Re: [Fink-users] Problems with installing certain packages (64 bit) on Mac OS 10.6
Hi Claus, I've just run into the same troubles with inkscape... On 20 Sep 2009, at 13:52, Claus Atzenbeck wrote: ** inkscape * quartz_font.cc: In constructor 'Cairo::QuartzFontFace::QuartzFontFace(ATSUFontID)': quartz_font.cc:34: error: 'cairo_quartz_font_face_create_for_atsu_font_id' was not declared in this scope make[3]: *** [quartz_font.lo] Error 1 make[2]: *** [all-recursive] Error 1 make[1]: *** [all] Error 2 make: *** [all-recursive] Error 1 ### execution of make failed, exit code 2 this is actually a failure in building cairomm1, right? I've found a patch for disabling that font definition in 64bit-builds, which is apparently missing in Snow Leopard (originally for cairo-1.6.4, from http://lists.cairographics.org/archives/cairo-bugs/2008-April/002240.html) , and tried to apply it to cairomm1-1.8.2. This allows cairomm1 and inkscape to build - it is running, but that's about all that I tested, so proceed at your own risk... Cheers, Derek cairomm1.info Description: Binary data cairomm1.patch Description: Binary data -- 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-users mailing list Fink-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] matplotlib-py26 install trouble.
Hi, I only came upon this thread after having some trouble with matplotlib myself... On 18 Sep 2009, at 22:44, John K. Parejko wrote: On Wed, Sep 16, 2009 at 02:09:05PM -0400, John K. Parejko wrote: I'm trying to install the new version of matplotlib-py26, but I get the following error about it trying to overwrite files that belong to dateutil: I can't reproduce this on my 10.4/ppc machine with unstable activated (dateutil-py26-1.4-1 matplotlib-py26-0.99.0-1) and with all packages built locally by fink (vs downloaded precompiled .deb). After some digging, I remembered that I'd built a pre-release of matplotlib 0.98.something so that I could use some of the new 0.98.x features. When I did the build, I used the SVN version with "python setupegg.py develop" to make following SVN changes easier. This sets up some symlinks and references in /site-packages back to the build directory. I removed fink's python2.6 and all *-py26 packages and matplotlib now installs cleanly and seems to work fine. Let this be a warning to those who try to run on the bleeding edge! Just wondering - by this do you mean you removed all the fink python2.6-libraries and reinstalled them and afterwards succeeded in building matplotlib- py26, or did you then compile matplotlib independently of fink? The latter does not really sound like a solution to the fink package's problem... What I found, also perhaps trying to be more on the bleeding edge, was that matplotlib 0.99 and in particular 0.99.1 introduced a number of new configuration settings, among others setting a default to try to install its own pytz and dateutil in setup.cfg. I managed to put a working package for 0.99.1.1 afterwards, attached below. On this occasion, I am wondering if sending updated package descriptions to the list, as I've done before, is the appropriate way, or whether I should rather send them directly to the maintainer or simply be content to wait until they've checked in their own updates or... Whatever the correct upgrade path, I'd also like to make a suggestion at this point: Is it possible to remove the numeric-py and numarray-py dependencies, perhaps at least make them "Suggests:" or "Enhances:" instead? They are clearly no longer required to build or use matplotlib (in fact I am not even sure if they are still officially supported as backends, since that configuration option has been removed now), and I find it rather painful to still being forced to keep two long outdated libraries installed. I'd think anyone who still needs either of them for legacy routines or data would know so on their own - just my 2d. Cheers, Derek matplotlib-py.info Description: Binary data matplotlib-py.patch Description: Binary data -- 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-users mailing list Fink-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] matplotlib-py26 install trouble.
Hi,
Derek Homeier wrote:
>
> On this occasion, I am wondering if sending updated package
> descriptions
> to the
> list, as I've done before, is the appropriate way, or whether I should
> rather send
> them directly to the maintainer or simply be content to wait until
> they've checked in
> their own updates or...
>
> Whatever the correct upgrade path, I'd also like to make a
> suggestion at
> this point:
> Is it possible to remove the numeric-py and numarray-py dependencies,
> perhaps at
> least make them "Suggests:" or "Enhances:" instead? They are clearly
> no
> longer required
> to build or use matplotlib (in fact I am not even sure if they are
> still
> officially supported as
> backends, since that configuration option has been removed now), and I
> find it rather
> painful to still being forced to keep two long outdated libraries
> installed.
> I'd think anyone who still needs either of them for legacy routines or
> data would know
> so on their own - just my 2d.
On Oct 4, 2009, at 4:17 AM, Alexander Hansen wrote:
> You should consider sending your package description to the maintainer
> of the matplotlib package ("fink info matplotlib") because he would be
> the one to make any changes.
On Oct 5, 2009, at 5:18 AM, John K. Parejko wrote:
> No, I re-installed matplotlib-py26 from within fink. The problem was
> apparently due to the older libraries that were still being linked to.
>
> I would also support the removal of numeric and numarray dependencies.
yes, I guess I assumed Kurt Schwehr might notice anyway, since he was
posting in
this thread earlier, but I am CCing him directly now - probably safer
than hoping he
constantly monitors the list for all packages he's maintaining.
Thanks,
Derek
--
-
Derek Homeier Institut für Astrophysik Göttingen
Georg-August-Universität Phone: +49 551 39-7980
Friedrich-Hund-Platz 1Fax: +49 551 39-5043
D-37077 Göttingen, GermanyFeet: E.04.104
Web: http://www.astro.physik.uni-goettingen.de/~derek/
-
--
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-users mailing list
Fink-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] mmap and Python on Fink
Hi Alexander, Buz, I can confirm that this appears to be an issue with python2.5 on 10.6, and it's apparently an installation problem on the python side (64bit-installation in my case). > find /System/Library/Frameworks/Python.framework -name mmap* /System/Library/Frameworks/Python.framework/Versions/2.3/lib/python2.3/lib-dynload/mmap.so /System/Library/Frameworks/Python.framework/Versions/2.5/lib/python2.5/lib-dynload/mmap.so /System/Library/Frameworks/Python.framework/Versions/2.6/lib/python2.6/lib-dynload/mmap.so > find /sw/lib/python2.6 -name mmap* /sw/lib/python2.6/lib-dynload/mmap.so > find /sw/lib/python2.5 -name mmap* - nothing there. numpy.test() also fails on importing mmap. I could not track this much further down, embarrassingly I could not even reproduce how I had installed python2.5 in the first place, as there was a python25_2.5.4-102 installed, but the package descriptions both in stable and unstable state: Package: python25 Version: 2.5.4 Revision: 1 Epoch: 1 Distribution: 10.4, 10.5 (!) Nb. this is a distribution bootstrapped on Snow Leopard in 64bit, never updated from 10.5. I tried to investigate a bit further, so I updated the package in my local tree to 2.5.5 and, noticing that one difference between the python25 and python26 packages was the ConfigureParams: --enable-shared field, I added this flag to the python25 package as well. python-2.5.5 builds that way and appears to work (is there a generic .test() for the python interpreter itself?), but still does not install mmap.so. I am out or my depth here for now, so I am passing this one to whoever else might have another suggestion. HTH, Derek Python 2.5.4 (r254:67916, Apr 22 2010, 12:57:40) Type "copyright", "credits" or "license" for more information. IPython 0.8.2 -- An enhanced Interactive Python. ? -> Introduction and overview of IPython's features. %quickref -> Quick reference. help -> Python's own help system. object? -> Details about 'object'. ?object also works, ?? prints more. In [1]: import scipy In [2]: from scipy.io.array_import import read_array --- ImportError Traceback (most recent call last) /Users/buz/Documents/GTD/GTD Supporting Information/Work/Make an O2-tolerant Hydrogenase/Sequence 2_12_2/Homology Model/ in () /sw/lib/python2.5/site-packages/scipy/io/__init__.py in () 78 79 # netCDF file support ---> 80 from netcdf import netcdf_file, netcdf_variable 81 82 from recaster import sctype_attributes, Recaster /sw/lib/python2.5/site-packages/scipy/io/netcdf.py in () 16 17 import struct ---> 18 import mmap 19 20 from numpy import ndarray, zeros, array ImportError: No module named mmap In [3]: >>> -- ___ Fink-users mailing list Fink-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] issue with numpy-py26 1.3.0-7
Hi,
>> Developing our application here, based on python, and we start to see
>> this error:
>>
>> ValueError: can't handle version -1 of numpy.dtype pickle
>>
>> My partner uses MacPorts and solved this issue by updating numpy to 1.4.
>> I was wondering if numpy 1.4 for Fink is on the way.
>>
>> Thanks,
>>
>> Alan
>>
>> --
>> Alan Wilter S. da Silva, D.Sc. - CCPN Research Associate
>> Department of Biochemistry, University of Cambridge.
>> 80 Tennis Court Road, Cambridge CB2 1GA, UK.
http://www.bio.cam.ac.uk/~awd28<<
>>
>>
>>
>
> You might try checking with the maintainer ("fink info numpy-py26")
> since he would be the one who would be responsible for handling an update.
I have built and tested the updated numpy and scipy versions on my machines,
feel free to copy the adapted package files to your
/sw/fink/dists/local/main/finkinfo/sci
and give them a try.
Please note that I just modified them on my own, and Sebastien Maret, the
maintainer,
is not responsible for any problems you might encounter (though basically,
numpy.test() and scipy.test() did not produce any errors not already present in
the 1.3 and 0.7.0 versions on 10.5/6, python2.5/6, powerpc/i386/x86_64).
Cheers,
Derek
numpy-py.info
Description: Binary data
numpy-py.patch
Description: Binary data
scipy-py.info
Description: Binary data
--
___
Fink-users mailing list
Fink-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] mmap and Python on Fink
Hi Dan, On 14.05.2010, at 5:11AM, Daniel Macks wrote: > Both platforms' autodetect and determine to build the mmap module, but > on 10.6 the actual compiling fails. Compare starting at the line: > > building 'mmap' extension > [compiler command looks about the same] > [warnings about winsize look the same] > [on 10.6 but not on 10.5 the following error...] > /sw32/src/fink.build/python25-2.5.4-102/Python-2.5.4/Modules/mmapmodule.c: > In function 'new_mmap_object': > > /sw32/src/fink.build/python25-2.5.4-102/Python-2.5.4/Modules/mmapmodule.c:919: > warning: implicit declaration of function 'open' > > /sw32/src/fink.build/python25-2.5.4-102/Python-2.5.4/Modules/mmapmodule.c:919: > error: 'O_RDWR' undeclared (first use in this function) > > /sw32/src/fink.build/python25-2.5.4-102/Python-2.5.4/Modules/mmapmodule.c:919: > error: (Each undeclared identifier is reported only once > > /sw32/src/fink.build/python25-2.5.4-102/Python-2.5.4/Modules/mmapmodule.c:919: > error: for each function it appears in.) > > Like it's missing #include or something. Probably some > unrelated platform difference triggers some difference in #include > spaghetti somewhere? See if you can patch mmapmodule.c to add fcntl.h > > I vaguely remember other cases where python-module packages where the > standard build system don't abort when a fatal compiler error occurs. > Makes debugging/packaging "interesting":( thanks for the pointer, you're quite right! Patching mmapmodule.c as you suggested worked. I am attaching the updated info and patch files (including the update to 2.5.5, as this is only a security fix release, there don't seem to be many other changes). Checking the build log revealed three other errors bypassed in the same way: ctypes, bsddb and resource are broken as well (all three do work in python2.5 on 10.5 and python2.6 on 10.6. ctypes and resource also work in OS X's /usr/bin/python2.5, but bsddb doesn't seem to be enabled - there probably is no Berkeley DB included in the system to begin with...) For the moment, mmap at least would work, including scipy read_array. The rest is above my head for now, I am afraid. Cheers, Derek python25-10.6.info Description: Binary data python25-10.6.patch Description: Binary data 1: building '_ctypes' extension creating build/temp.macosx-10.6-i386-2.5/scratch.noindex/fink.build/python25-2.5.5-2/Python-2.5.5/Modules/_ctypes creating build/temp.macosx-10.6-i386-2.5/scratch.noindex/fink.build/python25-2.5.5-2/Python-2.5.5/Modules/_ctypes/darwincreating build/temp.macosx-10.6-i386-2.5/scratch.noindex/fink.build/python25-2.5.5-2/Python-2.5.5/Modules/_ctypes/libffi creating build/temp.macosx-10.6-i386-2.5/scratch.noindex/fink.build/python25-2.5.5-2/Python-2.5.5/Modules/_ctypes/libffi/src creating build/temp.macosx-10.6-i386-2.5/scratch.noindex/fink.build/python25-2.5.5-2/Python-2.5.5/Modules/_ctypes/libffi/src/x86 creating build/temp.macosx-10.6-i386-2.5/scratch.noindex/fink.build/python25-2.5.5-2/Python-2.5.5/Modules/_ctypes/libffi/src/powerpc gcc -fno-strict-aliasing -mno-fused-madd -fno-common -DNDEBUG -g -fwrapv -O3 -Wall -Wstrict-prototypes -I_ctypes/darwin -I. -I/scratch.noindex /fink.build/python25-2.5.5-2/Python-2.5.5/./Include -Ibuild/temp.macosx-10.6-i386-2.5/libffi/include -Ibuild/temp.macosx-10.6-i386-2.5/libffi -I/scratch.noindex/fink.build/python25-2.5.5-2/Python-2.5.5/Modules/_ctypes/libffi/src -I. -IInclude -I./Include -I/sw/include -I/scratch.noin dex/fink.build/python25-2.5.5-2/Python-2.5.5/Include -I/scratch.noindex/fink.build/python25-2.5.5-2/Python-2.5.5 -c /scratch.noindex/fink.build/python25-2.5.5-2/Python-2.5.5/Modules/_ctypes/_ctypes.c -o build/temp.macosx-10.6-i386-2.5/scratch.noindex/fink.build/python25-2.5.5-2/Pytho n-2.5.5/Modules/_ctypes/_ctypes.o In file included from /scratch.noindex/fink.build/python25-2.5.5-2/Python-2.5.5/./Include/Python.h:57, from /scratch.noindex/fink.build/python25-2.5.5-2/Python-2.5.5/Modules/_ctypes/_ctypes.c:107: /scratch.noindex/fink.build/python25-2.5.5-2/Python-2.5.5/./Include/pyport.h:547: warning: 'struct winsize' declared inside parameter list /scratch.noindex/fink.build/python25-2.5.5-2/Python-2.5.5/./Include/pyport.h:547: warning: its scope is only this definition or declaration, which is probably not what you want /scratch.noindex/fink.build/python25-2.5.5-2/Python-2.5.5/./Include/pyport.h:548: warning: 'struct winsize' declared inside parameter list In file included from /scratch.noindex/fink.build/python25-2.5.5-2/Python-2.5.5/Modules/_ctypes/_ctypes.c:110: build/temp.macosx-10.6-i386-2.5/libffi/include/ffi.h:161: error: expected specifier-qualifier-list before 'ffi_abi' build/temp.macosx-10.6-i386-2.5/libffi/include/ffi.h:183: error: expected specifier-qualifier-list before 'ffi_sarg'build/temp.macosx-10.6-i386-2.5/libffi/include/ffi.h:267: error: expected declaration specifiers or '...' bef
Re: [Fink-users] trouble using tcptraceroute
On Tuesday, May 27, 2003, at 05:41 PM, Justin C. Walker wrote: In fact, I'm even more puzzled. While tcptraceroute, compiled as above, gets to the point of sending TTL-limited packets, it seems to time-out on each one. Looking at the network with tcpdump, I see the ICMP packets returning as expected, but tcptraceroute doesn't seem to see them. Perhaps it's a 'libnet' problem. Has anyone had success with either tcptraceroute or libnet-1.0.2a in Fink? I'm on Mac OS X, 10.2.6. The old version (1.2-3) seems to be broken, too (ww/ libnet 1.1.0-3), since it gets lost even on tracing a machine on the same switch: Warning: Could not determine appropriate device; resorting to pcap_lookupdev() Selected device en0, address 128.192.19.18, port 58313 for outgoing packets Tracing the path to homer (128.192.19.12) on TCP port 80 (http), 30 hops max 1 * * * 2 * * * After purging libnet and updating to tcptraceroute 1.4-2, it fails just as Joel described. Cheers, Derek --- This SF.net email is sponsored by: Etnus, makers of TotalView, The best thread debugger on the planet. Designed with thread debugging features you've never dreamed of, try TotalView 6 free at www.etnus.com. ___ Fink-users mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-users
[Fink-users] PGPLOT
Hi, has anyone been able to link to the fink-pgplot library with the current g77? I have g77-3.3-2, and it does not even build the cpgdemo when I rebuild pgplot. First of all, the linker flags obviously need to be changed to contain "-Xlinker -framework -Xlinker Foundation -Xlinker -framework -Xlinker AppKit", but with this it still returns a /usr/bin/ld: Undefined symbols: restFP saveFP Any idea where to find these? Thanks, Derek --- This SF.Net email sponsored by: Free pre-built ASP.NET sites including Data Reports, E-commerce, Portals, and Forums are available now. Download today and enter to win an XBOX or Visual Studio .NET. http://aspnet.click-url.com/go/psa0016ave/direct;at.asp_061203_01/01 ___ Fink-users mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] PGPLOT
On Wednesday, July 2, 2003, at 07:29 PM, Martin Costabel wrote: /usr/bin/ld: Undefined symbols: restFP saveFP Any idea where to find these? They are in /usr/lib/libgcc.a. Usually when this bug comes up, people have some libgcc.dylib in /usr/local/lib that shadows the libgcc.a. The usual suspect, eh? However, I have no such thing, instead g77 seems to come with its own: find /sw/lib -name libgcc* /sw/lib/gcc-lib/powerpc-apple-darwin6.6/3.3/libgcc.a /sw/lib/gcc-lib/powerpc-apple-darwin6.6/3.3/soft-float/libgcc.a The full linker command is /sw/lib/gcc-lib/powerpc-apple-darwin6.6/3.3/collect2 -dynamic -o pgdemo1 -lcrt1.o -lcrt2.o -L/Volumes/scratch/derek/BUILD/pgplot -L/sw/lib -L/usr/X11R6/lib -L/sw/lib/gcc-lib/powerpc-apple-darwin6.6/3.3 -L/sw/lib/gcc-lib/powerpc-apple-darwin6.6/3.3/../../.. /scratch/bart/derek/tmp/ccvATHxQ.o -lpgplot -lpng -lz -lX11 -framework Foundation -framework AppKit -lfrtbegin -lg2c -lgcc -lSystem | c++filt3 /usr/bin/ld: Undefined symbols: restFP saveFP I do not have the latest Devtools update installed; since fink is not working with gcc3.3 yet, according to the website, I wonder how g77_3.3 is supposed to work at all... FWIW, after downgrading to g77_3.1-20020420-6 this works just fine! Thanks, Derek --- This SF.Net email sponsored by: Free pre-built ASP.NET sites including Data Reports, E-commerce, Portals, and Forums are available now. Download today and enter to win an XBOX or Visual Studio .NET. http://aspnet.click-url.com/go/psa0016ave/direct;at.asp_061203_01/01 ___ Fink-users mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] PGPLOT
On Thursday, July 3, 2003, at 07:51 AM, [EMAIL PROTECTED] wrote: The workaround for this is to add -lcc_dynamic to the g77 link line. Thanks, yes, that made it work. I do not have the latest Devtools update installed; since fink is not working with gcc3.3 yet, according to the website, I wonder how g77_3.3 is supposed to work at all... It works fine with this fix. In fact, better in may respects (it works if MACOS_DEPLOYMENT_TARGET is set to 10.2, and it compiles more of the testsuite). One other problem which you will face linking PGPLOT is that g77 doesn't understand -framework. To link a framework (such as vecLib or Foundation), use cc as the linker. Well, I just added the -Xlinker flags as in my original post. All the pgdemos seemed to compile, and basically work, too, but there is another problem (feature?) with the current xwin driver: the plot window is only updated when it's focused, which makes the "animated" features (e.g. pgdemo11) rather pointless. With the aquaterm driver this works, but it seems to be extremely slow! Thanks, Derek --- This SF.Net email sponsored by: Free pre-built ASP.NET sites including Data Reports, E-commerce, Portals, and Forums are available now. Download today and enter to win an XBOX or Visual Studio .NET. http://aspnet.click-url.com/go/psa0016ave/direct;at.asp_061203_01/01 ___ Fink-users mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-users
[Fink-users] xephem compile fails with BuildJobs>1
Hi, building xephem-3.7.5-2 failed on a 4-core iMac running Mountain Lion due to concurrent make tasks. Adding UseMaxBuildJobs: False solved the problem for me. Cheers, Derek -- Got visibility? Most devs has no idea what their production app looks like. Find out how fast your code is with AppDynamics Lite. http://ad.doubleclick.net/clk;262219671;13503038;y? http://info.appdynamics.com/FreeJavaPerformanceDownload.html ___ Fink-users mailing list Fink-users@lists.sourceforge.net List archive: http://news.gmane.org/gmane.os.macosx.fink.user Subscription management: https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] unresolved dependency
> For Xcode versions before Apple decided that the command line tools > should be separate, a versioned dependency on "xcode" generally meant > that a package was looking for a fixed command-line element, e.g. a fix > to the linker, rather than anything referring to the Xcode app. > > To be consistent with past practices, we decided that "xcode" should > refer to the command-line tools in Xcodes with separate command-line > packages. Admittedly, this can be a little confusing. > > And unfortunately I'm pretty sure we're not allowed to repackage Xcode > and/or the command line tools so that Fink can install them. I just ran into a very similar problem after updating to Xcode 4.5. Noting that fink list showed xcode.app as version 4.5.1, while xcode was still at 4.4.0 allowed me to realise that the command line tools had not been updated along with the main app (which really is extremely annoying - Apple might have a right to think not everyone using Xcode would want or need the command line tools; but if they ARE already selected and installed, to omit them from the update is just utterly stupid!). However even though I had been through the whole process finding my way to go to the preferences and select the right tab etcetc. took a while, and I think one can easily get lost there… Even if fink certainly cannot install the tools on its own, perhaps it would help to include the information given by Martin to xcode.info as something like This package represents the C/C++/ObjC developer tools provided by Apple. If it does not show as installed, you can download Xcode (for OS 10.5, 10.6, and 10.7) or the Command Line Tools For Xcode (OS 10.7 or later) from Apple at:… If you have already installed the Xcode toolset (xcode.app), these Tools are also embedded within the Xcode IDE and can be installed using the Downloads pane within the Xcode Preferences. You need to repeat this step after every update of the Xcode toolset. [Ideally, `fink list xcode` should perhaps show xcode as outdated if xcode.app is installed in a newer version, but I don't know how feasible that is… Even after updating the command line tools, xcode still shows up as 4.5.0.0.1…. vs. xcode.app 4.5.1] Cheers, Derek -- Don't let slow site performance ruin your business. Deploy New Relic APM Deploy New Relic app performance management and know exactly what is happening inside your Ruby, Python, PHP, Java, and .NET app Try New Relic at no cost today and get our sweet Data Nerd shirt too! http://p.sf.net/sfu/newrelic-dev2dev ___ Fink-users mailing list Fink-users@lists.sourceforge.net List archive: http://news.gmane.org/gmane.os.macosx.fink.user Subscription management: https://lists.sourceforge.net/lists/listinfo/fink-users
[Fink-users] pil-py*-1.7.8 install failure on case-sensitive filesystems
Hi, see subject - the documentation installation step fails cp -R Docs/ /scratch.noindex/fink.build/root-pil-py26-1.7.8-1/sw/share/doc/pil-py26/html cp: Docs/: No such file or directory ### execution of /tmp/fink.hYPWQ failed, exit code 1 because in Pillow-1.7.8 the directory is called [lowercase] docs. TIA, Derek -- Symantec Endpoint Protection 12 positioned as A LEADER in The Forrester Wave(TM): Endpoint Security, Q1 2013 and "remains a good choice" in the endpoint security space. For insight on selecting the right partner to tackle endpoint security challenges, access the full report. http://p.sf.net/sfu/symantec-dev2dev ___ Fink-users mailing list Fink-users@lists.sourceforge.net List archive: http://news.gmane.org/gmane.os.macosx.fink.user Subscription management: https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] pil-py*-1.7.8 install failure on case-sensitive filesystems
On 12.03.2013, at 3:17AM, Daniel Johnson wrote: > You must have a case sensitive file system. It should be fixed now. indeed I do (see subject again ;-). Rsync'ed 1.7.8-1 (still the same revision) again, and it builds here now. Thanks, Derek -- Everyone hates slow websites. So do we. Make your web apps faster with AppDynamics Download AppDynamics Lite for free today: http://p.sf.net/sfu/appdyn_d2d_mar ___ Fink-users mailing list Fink-users@lists.sourceforge.net List archive: http://news.gmane.org/gmane.os.macosx.fink.user Subscription management: https://lists.sourceforge.net/lists/listinfo/fink-users
[Fink-users] libzvbi0 install failure on case-sensitive filesystems
Hi, see subject, libzvbi0-shlibs-0.2.35-2 fails on installing a DocFile /bin/cp -r Changelog /scratch.noindex/fink.build/root-libzvbi0-shlibs-0.2.35-2/sw/share/doc/libzvbi0-shlibs/ cp: Changelog: No such file or directory The actual filename is ChangeLog (same for all Splitoffs). Cheers, Derek -- Android is increasing in popularity, but the open development platform that developers love is also attractive to malware creators. Download this white paper to learn more about secure code signing practices that can help keep Android apps secure. http://pubads.g.doubleclick.net/gampad/clk?id=65839951&iu=/4140/ostg.clktrk ___ Fink-users mailing list Fink-users@lists.sourceforge.net List archive: http://news.gmane.org/gmane.os.macosx.fink.user Subscription management: https://lists.sourceforge.net/lists/listinfo/fink-users
[Fink-users] bzr-git-py 0.6.11 signature error
Hi, the bzr-git source has an incorrect MD5SUM and appears to be missing from each and every mirror for weeks (seems this package is not really used a lot…): curl --connect-timeout 30 -f -L -A 'fink/0.36.0' -O http://distfiles.sjc.ca.us.finkmirrors.net/bzr-git-0.6.11.tar.gz % Total% Received % Xferd Average Speed TimeTime Time Current Dload Upload Total SpentLeft Speed 0 00 00 0 0 0 --:--:-- --:--:-- --:--:-- 0 curl: (22) The requested URL returned error: 404 Not Found ### execution of curl failed, exit code 22 Downloading the file "bzr-git-0.6.11.tar.gz" failed. (1) Give up (2) Retry the same mirror (3) Retry another mirror from your continent (4) Retry another mirror (5) Retry using original source URL Default answer will be chosen in 120 seconds... How do you want to proceed? [5] curl --connect-timeout 30 -f -L -A 'fink/0.36.0' -O http://samba.org/\~jelmer/bzr/bzr-git-0.6.11.tar.gz % Total% Received % Xferd Average Speed TimeTime Time Current Dload Upload Total SpentLeft Speed 100 329 100 3290 0903 0 --:--:-- --:--:-- --:--:-- 903 100 117k 100 117k0 0 66317 0 0:00:01 0:00:01 --:--:-- 146k The checksum of the file is incorrect. The most likely cause for this is a corrupted or incomplete download Expected: 8d55b1a6bfaf23394dc468c6a4014c90 Actual: MD5(105eb8c52dca00e3eaf27a45929e4fab) SHA1(211f0ee32ee30a96c8be9f09432ac66b24eb597c) Downloading the file "bzr-git-0.6.11.tar.gz" failed. When changing the MD5 manually, it builds on 10.8, but with a couple more than the mentioned 3 test failures: bzr selftest: /sw/bin/bzr-py27 /sw/lib/python2.7/site-packages/bzrlib bzr-2.6.0 python-2.7.6 Darwin-12.5.0-x86_64-i386-64bit ... -- Ran 244 tests in 5.233s FAILED (errors=15, known_failure_count=1) 1 test skipped Missing feature 'Missing tdb' skipped 6 tests. There is a version 0.6.12 on the author's page (md5sum aa4dcd4fa03631c4802bc7795af7932b), which fares a bit better: -- Ran 244 tests in 6.009s FAILED (errors=3, known_failure_count=2) 1 test skipped Missing feature 'Missing tdb' skipped 6 tests. Cheers, Derek -- DreamFactory - Open Source REST & JSON Services for HTML5 & Native Apps OAuth, Users, Roles, SQL, NoSQL, BLOB Storage and External API Access Free app hosting. Or install the open source package on any LAMP server. Sign up and see examples for AngularJS, jQuery, Sencha Touch and Native! http://pubads.g.doubleclick.net/gampad/clk?id=63469471&iu=/4140/ostg.clktrk ___ Fink-users mailing list Fink-users@lists.sourceforge.net List archive: http://news.gmane.org/gmane.os.macosx.fink.user Subscription management: https://lists.sourceforge.net/lists/listinfo/fink-users
[Fink-users] numpy-py31 update failure
Hi,
last bug report for the day, I think. numpy 1.8 no longer supports Python3.1:
/sw/bin/python3.1 setup.py build
Traceback (most recent call last):
File "setup.py", line 27, in
raise RuntimeError("Python version 2.6, 2.7 or >= 3.2 required.")
RuntimeError: Python version 2.6, 2.7 or >= 3.2 required.
### execution of /sw/bin/python3.1 failed, exit code 1
A Python3.3 version can be built without any other changes, so the Type: python
()
could simply be updated (and numpy 1.7.2 will still support 2.4-3.1).
Cheers,
Derek
--
Shape the Mobile Experience: Free Subscription
Software experts and developers: Be at the forefront of tech innovation.
Intel(R) Software Adrenaline delivers strategic insight and game-changing
conversations that shape the rapidly evolving mobile landscape. Sign up now.
http://pubads.g.doubleclick.net/gampad/clk?id=63431311&iu=/4140/ostg.clktrk
___
Fink-users mailing list
Fink-users@lists.sourceforge.net
List archive:
http://news.gmane.org/gmane.os.macosx.fink.user
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] bzr-git-py 0.6.11 signature error
On 18.11.2013, at 11:51PM, Daniel Johnson wrote: > Ok, version 0.6.12 builds for me and I just checked it in. The 3 test errors > are the same ones I've seen for a while. I think it's actually a mistake in > the test. It doesn't set a committer name during the test and newer versions > of Dulwich don't like that. Just ignore it. As I mentioned, I got a lot more errors with 0.6.11, but I could check out 0.6.12 now (after learning that only one selfupdate mirror is working right now), same 3 errors here. Thanks, Derek -- Shape the Mobile Experience: Free Subscription Software experts and developers: Be at the forefront of tech innovation. Intel(R) Software Adrenaline delivers strategic insight and game-changing conversations that shape the rapidly evolving mobile landscape. Sign up now. http://pubads.g.doubleclick.net/gampad/clk?id=63431311&iu=/4140/ostg.clktrk ___ Fink-users mailing list Fink-users@lists.sourceforge.net List archive: http://news.gmane.org/gmane.os.macosx.fink.user Subscription management: https://lists.sourceforge.net/lists/listinfo/fink-users
[Fink-users] tree-dag-node-pm-1.18-1 failure on case-sensitive FS
Hi, little spelling problem here; the package's Docfile is capitalised "Changes“, thus not found on a case-sensitive system: /bin/cp -r CHANGES /scratch.noindex/fink.build/root-tree-dag-node-pm-1.18-1/sw/share/doc/tree-dag-node-pm/ cp: CHANGES: No such file or directory ### execution of /bin/cp failed, exit code 1 ### execution of /tmp/fink.LYK1S failed, exit code 1 Cheers, Derek -- CenturyLink Cloud: The Leader in Enterprise Cloud Services. Learn Why More Businesses Are Choosing CenturyLink Cloud For Critical Workloads, Development Environments & Everything In Between. Get a Quote or Start a Free Trial Today. http://pubads.g.doubleclick.net/gampad/clk?id=119420431&iu=/4140/ostg.clktrk ___ Fink-users mailing list Fink-users@lists.sourceforge.net List archive: http://news.gmane.org/gmane.os.macosx.fink.user Subscription management: https://lists.sourceforge.net/lists/listinfo/fink-users
[Fink-users] qt3 Error: Pathname collision on case-insensitive filesystems
Hi again, this time a case of case-insensitive problems… nothing critical, I can install the package, but building with ‘-m’ fails on validating all manpages (on case-sensitive HFS+): Validating .deb dir /scratch.noindex/fink.build/root-qt3-3.3.8-2002... Error: Pathname collision on case-insensitive filesystems Offending file: /sw/share/man/man3/qaccel.3qt Offending file: /sw/share/man/man3/qaccessible.3qt Offending file: /sw/share/man/man3/qaccessibleinterface.3qt Offending file: /sw/share/man/man3/qaccessibleobject.3qt Offending file: /sw/share/man/man3/qaction.3qt ... Offending file: /sw/share/man/man3/qxmlreader.3qt Offending file: /sw/share/man/man3/qxmlsimplereader.3qt The files are all installed with capitalised names (QAccel.3pt, …, QXmlSimpleReader.3qt); don’t know why dpkg thinks they should be lowercase, or there would be files of conflicting name around. Cheers, Derek -- CenturyLink Cloud: The Leader in Enterprise Cloud Services. Learn Why More Businesses Are Choosing CenturyLink Cloud For Critical Workloads, Development Environments & Everything In Between. Get a Quote or Start a Free Trial Today. http://pubads.g.doubleclick.net/gampad/clk?id=119420431&iu=/4140/ostg.clktrk ___ Fink-users mailing list Fink-users@lists.sourceforge.net List archive: http://news.gmane.org/gmane.os.macosx.fink.user Subscription management: https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] Subject: fail to update-all on OSX 10.4
On 18.03.2014, at 8:13AM, Ivan Kanis wrote: >>> "/sw/fink/dists/stable/main/finkinfo/10.4-EOL/base/libgettext8-shlibs.patch" >>> checksum does not match! >>> Actual: 0fa124a9e56f702eb9f817ebd10aae5f >>> Expected: 34a11cf58bde5e40fd110004731fb574 > >> My guess is that you can edit >> /sw/fink/10.4/stable/main/finkinfo/10.4-EOL/base/libgettext8-shlibs.info >> and put the "Actual" value above in its PatchFile-MD5 field. > > The info file looks correct: > > Source-MD5: 0fa124a9e56f702eb9f817ebd10aae5f > # libgettext8-shlibs.patch must be regenerated for every update since > # dpkg uses it and needs the correct full file headers for patch -p0 > PatchFile: %n.patch > : 0fa124a9e56f702eb9f817ebd10aae5f > > It seems that fink is looking at Source-MD5 to check for the patch. If > you could tell me what script is doing this I can take it from there. No, this does not look correct - the source file, probably this md5sum /sw/src/gettext-0.18.3.1.tar.gz 3fc808f7d25487fc72b5759df7419e02 /sw/src/gettext-0.18.3.1.tar.gz and the patch definitely should not have the same md5sum. But it is a mystery to me why it would expect a completely different MD5, and why this did not fail right at the beginning of the build with a wrong MD5 for the source file. But apparently the package build was completed and then only the buildlock removal failed, maybe something got messed up in fink's internal database… Can you check if it's still there with fink list fink-buildlock-gettext-tools-0.18.1.1-1 ^^^ Oh, just noticed this, that looks definitely odd - the PatchFile-MD5 in your info file is for libgettext 1.18.3.1-2 (and for 1.18.2.1-1 it was 0fc008bcbe67004c6f2bcbf5581ef1c3) but the buildlock claims it's for 1.18.1.1-1, looks like it was left over from an old update and maybe the info files had been updated in the meantime. I'd try to remove it manually with fink remove fink-buildlock-gettext-tools-0.18.1.1-1 or "sudo dpkg -r …" and then try a "fink rebuild libgettext8-shlibs", see if that helps. (But you want to make sure the Source-MD5 in the info file is correct for that). Derek -- Learn Graph Databases - Download FREE O'Reilly Book "Graph Databases" is the definitive new guide to graph databases and their applications. Written by three acclaimed leaders in the field, this first edition is now available. Download your free book today! http://p.sf.net/sfu/13534_NeoTech ___ Fink-users mailing list Fink-users@lists.sourceforge.net List archive: http://news.gmane.org/gmane.os.macosx.fink.user Subscription management: https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] Sites not responding or timestamps older messages
On 21 Mar 2014, at 04:50 am, Alexander Hansen wrote: > The second is something we can work with. The central site where most > of the rest get their data is stuck for some reason, and I tried to > contact an administrator today. > > You might want to change your rsync mirror > > rsync://ber.de.eu.finkmirrors.net/finkinfo/ > > since that site updates independently and is currently up. BTW is it possible to modify something like the regional list of mirrors as well? Currently some sites,most notoriously distfiles.ams.nl.eu.finkmirrors.net and distfiles.hel.fi.eu.finkmirrors.net are down almost all the time, yet (at least with MirrorOrder Master-First) fink keeps going back to them to download sources and just gets no response, so on a larger update it is waiting every time for the timeout and then for the timeout for the default next mirror choice unless one is sitting by and intervenes manually during the download phase. Cheers, Derek -- Learn Graph Databases - Download FREE O'Reilly Book "Graph Databases" is the definitive new guide to graph databases and their applications. Written by three acclaimed leaders in the field, this first edition is now available. Download your free book today! http://p.sf.net/sfu/13534_NeoTech ___ Fink-users mailing list Fink-users@lists.sourceforge.net List archive: http://news.gmane.org/gmane.os.macosx.fink.user Subscription management: https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] emacs24?
On 30 May 2014, at 05:26 am, Hanspeter Niederstrasser wrote: >>> However, on emacs, the current package is quite a bit behind: >>> *fink offers 23.4, whereas >>> *at gnu they are at 24.3. >>> >>> Is there any hope for an updated package? >> >> Please download the emacs24.info and emacs24.patch files to >> /sw/fink/dists/local/main/finkinfo and try to build and run emacs24. For >> any one who actually understands emacs and install-info packaging, I had >> to comment out the .info file games because I couldn't understand what >> the point of the mangle-info script was. Feel free to fix as needed. >> For now it builds for me and starts up, but that's the extent of the >> testing I can do. > > Downloads are at > http://cvs.snaggledworks.com/viewvc.cgi/fink/3rdparty/main/finkinfo/editors/ Thanks for getting this this far to work! I still could not install the .deb though because the info file names were not matching, so I figured out the mangle-info script apparently serves to point the entry in the “dir” file to the fink-modified path. It would not work in gzipped files; as I did not have the nerve to polish my perl skill to do that with "use IO::Zlib” and stuff, I fixed it the primitive way and also fixed the list of info files. I was able to install the package now, browse some info files and load various modes and auctex successfully. There is an apparently non-serious font error message when starting auctex, but that can be fixed by upgrading to auctex 11.87 (can submit that to the package tracker later). Cheers, Derek emacs24.info Description: Binary data emacs24.patch Description: Binary data -- Learn Graph Databases - Download FREE O'Reilly Book "Graph Databases" is the definitive new guide to graph databases and their applications. Written by three acclaimed leaders in the field, this first edition is now available. Download your free book today! http://p.sf.net/sfu/NeoTech___ Fink-users mailing list Fink-users@lists.sourceforge.net List archive: http://news.gmane.org/gmane.os.macosx.fink.user Subscription management: https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] emacs24?
Hi Sven, > I had probs with an older auctex. When I tried using your files, fink > complains about a mismatch of checksum for the patchfile. > > under the assumption that you attached patch was the right one, i fixed the > checksum in your info-file accordingly. > > thanks. it seems that your mail client has converted the line endings to DOS on saving the files; they should still work, but of course the checksum would have changed. I’ve submitted the auctex update to https://sourceforge.net/p/fink/package-submissions/4411/ maybe downloading from there will preserve the text format. Cheers, Derek -- Learn Graph Databases - Download FREE O'Reilly Book "Graph Databases" is the definitive new guide to graph databases and their applications. Written by three acclaimed leaders in the field, this first edition is now available. Download your free book today! http://p.sf.net/sfu/NeoTech ___ Fink-users mailing list Fink-users@lists.sourceforge.net List archive: http://news.gmane.org/gmane.os.macosx.fink.user Subscription management: https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] Failed: phase compiling: h5py-py34-2.3.0-2 failed
On 23.06.2014, at 10:11, domi...@lps.ens.fr (Dominique Dhumieres) wrote:
> h5py/h5fd.c:1018:39: error: use of undeclared identifier 'H5FD_MPIPOSIX'
> __pyx_t_1 = __Pyx_PyInt_to_py_hid_t(H5FD_MPIPOSIX); if
> (unlikely(!__pyx_t_1)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 34;
> __pyx_cl...
> ^
> 1 warning and 1 error generated.
> error: command 'gcc' failed with exit status 1
> ### execution of /sw64/bin/python3.4 failed, exit code 1
> ### execution of /tmp/fink.ZrO4J failed, exit code 1
> Removing runtime build-lock...
> Removing build-lock package...
> /sw64/bin/dpkg-lockwait -r fink-buildlock-h5py-py34-2.3.0-2
> (Reading database ... 335492 files and directories currently installed.)
> Removing fink-buildlock-h5py-py34-2.3.0-2 ...
> Failed: phase compiling: h5py-py34-2.3.0-2 failed
Just discussed this on the devel list: there's an incompatibility with
hdf5.8-18 that's been fixed in the .pyx/.pxd files, but not in the generated
h5fd.c.
If you install cython-py34, it should work for now.
Cheers,
Derek
--
HPCC Systems Open Source Big Data Platform from LexisNexis Risk Solutions
Find What Matters Most in Your Big Data with HPCC Systems
Open Source. Fast. Scalable. Simple. Ideal for Dirty Data.
Leverages Graph Analysis for Fast Processing & Easy Data Exploration
http://p.sf.net/sfu/hpccsystems
___
Fink-users mailing list
Fink-users@lists.sourceforge.net
List archive:
http://news.gmane.org/gmane.os.macosx.fink.user
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] Cannot install glib2-dev glib2-shlibs MBP OSX 10.9.5
On 25 Sep 2014, at 07:20 am, Andreas Skarlatoudis wrote: > The python version I have is 3.4.1 and is installed through fink. I’ve made > /usr/bin/python a symbolic link pointing to /opt/sw/bin/python > in case this helps you reproduce the error. > It explains the error. pykg-config is a Python2 package and not Python3-ready, as the syntax error shows, because all OS X system python versions are 2.7.x so far. You should under NO circumstances replace system files (anything in /usr/bin | /usr/lib | /usr/include) with 3rd-party installations (including fink or anything installing under /usr/local, /opt etc…), or there is no telling what on your system will break! The best fix should be to restore the original /usr/bin/python from a backup; if you don’t have one the next best thing might be to cp /System/Library/Frameworks/Python.framework/Versions/2.7/bin/python2.7 to /usr/bin/python - the two binaries have different sizes on my system, but they are both 2.7.5, so I’d hope this could work perhaps until the next software update. HTH, Derek -- Meet PCI DSS 3.0 Compliance Requirements with EventLog Analyzer Achieve PCI DSS 3.0 Compliant Status with Out-of-the-box PCI DSS Reports Are you Audit-Ready for PCI DSS 3.0 Compliance? Download White paper Comply to PCI DSS 3.0 Requirement 10 and 11.5 with EventLog Analyzer http://pubads.g.doubleclick.net/gampad/clk?id=154622311&iu=/4140/ostg.clktrk ___ Fink-users mailing list Fink-users@lists.sourceforge.net List archive: http://news.gmane.org/gmane.os.macosx.fink.user Subscription management: https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] Cannot install glib2-dev glib2-shlibs MBP OSX 10.9.5
On 25 Sep 2014, at 03:40 pm, Derek Homeier wrote: > On 25 Sep 2014, at 07:20 am, Andreas Skarlatoudis wrote: > >> The python version I have is 3.4.1 and is installed through fink. I’ve made >> /usr/bin/python a symbolic link pointing to /opt/sw/bin/python >> in case this helps you reproduce the error. >> > It explains the error. pykg-config is a Python2 package and not > Python3-ready, as the syntax error shows, > because all OS X system python versions are 2.7.x so far. > You should under NO circumstances replace system files (anything in /usr/bin > | /usr/lib | /usr/include) with > 3rd-party installations (including fink or anything installing under > /usr/local, /opt etc…), or there is no > telling what on your system will break! > The best fix should be to restore the original /usr/bin/python from a backup; > if you don’t have one the next > best thing might be to cp > /System/Library/Frameworks/Python.framework/Versions/2.7/bin/python2.7 to > /usr/bin/python > - the two binaries have different sizes on my system, but they are both > 2.7.5, so I’d hope this could work > perhaps until the next software update. Actually /usr/bin/python and /usr/bin/pythonw are identical binaries, so the latter would be a better recovery option. Derek -- Meet PCI DSS 3.0 Compliance Requirements with EventLog Analyzer Achieve PCI DSS 3.0 Compliant Status with Out-of-the-box PCI DSS Reports Are you Audit-Ready for PCI DSS 3.0 Compliance? Download White paper Comply to PCI DSS 3.0 Requirement 10 and 11.5 with EventLog Analyzer http://pubads.g.doubleclick.net/gampad/clk?id=154622311&iu=/4140/ostg.clktrk ___ Fink-users mailing list Fink-users@lists.sourceforge.net List archive: http://news.gmane.org/gmane.os.macosx.fink.user Subscription management: https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] Cannot install glib2-dev glib2-shlibs MBP OSX 10.9.5
Hi Andreas, > unfortunately I've erased every instance of Python2.X from everywhere in the > system including /System/Library/Frameworks/Python.framework > that was thorough. > The solution I got from apple support is to re-install the system. If you > have a better one I'd be glad to hear it. > Probably safest, and I guess it should not be too painful if you restore your user settings and/or have a full Tim Machine backup. Apple does have the sources available here https://opensource.apple.com/tarballs/python/ so I guess in principle it might be possible to do an xcodebuild of the latest version, but I’d certainly search for additional instructions if you are inclined to experiment with this path. Cheers, Derek -- Meet PCI DSS 3.0 Compliance Requirements with EventLog Analyzer Achieve PCI DSS 3.0 Compliant Status with Out-of-the-box PCI DSS Reports Are you Audit-Ready for PCI DSS 3.0 Compliance? Download White paper Comply to PCI DSS 3.0 Requirement 10 and 11.5 with EventLog Analyzer http://pubads.g.doubleclick.net/gampad/clk?id=154622311&iu=/4140/ostg.clktrk ___ Fink-users mailing list Fink-users@lists.sourceforge.net List archive: http://news.gmane.org/gmane.os.macosx.fink.user Subscription management: https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] Failed: PatchFile "/sw/fink/dists/unstable/main/finkinfo/sci/matplotlib-py.patch" checksum does not match!
Hi there, any progress with the matplotlib package? I am still also getting a patch file with wrong md5sum on selfupdate, but I suspect it might be rather the updated patch file is still missing from the repo than the patch updated without the checksum. On 5 Aug 2010, at 16:34, Alexander Hansen wrote: You're much better off running "fink index -f" first if you think you have an incorrect package description rather than removing files, because fink normally uses a stored cache rather than the physical files. Occasionally there are situations where the cache doesn't get updated even after the files change. In this case, the package fails the md5 test because the patchfile got updated without a corresponding update in the .info file. The appropriate thing to do in such cases is to contact the package maintainer, since they are the ones who break and fix packages. :-) (or in principle you could edit /sw/fink/dists/unstable/main/finkinfo/sci/matplotlib-py.info and change the PatchFile-MD5 entry) However, you probably don't want to do that, because I found some problems with the patch (maintainer cc'ed). After changing the PatchFile-MD5 I got: [ -r /sw/fink/dists/unstable/main/finkinfo/sci/matplotlib-py.patch ] sed 's|@PREFIX@|/sw|g' < /sw/fink/dists/unstable/main/finkinfo/sci/matplotlib-py.patch | patch -p1 patching file setup.cfg patching file setup.py Hunk #1 succeeded at 266 with fuzz 2 (offset 11 lines). patching file setupext.py patch: malformed patch at line 27: # make -f make.osx fetch deps mpl_build mpl_install Looks like the patch is not complete yet. I also found when trying to update on my own, that mpl 1.0.0 needs an extra patch to build against fink's python rather than the framework-based system python. The attached one allows it to build (on 10.5 + 10.6 anyway), but of course the warning might still apply, so YMMV. I also noted that matplotlib-py27, while it builds and runs this way, is not running very well, i.e. mouse input in the plot window has basically no effect at all on powerpc, and produces loads of errors and ipython crashes on intel. Maybe I am missing some other fix for py27. One last request to the maintainer - could you consider changing the Depends: numeric, numarray to Suggests: ... or something the like? They have not been required for running matplotlib for some time, and I think anyone who needs to hold on to these packages at this time should know so on their own... HTH, Derek matplotlib-py.info Description: Binary data matplotlib-py.patch Description: Binary data -- This SF.net email is sponsored by Make an app they can't live without Enter the BlackBerry Developer Challenge http://p.sf.net/sfu/RIM-dev2dev ___ Fink-users mailing list Fink-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/fink-users
[Fink-users] cython broken for Python2.[45]
Hi, the cython-py[xx]_0.14 package is broken for Python2.4 and 2.5 because setup.py only builds the debugger cygdb for Python versions > 2.5. I am attaching a modified info script (with 0.14.1 update) to take this into account. HTH, Derek cython-py.info Description: Binary data -- Xperia(TM) PLAY It's a major breakthrough. An authentic gaming smartphone on the nation's most reliable network. And it wants your games. http://p.sf.net/sfu/verizon-sfdev___ Fink-users mailing list Fink-users@lists.sourceforge.net List archive: http://news.gmane.org/gmane.os.macosx.fink.user Subscription management: https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] Gnuplot mode for Emacs 23.1.1?
On 6 Apr 2011, at 21:44, Ingo Thies wrote:
> Ah thanks. And how do I make emacs use this file? I never had to
> customize
> such things before...
with my Emacs 23.1.1, it's automatically loaded when opening a script
e.g.
with a .gp extension. Otherwise 'M-x gnuplot-mode' should do it.
Ah, wait - back in the time when I was using gnuplot, I had actually put
the following into my .emacs (not sure if all of them are still
necessary):
;; these lines enable the use of gnuplot mode
(autoload 'gnuplot-mode "gnuplot" "gnuplot major mode" t)
(autoload 'gnuplot-make-buffer "gnuplot" "open a buffer in gnuplot
mode" t)
;; this line automatically causes all files with the .gp extension to
;; be loaded into gnuplot mode
(setq auto-mode-alist (append '(("\\.gp\\'" . gnuplot-mode)) auto-mode-
alist))
Cheers,
Derek
--
Xperia(TM) PLAY
It's a major breakthrough. An authentic gaming
smartphone on the nation's most reliable network.
And it wants your games.
http://p.sf.net/sfu/verizon-sfdev
___
Fink-users mailing list
Fink-users@lists.sourceforge.net
List archive:
http://news.gmane.org/gmane.os.macosx.fink.user
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] Fatal Python Error
Hi Yogesh, > > 4) If it's a script, what Python interpreter is it expecting? > This may > > be indicated an initial line. > > > >> The first line of this script says: > >> #!/sw64/bin/python2.5 > > > > 5) How are you running it? If it's a script, you might be able to > get > > it to work just by using "/sw64/bin/python " > > > >> I just tried this way but got the same Fatal Python error (version > >> mismatch) > > > > Can we see the full output from an attempt to run your executable? > This is all i get when i run this disp.py script either with ./ or > the python that is specified on the first line of this script: > $ ./disp.py lysb_buf-2.90s > Fatal Python error: Interpreter not initialized (version mismatch?) > Abort trap This is in fact the characteristic error from trying to load a module for a different Python version - I've just checked you can in fact load fink's numpy- py25 (as well as pylab and PIL) in /usr/bin/python2.5 after adapting the path, and vice-versa. It even passes all tests :-) So it seems there must be a version mismatch - could you provide the following information: echo $PYTHONPATH (just to make sure) insert the following lines into your script: import sys print sys.version print sys.path Also, is your system Python 2.5 or 2.6 (it's 2.5 on Leopard, not sure about Snow Leopard right now)? Cheers, Derek -- Benefiting from Server Virtualization: Beyond Initial Workload Consolidation -- Increasing the use of server virtualization is a top priority.Virtualization can reduce costs, simplify management, and improve application availability and disaster protection. Learn more about boosting the value of server virtualization. http://p.sf.net/sfu/vmware-sfdev2dev ___ Fink-users mailing list Fink-users@lists.sourceforge.net List archive: http://news.gmane.org/gmane.os.macosx.fink.user Subscription management: https://lists.sourceforge.net/lists/listinfo/fink-users
[Fink-users] pyfits versioning conflict
Hi, the pyfits-2.4 update breaks installations for multiple Python versions, as all of them want to install the /sw/bin/fitscheck utility. I am attaching a revised .info to handle this with update-alternatives. Cheers, Derek pyfits-py.info Description: Binary data -- Benefiting from Server Virtualization: Beyond Initial Workload Consolidation -- Increasing the use of server virtualization is a top priority.Virtualization can reduce costs, simplify management, and improve application availability and disaster protection. Learn more about boosting the value of server virtualization. http://p.sf.net/sfu/vmware-sfdev2dev___ Fink-users mailing list Fink-users@lists.sourceforge.net List archive: http://news.gmane.org/gmane.os.macosx.fink.user Subscription management: https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] Fatal Python Error
Hi Yogesh, On 19.04.2011, at 10:28PM, Alexander Hansen wrote: >> >>insert the following lines into your script: >> >>import sys >>print sys.version >>print sys.path >> >> Inserted all 3 lines in my script, still the same error. No output from those commands, even if you put them right at the beginning of the script? >> >>Also, is your system Python 2.5 or 2.6 (it's 2.5 on Leopard, not >>sure about >>Snow Leopard right now)? >> >> It was 2.6 (showed in the current version), to make it consistent i >> changed the current version to 2.5 using sudo ln -s command. >> >> >>Cheers, >> Derek >> >> >> >> >> -- >> > I don't think that will work. They're different versions. You'd be > better off installing the -py26 versions of the modules you need if > you're going to try to run them with the system python. > > And if you overwrote the system's /usr/bin/python with your "ln -s" > operation, we definitely don't recommend that. Yes! I think you could switch to Python2.5 using the python-config if needed, but it would probably still be preferable to track down why your script is not using the python interpreter specified, or not using its default module path. It would help to get the output from the sys.path and sys.version above. On 4/19/11 3:18 PM, Yogesh Gupta wrote: > First few lines of disp.py: > > #!/sw64/bin/python2.5 > > from exp_setup import * > import matplotlib as mpl > import sys > Since Apple's Python does not include a matplotlib installation, this could hardly induce the crash due to a version conflict; thus the problem is most likely somewhere inside the exp_setup, so I'd have a closer look into that module - I suspect it might be changing the sys.path settings... HTH, Derek -- Benefiting from Server Virtualization: Beyond Initial Workload Consolidation -- Increasing the use of server virtualization is a top priority.Virtualization can reduce costs, simplify management, and improve application availability and disaster protection. Learn more about boosting the value of server virtualization. http://p.sf.net/sfu/vmware-sfdev2dev ___ Fink-users mailing list Fink-users@lists.sourceforge.net List archive: http://news.gmane.org/gmane.os.macosx.fink.user Subscription management: https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] build of python27-2.7.2-3 failed
On 17.06.2011, at 2:16PM, Bruno Martin wrote:
> I've tried to build python27-2.7.2-3 which fails to build on a PPC/G5 running
> 10.5.
> The error is described below:
>
> Thanks for your help,
>
> Bruno
>
>
> Python build finished, but the necessary bits to build these modules were not
> found:
> bsddb185 linuxaudiodev ossaudiodev
> spwd sunaudiodev
> To find the necessary bits, look in setup.py in detect_modules() for the
> module's name.
> (Fink package build should have 5 missing)
>
That's not an error - it was printed at the end of my build as well (would not
even have noticed on my own), but installation succeeds just fine without
these. I guess bsddb185 would be some wrapper for Berkeley DB (if you need one,
you should be able to install bsddb3-py27 once your python build is fixed), the
next are obviously audio devices for various OSs you'd neither want nor need,
and no idea what spwd might be...
>
> Failed to build these modules:
> _tkinter
That's probably the show-stopper, but almost no information on the causes - are
there no further messages below? I'd then look further up in the output for
some error messages about failing to compile _tkinter.so or the like.
I just built that package on the same system version; without further
diagnostics I could only suggest to check your installation of tcltk and
tcltk-dev, maybe rebuild those packages ('fink rebuild tcltk').
HTH,
Derek
--
EditLive Enterprise is the world's most technically advanced content
authoring tool. Experience the power of Track Changes, Inline Image
Editing and ensure content is compliant with Accessibility Checking.
http://p.sf.net/sfu/ephox-dev2dev
___
Fink-users mailing list
Fink-users@lists.sourceforge.net
List archive:
http://news.gmane.org/gmane.os.macosx.fink.user
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] Serious problems with gnuplot and other installations
Hi Frank, as Alexander mentioned, fink is working under 10.6.8 in other places, e.g. I just recompiled openldap24 worked here with the same XCode version (although in 64bit, I admit). If I am correct in assuming the first error already is fatal: On 10 Oct 2011, at 01:19, Frank Schmidt wrote: > It seems fink has stopped working for MAC 10.6.8. Before it had been > very fragile. > > So where to begin? Here the first failing step to install gnuplot which > seems to depend > on openldap24 and other failing packages. I get: > > . > . > . > Entering subdirectory libraries > Making all in > /sw/src/fink.build/openldap24-2.4.25-1/openldap-2.4.25/libraries > Entering subdirectory liblutil > rm -f version.c > ../../build/mkversion -v "2.4.25" liblutil.a > version.c > cc -g -O2 -I../../include-I../../include > -I/sw/lib/system-openssl/include -I/sw/include/db5 -DBIND_8_COMPAT > -I/sw/include -I/sw/lib/system-openssl/include -I/sw/include/db5 > -DBIND_8_COMPAT -I/sw/include -c -o base64.o base64.c > In file included from /usr/include/arpa/nameser.h:59, > from ../../include/ac/socket.h:52, > from base64.c:46: > /usr/include/arpa/nameser_compat.h:54:2: error: #error "Undefined or > invalid BYTE_ORDER"; this would seem to be rooted in your /usr/include tree, as normally that macro - #include #if !defined(BYTE_ORDER) || \ (BYTE_ORDER != BIG_ENDIAN && BYTE_ORDER != LITTLE_ENDIAN && \ BYTE_ORDER != PDP_ENDIAN) /* you must determine what the correct bit order is for * your compiler - the next line is an intentional error * which will force your compiles to bomb until you fix * the above macros. */ #error "Undefined or invalid BYTE_ORDER"; #endif would be defined in i386/endian.h (via machine/endian.h), which sets both #define __DARWIN_BYTE_ORDER __DARWIN_LITTLE_ENDIAN #define LITTLE_ENDIAN __DARWIN_LITTLE_ENDIAN #define BYTE_ORDER __DARWIN_BYTE_ORDER so this rather looks like your XCode installation is broken or affected by 3rd-party installs, as already suggested - have you tried reinstalling the developer tools? I don't know if this also explains the bootstrap errors you got, I suspect you may need to run the build more verbosely or send more output... HTH, Derek -- All the data continuously generated in your IT infrastructure contains a definitive record of customers, application performance, security threats, fraudulent activity and more. Splunk takes this data and makes sense of it. Business sense. IT sense. Common sense. http://p.sf.net/sfu/splunk-d2dcopy1 ___ Fink-users mailing list Fink-users@lists.sourceforge.net List archive: http://news.gmane.org/gmane.os.macosx.fink.user Subscription management: https://lists.sourceforge.net/lists/listinfo/fink-users
[Fink-users] pygsl-py 0.9.5-2 broken
[sorry, I'm not subscribed to fink-devel] Package: pygsl-py%type_pkg[python] Version: 0.9.5 Revision: 2 Distribution: (%type_pkg[python] = 23) 10.4, (%type_pkg[python] = 24) 10.4, (%type_pkg[python] = 24) 10.5 Type: python(2.3 2.4 2.5 2.6 2.7) Description: Python interface for the GSL License: GPL Maintainer: None Depends: << python%type_pkg[python] | python%type_pkg[python]-nox, gsl-shlibs (= 1.14-1), numpy-py%type_pkg[python] (>= 1.3.0-3) << BuildDepends: gsl (= 1.15-1) Just a typo in Depends? Cheers, Derek -- RSA(R) Conference 2012 Save $700 by Nov 18 Register now http://p.sf.net/sfu/rsa-sfdev2dev1 ___ Fink-users mailing list Fink-users@lists.sourceforge.net List archive: http://news.gmane.org/gmane.os.macosx.fink.user Subscription management: https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] pygsl-py 0.9.5-2 broken
On 10 Nov 2011, at 19:23, Alexander Hansen wrote: > On 11/10/11 1:10 PM, Derek Homeier wrote: >> [sorry, I'm not subscribed to fink-devel] > > List moderators let those messages through. There's no requirement to > subscribe. > >> >> Package: pygsl-py%type_pkg[python] >> Version: 0.9.5 >> Revision: 2 >> Distribution: (%type_pkg[python] = 23) 10.4, (%type_pkg[python] = 24) 10.4, >> (%type_pkg[python] = 24) 10.5 >> Type: python(2.3 2.4 2.5 2.6 2.7) >> Description: Python interface for the GSL >> License: GPL >> Maintainer: None >> Depends:<< >> python%type_pkg[python] | python%type_pkg[python]-nox, gsl-shlibs (= >> 1.14-1), >> numpy-py%type_pkg[python] (>= 1.3.0-3) >> << >> BuildDepends: gsl (= 1.15-1) >> >> Just a typo in Depends? >> >> Cheers, >> Derek >> >> > > You're quite correct. I just uploaded a fix. Thanks, it builds correctly now! Derek -- RSA(R) Conference 2012 Save $700 by Nov 18 Register now http://p.sf.net/sfu/rsa-sfdev2dev1 ___ Fink-users mailing list Fink-users@lists.sourceforge.net List archive: http://news.gmane.org/gmane.os.macosx.fink.user Subscription management: https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] On OSX 10.7 wget's dependencies just exploded in number: is that intentional?
On 10 Nov 2011, at 20:19, Hanspeter Niederstrasser wrote: >> Nonetheless, I find it somewhat annoying to have to install all the >> dependencies, including a lot which seem to have to do with font rendering >> and doc building, to install a command line utility and what appear to be >> its required libraries. 8P > > For the record, it seems that libidn (a new Dep for wget with the new > gnutls requirement) is the 'culprit' for bringing in all those > dependencies since it BuildDepends on gtk-doc. If there is a "fix" for > this, it would be to patch libidn to not need gtk-doc - have the libidn > docs as a separate package - but I have no idea if that is possible at all. Easy enough - just remove the ' --enable-gtk-doc' from the configure line (and the gtk-doc BuildDepends of course) - just tested it. Haven't checked though if the documentation is looking uglier now... HTH, Derek -- RSA(R) Conference 2012 Save $700 by Nov 18 Register now http://p.sf.net/sfu/rsa-sfdev2dev1 ___ Fink-users mailing list Fink-users@lists.sourceforge.net List archive: http://news.gmane.org/gmane.os.macosx.fink.user Subscription management: https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] build error in sphinx-py27 on 10.6/i386/Xcode4.2 was Re: Problems installing tortoisehg-py27
On 11 Nov 2011, at 22:38, Alexander Hansen wrote: > I tried the 10.7 package description on my 10.5 machine and sphinx-py27 > appears to work there. > tortoise-py27 builds on 10.6 as well, but does not run: ... File "/sw/lib/python2.7/site-packages/tortoisehg/util/hglib.py", line 22, in from mercurial.patch import updatedir ImportError: cannot import name updatedir Apparently you have to upgrade to 2.2 for compatibility with mercurial (>= 1.9): Version: 2.2 Revision: 1 Type: python (2.5 2.6 2.7) Source: http://bitbucket.org/tortoisehg/targz/downloads/tortoisehg-%v.tar.gz Source-MD5: 6f25086aa31238db25bc065ef7c25442 Builds and runs (with some slightly annoying debugging output ;-) nicely. Looks like a nice tool, thanks for the package! Cheers, Derek -- All the data continuously generated in your IT infrastructure contains a definitive record of customers, application performance, security threats, fraudulent activity, and more. Splunk takes this data and makes sense of it. IT sense. And common sense. http://p.sf.net/sfu/splunk-novd2d ___ Fink-users mailing list Fink-users@lists.sourceforge.net List archive: http://news.gmane.org/gmane.os.macosx.fink.user Subscription management: https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] Circular dependencies in Perl 5.10 modules
Hi Daniel, >> >> I also noted that the following packages fail their test suites on >> 10.6/32bit: >> compress-raw-bzip2-pm5100-2.037-1 io-compress-pm5100-2.037-1 >> module-find-pm5100-0.10-1 >> compress-raw-zlib-pm5100-2.037-1 io-socket-inet6-pm5100-2.67-1 >> socket6-pm5100-0.23-1 >> encode-pm5100-2.44-1 locale-gettext-pm5100-1.05-2 >> >> Let me know if you want the detailed error output. >> >> Cheers, >> Derek >> > > Bleh, I hate fat perl. This is not an unusual issue with perl tests on > 10.6/i386. Make sure that test-simple-pm5100 is installed and if not try > again. Other than that, someone else will have to look at it since I don't > have a 10.6 system anymore to test with. yes, that did it! It had been installed along with one of the other packages (all part of the texlive dependencies), but apparently not the first time I tried the build. Just a missing TestDepends in those packages? Thanks, Derek -- This SF email is sponsosred by: Try Windows Azure free for 90 days Click Here http://p.sf.net/sfu/sfd2d-msazure ___ Fink-users mailing list Fink-users@lists.sourceforge.net List archive: http://news.gmane.org/gmane.os.macosx.fink.user Subscription management: https://lists.sourceforge.net/lists/listinfo/fink-users
[Fink-users] Circular dependencies in Perl 5.10 modules
Hi, when I am trying to build and install any of the 6 modules below in a clean fink tree with testing turned on, it fails due to unresolved dependencies, but gives no information on the dependencies. > fink --validate=on --tests=on --build-as-nobody update io-socket-ssl-pm5100 > Information about 11389 packages read in 1 seconds. Validating package file /sw32/fink/dists/stable/main/finkinfo/libs/perlmods/io-socket-ssl-pm.info... Package looks good! The following package will be installed or updated: io-socket-ssl-pm5100 The following 5 additional packages will be installed: libwww-pm5100 module-cpants-analyse-pm5100 net-http-pm5100 net-ssleay-pm5100 test-kwalitee-pm5100 Do you want to continue? [Y/n] Failed: Problem resolving dependencies. Check for circular dependencies. Since it works after installing one of them without --tests=on, I realize it must be due to the additional test dependencies. So I don't know whether this should be considered a bug, but thought you might want to know. I also noted that the following packages fail their test suites on 10.6/32bit: compress-raw-bzip2-pm5100-2.037-1 io-compress-pm5100-2.037-1 module-find-pm5100-0.10-1 compress-raw-zlib-pm5100-2.037-1 io-socket-inet6-pm5100-2.67-1 socket6-pm5100-0.23-1 encode-pm5100-2.44-1 locale-gettext-pm5100-1.05-2 Let me know if you want the detailed error output. Cheers, Derek -- This SF email is sponsosred by: Try Windows Azure free for 90 days Click Here http://p.sf.net/sfu/sfd2d-msazure ___ Fink-users mailing list Fink-users@lists.sourceforge.net List archive: http://news.gmane.org/gmane.os.macosx.fink.user Subscription management: https://lists.sourceforge.net/lists/listinfo/fink-users
[Fink-users] passwd PostInstall issues
Hi, when trying to install with "update" in a clean fink tree, passwd stumbles in the PostInstall phase: > fink update passwd > Information about 11107 packages read in 2 seconds. The following package will be installed or updated: passwd Reading buildlock packages... /sw64/bin/dpkg-lockwait -i /sw64/fink/dists/stable/main/binary-darwin-x86_64/base/passwd_20100706-3_darwin-x86_64.deb (Reading database ... 38914 files and directories currently installed.) Preparing to replace passwd 20100706-3 (using .../passwd_20100706-3_darwin-x86_64.deb) ... Unpacking replacement passwd ... Setting up passwd (20100706-3) ... /sw64/var/lib/dpkg/info/passwd.postinst: line 6: update-passwd: command not found /sw64/bin/dpkg: error processing passwd (--install): subprocess post-installation script returned error exit status 127 Errors were encountered while processing: passwd ### execution of /sw64/bin/dpkg-lockwait failed, exit code 1 Failed: can't install package passwd-20100706-3 "fink install passwd" works, but since I habitually use "update" in place of "install" to make sure I get the newest version (or, in this case, the install was triggered as a dependency from a "fink update" of another package), I thought this might be worth pointing out. Cheers, Derek -- This SF email is sponsosred by: Try Windows Azure free for 90 days Click Here http://p.sf.net/sfu/sfd2d-msazure ___ Fink-users mailing list Fink-users@lists.sourceforge.net List archive: http://news.gmane.org/gmane.os.macosx.fink.user Subscription management: https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] [fink-core] passwd PostInstall issues
Hi Max, > thanks for the report, I think I spotted and fixed the cause of this issue. > > But let me point out one thing, lest people get confused about it: There is > absolutely no difference between "fink install FOO" and "fink update FOO". > The "update" command is simply an alias for "install", and is 100% identical > in its functionality. So this was not the cause of the problem you were > seeing. thanks for the clarification; that's more or less what I assumed as well and just was using "update" for the benefit of doubt. Now I am rather confused as to what happened, possibly it was actually fixed by a rebuild of the package. I thought I had first tried unsuccessfully again the "fink update" after the rebuild, but apparently I did not... I was wondering as well whether package being present in several fink installations simultaneously could cause problems (in addition to /sw I have installations both in /sw32 and /sw64 for testing). Each instance seems just to happily add and remove entries to NetInfo on its own... BTW, has finkproject.org been hijacked or something? I was just trying to look up at what step of the package install PostInst is actually executed, but I am immediately redirected from any URL under finkproject.org to some advertising sites! Cheers, Derek -- This SF email is sponsosred by: Try Windows Azure free for 90 days Click Here http://p.sf.net/sfu/sfd2d-msazure ___ Fink-users mailing list Fink-users@lists.sourceforge.net List archive: http://news.gmane.org/gmane.os.macosx.fink.user Subscription management: https://lists.sourceforge.net/lists/listinfo/fink-users
[Fink-users] gtk-doc build failure on 10.6/x86_64
Hi, on a clean 64bit tree under Snow Leopard gtk-doc-1.18-1 fails to build with the following error: checking for dblatex... (cached) no checking for fop... (cached) no configure: WARNING: neither dblatex nor fop found, so no pdf output from xml checking for XML catalog (/sw64/etc/xml/catalog)... found checking for xmlcatalog... /sw64/bin/xmlcatalog checking for DocBook XML DTD V4.3 in XML catalog... not found configure: error: could not find DocBook XML DTD V4.3 in XML catalog ### execution of ./configure failed, exit code 1 Removing runtime build-lock... Removing build-lock package... /sw64/bin/dpkg-lockwait -r fink-buildlock-gtk-doc-1.18-1 (Reading database ... 158744 files and directories currently installed.) Removing fink-buildlock-gtk-doc-1.18-1 ... Failed: phase compiling: gtk-doc-1.18-1 failed Since I have gtk-doc on my old x86_64 fink installation (updated over time) I suspected a missing dependency, but building it in an also fresh 32bit tree on the same machine works. I recompiled libxml2 just in case, but to no effect. Cheers, Derek -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ Fink-users mailing list Fink-users@lists.sourceforge.net List archive: http://news.gmane.org/gmane.os.macosx.fink.user Subscription management: https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] gtk-doc build failure on 10.6/x86_64
On 15.05.2012, at 1:15AM, Alexander Hansen wrote: > Check /sw/src/fink.build/gtk-doc-1.18-1/gtk-doc-1.18/config.log . I have: > > configure:11372: checking for XML catalog (/sw/etc/xml/catalog) > configure:11375: result: found > configure:11386: checking for xmlcatalog > configure:11404: found /sw/bin/xmlcatalog > configure:11417: result: /sw/bin/xmlcatalog > configure:11436: checking for DocBook XML DTD V4.3 in XML catalog > configure:11439: $XMLCATALOG --noout "$XML_CATALOG_FILE" "-//OASIS//DTD > DocBook XML V4.3//EN" >&2 > file:///sw/share/xml/dtd/docbookx/4.3.0/docbookx.dtd Ah, that ^ file was missing - it appears I had a broken docbook installation that had installed into /sw instead of the %p for that tree, /sw64. /sw64/bin/fink rebuild docbook-dtd fixed the problem. Thanks for your help, Derek -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ Fink-users mailing list Fink-users@lists.sourceforge.net List archive: http://news.gmane.org/gmane.os.macosx.fink.user Subscription management: https://lists.sourceforge.net/lists/listinfo/fink-users
[Fink-users] libsigc++2 build failure on Leopard
Hi, with Fink 0.33.3.1 on i386/10.5 trying to build libsigc++2-2.2.8 fails with the following error on creating the package, after the actual build has succeeded: Writing control file... Removing runtime build-lock... Removing build-lock package... /sw/bin/dpkg-lockwait -r fink-buildlock-libsigc++2-2.2.8-1002 (Reading database ... 243798 files and directories currently installed.) Removing fink-buildlock-libsigc++2-2.2.8-1002 ... Failed: Nested quantifiers in regex; marked by <-- HERE in m/^/scratch.noindex/fink.build/root-libsigc++ <-- HERE 2-2.2.8-1002// at /sw/lib/perl5/Fink/PkgVersion.pm line 4431, line 1. I've tried reinstalling perl and rebuilding fink, but to no avail - any idea if this is a problem with the fink database or what else might be done to fix it? Thanks, Derek -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ Fink-users mailing list Fink-users@lists.sourceforge.net List archive: http://news.gmane.org/gmane.os.macosx.fink.user Subscription management: https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] libsigc++2 build failure on Leopard
On 06.07.2012, at 11:19PM, Daniel Macks wrote: > On Fri, 06 Jul 2012 14:59:23 -0700, Alexander Hansen > wrote: > On 7/6/12 1:49 PM, Derek Homeier wrote: >> > Hi, >> > > with Fink 0.33.3.1 on i386/10.5 trying to build libsigc++2-2.2.8 fails >> > > with the following error > on creating the package, after the actual >> > > build has succeeded: >> > > Writing control file... > Removing runtime build-lock... > Removing >> > > build-lock package... > /sw/bin/dpkg-lockwait -r >> > > fink-buildlock-libsigc++2-2.2.8-1002 >> > (Reading database ... 243798 files and directories currently installed.) >> > Removing fink-buildlock-libsigc++2-2.2.8-1002 ... > Failed: Nested >> > quantifiers in regex; marked by <-- HERE in >> > m/^/scratch.noindex/fink.build/root-libsigc++ <-- HERE 2-2.2.8-1002// at >> > /sw/lib/perl5/Fink/PkgVersion.pm line 4431, line 1. > > I've tried >> > reinstalling perl and rebuilding fink, but to no avail - any idea if this >> > is a problem > with the fink database or what else might be done to fix it? >> > > Thanks, >> > Derek >> > > The line the error mentions has the following contents: >> >> $md5file =~ s/^$destdir\// /; >> >> I assume you have >> >> Buildpath: /scratch.noindex/fink.build >> >> set in fink.conf? I tried setting that but was unable to reproduce the >> problem on 10.7. >> I'll see if I can reproduce the problem on 10.5. > > Probably platform-independent but affects all packages with "+" in their name > or version fields. It's a regex metachar but the perl line is expecting to > match it as a literal string. See if my commit to master fixes it: > > https://github.com/fink/fink/commit/eedeb59f79616d281d0282e5808d6a14a6093739 Yep, that fixes it, thanks! I suspected something with the "+" being interpreted as regex, but my perl skills were too rusty to take me any further. I could not reproduce it on 10.7 either, or 10.6 AFAIR (can't check right now), and had libsigc++2-2.2.8 installed on 10.5/ppc, but trying to rebuild it there (without your patch) just produced the same error; so I seem to have it built there with an earlier fink version before this became broken. And then it apparently depends on the system perl version how it handles this (5.8.8 vs. 5.10+). Thanks again, Derek -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ Fink-users mailing list Fink-users@lists.sourceforge.net List archive: http://news.gmane.org/gmane.os.macosx.fink.user Subscription management: https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] copying fink
On Wednesday, January 16, 2002, at 03:40 AM, Martin Costabel wrote: > > sudo rsync -az -e ssh /sw othermachine:/ > > There are some more caveats: Some packages need to run postinstall > scripts. Some are installing stuff outside /sw (the xfree packages, of > course; also g77; and anything that uses daemonic, most notably > openssh). And if your /sw is not /sw but for example a symlink, the real > path may be hardcoded somewhere if you are not careful. So you may > expect some more or less subtle breakage. Nothing that can't be > repaired, usually by a "fink reinstall" or else "fink rebuild", but be > warned. I have done it twice, it's possible but maybe not completely > painless. > An alternative were to install the base system and copy over the /fink/dists tree from your existing installation, then fink should just use your local .deb files for setting up the other packages. Derek ___ Fink-users mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] copying fink
On Wednesday, January 16, 2002, at 01:56 PM, Stephan Jaeger wrote: > I am almost sure that /sw/fink/debs/ has to be copied, as well (or the > links have to be set manually). > > I know from my own experience that this directory is a kind of a > reference for fink. Without the links, fink starts to build the packages > again, although the debs are in the dists directory. > That's entirely possible, I think I had this problem sometimes. I've actually switched over to do a "fink scanpackages" and use apt-get for transferring the installation to other machines now, but I thought for a one-time installation copying the /sw/fink might still be the easiest way to go. (I guess with "rsync -a" the symlinks in ./debs should be preserved. ) Thanks, Derek ___ Fink-users mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] reusing .deb's
On Saturday, February 2, 2002, at 06:40 PM, Martin Costabel wrote: > Or you can mount your partition from cube2 via Appleshare and then do > the rsync "locally", i.e. from the mounted partition, like > > rsync -auv /Volumes/Cube2Software/sw/fink/dists /sw/fink/ > In that case you could also add a line like deb file:/Volumes/Cube2Software/sw/fink/dists unstable main crypto (or whatever branch you have installed) to /sw/etc/apt/sources.list on cube1, and then install the packages with sudo apt-get update sudo apt-get install without even having to copy the .debs over to cube1 (provided you did the "fink scanpackages" on cube2). HTH, Derek ___ Fink-users mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] Weird users installed
On Tuesday, February 5, 2002, at 02:52 PM, Finlay Dobbie wrote: >> (I already deleted them via the pane and my system refused to recognize >> my Admin account!!)? > > What? It probably isn't wise to delete them via the Users pane because > System Preferences is idiotic. I don't know what you're talking about > with your Admin account, you just seem to be talking nonsense. > I don't think so, wasn't this a known problem of the User's prefs panel being dumb? IIRC it just transfers ownership of a deleted user's home dir to some other (admin) user, and if that home dir happens to be / or /usr as probably the case for some of the special users, you're in trouble. Which of course boils down again to "don't use the Users panel to delete users!" On Tuesday, February 5, 2002, at 10:28 AM, Erik Price wrote: > But this makes me wonder -- what criteria does Mac OS X use to hide its > own default "extra" users from the login prompt? I know I can switch to > a login screen that requires one to type the login name manually rather > than picking from a list of icons, but if I want to keep the list format, > how do I hide the rest (the same way Darwin hides "www", "nobody", > "daemon", "root" etc)? Or does it automatically hide login names with > UID < 100 ? > Yes, I think UID<100 is exactly the criterion (well, maybe <=100, dunno ;- ). I have fared very well on my system with just reducing all UIDs and GIDs of the passwd-fink users by 200. Which of course raises the question: what prevents fink from using these IDs per default? I am of course aware that IDs in that range are somewhat more queasy than >200, I just haven't noticed anything that wants to install users in this range. Is there any common software around using the 50-60 IDs, or is there an official Apple policy that 3rd-party software shall not install users in the special range? Just curious, Derek ___ Fink-users mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] Weird users installed
On Wednesday, February 6, 2002, at 09:16 AM, Erik Price wrote: > Hmm... I just installed the 'passwd' package for the first time the other > day. And you know what's strange? The 'dummy' users aren't showing up > in my Login window (at startup) after all! I didn't make any changes to > the dummy users' UIDs, either, they are the default that are specified by > the 'passwd' package (250-255 I think). > > Did Apple make a change in 10.1.2 that hides UID < 300, or did the > package 'passwd' itself somehow compensate for this, or is something > wrong? These users definitely exist on my system -- > I have noticed something similar on one machine, and I think the login window is just hiding users that have never been logged in. It did not show some regular users (UID>500) either, but after a certain user had logged in for the first time, he appeared in the list. You will probably find that the 'passwd' users still show up in the System prefs panel, but at least that's less annoying (just don't get tempted to remove them there...). On Wednesday, February 6, 2002, at 12:13 PM, Thomas Peters II wrote: > I tried reducing the UID to <100 a long time ago, but it didn't have any > affect in hiding them in Users Preferences, so I put them back to the > original UIDs. Did anyone actually succeed in hiding them, or was it > just pure speculation? Actually I was pretty sure they were _not_ in my Prefs, but they are there now! Either I am just confused or maybe they only keep out of there if they have never been installed with a UID>99, or maybe this also changed with 10.1.2... Derek ___ Fink-users mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] vim-nox and syntax highlighting
On Wednesday, March 13, 2002, at 04:46 AM, Bertrand Lupart wrote: > I use syntax highlighting with vim-nox and it works great despite the > fact that the text isn't colored but bolded and underlined, making it > quite unreadable :) > > It does it both in local and remote. > > Since ls --color gives me colors in both local and remote, i think it > isn't a terminal issue. > Not an issue of Terminal itself, probably rather that vim isn't aware it is colour-capable. Try this: vim -T linux HTH, Derek ___ Fink-users mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] Still Can't Install wxgtk Stuff
On Friday, March 15, 2002, at 06:27 PM, Dan Shafer wrote: > make: *** [lib/libwx_gtk-2.3.dylib.1.0.0] Error 1 > ### (cd failed, exit code 2 > Failed: installing wxgtk-2.3.2-3 failed > > I have checked and confirmed that there is in fact no such directory as > lib/libwx_gtk-2.3.dylib.1.0.0 anywhere in my OS X directory structure, > which would explain why an attempt to cd to it doesn't work. > > BTW, after I copied the two files Bill Baumgarner sent me into the > appropriate directories, it was his impression that I should be able to > fink install without it going through a download/build cycle. It did not. > > I would give up but now I'm ticked. If one person can get this working,I > should be able to as well. > > So I'm sending another SOS here. > I did not have any problems with that particular library, but I could not build wxpython-wxgtk in the first place with the current glib, because glib-1.2 installs glib.h in /sw/include/glib-1.2, while the patch to wxpython-wxgtk puts /sw/lib/glib/include into the include-dirs. This caused the compile to fail at a very early stage. Adding the new path to the patch made it work. Maybe this has nothing to do with your troubles, but I'd suggest checking for the output where the actual compilation of the missing .o files in your build failed. Derek ___ Fink-users mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-users
[Fink-users] control-center fails to link Xext
Hi all, for some time now control-center is failing to build in my cvs installation, apparently because the gnomecc compile is missing a -L/usr/X11R6/lib: cc -I/sw/include/gdk-pixbuf-1.0 -I/sw/include/gtk-1.2 -I/sw/include/glib-1.2 -I/sw/lib/glib/include -I/usr/X11R6/include-O3 -Wall -Wunused -o .libs/gnomecc main.o tree.o callbacks.o corba-glue.o gtk-multiview.o splash.o capplet-manager.o -lm -L/sw/lib -lgnorba -lORBitCosNaming -lORBit -lIIOP -lORBitutil -lglib -lm -lglib -lm -lm -lglib -lm -lgnomeui -lesd -laudiofile -lm -laudiofile -lm -lglib -lgdk_imlib -lgtk -lgdk -lgmodule -lglib -ldl -lintl -lXext -lX11 -lm -lSM -lICE -lgtk -lgdk -lgmodule -lglib -ldl -lintl -lXext -lX11 -lm -lm -lart_lgpl -lm -lgdk_imlib -lgmodule -lglib -ldl -lSM -lICE -lgtk -lgmodule -lglib -ldl -lXext -lX11 -lm -lintl -lgdk -lgmodule -lglib -ldl -lXext -lX11 -lm -lintl -lgmodule -ldl -ldl -lintl -liconv -lXext -lX11 -lgnome -lglib -lm -lm -lgnomesupport -lz -lm -lesd -laudiofile -lm -lm -laudiofile -lm -lm -lORBitCosNaming -lORBit -lglib -lm -lIIOP -lORBitutil -lglib -lm /usr/bin/ld: can't locate file for: -lXext make[2]: *** [gnomecc] Error 1 make[1]: *** [all-recursive] Error 1 make: *** [all-recursive-am] Error 2 ### make failed, exit code 2 Failed: compiling control-center-1.4.0.4-2 failed bart:704> ls -1 /usr/X11R6/lib/libXext.* /usr/X11R6/lib/libXext.6.4.dylib /usr/X11R6/lib/libXext.6.dylib /usr/X11R6/lib/libXext.a /usr/X11R6/lib/libXext.dylib This has been preventing me from updating most of the gnome-packages so far. Any ideas? (In case this might matter, /usr/X11R6 is actually a link to /sw/X11R6, which lives on its own partition, like with several other people here, but this has never caused any problems yet.) Thanks, Derek ___ Fink-users mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-users
[Fink-users] sh-mode bug in vim-6.0-4
Hi all, I see a probably completely harmless, but annoying bug in vim: when editing shell scripts (sh/zsh/ksh at least), I frequently get the following error on startup: Error detected while processing /sw/share/vim/vim60/syntax/sh.vim: line 321: E394: Didn't find region item for shRepeat Invalid argument: shUntilSync^Igrouphere^IshRepeat^I"\" line 322: E394: Didn't find region item for shRepeat Invalid argument: shWhileSync^Igrouphere^IshRepeat^I"\" Hit ENTER or type command to continue It works smoothly after that, with syntax highlighting and colours etc. The version is vim-nox 6.0-4 (not even in unstable, it seems), though I think the same happened with the regular vim. Anyone knows what it means (or even better, how to fix it)? Cheers, Derek ___ Fink-users mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] Fink on different partition
On Friday, March 22, 2002, at 04:58 PM, Chris Devers wrote: > >> With other Unix systems you'd simply have an empty directory called /sw >> which would be a mount point for your partitions. The all the paths are >> still the default /sw... MacOS X insists on mounting everything under >> /Volumes but you can still have /sw be a symlink to /Volumes/SomeDisk >> and all should be happy. > > I *think* he meant using a separate fixed partition, which ...well I don' > t > really know if that plays into your point or not I guess, but in any event > I don't think it's a big deal that secondary partitions go in /Volumes. > Actually, mounting it under /sw works just fine if you add it to /etc/fstab and ". /mounts" in netinfo -- provided it is an UFS partition, because HFS+ partitions can't be checked by fsck, I think autodiskmount takes care of that. I have not figured out how to have autodiskmount check the filesystems at the right time (in /etc/rc), so I gave up on trying to mount HFS partitions anywhere else. Even UFS partitions often failed to mount with the "mount -vat ufs" in /etc/rc since 10.1 or so, so I have added softlinks to everything in /Volumes/Fink to /sw (the mount point on /) as a fallback. If everything goes right, the fink partition is just mounted over the symlinks. Coming to think of it, I just checked that there actually is no fsck call in /etc/rc, maybe adding that would be a good idea. DOH! As a final remark, this setup still forced me to link /sw/src to yet another partition with a HFS+ filesystem, since the UFS implementation in Darwin is still so crappy that it doesn't survive an Xfree86 compile! Of course, all this is entirely at your own risk (especially changing /etc/ rc and /etc/fstab) and probably not for the uninitiated. You might just fare better with the symlink approach, and I can't see any real disadvantages to it (except that it looks ugly ;-). Cheers, Derek ___ Fink-users mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] sh-mode bug in vim-6.0-4
On Friday, March 22, 2002, at 06:50 PM, Chris Devers wrote:
>> Anyone knows what it means (or even better, how to fix it)?
>
> Nope, sorry. Off the top of my head I think that I was getting this error
> both with Fink's vim and the one that I built myself, though it's been a
> few months and I wouldn't swear to that at this point. :/
>
It certainly seemed like a problem already present in the source
distribution. Heh, I can't say I really understand the syntax, but this
seems to fix it: change lines 321,322 in /sw/share/vim/vim60/syntax/sh.vim
to
+if exists("b:is_kornshell") || exists("b:is_bash")
syn sync match shUntilSync grouphere shRepeat "\"
syn sync match shWhileSync grouphere shRepeat "\"
+endif
(the error apparently originated from the fact that shRepeat was also only
defined under that condition. I think it also did not occur with scripts
having a #!/bin/ksh or #!/sw/bin/bash at the beginning).
Jeff, want to include that patch or should this rather be sent upstream to
the vim developers? It might not occur on other systems, maybe "b:is_bash"
is always set on Linux or whatever, but this still seems like a general
bug.
Cheers,
Derek
___
Fink-users mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] checking for working makeinfo... missing
On Friday, March 22, 2002, at 11:06 PM, John Melby wrote: > Thanks for the advice--but I'm afraid it doesn't apply in this case, > since I've been using the Mac OS X Developer Tools from day one of their > availability with the Public Beta, and I've been using XFree86 with Fink > for quite some time, having installed, among other things, Gimp, abiword, > g77, etc., etc. with it. I'm currently running OroborOSX with XFree86 4. > 2.0 and XDarwin. Fink has been behaving consistently until today, which > perhaps not coincidentally follows immediately upon the release of > version 0.9.9. The problem arises when makeinfo turns up missing. > Everything else is there. And just for the record, when I use Apple's cc, > my programs compile with no trouble at all. > > On Friday, March 22, 2002, at 09:22 PM, Chris Devers wrote: > >> On Fri, 22 Mar 2002, John Melby wrote: >> >>> checking for gcc... gcc >>> checking whether the C compiler (gcc -L/sw/lib) works... no >>> configure: error: installation or configuration problem: C compiler >>> cannot create executables. > Well, good point: what gcc? Fink, per default _does_ use Apple's /usr/bin/ cc, and I'm pretty sure it does not install any gcc (of course it is gcc, really, but this seems to hint at something unusual in your setup). Derek ___ Fink-users mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] control-center fails to link Xext
On Friday, March 22, 2002, at 06:00 PM, Derek Homeier wrote: > Hi all, > > for some time now control-center is failing to build in my cvs > installation, > apparently because the gnomecc compile is missing a -L/usr/X11R6/lib: > Looks like this is settled, it seemed to be another instance of persistent problems with ksh and libtool: whenever there exists a ksh in /bin (I have normally copied the fink binary to that place, because we have just many shell scripts around that expect it there), libtool keeps creating its scripts with /bin/ksh instead of /bin/sh. Obviously the pdksh port just is not up to that job yet, because on long builds like for gnome stuff I keep seeing it crash with a memory fault. This sometimes also happens with very long job scripts, where the shell actually doesn't do much, but has to wait for some hours until the main program has finished. But this was the first time I saw it just failing on creating a linker flag. Cheers, Derek ___ Fink-users mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] Re: Compiler error
On Sun, Mar 24, 2002 at 03:24:42PM -0600, Chris Devers wrote: > On Sun, 24 Mar 2002, Ben Hines wrote: > > > At 2:59 PM -0600 3/24/02, John Melby wrote: > > > > >>checking whether the C compiler (cc ) works... no > > >>configure: error: installation or configuration problem: C compiler > > >>cannot create executables. > > > > > >However, when I use cc to compile a C program, everything works perfectly. > > > > > >Is it possible that only you and I are having this problem? Or is > > >something rotten in the world of XFree86dom? > > > > You likely need the December 2001 devtools, not the ones that came with OS X. > > When this came up before, he swore up & down that he was using the current > [Dec] version of the devtools. If that's really the case -- and I still > have my doubts, but John is the one that would know -- then I'm wondering > if something else was installed that mucked around with things. What, I > have no idea. GNU-Darwin? MacGimp? Beats me, but the signature of this > problem really does point to a missing or corrupted devtools installation. > > The error lines above come from the './configure' phase of building, > right? And configure scripts, as far as I know, interact with the system's > autoconf command/library/thingy. So *maybe* there's a problem there, but > again I'm not quite sure what. Did you try installing Fink's "autoconf", > and did it work? Can you download & build other autoconf/configure using > packages? (Before using Fink, I had installed my own copies of Vim, Pine, > and a couple of others -- maybe something like that is worth a shot over > in /usr/local.) > > Something about the interaction between autoconf (etc) and cc/gcc isn't > right, but I can't say which end is causing the problem, or if it's > something that comes in between them. You [John] say that you can get > programs with straight cc to work: are these "hello world" type programs > or can you do anything more complex than that? Rather than righteous > indignation, maybe you could offer more diagnostic details... :) > I seem to remember from John's original posting that configure found a gcc in the path, and that would definitely indicate that there is another executable around which is probably not from the current devtools. John, I would suggest doing a `which gcc`, and if that returns anything that is not a symlink to /usr/bin/cc, it likely is the cause of the problems. Try a 'cmp $(which cc) /usr/bin/cc' to see if these are actually different executables, or find out where it comes from (GNU-Darwin in fact comes to mind), but if in doubt I would just remove it or at least move it out of the way. All this is assuming that the above error message came from Wayne's posting, who might in fact have to check his devtools version. HTH, Derek ___ Fink-users mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] control-center fails to link Xext
On Sun, Mar 24, 2002 at 05:42:34PM -0500, Philip McDunnough wrote: > Hmm...how is this settled? What does one have to do to get it to compile? > OK, not exactly settled, rather worked around: I manually remove /bin/ksh on the build machine before a major fink update. Since I manually installed it there I would not necessarily expect fink to support that setup. Still, it would be nice if ksh would not crash, or one could persuade libtool to stick with /bin/sh, but the erratic nature of the ksh segfaults make it somewhat difficult to diagnose the error (or does any of the developers recognize this as a common problem?). And I don't know enough about libtool to have any remedy on that side. Cheers, Derek > > On Saturday, March 23, 2002, at 06:19 PM, Derek Homeier wrote: > > > On Friday, March 22, 2002, at 06:00 PM, Derek Homeier wrote: > > > >> Hi all, > >> > >> for some time now control-center is failing to build in my cvs > >> installation, > >> apparently because the gnomecc compile is missing a -L/usr/X11R6/lib: > >> > > Looks like this is settled, it seemed to be another instance of > > persistent problems with ksh and libtool: whenever there exists a ksh > > in /bin (I have normally copied the fink binary to that place, because > > we have just many shell scripts around that expect it there), libtool > > keeps creating its scripts with /bin/ksh instead of /bin/sh. Obviously > > the pdksh port just is not up to that job yet, because on long builds > > like for gnome stuff I keep seeing it crash with a memory fault. This > > sometimes also happens with very long job scripts, where the shell > > actually doesn't do much, but has to wait for some hours until the main > > program has finished. But this was the first time I saw it just failing > > on creating a linker flag. > > > > Cheers, > > Derek > > ___ Fink-users mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-users
[Fink-users] wxpython-wxgtk and glib again
Hi, > wxpython does not compile with the current glib, because glib-1.2 installs > glib.h in /sw/include/glib-1.2, while the patch to wxpython-wxgtk puts > /sw/lib/glib/include into the include-dirs. Adding the new path to the > patch seems to make it work. Thanks for working on this (I assume you did), but the solution of finding the include path with glib-config does not work here: creating build/temp.darwin-5.3-PowerMacintosh-2.2/src cc -DNDEBUG -g -O3 -Wall -Wstrict-prototypes -no-cpp-precomp -DSWIG_GLOBAL -DHAVE_CONFIG_H -DWXP_USE_THREAD=1 -Isrc -I/sw/include/python2.2 -c src/helpers.cpp -o build/temp.darwin-5.3-PowerMacintosh-2.2/src/helpers.o -I/sw/lib/wx/include/gtk-2.3 -I/sw/include -DGTK_NO_CHECK_CASTS -D__WXGTK__ -fno-rtti -fno-exceptions -I/usr/X11R6/include `glib-config --cflags` cc: cannot specify -o with -c or -S and multiple compilations error: command 'cc' failed with exit status 1 ### python failed, exit code 1 Failed: compiling wxpython-wxgtk-2.3.2.1-2 failed bart:787> glib-config --cflags -I/sw/include/glib-1.2 -I/sw/lib/glib/include The same command issued from the command line compiles fine, and putting the include path explicitly into the SetCFLAGS field also works. I guess the problem is that the build script is interpreted by python, which does not understand the shell backquotes. Cheers, Derek ___ Fink-users mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-users
[Fink-users] scipy does not find X11 includes
Hi, another python-related build problem: scipy fails to compile with the following error: cc -DNDEBUG -g -O3 -Wall -Wstrict-prototypes -no-cpp-precomp -DSTDC_HEADERS=1 -c /Volumes/Galapagos/src/scipy-20020305-1/scipy-20020305/xplt/gist/dispax.c -o build/temp.darwin-5.3-PowerMacintosh-2.2/dispax.o In file included from /Volumes/Galapagos/src/scipy-20020305-1/scipy-20020305/xplt/gist/dispax.c:12: /Volumes/Galapagos/src/scipy-20020305-1/scipy-20020305/xplt/gist/dispax.h:40: X11/Xlib.h: No such file or directory /Volumes/Galapagos/src/scipy-20020305-1/scipy-20020305/xplt/gist/dispax.c:14: X11/Xutil.h: No such file or directory /Volumes/Galapagos/src/scipy-20020305-1/scipy-20020305/xplt/gist/dispax.c:17: X11/Xresource.h: No such file or directory error: command 'cc' failed with exit status 1 ### python failed, exit code 1 Failed: compiling scipy-20020305-1 failed Anyone knows how to make it know it should use "-I/usr/X11R6/include"? (OK, I want to learn python, but sometimes I wish they just sticked to configure/make for the build process!) TIA, Derek ___ Fink-users mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] Re: scipy does not find X11 includes
On Sun, Mar 24, 2002 at 06:49:07PM -0700, Jeff Whitaker wrote: > Fixed in unstable CVS yesterday. > Affirmative! I should have updated yesterday, but had some problems with the cvsupdate, tons of messages like cvs server: dists/stable/main/finkinfo/gnome/gnome-libs-1.4.1.2-4.info is no lon ger in the repository which should not be fatal, I assume, but ended in ### su failed, exit code 1 Failed: Updating using CVS failed. Check the error messages above. Manually deleting all local .info and .patch files finally fixed that. For the (un)stable record: I have installed the current cvs versions of python and numeric-atlas, and as far as I have used them yet (not very much) found them to work well. Except for the interactive help in the python interpreter, which frequently causes python to crash with a Bus error. Happens especially when the search facility is used, a la "modules dinsdale". Thanks, Derek ___ Fink-users mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-users
[Fink-users] Python issues
On Sunday, March 24, 2002, at 11:17 PM, Derek Homeier wrote: > For the (un)stable record: I have installed the current cvs versions of > python and numeric-atlas, and as far as I have used them yet (not > very much) found them to work well. Except for the interactive help in > the python interpreter, which frequently causes python to crash with > a Bus error. Happens especially when the search facility is used, a la > "modules dinsdale". > I had a few other minor problems: Some packages, such as scientificpython and pil install into /sw/lib/python2.1/site-packages/ , though they don't explicitly require python2.1. They are thus invisible to python2.2 unless you create symlinks into /sw/lib/python2.2/site-packages. Then, I found that env would fail to execute "#!/usr/bin/env python" scripts with "env: python : No such file or directory". Of course, /sw/bin is in my path, but how do I make that known to env? Thanks, Derek ___ Fink-users mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-users
[Fink-users] Re: Python issues
On Monday, March 25, 2002, at 08:39 PM, Jeff Whitaker wrote: > All extension modules need to be rebuilt with each new minor revision of > python. When you rebuild, they will be put into the correct directory for > the version of python you have installed. This is what Guido recommends, > mainly because you can't assume the bytecode will be compatible from one > version to the next. > > We probably should make these packages depend on a specific version of > python though, for the sake of the binary distribution. > Yep, that worked. Didn't remember I had those packages already around that long. gnome-python still claims to install libraries in /sw/lib/python2.1/site-packages though, but it doesn't. >> >> Then, I found that env would fail to execute "#!/usr/bin/env python" >> scripts with "env: python : No such file or directory". >> Of course, /sw/bin is in my path, but how do I make that known to env? > > Works for me. Are you sure /sw/bin is in your path (in the environment > that you're trying to execute the python script)? > Hmm, in fact I forgot I could execute some of the setup.py scripts in the scipy build tree just fine. Seems to have been a formatting problem, those were files I had downloaded as DOS-text. Converting the line returns and replacing some space characters in the first line solved it, apparently env did not like those (well, it has my sympathy!). Thanks, Derek ___ Fink-users mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-users
[Fink-users] Re: [Fink-users] Rép : [Fink-users] Failed: compiling xfree86-rootless-4.2.0-2
On Wednesday, March 27, 2002, at 11:32 AM, Marc Champesme wrote: >> Le mardi 26 mars 2002, à 03:14 PM, Jeff Whitaker a écrit : >> >> Looks like you have a UFS filesystem - building XFree86 fails due to file >> corruption on UFS for some unknown reason. Try building it on an HFS+ >> partition. > > > Thanks a lot ! > I have reformated my UFS partition to HFS an now xfree86-rootless-4.2.0-2 > build fine... but XDarwin crashes immediatly after launch (see below for > the error messages). > Concerning the first error message "error opening security policy file > /usr/X11R6/lib/X11/xserver/SecurityPolicy", it appears that directory > "xserver" doesn't exists in /usr/X11R6/lib/X11 > You should also rebuild xfree86-base. The file system corruption does not always make XFree86 builds fail completely, it's often just missing random components |-(. /usr/X11R6/lib/X11/xserver is part of xfree86-base-4.2.0, the fonts too. HTH, Derek ___ Fink-users mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] Can't "fink selfupdate"
On Thursday, March 28, 2002, at 05:44 PM, Hester, Jeffrey W. wrote:
> Per Alexander's recommendation I did a "fink remove packagename" on all
> those listed with a "C" in the first column. What's strange is that only
> "ddd" was actually removed. All of the other packages gave the following
> error:
>
> dpkg --remove blackbox-rootless
> dpkg - warning: ignoring request to remove blackbox-rootless which isn't
> installed.
> It's like they weren't installed. Which they shouldn't be because many I
> didn't recognize and many I didn't even download (even as dependencies).
> These included: ddd, ghx, hx, links-ssl, ee, f2c, gnut, rsync, hdf, hdf5,
> netcdf, netcdf-absoft, rasmol, scigraphica, ispell, tex2im, xless,
> keychain,
> kinput, links, basiliskii, eterm, blackbox, and blackbox-rootless.
>
> Sounds like I'm really hosed. Is my only option to trash fink (ie /sw)
> and
> start from scatch? I hope I don't have to reload all of those packages.
> If
> I do, how can I prevent this from getting messed up again. I'm not one
> that
> mucks around with the files except for using dselect and fink.
>
Did not read this thread before, but I had the exact same problems with a
recent selfupdate (and posted it here, too). I think fink just barfs on
the number of error msgs from the packages removed from the repository, a
la
>> cvs server: conflict: dists/stable/main/finkinfo/net/ghx-0.1.39-1.info is
>> modified but no longer in the repository
My solution for this was not to remove any packages (the errors occurred
for packages I never installed as well), but remove the old package
_descriptions_ from the /sw/fink/dists tree. A complete removal (or
renaming) of the "finkinfo" directories would probably work, but to play
it safe you might try
find /sw/fink/dists -name '*.info' -exec sudo rm {} \;
find /sw/fink/dists -name '*.patch' -exec sudo rm {} \;
selfupdate-cvs will normally install a full new set of package
descriptions, so you should be fine after that -- at least it worked for
me, and I had the exact same
### su failed, exit code 1
Failed: Updating using CVS failed. Check the error messages above.
error before.
HTH,
Derek
___
Fink-users mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] Fink Cannot Self Update
On Sunday, March 31, 2002, at 05:04 PM, Chris Devers wrote: > Yes, it's a solved problem. If you're manually moving things around in > the file trees that Fink/CVS wants to update, then it's bound to happen > sooner or later, and fixing it is no big deal. As the error messages you > got indicate, the solution is to move (or just delete) the files that are > in the way of what CVS needs to update. > >> I noted that these were all packages, formerly in unstable, that I >> updated last >> week (when I update gnome). > > Very astute. This is indeed the usual way for this problem to crop up. > >> I am going to try moving them to the local branch and see what happens. > > Good solution in this case. It should fix the problem. > >> Perhaps this is what has happened to other people? > > Yep :) > Not exactly! The selfupdate problems here occurred on a machine that had been running unstable from cvs for months, so there certainly wasn't anything moved manually into stable. It also did not print out that helpful advice to remove the .info and .patch files, just errors of the kind cvs server: dists/stable/main/finkinfo/gnome/gnome-libs-1.4.1.2-4.info is no longer in the repository The same seems to have happened for at least one other user. However, removing the files solved the problem nonetheless. It did happen after I had not updated for a relatively long time, so I believe the updating script (or cvs?) can only digest so many of these errors, or it somehow loses track of old versions after some time. I would still consider it a bug of some sort. Derek ___ Fink-users mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] Fink Cannot Self Update
On Monday, April 1, 2002, at 03:00 PM, Martin Costabel wrote: >> cvs server: dists/stable/main/finkinfo/gnome/gnome-libs-1.4.1.2-4.info is >> no longer in the repository > > These are not errors. The cvs server informs you about files that have > been removed from the cvs tree, usually because they are obsolete. > Normally cvs should then remove them also from your tree. > That's what I assumed, too, but I was puzzled by getting a ### su failed, exit code 1 Failed: Updating using CVS failed. Check the error messages above. after some 80 of these messages. On second inspection, though, it turned out there _was_ another problem: cvs server: conflict: dists/unstable/main/finkinfo/languages/wxpython-wxgtk-2.3.2.1-1.patch is modified but no longer in the repository cvs server: conflict: dists/unstable/main/finkinfo/languages/wxpython-wxgtk-2.3.2.1-2.info is modified but no longer in the repository which I had indeed changed myself to fix the include path. So it seems it' s just me who can only digest so many warning messages ;-) Derek ___ Fink-users mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-users
[Fink-users] qt3-shlibs linking error
Greetings, I have been unable to build qt3 (or rather qt3-shlibs, I assume), because of a failure to link to libfreetype: mv libqui.1.0.0.dylib libqui.dylib libqui.1.dylib libqui.1.0.dylib ../../../lib/ cd designer && make -f Makefile /Volumes/Galapagos/src/qt3-3.0.4-3/qt-x11-free-3.0.4/bin/uic dbconnections.ui -o dbconnections.h dyld: /Volumes/Galapagos/src/qt3-3.0.4-3/qt-x11-free-3.0.4/bin/uic version mismatch for library: /sw/lib/libfreetype.6.dylib (compatibility version of user: 6.2.0 greater than library's version: 0.0.0) make[4]: *** [dbconnections.h] Error 61 make[3]: *** [sub-designer] Error 2 ... I have seen dyld bitching like this before, but never found a consistent solution for it. I have rebuilt freetype2-shlibs, freetype2 from latest cvs and uninstalled qt and dependent packages. This is a 10.1.4 system with Dec. 2001 Devtools. Ideas? Derek ___ Have big pipes? SourceForge.net is looking for download mirrors. We supply the hardware. You get the recognition. Email Us: [EMAIL PROTECTED] ___ Fink-users mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] qt3-shlibs linking error
On Sat, May 11, 2002 at 10:08:27AM +0200, Martin Costabel wrote: > Derek Homeier wrote: > > > dyld: /Volumes/Galapagos/src/qt3-3.0.4-3/qt-x11-free-3.0.4/bin/uic version >mismatch for library: /sw/lib/libfreetype.6.dylib (compatibility version of user: >6.2.0 greater than library's version: 0.0.0) > > Strange. "otool -L /sw/lib/libfreetype.6.dylib" shows me compatibility > version 9.0.0. What version is your freetype2 fink package? > bart:316> otool -L /sw/lib/libfreetype.6.dylib /sw/lib/libfreetype.6.dylib: /sw/lib/libfreetype.6.dylib (compatibility version 0.0.0, current version 0.0.0) /usr/lib/libSystem.B.dylib (compatibility version 1.0.0, current version 55.0.0) bart:317> dpkg -p freetype2 Package: freetype2 Maintainer: Jeffrey Whitaker <[EMAIL PROTECTED]> Architecture: darwin-powerpc Source: freetype2 Version: 2.0.8-4 Depends: freetype2-shlibs (= 2.0.8-4) Size: 764960 the newest I could get... > I have a different problem with qt3: It builds completely, but does not > install, due to > > qt3-shlibs conflicts with qt-shlibs > qt-shlibs (version 3.0.3-3) is installed. > dpkg: error processing > >/sw/fink/dists/unstable/main/binary-darwin-powerpc/x11/qt3-shlibs_3.0.4-3_darwin-powerpc.deb > (--install): > conflicting packages - not installing qt3-shlibs > > This kind of defeats the whole purpose of the shlibs splitoff movement. > I had similar problems with updating gnuplot, which would not update without aquaterm, which would not install because it conflicts with the older gnuplot version... No way out but 'dpkg -r --force-depends' :-( On Sat, May 11, 2002 at 01:12:45PM -0400, Dave Vasilevsky wrote: > > Are you by any chance using bash (or another shell besides zsh) as > /bin/sh? This problem may come up in such situations. > Oh no, not again! My /bin/sh is the original zsh, but I have /bin/ksh (copied from fink pdksh), and this used to freak out libtool with every larger build. The latest update to pdksh actually seems to have solved that problem, but I never noticed any problems with dyld (well, I had problems, but could never relate anything to ksh being installed in /bin). I will try today's update to qt3 (though it is unlikely that that could change the version of libfreetype, coming to think of it), and then I will try once more with temporarily removing ksh from /bin. This does not have any immediate effect on the output of 'otool -L', though... Thanks for the tips, Derek ___ Have big pipes? SourceForge.net is looking for download mirrors. We supply the hardware. You get the recognition. Email Us: [EMAIL PROTECTED] ___ Fink-users mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] qt3-shlibs linking error
> > What matters isn't whether ksh is installed now, but whether it was > installed (and used) when you built freetype2. You might want to remove > other shells besides zsh and sh from /bin, and then do 'fink rebuild > freetype2'. See if that changes the results of otool -L. Note that this > problem (ksh and bash and other shells giving libraries version 0.0.0) > is a known bug in libtool, and we're in the process of changing packages > to work around it. > Right, occurred to me too after I sent the mail, and it also takes much less time than randomly trying to build qt3 ;-). This worked: /sw/lib/libfreetype.6.dylib: /sw/lib/libfreetype.6.dylib (compatibility version 9.0.0, current version 9.0.0) Good to hear that someone is working on the libtool issues, though. We are using a couple of scripts etc. here that expect ksh to be in /bin or /usr/bin. Thanks, Derek ___ Have big pipes? SourceForge.net is looking for download mirrors. We supply the hardware. You get the recognition. Email Us: [EMAIL PROTECTED] ___ Fink-users mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] 'gzip is not the right format' error
On Monday, August 12, 2002, at 05:54 PM, Justin Hallett wrote: > fink should never have issued that command. If it did there is likely > something wrong with your fink installing. And you should wait for > someone else to help you, since you where installing fink-0.10 I finigured > it was a new isntall and you issued that command. > > [EMAIL PROTECTED] writes: >> Thanx. >> Is there any way to configure fink to use the right (gzip/Tar) command? >> >> Now I will have to continue the install update procedure manually, and >> that >> might prove a bit tricky for me. >> There's nothing wrong with that command! > Justin Hallett wrote: > > this looks like a bzip2 line > > bzip2 -dc foo.tar.bz2 | tar -xf - > > gzip files like foo.tar.gz are extracted like this > > tar -xzf foo.tar.gz > > the z tells tar that the file is also gziped and you don't have to call > gunzip directly this way. > > [EMAIL PROTECTED] writes: >> gzip -dc packages-0.4.0.tar.gz | tar -xvf - > This does exactly the same as "tar xvzf packages-0.4.0.tar.gz", just as bzip2 -dc foo.tar.bz2 | tar -xf - does the same as tar xjf foo.tar.bz2 (with fink's tar, the flag for bzip'ed files is somewhat inconsistent ;). If gzip complains about a .gz file not being in gzip format, then that file is corrupt, probably from a botched download -- most likely it is a "file not found" html-message or something, check it with less packages-0.4.0.tar.gz What fink _should_ have done actually is tell you about the corrupt file and ask if it should delete and try to download it again. Since it has not done so, I suggest just to delete the packages file manually and run selfupdate again, or try to download packages-0.4.0.tar.gz manually in case there are still problems with the download. HTH, Derek --- This sf.net email is sponsored by: Dice - The leading online job board for high-tech professionals. Search and apply for tech jobs today! http://seeker.dice.com/seeker.epl?rel_code=31 ___ Fink-users mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] fink cleanup
On Tuesday, August 13, 2002, at 05:40 PM, mathias meyer wrote: > ah yes, now i found it. i was complaining (well, not really, just > mentioning) that fink cleanup has deleted some of my recent .deb's some > days ago. here's the scenario: > > 1) The .info file of package A is in unstable > 2) Package A builds in its .deb in unstable > 3) After a fink selfupdate-cvs the .info file of A is moved to stable. > the .deb of package A stays in unstable. no newer version of A is > introduced, no .info file of A left in unstable (really moved from > unstable to stable). the fink selfupdate-cvs does a fink index. > The same thing seems to happen when an updated .info file is available, but the .deb has not been rebuilt, e.g.: xfree86-base_4.2.0-11 is in cvs xfree86-base_4.2.0-10 is the most recent build -- xfree86-base_4.2.0-10_darwin-powerpc.deb gets deleted. Since I do not always keep up with every single version of packages like xfree86 or gimp, and some just won't build in the latest release (e.g. currently abiword and openldap), I'd consider it preferable to keep at least the latest .deb of any package. Otherwise good work, though in general I'd feel more comfortable if there was some sort of "verbose" option to tell me what files it actually removes... Thanks, Derek --- This sf.net email is sponsored by: OSDN - Tired of that same old cell phone? Get a new here for FREE! https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390 ___ Fink-users mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] Re: DORBIT_DEF_CORBA_sequence_CORBA_string
On Sunday, August 18, 2002, at 05:56 AM, Philip Mötteli wrote: > Am Sonntag den, 11. August 2002, um 17:15, schrieb Masanori Sekino: >>> configure: running /bin/ksh './configure' --prefix=/sw >>> 'CPPFLAGS=-no-cpp-precomp >> >> Is your /bin/sh a symlink to /bin/ksh? Or do you have any special >> setting about /bin/ksh? Currently, fink only supports zsh as /bin/sh and >> configure script should use /bin/sh. > > Interesting, ls gives the following output: > > -rwxr-xr-x1 rootwheel 449616 Aug 3 04:29ksh > -r-xr-xr-x1 rootwheel 449616 Aug 3 04:28sh > > So ksh and sh have the same size, but are not links. > So it's probably something in my shell settings? > Something's wrong there, yes, though the sizes correspond exactly to the regular sh (==zsh, try a ls -l /bin/zsh!). Still, it is possible that zsh behaves differently when called as "ksh", although I don't understand why it does not pass the string of arguments properly. I'd agree that it is probably the best solution to remove that ksh, since it seems to be a simple copy (you may verify this with cmp beforehand). Also, if for any reason you need a ksh you can always try fink's pdksh or ksh93 for the real thing, and if you should absolutely need it in /bin you can cp it from /sw/bin (though, as Masanori said, this can break some builds with fink, so you might need to rm it again when compiling, but I did not have any problems of that kind with pdksh recently...). HTH, Derek --- This sf.net email is sponsored by: OSDN - Tired of that same old cell phone? Get a new here for FREE! https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390 ___ Fink-users mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] Xterm with Scrollbar in XFree
On Wednesday, August 21, 2002, at 02:48 PM, Allen K. Smith wrote: > Not in front of my mac, but ctrl-middle mouse button, first menu option is > Enable ScrollBar. > > Or add -sb to your launcher. > > > On Wednesday 21 August 2002 11:38 am, Samuel Smith wrote: >> Anyone know of a more mac like xterm with a scroll bar that works in >> XFree >> 86. The XDarwin's rootless >> window server xterm doesn't have a scroll bar. I tried OroboOSX 0.8 and >> although its xterm has a scroll bar it doesn't drag scroll AFAIK and I >> prefer the scroll bar to be on the RHS of the window. >> Eterm also permits drag scroll and allows you to put the scroll bar on the right side. aterm might do the same... Cheers, Derek --- This sf.net email is sponsored by: OSDN - Tired of that same old cell phone? Get a new here for FREE! https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390 ___ Fink-users mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] man pages not working
On Monday, August 26, 2002, at 05:01 AM, Martin Costabel wrote: > For a temporary fix, until the Jaguar stuff is completely sorted out, you > can probably safely remove /sw/bin/man. Another possibility would be to > install the 'man' package instead of the 'manconf' package, but this will > probably force you to install a bunch of other packages you won't want. > Or rather, it should. The package is missing a dependency on groff. without which it dies as well. /usr/bin/man works fine for me, however (both in 10.1 and 10.2). Cheers, Derek --- This sf.net email is sponsored by: OSDN - Tired of that same old cell phone? Get a new here for FREE! https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390 ___ Fink-users mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] Re: Broken vim
On Sat, Aug 31, 2002 at 09:31:41PM -0400, Ian L. wrote: > On 8/31/02 8:59 PM, Thomas Harrington enlightened us with: > > > Those are the same errors that I had. For what it's worth, you get the > > same errors if you grab the Unix source from www.vim.org and try to > > build it yourself. So it's not specifically a fink problem, I think. > > You are correct, Tom. The different errors I received when I tried to > compile Vim via Fink both with X and without Xwindows were the same when I > downloaded the package and compiled it on my own. > > (If you've just tuned in, this was fixed by temporarily renaming > /usr/local/include to something else.) > I can confirm this, it's the same for vim-nox. FWIW, it seems to be a error related to ./configure or gcc 3.1 (is there a new autoconf in the July DevTools?). I got loads of warnings about "user-specified include dir -I/usr/local/include moved in priority since it is already in system include path" or something to that end, with every gcc call. What happens is apparently that only /usr/local/include is in the compiler's own include path, not /usr/include, and consequently some essential .h files are missing in the course of the compile. It looks like configure sets it up this way whenever there is a /usr/local/include present, which seems rather braindead, of course. On second thought, maybe this is a remnant from a developer-gcc3 setup to allow gcc3 to be installed independently from gcc 2.95... Any thoughts? Derek --- This sf.net email is sponsored by: OSDN - Tired of that same old cell phone? Get a new here for FREE! https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390 ___ Fink-users mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] 10.2 -- How to test upgrade fink
On Wednesday, September 4, 2002, at 04:34 AM, Peter Kostka wrote: >>> On Monday, September 2, 2002, at 09:04 PM, Ben Hines wrote: >>> This should do it: cvs -d:pserver:[EMAIL PROTECTED]:/cvsroot/fink login cvs -d:pserver:[EMAIL PROTECTED]:/cvsroot/fink co experimental cd experimental/dmrrsn/fink ./inject.pl >> >> I think a crucial step is missing here: >> >> fink selfupdate-cvs >> Also the unstable tree must be enabled. > > All went smoothly until "fink update-all". > It started promising: > > apt-dev_0.5.4-5_darwin-powerpc.deb apt-shlibs_0.5.4-5_darwin- > powerpc.deb apt_0.5.4-5_darwin-powerpc.deb dpkg_1.9.20-4_darwin- > powerpc.deb ncurses_5.2-8_darwin-powerpc.debf2c_20020123-3_darwin- > powerpc.deb fort77_1.18-4_darwin-powerpc.debg77_3.1-20020420- > 1_darwin-powerpc.deb libsigc++-shlibs_1.0.4-2_darwin-powerpc.deb > libsigc++_1.0.4-2_darwin-powerpc.deb db3-bin_3.3.11-8_darwin- > powerpc.deb db3-doc_3.3.11-8_darwin-powerpc.deb db3-shlibs_3.3.11- > 8_darwin-powerpc.deb db3_3.3.11-8_darwin-powerpc.deb libcapsinetwork- > dev_0.1.1-2_darwin-powerpc.deblibcapsinetwork_0.1.1-2_darwin- > powerpc.deb libmath++-dev_0.0.3-2_darwin-powerpc.deblibmath++_0.0.3- > 2_darwin-powerpc.deb libsigc++-shlibs_1.0.4-3_darwin-powerpc.deb > libsigc++_1.0.4-3_darwin-powerpc.deb ispell_3.2.06-3_darwin-> powerpc.deb > > But > compiling gtkmm-1.2.10 I got the following: > -snip- > Making all in tests > c++ -DHAVE_CONFIG_H -I. -I. -I../.. -I../.. -I../.. -I./.. > -I/sw/include/gtk-1.2 -I/sw/include/glib-1.2 -I/sw/lib/glib/include > -I/usr/X11R6/include -no-cpp-precomp -I/sw/include -g -O2 -Wall > -Wno-unused -c hello.cc > /bin/sh ../../libtool --mode=link c++ -g -O2 -Wall -Wno-unused > -L/sw/lib -o hello hello.o ../gdk--/libgdkmm.la -L/sw/lib > -L/usr/X11R6/lib -lgtk -lgdk -lgmodule -lglib -ldl -lintl -lXext -lX11 > -lm > mkdir .libs > c++ -g -O2 -Wall -Wno-unused -o .libs/hello hello.o -L/sw/lib > -L../gdk--/.libs -lgdkmm -L/usr/X11R6/lib -lgtk -lgdk -lgmodule -lglib > -ldl -lintl -liconv -lXext -lX11 -lm > ld: Undefined symbols: > Gdk_Window::show() > Gdk_Window::Gdk_Window[in-charge](_GdkWindow*) > I think the update is not always smart enough about rebuilding all necessary dependencies. "fink rebuild gtk+ gtk+-shlibs" might help here, though I am not sure about this particular one (but I had a similar problem with gtk+/gdk libs not linking). HTH, Derek --- This sf.net email is sponsored by: OSDN - Tired of that same old cell phone? Get a new here for FREE! https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390 ___ Fink-users mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] dpkg errors prevent update-all
On Wednesday, September 4, 2002, at 05:41 PM, Ronald Florence wrote: > With fink configured in the experimental 10.2 mode, and with the > unstable trees added to my configuration, any effort at `fink > update-all' or `fink install gimp' produces errors like these: > > dpkg -i > /sw/fink/dists/unstable/main/binary-darwin-powerpc/gnome/gnome-libs- > shlibs_1.4.1.7-4_darwin-powerpc.deb > /sw/fink/dists/unstable/main/binary-darwin-powerpc/gnome/gnome-libs- > dev_1.4.1.7-4_darwin-powerpc.deb > (Reading database ... 12350 files and directories currently installed.) > Unpacking gnome-libs-shlibs (from > .../gnome-libs-shlibs_1.4.1.7-4_darwin-powerpc.deb) ... > dpkg: error processing > /sw/fink/dists/unstable/main/binary-darwin-powerpc/gnome/gnome-libs- > shlibs_1.4.1.7-4_darwin-powerpc.deb (--install): > trying to overwrite `/sw/lib/libart_lgpl.2.2.0.dylib', which is also > in package gnome-libs > dpkg-deb: subprocess paste killed by signal (Broken pipe) > Preparing to replace gnome-libs-dev 1.4.1.7-4 (using > .../gnome-libs-dev_1.4.1.7-4_darwin-powerpc.deb) ... > Unpacking replacement gnome-libs-dev ... > dpkg: dependency problems prevent configuration of gnome-libs-dev: > gnome-libs-dev depends on gnome-libs-shlibs (= 1.4.1.7-4); however: > Package gnome-libs-shlibs is not installed. > dpkg: error processing gnome-libs-dev (--install): > dependency problems - leaving unconfigured > Errors were encountered while processing: > > /sw/fink/dists/unstable/main/binary-darwin-powerpc/gnome/gnome-libs- > shlibs_1.4.1.7-4_darwin-powerpc.deb > gnome-libs-dev > ### execution of dpkg failed, exit code 1 > Failed: can't batch-install packages: > /sw/fink/dists/unstable/main/binary-darwin-powerpc/gnome/gnome-libs- > shlibs_1.4.1.7-4_darwin-powerpc.deb > /sw/fink/dists/unstable/main/binary-darwin-powerpc/gnome/gnome-libs- > dev_1.4.1.7-4_darwin-powerpc.deb > > It seems the problem here is some files have wandered from gnome-libs in the previous package version to gnome-libs-shlib in the current one, a problem sometimes occurring due to the split-off of the -shlibs packages. One thing you could try is "sudo dpkg -r --force-depends gnome-libs-dev gnome-libs gnome-libs-shlibs" to purge the old packages, and then do a "fink update gnome-libs-dev". HTH, Derek --- This sf.net email is sponsored by: OSDN - Tired of that same old cell phone? Get a new here for FREE! https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390 ___ Fink-users mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] Re: fink and gimp
On Wednesday, September 4, 2002, at 07:57 AM, Martin Costabel wrote: >>[auda:fink/10.2/CVS] ron% sudo fink install gimp >>Information about 1129 packages read in 1 seconds. >>Failed: Internal error: node for db3-shlibs already exists >> I'm still eager to install gimp, and would welcome suggestions. > > You should rebuild db3 first (and fink reinstall db3-shlibs if it > doesn't do it automatically). > > As I understand the "node for ... already exists" problem, the > following happens: > gimp BuildDepends on gnome-libs-dev which in turn BuildDepends on "db3 > (>= 3.3.11-8)". You probably have an earlier version of db3-shlibs > installed, and when fink constructs its dependency list, it first > encounters a dependence on db3 without specified version, so it takes > the one you already have installed. Later on it sees the dependence on > version >=3.3.11-8, but it is not intelligent enough to replace the > earlier dependence on an older version by the newer one. > > In any case, the new gimp will need db3-shlibs-3.3.11-8. > Thanks for that bit of background info, but that does not seem to be the whole story. On larger updates, I have frequently seen that error without any conceivable dependency problem. Typical scenario: fink update-all ... building libsigc++ fails for reasons I have not investigated yet (this is on experimental). To got the update going, I manually select all other missing packages and do a 'fink update ...' on them. > Failed: Internal error: node for libsigc++ already exists It finally can be brought to work again when cutting the number of packages down to a sufficiently small number, but the error seems to occur whenever a very large number of packages is to be updated. I think Max actually mentioned this as a bug. So the update needs to be split up into many small chunks whenever a lot of packages need updating, making the process rather tedious of course. It would be really helpful if packages that don't build successfully could be put on "hold", so that "update-all" would just ignore them. Cheers, Derek --- This sf.net email is sponsored by: OSDN - Tired of that same old cell phone? Get a new here for FREE! https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390 ___ Fink-users mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] Congigure Error
On Monday, September 23, 2002, at 11:05 AM, Myers, Parrish wrote: >> On Monday, September 23, 2002, at 11:01 PM, Myers, Parrish wrote: >>> >>> I have 10.2 installed... I also have a 500MB UFS partition on the boot >>> harddrive. The system automatically mounts the driver under >>> "/Volumes/Carbon\ UFS\ HD/"... I have read the archives, and they >>> seem to >>> indicate that simply symlinking the "Volumes" mount to "/sw" will >>> work. >>> >> >> You should rename your ufs partition so it contains no spaces.. Many >> unix utilities and shell scripts have problems with spaces in paths. >> You could also make it mount directly under /sw by adding a line like this in /etc/fstab LABEL= /sw ufs rw 1 1 and running 'niload fstab . < /etc/fstab'. Autodiskmount then will take care of the rest. Derek --- This sf.net email is sponsored by:ThinkGeek Welcome to geek heaven. http://thinkgeek.com/sf ___ Fink-users mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] Re: KDE
On Sunday, September 29, 2002, at 12:32 AM, Joel Hacker wrote: > > On Saturday, September 28, 2002, at 01:37 AM, Steve Wall wrote: >> Can't speak for Joel but I don't have Norton SystemWorks installed and >> I can't >> get aalib to compile either. I posted about this on the 18th >> (Subject: aalib Update >> for 10.2 Fails) and got no response. I just did a "fink >> selfupdate-cvs" and tried >> again and it still won't compile. Errors are identical to what I saw >> the last time. >> >> Anyone got any clues? If I am interpreting the description of aalib >> correctly it is >> an ascii-art library and it can't be that critical to GIMP. So I >> should be able >> to copy the .info file in my local finkinfo directory tree and remove >> the dependency >> without breaking GIMP too much, right? >> >> Steve Wall >> > > I finally got Gimp and aalib to install (and all the parts of KDE that > I wanted, namely the base and Koffice and KDEgames, don't need the > whole bundle). As far as I can tell it isn't the whole SystemWorks > suite that is causing the prob but Anti-virus auto-protect > specifically. Try turning off any and all unnecessary background > processes first. Anti-virus, fax listeners, special hardware drivers > that "listen" for events like mouse drivers, etc. Then make note of > everything that you turned off and try again. For those things with no > gui option to turn them off you can either use top to get their process > id's and then sudo kill (process id #) or use the Process Viewer > application to kill them. Some of them may require you to log in as > root to kill them via Process Viewer. Be careful what you kill though, > you could crash you system, but it isn't likely that you could do any > lasting damage. > That probably wouldn't help much as he said he is not using any Norton tools in the first place :-). Moreover, Steve's problem is a simple linker error, no crashing patch or kernel panics, if I read him correctly. Steve, the symbols ld is missing in your installation belong to libslang, not to ncurses, so I'd suggest to check if slang is installed, and possibly rebuilt under 10.2. It seems it did not include -lslang in the linker commands at all, though it obviously included the slang headers, so something must have gone wrong there. If in doubt just try to rebuild slang, the 10.2 updater might not always detect which packages need to be recompiled. HTH, Derek --- This sf.net email is sponsored by:ThinkGeek Welcome to geek heaven. http://thinkgeek.com/sf ___ Fink-users mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] Re: KDE
On Monday, September 30, 2002, at 12:30 AM, Joel Hacker wrote: > Actually, he didn't say he wasn't using ANY Norton tools, just that he > wasn't using SystemWorks specifically. I tend to take people at their > literal word and endevour not to draw too many inferences. My > thought is that if auto-protect can cause probs then it is possible > that there may be some other "always on" processes that can cause > similar probs. I would be deeply interested in knowing about any and > all of them, if indeed more than just NAV auto-protect can cause this > phenomenon. > Right, I just wasn't aware of other tools that might destabilize your system in a similar way, for a virus scanner it's somewhat conceivable, though I would not want anything like this on my system... My main point anyway was that this was not a case of a system-crashing compile failure. >> Steve, the symbols ld is missing in your installation belong to >> libslang, not to ncurses, so I'd suggest to check if slang is >> installed, and possibly rebuilt under 10.2. > > Would "fink selfupdate-cvs" fix this? If not, are there any other > probs that may arise as a result of non-updated libraries that you > know of that the rest of us should know about? > Well, this case was apparently a problem of linking to an internal library of the package, not of another package, but it happened to me in a few instances. I assume these problems partly occurred because I had the experimental 10.2 tree installed before the official upgrade recipe was around, but it's possible this could happen to someone else too, as the upgrade script might not be tested for every possible configuration. As I understand it, fink update will not automatically recognize if a package needs to be recompiled for Jaguar -- selfupdate will install the package descriptions into the 10.2 tree, but if the version number did not increase, an update-all or 'update package' will not rebuild the package if it is already installed. This is why the update-fink.sh script needs to be executed before other packages are updated to 10.2. So, whenever I found a package that would not build under Jaguar, I looked up the Dependencies and checked if they had already been rebuilt since the system upgrade, and if not, an explicit 'rebuild package' would often fix the problem. Again, if you have upgraded fink through the regular path, you probably should not see such problems. Derek --- This sf.net email is sponsored by:ThinkGeek Welcome to geek heaven. http://thinkgeek.com/sf ___ Fink-users mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] Fink selfupdate failed
On Monday, September 30, 2002, at 08:32 PM, Justin Wright wrote: >> tcsh% fink selfupdate-cvs |& tee ~/fink_selfupdate-cvs >> tcsh% grep 'no longer in the repository' ~/fink_selfupdate-cvs >> >> Note the '|&' construct -- in the tcsh shell [OSX default], this sends >> both standard output & standard error down the pipe. If you don't >> capture >> stderr then chances are you won't capture error messages. The same >> trick >> can be donw in bash/sh but I forget the syntax on that side... > >bash$ fink selfupdate-cvs >fink.log 2>&1 > > is one easy way to do it. '>' will redirect stdout, '2>' for stderr. > For catching the output in the terminal as well, that would be /sw/fink> fink selfupdate-cvs 2>&1 | tee ~/fink.log or alternatively, just redirect the output as above: /sw/fink> fink selfupdate-cvs >fink.log 2>&1 & and watch the logfile with: tail -f ~/fink.log Cheers, Derek --- This sf.net email is sponsored by:ThinkGeek Welcome to geek heaven. http://thinkgeek.com/sf ___ Fink-users mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] fink permissions
On Tuesday, October 1, 2002, at 06:07 PM, Chris Devers wrote: > On Tue, 1 Oct 2002, Andrew Hartung wrote: > >> I live alone and have a firewall up. > > Like winter clothing, security practices work best in layers. If your > computer is attached to the internet, it's never really alone. Firewalls > may block certain kinds of unwanted traffic, but you never know when > someone will find a hole -- all you can do is assume that, sooner or > later, someone you don't want to be there will get inside. With that in > mind, it's still a good idea to compartmentalize how much damage an > intruder can do, and leaving in reasonable permissions -- certainly not > global write access everywhere -- is a big step towards doing that. > > If you really want to be able to create & delete system files in the > GUI, > I'm sure that there are shareware tools for doing this. Seems like your > options are to get used to the shell, pay for shareware, or just accept > that sooner or later someone will get past your firewall and, when that > happens, it will have been your main/only defence. > > *shrug* > Just to embark on the layer concept, for the most part it should be sufficient to make things group-writable by wheel, since you are naturally admin user (not that there's a guarantee against someone intruding under an admin user account...) >> The reason I want it that way is to make it easier to delete old >> source, >> builds, etc... using the finder. > > ...if that's all you want to do, you can make a cron job out of it. :) > Furthermore, that would basically just require the permissions of /sw/src to be changed. There probably is not much to be said against that (unless someone should actually install a spoofed tarball, but that's probably a bit far-fetched, and many packages also test the checksum before compiling). As for anything else, there's also the point of protection against yourself. You don't really want to delete anything inside /sw/bin, /sw/lib etc. from the finder, so it's best not to enable it in the first place! --- This sf.net email is sponsored by: DEDICATED SERVERS only $89! Linux or FreeBSD, FREE setup, FAST network. Get your own server today at http://www.ServePath.com/indexfm.htm ___ Fink-users mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] fink permissions
On Tuesday, October 1, 2002, at 10:18 PM, Joel Hacker wrote: > On Tuesday, October 1, 2002, at 06:29 PM, Derek Homeier wrote: >> Just to embark on the layer concept, for the most part it should be >> sufficient to make things group-writable by wheel, since you are >> naturally admin user (not that there's a guarantee against someone >> intruding under an admin user account...) >> > > Or perhaps change ownership of everything to you and leave group and > all access "off". > > chown -R YourUserId:wheel /sw > > That should do the trick. However, I recommend leaving it as-is and > using the command-line. > I'd not only recommend that but strictly warn against doing otherwise, as it would probably corrupt your fink installation even worse than manually removing files! A number of programs depend on being installed under a certain UID, and at best, will not work if owned by anyone else -- at worst, they might become unpredictable security risks. After all, having sensitive programs, sshd for example, not easily modified by regular users, even if it is your own account, was the basic idea of privilege separation in the first place. Actually, some programs will refuse to run not only if owned by someone else than they are supposed to, but even if they live in a world- or group-writable directory -- just remembering the notorious sendmail problem on MacOS X. > Did you know that you could use the command-line to type the command > and then drag any folder into the terminal window from the finder to > select the files? > > try this: type "sudo rm -rf " in the command-line (without the > quotation marks of course, be sure to include the space at the end) and > then drag however many folders and files you want from the finder into > the terminal window and let them go. Instantly the path(s) to the > file(s) will appear in the terminal and then you hit return, enter your > password, and your files are gone forever, no matter who owns them! If > That's a nice feature of Terminal :-) Though personally, I'll probably stick to using wildcards from the shell. Cheers, Derek --- This sf.net email is sponsored by: DEDICATED SERVERS only $89! Linux or FreeBSD, FREE setup, FAST network. Get your own server today at http://www.ServePath.com/indexfm.htm ___ Fink-users mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] tex/doxygen build issues
On Thursday, October 3, 2002, at 11:34 AM, Chris Zubrzycki wrote: > On Thursday, October 3, 2002, at 09:38 AM, Viktor Haag wrote: > >> Rohan Lloyd writes: >>> On Thursday, October 3, 2002, at 01:50 AM, Viktor Haag wrote: >> I suspect the hitch is you need to install tetex-macosx: I'm pretty sure that it's required to build doxygen (it and its dependencies provide some of the hyperlinking stuff that's required?). >>> >>> Yep, that fixed it. >>> >>> Looks like a missing dependency in the doxygen package. >> >> I should note that I am not installing Doxygen through Fink. I'm >> building and installing it on my own from source in >> /usr/local. (However, using some tools installed from fink, >> i.e. tetex.) >> >> That may make a difference. I notice that Fink's doxygen package >> is still using the 1.2.16 build. The latest available version is >> 1.2.18, and I'm building from CVS (although I'm currently *using* >> 1.2.16 because of various features/bugs in the later versions). > > for me also installing tetex-macosx fixed the problems i was having. i > cc'ed the maintainer so he knows that it needs the added dependancy and > a revision bump. > Just for clarification, the actual problem is, as noted in a previous post, that pdflatex cannot use the graphics in eps format. tetex-macosx includes the "epstopdf" script, which is then probably automatically found and used by doxygen, so this fix will probably work both for fink and non-fink builds. Actually, I noticed tetex-macosx also provides an alternative pdflatex which can directly include eps graphics, so maybe that's what is used. Might be the reason why epstopdf is not included with pdflatex, too. HTH, Derek --- This sf.net email is sponsored by:ThinkGeek Welcome to geek heaven. http://thinkgeek.com/sf ___ Fink-users mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] tex/doxygen build issues
On Thursday, October 3, 2002, at 01:51 PM, Derek Homeier wrote: > Just for clarification, the actual problem is, as noted in a previous > post, that pdflatex cannot use the graphics in eps format. tetex-macosx > includes the "epstopdf" script, which is then probably automatically > found and used by doxygen, so this fix will probably work both for fink > and non-fink builds. > Actually, I noticed tetex-macosx also provides an alternative pdflatex > which can directly include eps graphics, so maybe that's what is used. > Might be the reason why epstopdf is not included with pdflatex, too. > The altpdflatex is in fact just a wrapper script to call ps2pdf which has two disadvantages compared two pdftex: 1. Any hypertext structure of the tex/dvi files is lost during the postscript conversion, thus there won't be any links to chapters, figures, citations etc. in the output. 2. The script contains a hardcoded path to gs and ps2pdf in /usr/local/bin and thus does not work with the fink ghostscript distribution. This should be fixed in the fink package. Cheers, Derek --- This sf.net email is sponsored by:ThinkGeek Welcome to geek heaven. http://thinkgeek.com/sf ___ Fink-users mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] Any one ported PVM or MPI yet?
On Friday, October 4, 2002, at 12:51 AM, Stephen Hocking wrote: > Inquiring minds and all that - the Xserve's significantly reduced power > consumption and heat production over P4s is making it rather > interesting... > Using LAM-MPI here, although on a cluster of mere Quicksilvers. The port was not very difficult, I have not ventured on making a fink package of it yet since we use Absoft f90 (though this could probably done as a source package dependent on system-fortran). Meanwhile, I can send you the patch if you are interested. Cheers, Derek --- This sf.net email is sponsored by:ThinkGeek Welcome to geek heaven. http://thinkgeek.com/sf ___ Fink-users mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] Any one ported PVM or MPI yet?
On Monday, October 7, 2002, at 09:51 AM, Brian W. Barrett wrote: > Developing LAM/MPI on OS X here :). Right now, there are some bugs in > the > stable release of LAM/MPI (6.5.6) that can cause issues on OS X. > Namely: > > * Our cleanup script can't remove our temporary directory properly, as > OS X doesn't allow remove() to be called on a directory That seemed to be fixed already, at least I never saw that error since upgrading to 6.5.6. > * A strange behavior in select() was causing some internal acocunting > errors that would only allow you to run about 60 jobs before having > to > clean up the environment (via the lamclean command) Haven't seen that one yet, maybe I've not run quite the magical number of jobs yet. > * A couple of case-sensitivity issues existed (like the standard > wrapper > compilers for MPI are generally called mpicc and mpiCC). > Yep, remember that one, I've got around it by installing the executables on a UFS partition. Everything else seems to work flawlessly, indeed; the full MPI testsuite completes without errors, and I've even run the extra LAM tests from the 6.6-beta with partial success. One thing to consider when running jobs across multiple machines is MacOS X's awkward NFS naming scheme. Thanks for the great work! Derek --- This sf.net email is sponsored by:ThinkGeek Welcome to geek heaven. http://thinkgeek.com/sf ___ Fink-users mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] Capturing Distro to CD
On Tuesday, October 15, 2002, at 09:44 AM, Josh Kuperman wrote: > I upgraded my G3 iMac and just copied over /sw from my powerbook. I > ran a resinstall of xfree86 and everything seems to work equally well > on both machines. So that at least was simple. > > I was unable to figure out a good way to simply capture my install to > CDs. I realized what would be very handy was a truly (sanely) > compressed version that could be used for a binary install. I have a > decent broadband connection, but I know some people who don't. Has > anyone thought of a decent way to package up all of fink for > distribution this way. > I found the most convenient way of sharing a fink installation or backing it up on CD, and still getting a clean install of the software, is to do a 'fink scanpackages' and just copy the whole content of /sw/fink/dists (you can omit the "finkinfo" directories, but that's insignificant in terms of storage space) over to the CD (or export it over NFS, AFP or whatever). You then just need to point your /sw/etc/apt/sources.list to that archive and can do a binary installation of all the packages. The .deb files are also fairly compact (effectively, the contents are compressed about as well as with tar.gz), so you can get a fairly complete distribution on one CD (I don"t have a lot of KDE stuff installed here, if you add that you'd probably need a second disk). To get a list of the packages that are actually installed on the original system, "dpkg --get-selections" is quite useful, too. HTH, Derek --- This sf.net email is sponsored by:ThinkGeek Welcome to geek heaven. http://thinkgeek.com/sf ___ Fink-users mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-users
[Fink-users] CVS problem
I seem to have a problem with my cvs setup/permissions in /sw/fink since upgrading to the 10.2 tree. I did not run the upgrade script on that system since I had already been using the experimental 10.2 tree before. So maybe someone can give me a hint what exactly should be the ownerships and permissions in the /sw/fink hierarchy. The fink installation itself works flawlessly, otherwise, the only problem I have is that every 'selfupdate-cvs' will fail with the dreaded cvs server: Updating 10.1/unstable/main/finkinfo/x11-wm cvs server: Updating 10.2 cvs server: Updating 10.2/stable cvs update: move away 10.2/stable/.cvsignore; it is in the way C 10.2/stable/.cvsignore cvs server: Updating 10.2/stable/crypto cvs update: move away 10.2/stable/crypto/.cvsignore; it is in the way C 10.2/stable/crypto/.cvsignore cvs server: Updating 10.2/stable/crypto/finkinfo cvs server: Updating 10.2/stable/main cvs update: move away 10.2/stable/main/.cvsignore; it is in the way C 10.2/stable/main/.cvsignore cvs server: Updating 10.2/stable/main/finkinfo cvs server: Updating 10.2/stable/main/finkinfo/base cvs update: move away 10.2/stable/main/finkinfo/base/apt-0.5.4-5.info; it is in the way C 10.2/stable/main/finkinfo/base/apt-0.5.4-5.info cvs update: move away 10.2/stable/main/finkinfo/base/apt-0.5.4-5.patch; it is in the way C 10.2/stable/main/finkinfo/base/apt-0.5.4-5.patch cvs update: move away 10.2/stable/main/finkinfo/base/base-files-1.6-1.info; it is in the way error for every single .info and .patch file in the 10.2 tree (none in the 10.1 tree!). Only way to get the update I found is to remove all infos, patches and .cvsignores in 10.2 before the selfupdate. So I have tried to figure out if cvs is unhappy with the permissions in any way, it turned out that initially most of the subdirectories were owned by root:wheel, which cvs could not update at all (that seems reasonable, since it is operating as user "derek" at that time). So I made sure all ownerships were changed to "derek" with the exception of the binary-* directories and the .deb files themselves. Result see above -- permissions are exactly the same as in the 10.1 dir, none of the info/patch files has even been touched since the last selfupdate, yet cvs invariably runs amok. Any idea what's screwed up here, or any other settings for the selfupdate I might have to look into? TIA, Derek --- This sf.net email is sponsored by:ThinkGeek Welcome to geek heaven. http://thinkgeek.com/sf ___ Fink-users mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] CVS problem
>> I seem to have a problem with my cvs setup/permissions in /sw/fink
>> since upgrading to the 10.2 tree. I did not run the upgrade script on
>> that system since I had already been using the experimental 10.2 tree
>> before. So maybe someone can give me a hint what exactly should be
>> the ownerships and permissions in the /sw/fink hierarchy.
>> The fink installation itself works flawlessly, otherwise, the only
>> problem I have is that every 'selfupdate-cvs' will fail with the
>> dreaded
>>
>
> Try cleaning your cvs with:
>
> cd /sw/fink
> sudo find . -name 'CVS' -type d -exec rm -rf {} \;
>
> Then fink selfupdate-cvs will prompt you for your settings again.
>
Thanks, Ben, that fixed it! It also skips searching the 10.1 tree
altogether now.
Cheers,
Derek
---
This sf.net email is sponsored by: viaVerio will pay you up to
$1,000 for every account that you consolidate with us.
http://ad.doubleclick.net/clk;4749864;7604308;v?
http://www.viaverio.com/consolidator/osdn.cfm
___
Fink-users mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] Capturing Distro to CD
On Saturday, October 19, 2002, at 10:54 AM, Mike wrote: On Tuesday, October 15, 2002, at 09:44 AM, Josh Kuperman wrote: I couldn't even figure out how to get a decent tar archive, let alone a multidisk tar archive. (My first foolish attempt to make a .tgz or /sw/src resulted in a file larger than simply copying -- and was too largre to burn to a single CD anyhow. Copying the /sw/fink/dists archive and using dpkg --get-selections is a much cleaner solution than tarballing the whole directory structure is the cleanest solution, but if you really want the whole Fink tree backed up, here are some useful tar tips: If you pass tar the -z flag, as opposed to gzipping the resulting tar file separately, you wind up with a .tgz, which is different than a tar.gz. The first format contains each file individually gzipped, while the latter contains everything archived into a single .tar file, which is then gzipped as a whole, which will almost always result in greater compression. Pardon me, but that's nonsense! The 'z' flag just pipes the (archived) output internally through gzip, which results in exactly the same compressed file as doing it manually. Wether it is called .tgz or .tar.gz depends solely on what filename you specify after the 'f'. You can simply ungzip the .tgz to get a regular .tar file, but you cannot untar it to get the individually gzipped contents. I have largely switched to using the .tar.gz naming convention, mostly for clarity, but also because some MacOS apps can't make sense of the .tgz suffix -- IIRC, StuffitExpander does something utterly stupid like first uncompressing it into a ".tgz.tar" archive and then untarring that... Just some data: miranda:6039> tar cpzf /tmp/manpages.tgz sw/share/man miranda:6040> tar cpf - sw/share/man | gzip -c > /tmp/manpages.tar.gz miranda:6041> ll /tmp/manpages* -rw-r-1 derekwheel 3456211 Oct 19 15:49 /tmp/manpages.tar.gz -rw-r-1 derekwheel 3456211 Oct 19 15:41 /tmp/manpages.tgz Any reduction in files size you can get from doing the gzip manually is by giving it the '-9' flag to enforce higher compression, but the effect is rather small: miranda:6042> tar cpf - sw/share/man | gzip -9c > /tmp/manpages9.tar.gz -rw-r-1 derekwheel 3432947 Oct 19 15:51 /tmp/manpages9.tar.gz If you want really better compression, you can use the 'j' option with fink's tar to create a bzip2 archive, which typically gives about 30% reduction vs. gzip (and also is significantly slower than even gzip -9): miranda:6047> /sw/bin/tar cpjf /tmp/manpages.tar.bz2 sw/share/man -rw-r-1 derekwheel 2680388 Oct 19 15:54 /tmp/manpages.tar.bz2 It also is somewhat less compatible, but bzip2 now even comes preinstalled with 10.2, so there shouldn't be a great problem with getting it unarchived. HTH, Derek --- This sf.net email is sponsored by: Access Your PC Securely with GoToMyPC. Try Free Now https://www.gotomypc.com/s/OSND/DD ___ Fink-users mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-users
[Fink-users] linking error in mjpegtools-1.6.0-4
On 10.2.1, current cvs, updating fails at this point: gcc -I/sw/include -Wall -Wunused -o lav2yuv lav2yuv.o lav_common.o -L/sw/lib -L/Volumes/Galapagos/src/mjpegtools-1.6.0-4/mjpegtools-1.6.0/utils -lmjpegutils ./.libs/liblavfile.a -L/sw/include/quicktime -lquicktime -ldl -lpthread -lpng -lz -lm -ldl -L/sw -lm -lm -lpthread -ldl ./.libs/liblavjpeg.a -ljpeg -ldv -lpopt -lm -lSDLmain -lSDL -lm -lglib -lpthread -ldl ld: Undefined symbols: _y4m_dar_4_3 _y4m_guess_sar make[2]: *** [lav2yuv] Error 1 Anyone knows where to find these? Thanks, Derek --- This sf.net email is sponsored by: Influence the future of Java(TM) technology. Join the Java Community Process(SM) (JCP(SM)) program now. http://ads.sourceforge.net/cgi-bin/redirect.pl?sunm0003en ___ Fink-users mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] wxpython-wxgtk 2.3.3.x: when will I be able to get it with Fink?
On Thursday, October 24, 2002, at 09:34 PM, David Knapik wrote: I'm currently running wxpython-wxgtk 2.3.2.1-5 on Mac OS 10.1.5 and I need to upgrade to 2.3.3.x (because one of my apps' new version uses it), which is mentioned extensively on www.wxpython.org, but which I can't seem to get using Fink. That would be okay, and I'd just try it out directly from them (especially since they have a Mac OS X version now -- 2.3.3.1) but I don't know where to move the files their installer gives me in order for them to work (i.e. where do I move them in relation to my Fink /sw folder?). Their installer, of course, installs into my /users/mylastname directory, but my apps still seem to run off what's in the /sw folder. any advice? I am not sure if that binary they provide corresponds to the wxpython-wxgtk fink provided, and I am also not sure if it is 100% Jaguar-clean (it seems to be basically a 10.1 version, and their own /usr/local/bin/python you need to install with the MachoPython package seems to have some problems loading dylibs. It should also be unnecessary, since 10.2 provides python in /usr/bin). However, it basically appears to work with the occasional error message about unsupported locales etc., and maybe it will crash at some point... Actually, it should not install into your home directory but into /Library and /usr/local. So to use it with your fink python there are basically two options: set PYTHONHOME=/sw:/Library/Frameworks/Python.framework/Versions/2.2 or add /Library/Frameworks/Python.framework/Versions/2.2/lib/python2.2/site- packages to your PYTHONPATH variable. The former has the disadvantage that PYTHONHOME appears to accept only two elements, since I have it normally set to "/usr:/sw" to use both the system-installed and the fink modules, I cannot add another one. I don't know if it includes Gtk support, though. I have tried to use this wxPython installation to load the plt module from scipy, which works, but trying to plot only produces an empty window and an endlessly spinning rainbow wheel. I was in fact not able to kill that window by any other way than exiting python. So, your mileage in getting this to work with your applications may vary, but the above should at least allow you to load it. HTH, Derek --- This sf.net email is sponsored by: Influence the future of Java(TM) technology. Join the Java Community Process(SM) (JCP(SM)) program now. http://ads.sourceforge.net/cgi-bin/redirect.pl?sunm0004en ___ Fink-users mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-users
