Daniel Macks | 24 Aug 19:29 2015

applesystemfonts behavior and pathname problems

Minor: the package creates all its files via PostInstScript including the directory in which they reside.
That makes it difficult for sysadmins to determine what package owns those files because nothing about
them is managed by dpkg. It appears that creating the files at that stage makes sense, but that the
directory-name would be known at InstallScript (it's in %p and based on data published by another quite
stable package). Having the directory in the .deb would at least make *that* level known to dpkg, giving a
correct clue to the package that handes the files within. Easily fixable, will do so in a day or two unless
someone complains. 

Major 1: the PostInstScript hardcodes full paths to several executables in /usr/X11R6/bin, an x11 prefix
that is not the real x11 prefix on any supported platform (and hasn't been for many years). It remains as a
symlink (when the installer doesn't screw up) on currently supported platforms, but who knows what might
happen in the future. Fink's init.sh is designed to add the correct x11 paths to PATH and is always loaded
for all dpkg package scripts, so there's no need to hardcode at all. That avoids having to update or fork the
package for different x11 vendors (of which 10.7 is different than 10.8+). Easily fixable, will do so in a
day or two unless someone complains. 

Major 2: the PostInstScript, running as root, silently seems to alter files in all users' homedirs. That's
not very nice! And it links them all into a public (in %p) location. Not good to expose user's private
filenames to each other (even if the files themselves are protected by permissions). And it creates all
those links in a single location, so multiple users' file-links would overwrite each other. That means
it's broken for all users except the "last" one, since presumably those whose were overwritten would not
have permission to the last-one's files. This all might make sense on a single-user machine, but not
beyond that. This is all badly broken by design and I don't know a correct solution. Is there a way to have the
users' data cached in each user's own space, or at least in subdi
 rs?

dan

--
Daniel Macks
(Continue reading)

Daniel Macks | 16 Aug 18:20 2015

Perl-module variant upgrades

We're missing lots of -pm5182 variants, and even some -pm5162. Over the next day or two, I will be adding lots
of variants, with testing limited to: buildability on 10.10, dep-tree self-consistency. I will not be
making any version upgrades or other changes at this point unless things don't build or have noticeable
missing dependencies, except for adjusting C/R if there are variant-mutually-exclusive sets. If you do
not want me to touch your packages, speak now and/or add a note to your .info files. 

dan

--
Daniel Macks
dmacks <at> netspace.org

------------------------------------------------------------------------------
_______________________________________________
Fink-devel mailing list
Fink-devel <at> lists.sourceforge.net
List archive:
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

David Reiser | 7 Aug 22:21 2015

TheCaptain branch looking for patch files in the wrong place

I’m trying to build the experimental glib2-shlibs. The info file is in:

/sw/fink/10.7/local/main/finkinfo/glib2-shlibs.info

The patch files are in the same directory, but trying to build glib2-shlibs ends with an error:

Failed: Cannot read PatchFile "/sw/fink/dists/local/main/finkinfo/glib2-shlibs.patch”

Is this a bug, or did I do something odd? Fink is finding the local info file in the
usual-for-prior-MacOS-versions, but it is not looking for the patch file in the same directory in which
it found the info file.

Despite running fink configure, I’m also usually getting messages that:

Couldn't change value of PrimaryGroupID in DirectoryServices entry /Groups/fink-bld.

During the fink configure run, it thought that it succeeded making fink-bld user 601

Dave
--
Dave Reiser
dbreiser <at> icloud.com

------------------------------------------------------------------------------
_______________________________________________
Fink-devel mailing list
Fink-devel <at> lists.sourceforge.net
List archive:
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
(Continue reading)

Derek Homeier | 4 Aug 22:20 2015
Picon

openmpi-1.8.7-1 build error on case-sensitive FS

Hi Jack,

the 1.8 versions of Open-MPI have a couple of new files with conflicting filenames,
causing the build on case-sensitive HFS+ to fail (and probably some of those files
to be lost on case-insensitive ones. The vtCC file versions should be added to the
case-sensitive fix, I think:

Validating .deb dir /scratch.noindex/fink.build/root-openmpi-1.8.7-1...
Error: Pathname collision on case-insensitive filesystems
	Offending file: /sw/bin/vtcc
	Offending file: /sw/share/doc/vampirtrace/vtcc-wrapper-data.txt

Cheers,
							Derek

Attachment (openmpi.info): application/octet-stream, 5429 bytes
------------------------------------------------------------------------------
_______________________________________________
Fink-devel mailing list
Fink-devel <at> lists.sourceforge.net
List archive:
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel
Alexander Hansen | 2 Aug 17:39 2015
Picon

Re: lyx2 and pre-10.9

Umm, I’ve been doing this for 14 years.  Even if I don’t always check what’s available on what
platform, I kind of know how to fix it.

> On Aug 2, 2015, at 08:33, Jack Howarth <howarth.at.fink <at> gmail.com> wrote:
> 
> Alexander,
>      The new lyx2 packaging needs to have a  Distribution: 10.9, 10.10 field added as boost1.58 doesn't exist on
the earlier OS X releases. You also need to add a version variant of the package to allow 10.7 and 10.8 to
build the previous lyx2 packaging against boost1.55 with a Distribution: 10.7, 10.8 field.
>              Jack

------------------------------------------------------------------------------
_______________________________________________
Fink-devel mailing list
Fink-devel <at> lists.sourceforge.net
List archive:
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel
Picon

Fink 0.38.99.git (TheCaptain) & 10.11 Beta-1

Greetings Fink Developers,

I wanted to share some of my experiences building Fink under El Capitan 10.11 Public Beta-1. Unlike all the
difficulties moving from 10.9 to 10.10 (due to the minor version becoming 2-digit) the bootstrap process
was painless. Most of the packages that I’ve built thus far have gone without a hitch. Here is a list of a
few specific issues I’ve run into:

0) as mentioned earlier, I got help2man-perl to build by modifying finkinfo/devel/help2man-perl.info
to include "(%type_pkg[custom] = 5182) 10.11”.

1) gcc5 built successfully using the XCode app (vs. CLI tools) with /sw/etc/fink.conf "MaxBuildJobs: 1”.

2) gcc49 did NOT build successfully; it was initially missing the .info & .patch files.

3) po4a required a finkinfo/devel/pora-10.11.info file, with "Distribution: 10.11” inside (simply
copied po4a-10.10.info & modified)

4) I forced cfitsio to build with gcc5 (since I couldn’t get gcc49 to build) by modifying
finkinfo/libs/cfitstio.info to have "BuildDepends: gcc5”

5) gmt5 built successfully when I chose unixodbc2-nox, because:

6) qt3 did NOT build successfully

7) with those workarounds, I have succeeded to build my nominal set of tools:
   cdo, curl, fftw3, gcc5, gv, gifsicle, gmt5, grib-api, lftp, nco, ncview, wget

All the best,
John

(Continue reading)

Hanspeter Niederstrasser | 6 Jul 19:46 2015

lib install_name changing from absolute path to <at> rpath

If a library changes its install_name from /absolute/path/to/foo.X.dylib
to  <at> rpath/foo.X.dylib (X is the same, c_v is bumped), does this need a new
libN package name? The file is installed to the same location, so packages
linking to the old absolute path dylib should still be able to find the
new library.

Hanspeter

--

-- 
More agile than a turtle, stronger than a mouse, nobler than a lettuce

------------------------------------------------------------------------------
Don't Limit Your Business. Reach for the Cloud.
GigeNET's Cloud Solutions provide you with the tools and support that
you need to offload your IT needs and focus on growing your business.
Configured For All Businesses. Start Your Cloud Today.
https://www.gigenetcloud.com/
_______________________________________________
Fink-devel mailing list
Fink-devel <at> lists.sourceforge.net
List archive:
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Xiaoguang Xue | 2 Jul 16:36 2015
Picon

Can not install apbs through Fink

Hi everyone!

After I update my OS X tp 10.10.4, I can not install apbs through Fink.

I got error related to CMake or fetk.

Can anyone help me?

Thanks!


This is the detail

CMake Error at CMakeLists.txt:217 (get_filename_component):

  get_filename_component called with incorrect number of arguments



CMake Warning at CMakeLists.txt:232 (message):

  Did not find fetk



-- External maloc library was found: /sw/lib/libmaloc.dylib

-- Looking for include file maloc/maloc.h

-- Looking for include file maloc/maloc.h - found

-- External maloc headers found

-- Inline functions enabled

libs: -L/sw/src/fink.build/apbs-1.4.1-2/apbs-pdb2pqr-d4e78c62e6a07f92c26924318d83a159fa6af063/apbs/lib -lapbs_geoflow-L/sw/src/fink.build/apbs-1.4.1-2/apbs-pdb2pqr-d4e78c62e6a07f92c26924318d83a159fa6af063/apbs/lib -LLIBGFORTRAN_PATH-NOTFOUND -lgfortran -ltabipb -lstdc++-lmaloc-lm

-- Adding apbs_generic

-- With source files nosh.c;mgparm.c;femparm.c;pbeparm.c;bemparm.c;geoflowparm.c;apolparm.c;vacc.c;valist.c;vatom.c;vpbe.c;vcap.c;vclist.c;vstring.c;vparam.c;vgreen.c

-- With external header files nosh.h;mgparm.h;femparm.h;pbeparm.h;bemparm.h;geoflowparm.h;apolparm.h;vacc.h;valist.h;vatom.h;vpbe.h;vcap.h;vclist.h;vstring.h;vparam.h;vgreen.h;vmatrix.h;vhal.h;vunit.h

-- With internal header files 

-- With library dependencies -L/sw/src/fink.build/apbs-1.4.1-2/apbs-pdb2pqr-d4e78c62e6a07f92c26924318d83a159fa6af063/apbs/lib -lapbs_geoflow;-L/sw/src/fink.build/apbs-1.4.1-2/apbs-pdb2pqr-d4e78c62e6a07f92c26924318d83a159fa6af063/apbs/lib -LLIBGFORTRAN_PATH-NOTFOUND -lgfortran -ltabipb -lstdc++;-lmaloc;-lm

-- Added apbs_generic

-- Adding apbs_pmgc

-- With source files buildAd.c;buildBd.c;buildGd.c;buildPd.c;cgd.c;gsd.c;matvecd.c;mgcsd.c;mgdrvd.c;mgsubd.c;mikpckd.c;mlinpckd.c;mypdec.c;newtond.c;newdrvd.c;powerd.c;smoothd.c;mgfasd.c

-- With external header files buildAd.h;buildBd.h;buildGd.h;buildPd.h;cgd.h;gsd.h;matvecd.h;mgcsd.h;mgdrvd.h;mgsubd.h;mikpckd.h;mlinpckd.h;mypdec.h;newtond.h;newdrvd.h;powerd.h;smoothd.h;mgfasd.h

-- With internal header files 

-- With library dependencies -L/sw/src/fink.build/apbs-1.4.1-2/apbs-pdb2pqr-d4e78c62e6a07f92c26924318d83a159fa6af063/apbs/lib -lapbs_geoflow;-L/sw/src/fink.build/apbs-1.4.1-2/apbs-pdb2pqr-d4e78c62e6a07f92c26924318d83a159fa6af063/apbs/lib -LLIBGFORTRAN_PATH-NOTFOUND -lgfortran -ltabipb -lstdc++;-lmaloc;-lm

-- Added apbs_pmgc

-- Adding apbs_mg

-- With source files vgrid.c;vmgrid.c;vopot.c;vpmg.c;vpmgp.c

-- With external header files vgrid.h;vmgrid.h;vopot.h;vpmg.h;vpmgp.h

-- With internal header files 

-- With library dependencies -L/sw/src/fink.build/apbs-1.4.1-2/apbs-pdb2pqr-d4e78c62e6a07f92c26924318d83a159fa6af063/apbs/lib -lapbs_geoflow;-L/sw/src/fink.build/apbs-1.4.1-2/apbs-pdb2pqr-d4e78c62e6a07f92c26924318d83a159fa6af063/apbs/lib -LLIBGFORTRAN_PATH-NOTFOUND -lgfortran -ltabipb -lstdc++;-lmaloc;-lm;apbs_generic;apbs_pmgc

-- Added apbs_mg

-- Adding apbs_geoflow

-- With source files cpbconcz2.cpp;modules.cpp;pbsolvercz.cpp;solvationsub.cpp;subfunction.cpp;surfconcz.cpp

-- With external header files cpbconcz2.h

-- With internal header files 

-- With library dependencies -L/sw/src/fink.build/apbs-1.4.1-2/apbs-pdb2pqr-d4e78c62e6a07f92c26924318d83a159fa6af063/apbs/lib -lapbs_geoflow;-L/sw/src/fink.build/apbs-1.4.1-2/apbs-pdb2pqr-d4e78c62e6a07f92c26924318d83a159fa6af063/apbs/lib -LLIBGFORTRAN_PATH-NOTFOUND -lgfortran -ltabipb -lstdc++;-lmaloc;-lm

-- Added apbs_geoflow

-- 

--  

-- -L/sw/src/fink.build/apbs-1.4.1-2/apbs-pdb2pqr-d4e78c62e6a07f92c26924318d83a159fa6af063/apbs/lib -lapbs_geoflow;-L/sw/src/fink.build/apbs-1.4.1-2/apbs-pdb2pqr-d4e78c62e6a07f92c26924318d83a159fa6af063/apbs/lib -LLIBGFORTRAN_PATH-NOTFOUND -lgfortran -ltabipb -lstdc++;-lmaloc;-lm apbs_generic;apbs_pmgc;apbs_mg;apbs_geoflow

--  

-- -std=c99

-- Supplemental tools enabled

-- Building mesh

-- libraries: apbs_generic;apbs_mg;apbs_pmgc

-- Building manip

-- libraries: apbs_generic;apbs_mg;apbs_pmgc

-- Configuring incomplete, errors occurred!

See also "/sw/src/fink.build/apbs-1.4.1-2/apbs-pdb2pqr-d4e78c62e6a07f92c26924318d83a159fa6af063/apbs/build/CMakeFiles/CMakeOutput.log".

### execution of /tmp/fink.qfgXm failed, exit code 1

### execution of /tmp/fink.fA8kI failed, exit code 1

Removing runtime build-lock...

Removing build-lock package...

/sw/bin/dpkg-lockwait -r fink-buildlock-apbs-1.4.1-2

(Reading database ... 20552 files and directories currently installed.)

Removing fink-buildlock-apbs-1.4.1-2 ...

Failed: phase compiling: apbs-1.4.1-2 failed


Before reporting any errors, please run "fink selfupdate" and try again.


If you continue to have issues, please check to see if the FAQ on Fink's 

website solves the problem.  If not, ask on one (not both, please) of

these mailing lists:


The Fink Users List <fink-users <at> lists.sourceforge.net>

The Fink Beginners List <fink-beginners <at> lists.sourceforge.net>,


with a carbon copy to the maintainer:


Jack Howarth <howarth.at.fink <at> gmail.com>


Note that this is preferable to emailing just the maintainer directly,

since most fink package maintainers do not have access to all possible

hardware and software configurations.


Please try to include the complete error message in your report.  This

generally consists of a compiler line starting with e.g. "gcc" or "g++"

followed by the actual error output from the compiler.


Also include the following system information:

Package manager version: 0.38.6

Distribution version: selfupdate-rsync Thu Jul  2 16:30:25 2015, 10.10, x86_64

Trees: local/main stable/main stable/crypto

Xcode.app: 6.4

Xcode command-line tools: 6.4.0.0.1.1435007323

apbs-1.4.1-2 is set to build with only one job.



--
Xiaoguang Xue, PhD student
Utrecht University
Crystal & Structural Chemistry
Padualaan 8. Room N807
3584 CH Utrecht
The Netherlands
Tel. +31-30-253-2383
------------------------------------------------------------------------------
Don't Limit Your Business. Reach for the Cloud.
GigeNET's Cloud Solutions provide you with the tools and support that
you need to offload your IT needs and focus on growing your business.
Configured For All Businesses. Start Your Cloud Today.
https://www.gigenetcloud.com/
_______________________________________________
Fink-devel mailing list
Fink-devel <at> lists.sourceforge.net
List archive:
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel
Jack Howarth | 29 Jun 18:24 2015
Picon

stricter BuildDepends/ Depends for ghc libs

Brendan,
     I've checked in another set of changes for the ghc library packages in libs/ghc which have Shlibs entries. Since these packages link against the shared libs in the ghc package using versioned install names, the addition of...

ghc (>= 7.8.3-1), ghc (<< 7.8.4-1)

to both the BuildDepends and Depends will insure that they both build and run against the proper ghc version.
      I'm not certain yet how cleanly this will work when the ghc is upgraded to ghc 7.10.1. It may be that ghc needs to be refactored to behave like the r-base and r-mods packages which can co-exist. The current situation (prior to my changes today) would allow fink to upgrade ghc and then proceed to upgrade the packages in libs/ghc without considering that the installed packages previously built against the older ghc will no longer have the required supporting shared libraries from the older ghc.
       This explains why upgrading ghc has always been such a rough proposition and I have almost always had 'fink remove --recursive ghc' to purge the previous installation of ghc and its ghc lib packages. We simply weren't considering the breakage caused to  installed ghc lib packages built against the older ghc upon the installation of the newer ghc.
          Jack
ps As I mentioned on fink-seed, the rootless feature will either require the current ghc 7.8.3 to be rebuilt from source (using a gcc-bootstrap package as MacPorts does) with their patch...


or ghc upgraded to the new 7.10.1 release which contains this change.
------------------------------------------------------------------------------
Monitor 25 network devices or servers for free with OpManager!
OpManager is web-based network management software that monitors 
network devices and physical & virtual servers, alerts via email & sms 
for fault. Monitor 25 devices for free with no restriction. Download now
http://ad.doubleclick.net/ddm/clk/292181274;119417398;o
_______________________________________________
Fink-devel mailing list
Fink-devel <at> lists.sourceforge.net
List archive:
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel
Hanspeter Niederstrasser | 27 Jun 14:17 2015

wxwidgets300-* hardcodes flag-sort into compiler identification

 From /sw/lib/wxwidgets300-gtk2/wx/config/gtk2-unicode-3.0:

[ -z "$output_option_cc"            ] || echo "flag-sort -r gcc"
[ -z "$output_option_cxx"           ] || echo "flag-sort -r g++"
[ -z "$output_option_ld"            ] || echo "flag-sort -r g++ -o"

$ wx-config --cc
flag-sort -r gcc

This then shows up in ./configure outputs with results like:
configure:3891: checking Checking wxWidgets for min version:
configure:3895: result: -r gcc

which messes up subsequent configure tests.

wxwidgets300-osxcocoa has a similar issue.

Hanspeter

------------------------------------------------------------------------------
Monitor 25 network devices or servers for free with OpManager!
OpManager is web-based network management software that monitors 
network devices and physical & virtual servers, alerts via email & sms 
for fault. Monitor 25 devices for free with no restriction. Download now
http://ad.doubleclick.net/ddm/clk/292181274;119417398;o
_______________________________________________
Fink-devel mailing list
Fink-devel <at> lists.sourceforge.net
List archive:
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel

Max Horn | 25 Jun 12:39 2015
Picon

SDL and X11

Hi there,

so Jack sent me a patch for sdl.info to avoid use of the X11 convenience symlinks, see e.g. here:

https://sourceforge.net/p/fink/package-submissions/4507/

However, that patch isn't quite enough. It would result in an unusable SDL on systems without the
convenience links, since, if X11 support is enabled in SDL
 (a) sdl.h always pulls X11 headers, even if the application using SDL is not using the X11 video driver,
 (b) "sdl-config --cflags" does not include something like -I/opt/X11/include

Result: You can't compile code using SDL, at least not without manually adjusting the CFLAGS.

I see two possible solutions:

(1) Just disable the X11 video driver, under the assumption that everyone is using the native Quartz driver
these days anyway.

(2) Modify sdl-config to point to /opt/X11

Option (1) is a lot easier and also reduces the dependencies for anything using SDL. But I am not sure if the
underlying assumption is right. I.e. are there packages which rely on the SDL X11 video driver? Does
anybody know?

If not, I can have a look at option (2). It certainly is "safer", but not nicer.

An additional option would be to provide a -nox variant of sdl / sdl-shlibs, and start migrating packages to
using that.

Cheers,
Max
------------------------------------------------------------------------------
Monitor 25 network devices or servers for free with OpManager!
OpManager is web-based network management software that monitors 
network devices and physical & virtual servers, alerts via email & sms 
for fault. Monitor 25 devices for free with no restriction. Download now
http://ad.doubleclick.net/ddm/clk/292181274;119417398;o
_______________________________________________
Fink-devel mailing list
Fink-devel <at> lists.sourceforge.net
List archive:
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel


Gmane