On Mon, September 7, 2009 4:12 am, Steffen Schmidt wrote:
> Hi Bill,
>
>
> On Sep 4, 2009, at 17:52 , William G. Scott wrote:
>
>> Hi Steffen:
>>
>>
>> On Sep 4, 2009, at 7:43 AM, Steffen Schmidt wrote:
>>
>>> Hi ccp4 build fails (see below) - I recompiled fink on snow leopard
>>> (core duo) using fink in 64bit only configuration. Besides there is
>>> also the problem that e.g. gd2 compilations fails and that it
>>> complains the the .la (here /usr/lib/libiconv.la) is missing...
>>>
>>> Steffen
>>>
>>> fort77 -o xloggraph msg_box.o ZDr2d.o EditString.o hardcopy_ctrl.o
>>> xloggraph.o log_file.o tom_fortran_interface.o bits_and_pieces.o
>>> graphics.o -g -Os -I/usr/include/X11R6 -L/usr/X11R6/lib -lXaw -lXmu
>>> -lXt -lSM -lICE -lXpm -lXext -lX11 -lm -lgfortran -L/sw64/lib/
>>> gcc4.4/lib
>>> ld: warning: in bits_and_pieces.o, file is not of required
>>> architecture
>>> ld: warning: in graphics.o, file is not of required architecture
>>
>> That's worrying, because it suggests for some reason it is building
>> the object files in 32-bit and therefore can't link.
>>
>> FWIW, it did work on my machine, and gives something having 64-bit
>> architecture:
>>
>> % file /sw64/share/xtal/ccp4-6.1.2/bin/xloggraph.exe
>> /sw64/share/xtal/ccp4-6.1.2/bin/xloggraph.exe: Mach-O 64-bit
>> executable
>>
>>
>> So it would be worth finding out what is going on.
>
> so I looked a bit more into the compilation log - I hope I extracted
> the important snippets - it starts with the compilations of XCCPJIFFY...
>
>
> # Make wrappers for xplot84driver and xloggraph, so rename the
> executables
>
> make msg_box.o ZDr2d.o EditString.o hardcopy_ctrl.o xloggraph.o
> log_file.o tom_fortran_interface.o bits_and_pieces.o graphics.o
> make: `msg_box.o' is up to date.
> make: `ZDr2d.o' is up to date.
> make: `EditString.o' is up to date.
> make: `hardcopy_ctrl.o' is up to date.
>
> gfortran -fno-second-underscore -c -o bits_and_pieces.o
> bits_and_pieces.f
> (without absolute path?)
>
> gfortran -fno-second-underscore -c -o graphics.o graphics.f
>
> make xdlmapman xdldataman
> /sw64/lib/gcc4.4/bin/gfortran -O2 -fno-second-underscore -c -o /sw64/
> src/fink.build/ccp4-6.1.2-2/ccp4-6.1.2/x-windows/xdl_gksubs.o /sw64/
> src/fink.build/ccp4-6.1.2-2/ccp4-6.1.2/x-windows/xdlgjk/xdl_gksubs.f
> (here with the absolute path?)
>
>
> In the end it fails - a file *.o in XCCPJIFFY returns:
>
> EditString.o: Mach-O 64-bit object x86_64
> ZDr2d.o: Mach-O 64-bit object x86_64
> back_font.o: Mach-O object i386
> bits_and_pieces.o: Mach-O object i386
> graphics.o: Mach-O object i386
> hardcopy_ctrl.o: Mach-O 64-bit object x86_64
> log_file.o: Mach-O 64-bit object x86_64
> msg_box.o: Mach-O 64-bit object x86_64
> plot84_file.o: Mach-O 64-bit object x86_64
> tom_fortran_interface.o: Mach-O 64-bit object x86_64
> xloggraph.o: Mach-O 64-bit object x86_64
> xplot84driver.o: Mach-O 64-bit object x86_64
>
> So, my wild guess is that the Makefile in XCCPJIFFY should contain the
> complete path in the FC variable
>
> That's really weird - I find only gfortran x86_64 on my system - there
> is a 32bit fink installed in /sw but without gcc44 installed
>
> Steffen
Hi Steffen:
There is definitely a problem building this in 32bit fink on a 64-bit
competent machine, which I need to fix as soon as I wake up. It might be
there is an inverse problem as well, which I can try to fix at the same
time (force upon it -32 or -m64 directives as appropriate).
Thanks for spotting this.
Bill
William G. Scott
Contact info:
http://chemistry.ucsc.edu/~wgscott/
------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
trial. Simplify your report design, integration and deployment - and focus on
what you do best, core application coding. Discover what's new with
Crystal Reports now. http://p.sf.net/sfu/bobj-july
_______________________________________________
Fink-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/fink-users
