On 18/08/2011 17:06, Jack Howarth wrote:
> On Wed, Aug 17, 2011 at 07:50:32PM -0400, Alexander Hansen wrote:
>> -----BEGIN PGP SIGNED MESSAGE-----
>> Hash: SHA1
>>
>> (Sorry. I wanted to handle this on -users, but my default reply-to is
>> - -beginners. It's corrected in this message.)
>>
>> On 8/17/11 11:11 AM, Luigi Ernesto Zanotti wrote:
>>> Hello, here is my machine cobfiguration:
>>>
>>> iMac with Mac OS X Lion (10.7) Fink manager version: 0.31.0
>>> Distribution version: selfupdate-rsync Mon Aug 15 21:26:49 2011,
>>> 10.7, x86_64 Trees: local/main stable/main Xcode Version: 4.1-1
>>>
>>> After the successful installation of "fink" I tried to install
>>> "cernlib", that I need to keep operating some old codes, still useful
>>> for my job.
>>>
>>> 1 - As indicated, I moved "/usr/local" to "/usr/local-old" 2 - I
>>> launched then "fink install cernlib", from "su". The procedure
>>> stopped after a time with the message error reported here:
>>>
>>> -W -Wall -Wwrite-strings -Wcast-qual -Wstrict-prototypes
>>> -Wmissing-prototypes -Wmissing-format-attribute -pedantic
>>> -Wno-long-long -Wno-variadic-macros -Wno-overlength-strings
>>> -Wold-style-definition -Wc++-compat -DHAVE_CONFIG_H -I. -Ifortran
>>> -I../../gcc-4.6.1/gcc -I../../gcc-4.6.1/gcc/fortran
>>> -I../../gcc-4.6.1/gcc/../include
>>> -I../../gcc-4.6.1/gcc/../libcpp/include -I/sw/include -I/sw/include
>>> -I../../gcc-4.6.1/gcc/../libdecnumber
>>> -I../../gcc-4.6.1/gcc/../libdecnumber/dpd -I../libdecnumber
>>> -I/sw/include -I/sw/include -DCLOOG_INT_GMP -DCLOOG_ORG
>>> -I/sw/include ../../gcc-4.6.1/gcc/fortran/module.c -o
>>> fortran/module.o ../../gcc-4.6.1/gcc/fortran/module.c:110:1: error:
>>> conflicting types for 'pointer_t'
>>> /usr/include/mach/vm_types.h:40:26: note: previous declaration of
>>> 'pointer_t' was here make[3]: *** [fortran/module.o] Error 1 make[3]:
>>> *** Waiting for unfinished jobs.... rm gfortran.pod gcc.pod make[2]:
>>> *** [all-stage2-gcc] Error 2 make[1]: *** [stage2-bubble] Error 2
>>> make: *** [all] Error 2 ### execution of /var/tmp/tmp.1.PlBSr4
>>> failed, exit code 2 Removing runtime build-lock... Removing
>>> build-lock package... /sw/bin/dpkg-lockwait -r
>>> fink-buildlock-gcc46-4.6.1-1000 (Reading database ... 8953 files and
>>> directories currently installed.) Removing
>>> fink-buildlock-gcc46-4.6.1-1000 ... Failed: phase compiling:
>>> gcc46-4.6.1-1000 failed
>>>
>>> 3 - I then checked the file "/usr/include/mach/vm_types.h", and
>>> renemed it, just to see what could happen, to
>>> "/usr/include/mach/vm_types-old.h": the procedure stopped again with
>>> the following message:
>>>
>>> ranlib: file: .libs/libgfortran.a(_dim_i16.o) has no symbols ranlib:
>>> file: .libs/libgfortran.a(_mod_i16.o) has no symbols libtool: link:
>>> ( cd ".libs"&& rm -f "libgfortran.la"&& ln -s "../libgfortran.la"
>>> "libgfortran.la" ) make: *** [all] Error 2 ### execution of
>>> /var/tmp/tmp.1.bpqXKH failed, exit code 2 Removing runtime
>>> build-lock... Removing build-lock package... /sw/bin/dpkg-lockwait
>>> -r fink-buildlock-gcc46-4.6.1-1000 (Reading database ... 8953 files
>>> and directories currently installed.) Removing
>>> fink-buildlock-gcc46-4.6.1-1000 ... Failed: phase compiling:
>>> gcc46-4.6.1-1000 failed
>>>
>>> The first two lines were preceded by many other similar ones.
>>>
>>> 4 - Being clearly the error due to tha absence of "vm_types.h", as a
>>> last attempt, I simply removed 'pointer_t' from "vm_types.h" and
>>> started again. This time the procedure ended without errors (at
>>> least, so it seems).
> I'm splitting your message up into two parts. This part will address
> whether your gcc46 build is, in fact, sane. I didn't have to do
> anything special on my own 10.7 system to get gcc46 to build.
>
> The gcc46 maintainer (cc'ed) would be the best one to debug the
> issue--you shouldn't ever have to edit system-provided headers.
> Alexander Hansen, Ph.D.
> Luigi,
> This looks exactly like...
>
> https://trac.macports.org/ticket/24541
>
> You probably have been editing and symlinking headers in /usr/include. This is
> totally unsupported and wrong behavior. Depending upon how badly you have been
> abusing the headers in /usr/include, you may need to do a clean installation
> of MacOS X to purge out the damage.
> Jack
>
On 18/08/2011 16:47, Remi Mommsen wrote:
> Hi,
>
> On Aug 18, 2011, at 1:52 AM, Alexander Hansen wrote:
>
>> -----BEGIN PGP SIGNED MESSAGE-----
>> Hash: SHA1
>>
>> On 8/17/11 11:11 AM, Luigi Ernesto Zanotti wrote:
>>> Hello, here is my machine cobfiguration:
>>>
>>> iMac with Mac OS X Lion (10.7) Fink manager version: 0.31.0
>>> Distribution version: selfupdate-rsync Mon Aug 15 21:26:49 2011,
>>> 10.7, x86_64 Trees: local/main stable/main Xcode Version: 4.1-1
>>>
>>> After the successful installation of "fink" I tried to install
>>> "cernlib", that I need to keep operating some old codes, still
>>> useful for my job.
>>>
>> <snip>
>>> 5 - To check the result, I looked for "pawX11" and "paw++", and
>>> tried them. Here the outputs:
>>>
>>> PAW++ [1] func/plot sin(x)/x Paw++> /FUNCTION/PLOT 'sin(x)/x' '-10'
>>> '10' 'C' CSALLO: pointer difference too large to be represented by
>>> integer. You probably need to link PAW statically on 64-bit systems.
>>>
>>>
>>> *** Application has died or is running in a loop. ***
>>>
>>> Kxterm has been disconnected from application. Select Exit from the
>>> File menu to exit Kxterm.
>>>
>>>
>>> PAW> func/plot sin(x)/x Lower limit (<CR>=) -10. Upper limit (<CR>=)
>>> 10. CSALLO: pointer difference too large to be represented by
>>> integer. You probably need to link PAW statically on 64-bit systems.
>>>
>>> At this point the absence of "pointer_t" in "vm_types.h" seems to be
>>> the origin of the malfunction, but I don't know how to bypass the
>>> conflict with the module
>>> "/sw/src/fink.build/gcc46-4.6.1-1000/gcc-4.6.1/gcc/fortran/module.c".
>>>
>>> Does anybody know how to get out of this problem?
>>>
>>> Thanks in advance. Yours Luigi Zanotti
>>>
>> I'll address the remainder of the message to the cernlib maintainer,
>> since he would be the most likely one to know if, in fact, the package
>> is currently working properly on Lion.
> paw is currently broken on Lion. I haven't had time to look into it. However,
> it runs fine under gdb, i.e. if you do 'gdb pawX11', it works.
>
> Remi
After some other useless attempts to get "gcc46" working, I decided to
follow your advice to reinstall the system on my iMac from the original
"Snow Leopard" install DVD. So I did, and upgraded the system up to
10.6.8 (Xcode 3.2.6).
I then installed "fink 0.31.0" and run successfully "fink install
cernlib", so I have now "paw" and the libraries that I needed
functioning, using "gfortran" and "c++-4" provided in "/sw/bin" (I have
a fake main in c++ to launch the fortran programs as subroutines, among
the tests).
I am now considering to try and go to "Lion" and "Xcode 4.1.1", after a
full backup of the system.
What should I expect now, starting with a "clean" system? Are there any
news?
Thank you again for the valuable help. Yours,
Luigi Z.
--
Prof. Luigi E. Zanotti
Physics Department "G. Occhialini"
University of Milano-Bicocca& INFN
Piazza della Scienza 3 - 20126 Milano Italy
Tel. +39 02 6448 2460
[email protected]
------------------------------------------------------------------------------
EMC VNX: the world's simplest storage, starting under $10K
The only unified storage solution that offers unified management
Up to 160% more powerful than alternatives and 25% more efficient.
Guaranteed. http://p.sf.net/sfu/emc-vnx-dev2dev
_______________________________________________
Fink-users mailing list
[email protected]
List archive:
http://news.gmane.org/gmane.os.macosx.fink.user
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-users
