Re: [Java-gnome-developer] Weird TreeModel / TreeView behaviour

[Nicholas R. <ni...@mn...>]

I just tried this and i see the same behavior.  However, this is using
the stable versions (2.4 and 2.8) that i installed with the gentoo
ebuilds.

There have been some fairly significant updates to the sorting code in
the CVS version (2.5).  It would be good to try this test with that
version to see if those changes fix at least problem #1.  This might be
a bit difficult at the moment, since libgnome-java is still undergoing
the 64-bit upgrade.  libgtk-java and libglade-java work however so if
you remove the gnome stuff you could test it right away.

also note that the TreeView#{set,get}SearchColumn methods have been
deprecated because they were broken.  use
TreeView#{set,get}SearchDataColumn instead.

nick


On Wed, 2004-12-15 at 23:35 +1100, Andrew Cowie wrote:
> Hey,
> 
> I've been experimenting with the TreeModel / TreeView APIs, and came
> across something strange.
> 
> I've attached some example code. The CLASSPATH is correct for where the
> [my] Gentoo ebuilds put the jars and .sos, but other than touching up
> that it should just work if you type `make java` or `make native`
> 
> The specific thing I was testing was the ability to generate more than
> one GtkTreeView based on a single underlying GtkTreeModel.
> 
> The strange behaviours I want to note are:
> 
> 1) When you sort in one window, the other window is affected.
> 
> Run the program. You get two windows. There are three data columns; each
> window (a TreeView) has two of them only. Click on a header, which
> activates it as a sort column. That set of rows sort. SO DO THE ROWS IN
> THE OTHER WINDOWS. Huh? Aren't the views supposed to be separate? Is
> there some mismatch / object identity violation somewhere between the
> Java layer and the underlying GObject layer?
> 
> 2) When you close one window, the event isn't propagated to the
> LifeCyleListener until the other window is also closed, when both events
> show up. Try it. X close or Alt-F4 close one window. The way the code
> is, that should cause an event, which calls Gtk.mainQuit, which should
> cause the app to exit. That doesn't happen right.
> 
> I'm not sure that either of these are RC bugs, but they are certainly
> anomolous.
> 
> I'd appreciate it if people could try this and let me know if they're
> having the same behaviour. If you are, then we have some things we need
> to track down.
> 
> AfC
> Sydney
>