Re: More on MinGW build

[Benjamin Lindner]

> Hi,
>
> I finally decided to give a try on building Octave on Windows and ...

Welcome on board, and you already found the most important fact:

> ai the thing is not easy.

No it isn't. It has much improved in the past, with gcc 4.5.0 on the
one side and contributors to octave on the other side.

> To start with I find the installation of Msys/MinGW a real mess.

True, and because of this I wrote my own mini-installers for setting up
building environments for building octave with mingw.
If you're interested, you can take a peek at
https://octave.svn.sourceforge.net/svnroot/octave/trunk/octave-forge/admin/Windows/mingw-w64/dependencies/

mind you, it's - as everything here - work in progress, so it's experts only.

> Note also that the README.MinGW file mention a Howto.txt file
>
> http://octave.svn.sourceforge.net/viewvc/octave/trunk/octave-forge/admin/Windows/mingw32/HOWTO.txt

This should read
http://octave.svn.sourceforge.net/svnroot/octave/trunk/octave-forge/admin/Windows/mingw32/HOWTO.txt
I guess.

> I would really love to be able to build Octave under windows

Then please hop on and improve the situation by contributing.
There has been (in fact it's still there) the build environment for
octave 3.2 using mingw gcc 4.4.0 at
https://octave.svn.sourceforge.net/svnroot/octave/trunk/octave-forge/admin/Windows/mingw32/
It's actually a bit messy on its own, and not updated to gcc 4.5.0
(which I highly recommend to use)
But if you are interested on how it was done, take a look.
I am reorganising it, as it proved to be not reliably useful, and
hopefully more robust one is under construction - it works for me at
the moment, and I have to put it into svn

> But to think
> that I will have to build all dependencies with the messy mingw is not
> really motivating for me.

Well, you'll have to do it. The beauty of free software is that you
*can* do it. It requires effort, no doubt, but it can be done.
And mingw is not that messy as it probably looks to you.
The distribution scheme the mingw project chose is IMO perfectly
suitable. There is a non-trivial activation energy required to get it
up & running, but it pays back.

benjamin