Re: removing or using AWT/Swing

[Tom Tromey <tromey at redhat dot com>]

>>>>> "Marco" == Marco Trudel <mtrudel@gmx.ch> writes:

Marco> I played a little with AWT/Swing with a 4.2.0 gcj (20060630, 
Marco> experimental). There raised two questions:

On Windows?  Or on Linux?

Marco> I compiled gcc with "--without-x --disable-java-awt", but I saw 
Marco> multiple lines that told about compiling AWT and Swing stuff.

--disable-java-awt merely disabled the AWT peer implementations from
being built -- AWT itself is still built.

It should be possible to disable most of AWT+Swing (with the exception
of AWTPermission) without an enormous effort.  However AFAIK nobody
has done this in a way suitable for submission...

Marco> I wanted to check the current status of AWT/Swing. I tried with a gcj 
Marco> that wasn't compiled with any special AWT flag. A simple AWT frame gave 
Marco> this exception (I'm working on linux, awt is the executable name):
Marco> Exception in thread "main" java.awt.AWTError: Cannot load AWT toolkit: 
Marco> gnu.java.awt.peer.gtk.GtkToolkit

You built without peers, so there aren't any.  If you're on Windows,
then you basically need to figure out how to build the Gtk peers
(unfortunately I don't think anybody has tried this on Windows yet).
Nobody has written native Windows peers yet.

Marco> So, am I doing anything wrong or is AWT/Swing broken in the used gcj 
Marco> version? (I think this gcj works with gnu classpath v0.9).

It definitely does work for me :-)

Marco> A general question to this: My impression of following the mailinglist 
Marco> and searching in the net is, that swingWT currently is more useful than 
Marco> AWT/Swing support from gnu classpath. Any other opinions on that?

That was probably written quite a while ago.  Our Swing has progressed
quite a lot and seems to be useful these days.  I'm not expert on the
exact state of it though.

Tom