On 2015-05-21 08:07-0000 Arjen Markus wrote:

> Hi Alan,
>
>
>
>> -----Original Message-----
>> From: Alan W. Irwin [mailto:ir...@beluga.phys.uvic.ca]
>
>>
>> The first error message for the case of the traditional build of the f95 
>> examples when
>> PLplot was statically built was as follows:
>>
>> /cygdrive/d/plplot-
>> svn/comprehensive_test_disposeable/static/install_tree/lib/libplplot.a(qt.cpp.o):qt.cpp
>> :(.text+0x5cc0):
>> undefined reference to std::ios_base::Init::Init()'
>>
>> std refers to symbols in the special library that the C++ compiler normally 
>> adds to the
>> linking information (in this case the libstdc++ library that is associated 
>> with g++, but
>> the name/location/version of that library might depend on which C++ compiler 
>> the
>> user decides to use).  Of course, in this case we must link the Fortran 
>> examples with
>> a Fortran compiler for the traditional build system for the installed 
>> example.
>> Therefore, there is a linking error here because libstdc++ is obviously not
>> automatically linked in by gfortran.  There are also similar potential 
>> problems for
>> other non-C or C++ examples that are compiled.
>>

> Ah, that is indeed a familiar problem. If I understand it correctly,
then C++ code requires a linker with libraries suitable for C++. Of
course Fortran and other languages require their own libraries and it
is usually the compiler (or the driver program) that adds the
appropriate ones to the link step. One can get around this problem by
explicitly supplying these libraries, but that is very much
compiler-specific and even configuration-specific (for instance: under
Windows the libraries for a debuggable program tend to be different
than those for a release program).

Exactly.  CMake handles finding the missing special libraries for each
different compiler with a lot of machinery under the hood that has
been developed over the years, but there is no way that our
traditional pkg-config + Makefile build system for the installed
examples can or should be that sophisticated.

> I will give this new set-up a try. Since it is the computer which does most 
> work, it should not be that time-consuming for me.

I saw your later post that reported complete success.  That was good
to see (!), and I will take a detailed look at that report tarball and
also post that good test result on our Wiki after I have had some
sleep.

Alan

__________________________
Alan W. Irwin

Astronomical research affiliation with Department of Physics and Astronomy,
University of Victoria (astrowww.phys.uvic.ca).

Programming affiliations with the FreeEOS equation-of-state
implementation for stellar interiors (freeeos.sf.net); the Time
Ephemerides project (timeephem.sf.net); PLplot scientific plotting
software package (plplot.sf.net); the libLASi project
(unifont.org/lasi); the Loads of Linux Links project (loll.sf.net);
and the Linux Brochure Project (lbproject.sf.net).
__________________________

Linux-powered Science
__________________________

------------------------------------------------------------------------------
One dashboard for servers and applications across Physical-Virtual-Cloud 
Widest out-of-the-box monitoring support with 50+ applications
Performance metrics, stats and reports that give you Actionable Insights
Deep dive visibility with transaction tracing using APM Insight.
http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
_______________________________________________
Plplot-devel mailing list
Plplot-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/plplot-devel

Reply via email to