Re: looking ahead to 3.6

[John Swensen]

On Feb 14, 2011, at 4:40 AM, Michael Goffioul wrote:

> Additional comments:
> 1) The terminal emulation problem under Windows exists in *all*
> toolkits, as there's no pty concept Windows. Whatever the choice,
> you'll have to fake terminal through pipes under Windows. To achieve
> that, I hacked VTE a lot. But this can be re-done for any other
> toolkit.
> 2) Despite the terminal emulation problem, I'd still go that way in a
> GUI, as this would provide the most consistent interface bewteen a
> console version and a GUI (like having the same key bindings...). If
> you want real hard-core developers ever to use a GUI, that's the right
> choice to make. If you need better integration between readline and a
> GUI, I think it would be possible to modify readline to add suitable
> hooks.

Michael,
Since you already took a look at the Window PTY/pipes issue once for GTK+, 
would you be willing to take a look at the QT terminal widget at
http://qt-apps.org/content/show.php/QTermWidget?content=82832
?

This was the work of a person who took Konsole and ripped out the terminal 
emulator portions to make it "KDE free".  I have been able to compile it on 
both Linux and OSX (???) and was wondering if you can assess how easy it would 
be to modify for use on Windows.  As a side note, some of the comments said it 
was also compilable on Solaris, so I assume any of the *NIX-y platforms would 
work without much modification. I tried to compile on Windows and it had 
problems with memory mapping functions, but I am sure that was probably just 
the tip of the iceberg (and I don't really know Windows programming).

If we chose to go this route, then this would probably be included with the 
Octave IDE sources and possible periodically updated to merge in bugfixes from 
Konsole.  With a working version of this terminal for all platforms, then we 
could push forward with a QT IDE and QT OpenGL backend.

John Swensen