Re: MacOSX: ismac() implementation

[Thomas Treichl]

Jordi Gutiérrez Hermoso schrieb:
> Btw, exactly what code is platform-specific? I remember that there
> were some efforts recently in this list to get code to compile under
> Windows with a non-gcc compiler, and that there were issues compiling
> with MacOS X precisely because that OS doesn't obey the FHS. Is that
> about it, getting Octave to compile?
>
> - Jordi G. H.

You're right Jordi, that's why I thought that Mac users should be part of the 
big isunix() community and just add ismac() because of compatibility to Matlab. 
On compilation side things are normally be solved platform-independent but 
library-depend (eg. does a system have the 'realpath' C-function or the 
'resolvepath' function and then create a #define that depends which function is 
available). There also has been the one or other topic in the past where an 
ismac() function would have made things easier to solve (eg. the drawnow.m 
function with the AquaTerm backend driver for Gnuplot on Mac systems only) and I 
also think that Soren's example would work if we just keep in mind to check for 
an ismac() (if somewhen really necessary) before checking for isunix() like that 
example:

if (ismac()) # always check before isunix()
  # Mac code here
elseif (isunix())
  ...

That's my thoughts about that...

  Thomas