Re: 0.99.9 problems on Solaris.

[Owen Taylor <owt1 cornell edu>]

Bj|rn Augustsson <d3august@dtek.chalmers.se> writes:

> Hi!
> 
> A few problems with the latest release:
> (All this is on Sun Sparc Solaris 2.5.1, with creator 3D graphics.)
> 
> * In the text widget test in testgtk, the last row seems to get rendered
>   with incorrect colors. 
>   The last row is "And even coloredtext", but only "colored" is visible, in
>   what seems to be a slightly darker shade of grey than the rest of the text.
>   If I select the text, I can see it all. (In the "expected" light grey on 
>   blue.)

You mean you can't read black text on a black background? If you toggle
the "editable" button, you'll see it very faintly...
 
> * The dnd test in testgtk: If I grab the button labelled "Drag me!", I get
>   hold of an image of what looks like a brick. If I drag the brick over
>   the "Drag me!" button, it sort of erases it, like what you'd expect an 
>   eraser in a drawing program to do. If I drop the brick and enter the
>   button with the mouse pointer, the button reappears.

Yes, this is a known problem. I've been unable to figure it out -
no expose events are being generated when the shape is dragged
around; yet all windows _except_ the drag source window don't
get wiped - possibly because the server is using backing store?

The only difference I can see for the grab source is that it
has the pointer grab, but that shouldn't affect expose events.

> * If I do the "test mainloop: test, I get a grey window, and the text
>   "create_test: start" is printed. If I close the window and then try to
>   exit testgtk (by pressing the "close" button), the text 
>   "create_test: done" is printed and the main window closes, but the process
>   is still there. CTRL-C gets rid of it.

That test was a bit strange - I think it is a remnant from when
GTK wasn't much more than a main loop. I've rewritten now so it 
is a bit more obvious to the user what is going on.
 
> * Oh, and once X krashed (!) when I exited the shapes test. I couldn't
>   reproduce that though so hopefully it was a solar spot or something.

The X server crashed? (?!?!)

That wouldn't really be our bug ;-), but if you can figure out how
to reproduce it, perhaps it could be worked around.

Regards,
                                        Owen