From maintainers-request at octave dot org Sat Feb 11 18:05:44 2006
Subject: Re: octave GUI proof of concept
From: Sebastien Loisel <sloisel at gmail dot com>
To: David Bateman <David dot Bateman at motorola dot com>
Cc: octave maintainers mailing list <maintainers at octave dot org>
Date: Sun, 12 Feb 2006 01:04:07 +0100

------=_Part_11384_28716636.1139702647134
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline

Ok, new version! Take another look at the screenshot
http://www.math.mcgill.ca/loisel/octave-gui/, there's now also a prototype
plotting widget. It can plot any line soup with Gouraud shading. I have a
helper function myplot.m that turns a [x1 y1;x2 y2; ... ; xn yn] matrix int=
o
a line soup and then calls the c++ internal_my_plot function. Also, I have
the ability of doing multiple plots; the plots rescale automatically, etc..=
.
Note that gnuplot is still available, if it works better for you.

Also, David said...

All of the patches necessary to build a mingw release of octave are in
> the CVS, its the dependencies that cause all of the problems. However,


That's great news.

as long as you have glob, which I can send you off-line, you can easily
> build a minimum version of octave. That is without atlas, fftw, etc,
> etc.. That should be largely sufficient to test your ideas out under
> mingw...
>
> I should have a dirty build finished by tomorrow, so if you are really


It sounds like you're going to have this MinGW thing licked soonish and
since I'm being so productive right now, do you suppose I can wait for your
binary tarball/msi/whatever?

I'm not sure which bits of octave I'm depending on right now, but the only
one that's really essential and that I don't know how it'll work in Windows
is mkoctfile.

In fact, qmake takes over make *and* the compiler, and octave wants to take
over the compiler. This is terrible! I don't understand why everybody wants
to be your compiler all the time! Anyway, by trial and error I found what
questions to pose to mkoctfile to get roughly correct compile options, and
by sheer force of will, I convinced qmake to build oct files using those
options. So if the MinGW octave can do the same and compile with g++, we're
in business.

Cheers,

Sebastien Loisel

------=_Part_11384_28716636.1139702647134
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline

<div><div><br>
Ok, new version! Take another look at the screenshot
<a href=3D"http://www.math.mcgill.ca/loisel/octave-gui/">http://www.math.mc=
gill.ca/loisel/octave-gui/</a>, there's now also a
prototype plotting widget. It can plot any line soup with Gouraud
shading. I have a helper function myplot.m that turns a [x1 y1;x2 y2;
... ; xn yn] matrix into a line soup and then calls the c++
internal_my_plot function. Also, I have the ability of doing multiple
plots; the plots rescale automatically, etc... Note that gnuplot is
still available, if it works better for you.<br>
<br>
Also, David said...<br>
</div><br><blockquote class=3D"gmail_quote" style=3D"border-left: 1px solid=
 rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">All of =
the patches necessary to build a mingw release of octave are in<br>the CVS,=
 its the dependencies that cause all of the problems. However,
</blockquote><div><br>
That's great news.<br>
</div><br><blockquote class=3D"gmail_quote" style=3D"border-left: 1px solid=
 rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">as long=
 as you have glob, which I can send you off-line, you can easily<br>build a=
 minimum version of octave. That is without atlas, fftw, etc,
<br>etc.. That should be largely sufficient to test your ideas out under<br=
>mingw...<br><br>I should have a dirty build finished by tomorrow, so if yo=
u are really</blockquote><div><br>
</div></div>It sounds like you're going to have this MinGW thing licked
soonish and since I'm being so productive right now, do you suppose I
can wait for your binary tarball/msi/whatever?<br>
<br>
I'm not sure which bits of octave I'm depending on right now, but the
only one that's really essential and that I don't know how it'll work
in Windows is mkoctfile.<br>
<br>
In fact, qmake takes over make *and* the compiler, and octave wants to
take over the compiler. This is terrible! I don't understand why
everybody wants to be your compiler all the time! Anyway, by trial and
error I found what questions to pose to mkoctfile to get roughly
correct compile options, and by sheer force of will, I convinced qmake
to build oct files using those options. So if the MinGW octave can do
the same and compile with g++, we're in business.<br>
<br>
Cheers,<br>
<br>
Sebastien Loisel<br>
<br>

------=_Part_11384_28716636.1139702647134--