Re: [Gcl-devel] Building GCL and Maxima plotting - Windows

[Mike Thomas]

Hi Camm.

> > Actually it doesn't use Tk support in GCL at all and Tk is not supported
yet
> > by GCL on Windows.  XMAXIMA is a Wish wrapper script which communicates
with
>
> Just wondering what the roadblock to gcl-tk on Windows was.

Just the part where the subprocess is spawned - it is fairly easily solvable
and I once started work on it but it has fallen onto the backburner.

>  I guess
> its not critical at present.

True.

> > ./configure --enable-custreloc --prefix c:/wherever/you/want/to/install
> >
>
> Do you need the --enable-custreloc, or will configure default to that
> on Windows?

It needs to be explicit as of CVS about a week ago:

$ ./configure --prefix c:/junk
loading site script /etc/config.site
loading cache ./config.cache
checking host system type... i686-pc-mingw32
host=i686-pc-mingw32
enable_machine=
Exactly one loader option must be chosen: dlopen=no statsysbfd=no
dynsysbfd=no locbfd=no custreloc=def_custreloc



> So it sounds as if both cygwin and mingw are active ongoing projects.

They sure are!  The MSYS fork of Cygwin is relatively new however and should
not be confused with the Mingw32 tool chain, which it is designed to host.


> Is there any reason to prefer one over the other with respect to gcl
> in the long run?

MSYS overcomes a number of configuration difficulties which can occur when
Mingw32 is hosted under Cygwin.  It is simply easier to use.


>  Cygwin doesn't happen to emulate elf, does it?

No.   It provides a fairly comprehensive Posix development environment and
emulation library.  As with all such animals, it is less efficient at
run-time than native libraries because of the emulation baggage hence
Mingw32.  Cygwin applications don't fit in very well with native Windows
tools because they use Posix paths.  Because of this a Mingw GCL is a better
long term solution for Windows users.

Cheers

Mike Thomas.