> - there is no indication which window is active, I suggest you add different
> window title background color for inactive window

I agree it would be more intuitive. Week ago, I actually considered 
implementing it, but in the end there was no time left to do it. Because of my
other commitments, I will not be able to work on graphics stack for at least
next two months. I might add this feature after this period, but the stack
will be probably merged into mainline by that time and I am not sure what
changes or improvements will be done during that process. But I am definitely
adding your suggestion to my TODO list.

> - when trying to drag an inactive window (title) using mouse, it does not
> work

Basically the same as above goes also for this suggestion. The complexity and
robustness of currently fairly simple mouse clicking handler (as well as 
button widget clicking handler) could be significantly improved to handle 
all the possible corner cases to behave more similarly as in GUIs of 
established operating systems.

> - when in kernel console (i.e. after pressing Alt+M), moving the mouse
> reveals

As strangely as it might seem, this is actually very useful "feature" when
debugging the graphics stack. Because both compositor and kernel console are
doing damage to the framebuffer, you can reproduce bugs while being able to
see debug printouts in the kernel console. As far as my memory goes, this
"cooperative" sharing of framebuffer is actually happening also in the 
original console server, which is currently in mainline (i.e. that even when
in kernel console, console server is still being able to do damage to 
framebuffer if instructed to do so). It was only much more subtle (i.e. at
some occasion, the bdsh command prompt popped out in the kernel console),
because the mouse pointer was handled by first tier of the stack (fb server). 


_______________________________________________
HelenOS-devel mailing list
[email protected]
http://lists.modry.cz/cgi-bin/listinfo/helenos-devel

Reply via email to