Re: Status of FLTK on MacOS X

[Ben Abbott]

On Jan 2, 2010, at 10:16 PM, Ben Abbott wrote:

> I compiled the current sources on MacOS 10.6.2 using a gfortran from …
> 
>       http://r.research.att.com/tools/
> 
> The site above provides a gfortran that is intended to be compatible with the 
> gcc compiler and libraries included with Apple's Xcode (version 4.2.x).
> 
> I did this in the hope that I'd be able to link to the Carbon and OpenGL 
> Frameworks.
> 
> The good news is that the experiment was a success, in that the resulting 
> Octave produces no errors with run, is able to use the FLTK backend, and 
> produces no errors when exited.
> 
> The bad news is that I encountered many failures during "make check"; as 
> summarized below.
> 
>  src/DLD-FUNCTIONS/eigs.cc .............................. PASS    0/149  FAIL 
> 149
>  src/data.cc ............................................ PASS  544/545  FAIL 
> 1
>  scripts/sparse/svds.m .................................. PASS    0/3    FAIL 
> 3
> 
>  PASS   5502
>  FAIL    155
> 
> The data.cc error is …
> 
> 1473   ***** assert(log2(complex(0,Inf)), Inf + log2(i));
> 1474 !!!!! test failed
> 1475 assert (log2 (complex (0, Inf)),Inf + log2 (i)) expected
> 1476 Inf + 2.266i
> 1477 but got
> 1478 Inf - NaNi
> 1479 NaNs don't match
> 
> … which I don't consider a significant problem.
> 
> The remaining errors appear to be due to using gfortran from gcc 4.4 to build 
> suitesparse and arpack.
> 
> I assume these errors result from mixing Apple's gcc 4.2 with gcc 4.4. As I 
> have time, I will try to verify that.
> 
> Ben

I discovered some typos in my build process.  With those fixed, I have one 
failure.

        >>>>> processing 
/Users/bpabbott/Development/mercurial/local_clone/src/data.cc
          ***** assert(log2(complex(0,Inf)), Inf + log2(i));
        !!!!! test failed
        assert (log2 (complex (0, Inf)),Inf + log2 (i)) expected
        Inf + 2.266i
        but got
        Inf - NaNi
        NaNs don't match

About half the time I run fntests.m from the command line it passes. I 
occasionally get a failure when testing chol.cc (increasing the error tolerance 
would fix that).

My dependencies are being handled by Fink (I've created a local package for 
qrupdate). The dependencies managed by Fink packages are all compiled with gcc 
4.4 (except for FLTK which is built with gcc 4.2). I complied and linked Octave 
using gcc 4.2. For MacOS 10.6 (10.5?), gcc 4.2 bundled with Xcode is required 
to link to Apple's OpenGL and Carbon Frameworks. If Octave is built with 
another version of gcc, the FLTK backend will not be functional.

I've attached three shell scripts which I use to build octave; 
pre_config-dev.sh, config-dev.sh, and post_config-sh. I expect those using 
MacPorts, or those who install the dependencies themselves, can modify the 
scripts to work.

Please note that these scripts are not intended to be used to actually install 
Octave. A Fink's more recent octave package is for version 3.0.5. Thus far I've 
been unsuccessful modifying it to work with current sources (to be  honest I 
haven't put much effort into that).

Ben

pre_config-dev.sh
Description: Binary data

config-dev.sh
Description: Binary data

post_config-dev.sh
Description: Binary data