomplete.o pcomplib.o syntax.o xmalloc.o  -lbuiltins -lsh 
> -lreadline -lhistory -lcurses -lglob -ltilde  -lintl -liconv -lc  -ldl
> Undefined symbols:
>"_rl_menu_completion_entry_function", referenced from:
>_rl_menu_completion_entry_function$non_lazy_ptr in bashline.o
>"_rl_completion_invoking_key", referenced from:
>_rl_completion_invoking_key$non_lazy_ptr in pcomplete.o
>"_rl_variable_value", referenced from:
>_command_word_completion_function in bashline.o
>_edit_and_execute_command in bashline.o
> ld: symbol(s) not found
> collect2: ld returned 1 exit status
> make: *** [bash] Error 1
> ### execution of make failed, exit code 2
> 
> Regarding which I found this thread: 
> http://www.nabble.com/-MacPorts---18621:-Error-with-bash-4.0-build-td22189056.html
> 
> Which confused me further because it indicates that this symbol is from 
> readline 6. It begs the question of how it was building before.

Before, on 10.5.8, it was building its own static libreadline.a and 
libhistory.a and linking to these. No idea why it doesn't on 10.6.

-- 
Martin

--
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day 
trial. Simplify your report design, integration and deployment - and focus on 
what you do best, core application coding. Discover what's new with 
Crystal Reports now.  http://p.sf.net/sfu/bobj-july
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] [Fwd: zsh as shell in compilescript on 10.6 breaks 32bit Fink]

[Martin Costabel]

William G. Scott wrote:
> 
> 
> On Sep 7, 2009, at 10:08 AM, Martin Costabel wrote:
> 
>> William G. Scott wrote:
>>>
>>> #!/bin/zsh -f
>>> source %p/bin/init.sh
>>> PATH=%p/bin:%p/sbin:${PATH} ; export PATH
>>> this appears to put everything right, at least in my experiment.
>>
>> But on 10.6 "gcc" then would still be "/usr/bin/gcc" (which produces 
>> 64bit binaries) and not "/usr/bin/gcc -arch i386" as it should. Hard 
>> to see how this can work.
> 
> Well, with apbs, it just did :)   (I switched it to /bin/bash for good 
> measure). 

Well, using /bin/bash, or /bin/sh for that matter, of course solves the 
problem. You don't need to source anything then, because fink already 
does it for you, and the startup scripts that are perhaps read by 
/bin/sh (I don't know if any is read in this situation) do not destroy 
the PATH as does the zsh startup script /etc/zshenv.

-- 
Martin


--
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day 
trial. Simplify your report design, integration and deployment - and focus on 
what you do best, core application coding. Discover what's new with 
Crystal Reports now.  http://p.sf.net/sfu/bobj-july
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] [Fwd: zsh as shell in compilescript on 10.6 breaks 32bit Fink]

[Martin Costabel]

William G. Scott wrote:
> 
> On Sep 7, 2009, at 5:23 AM, Martin Costabel wrote:
> 
>> Somehow one needs to place that path-prefix back in front of the PATH,
>> but since this is distribution and architecture dependent, to do so  
>> can
>> become quite a mess.
> 
> Hi Martin:
> 
> I am working on pdb2pqr, an apbs dependency that uses gfortran, gcc,  
> and g++
> 
> It is currently broken on 23-bit 10.6 but works fine on the other  
> platforms.
> 
> If the first three lines are
> 
> #!/bin/zsh -f
> source %p/bin/init.sh
> PATH=%p/bin:%p/sbin:${PATH} ; export PATH
> 
> this appears to put everything right, at least in my experiment.

But on 10.6 "gcc" then would still be "/usr/bin/gcc" (which produces 
64bit binaries) and not "/usr/bin/gcc -arch i386" as it should. Hard to 
see how this can work.

> Because of zsh's path anti-metastasis feature, I HAD to do this  

Are you saying that with zsh, /sw/bin/init.sh does not place Fink's 
paths in front of PATH?

-- 
Martin

--
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day 
trial. Simplify your report design, integration and deployment - and focus on 
what you do best, core application coding. Discover what's new with 
Crystal Reports now.  http://p.sf.net/sfu/bobj-july
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] [cvs] dists/10.4/unstable/main/finkinfo/sci mosflm.info, 1.23, 1.24

[Martin Costabel]

David R. Morrison wrote:
> Bill,
> 
> Isn't that going to break on powerpc?
> 
>-- Dave
> 
> 
> On Sep 6, 2009, at 11:09 PM, William Scott wrote:
> 
>> Update of /cvsroot/fink/dists/10.4/unstable/main/finkinfo/sci
>> In directory fdv4jf1.ch3.sourceforge.com:/tmp/cvs-serv9438
>>
>> Modified Files:
>>  mosflm.info
>> Log Message:
>> fix to force 32-bit compilation on 10.6 when appropriate
>>
>> Index: mosflm.info
>> ===
>> RCS file: /cvsroot/fink/dists/10.4/unstable/main/finkinfo/sci/ 
>> mosflm.info,v
>> retrieving revision 1.23
>> retrieving revision 1.24
>> diff -u -d -r1.23 -r1.24
>> --- mosflm.info  7 Sep 2009 02:22:22 -   1.23
>> +++ mosflm.info  7 Sep 2009 03:09:19 -   1.24
>> @@ -12,6 +12,10 @@
>> CompileScript: <<
>> #!/bin/zsh -efv
>> source %p/bin/init.sh
>> +if [[ "%m" != "x86_64" ]]; then
>> +perl -pi -e 's|gfortran|gfortran -m32|g' fink_gfortran.zsh
>> +perl -pi -e 's|gcc |gcc -m32 |g' fink_gfortran.zsh
>> +fi
>> source fink_gfortran.zsh
>> make
>> cd bin

One problem with zsh-based compilescripts on 10.6 is (see the zsh thread 
I started on this) that the zsh startup script on 10.6 clobbers fink's 
PATH.

I think what should be done is that Fink provides a script or perl 
function that can be called by the compilescript and that places the 
right fink path, in particular fink-path-prefix..., in front of PATH. 
Right now, this PATH surgery is buried deeply in the function 
Fink::PkgVersion::get_env.

This should take care of the right flag for gcc. As for the gfortran 
flag, Jack H. seems to assert that it shouldn't be necessary, but I am 
not sure that this is true.

-- 
Martin




--
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day 
trial. Simplify your report design, integration and deployment - and focus on 
what you do best, core application coding. Discover what's new with 
Crystal Reports now.  http://p.sf.net/sfu/bobj-july
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Re: [Fink-devel] zsh as shell in compilescript on 10.6 breaks 32bit Fink

[Martin Costabel]

Jack Howarth wrote:
[]
> 
> Martin,
>I am a bit confused. If you want gfortran to pass -m32, I would suspect 
> you are
> running i386 fink in which case the gcc44 package will have built a 32-bit 
> native
> compiler that generates i386 code. You shouldn't have to pass it anything at 
> all.

I have not yet had enough time to investigate, but I saw the following 
failure in the configure phase of apbs on 10.6/32bit:

It tries to link a fortran and a C module in order to find out the 
fortran name mangling scheme. This fails because the fortran-compiled 
*.o file is "of wrong architecture". I think for the C part it uses 
/usr/bin/gcc with the Fink-supplied automatic "-arch i386" and for the 
fortran part it uses gfortran from gcc44. But have to confirm this when 
I boot into 10.6 next time.

-- 
Martin


--
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day 
trial. Simplify your report design, integration and deployment - and focus on 
what you do best, core application coding. Discover what's new with 
Crystal Reports now.  http://p.sf.net/sfu/bobj-july
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel