Re: Problems building on OSX

[John Swensen]

John W. Eaton wrote:
> On 22-May-2006, John Swensen wrote:
>
> | Here is a patch for a change I had to make to get octave to compile on
> | OSX 10.4 with the Apple GCC/G++ and the fink G95.
>
> What version of GCC is that?
>   
powerpc-apple-darwin8-gcc-4.0.1 (GCC) 4.0.1 (Apple Computer, Inc. build
5250)
> | Before the patch, it
> | was choking and saying it didn't have an operator== that could compare a
> | streampos against a streampos.
>
> Can you please send the precise text of the error message?
>   
Well, it must have got fixed or I had something messed up in the first
place.  I moved my patched version out of the way, did a 'cvs update'
and recompiled with no problems.
> | Also, for some reason, the
> | Makeconf that was generated had an UGLY_DEFS that resulted in a DEFS
> | that contained an unmatched number of \ (slashes).  When trying to
> | compile any file that included oct-conf.h, it complained that
> | OCTAVE_CONF_DEFS has and illegal escape sequence of \040 (a space). 
> | This was because there was an unmatched \ before a space.
>
> What sed do you have?
>   
I was using the Apple OSX sed, but then switched to using the latest
fink sed.  Both sed's cause the same problem.
GNU sed version 4.1.4
> | Index: src/load-save.cc
> | ===================================================================
> | RCS file: /cvs/octave/src/load-save.cc,v
> | retrieving revision 1.213
> | diff -u -n -r1.213 load-save.cc
> | d1684 3
> | a1686 3
> |                   std::streamoff tmp1 = static_cast<std::streamoff> (0);
> |                   std::streamoff tmp2 =
> | static_cast<std::streamoff>((file.rdbuf ())->pubseekoff (0, std::ios::cur));
> |                   bool write_header_info = ( tmp2 == tmp1 );
> | d1706 3
> | a1708 3
> |                   std::streamoff tmp1 = static_cast<std::streamoff> (0);
> |                   std::streamoff tmp2 =
> | static_cast<std::streamoff>((file.rdbuf ())->pubseekoff (0, std::ios::cur));
> |                   bool write_header_info = ( tmp2 == tmp1 );
> | 
>
> Please send this as a diff -u (don't use -n).  Then there would be
> some context in the diff, not just the lines that are deleted or
> added.
>   
Since I'm not having that problem anymore, no sense in sending a patch. 
I will remember not to use the '-n' for future problems.
> Thanks,
>
> jwe
>
>
>