Ralph wrote:
> "shorter, pointed questions" (Thanks jim):
> 1) Why does dragging an icon from a panel then accross 
> Mozilla's title bar cause the dragged icon to freeze on the 
> title bar and the system to become virtually unresponsive?

Beats me :-) -- but if you can reliably duplicate this behavior, it may
be a valuable bug report!

> 2) Why does the mouse pointer still respond to the movements 
> of the mouse?

This I could understand. Someone else may correct my understanding,
but.. It is probably that the mouse input handling and cursor display is
independent of the X main loop .. the X event handler.

> 3) Why does the clock continue to tick along when everything 
> else seems frozen?

(see last comment) X may be frozen (and processes blocked on
non-forthcoming X-events) but the CPU and clock keep on ticking. Indeed,
as you have shown, everything outside of that X-instance still works.
Corollary: everything running within that X-instance is toast.

> 4) Why does the mouse pointer icon no longer change appearance 
> in accordance to the context of what it hovers over?

That depends on X's (and application's) response to mouse input (namely,
coordinates). And X is just not "paying [proper] attention"

> 5) Why do mouse button clicks no longer have any effect?

(see last remark)

> 6) Why does Alt-Tab, Alt-Esc, Ctrl-Alt-[->] (right arrow) no 
> longer work?

(ditto) But you have already seen that Ctrl-Alt-F1 (etc) do work, since
they are handled outside of X (at a lower level).

> 7) It seems like the system thinks that the dragging is still 
> taking place even tho the dragged icon is stuck on the title 
> bar and the mouse pointer moves independently from it, tho 
> keeping the appearance of the context of dragging.  How do I 
> tell the system to abort the drag?

Don't know, That's getting into the details of understanding the real
bug. You probably can't recover from whatever has happened -- although
the ultimate code fix _may actually be_ to recover from something that
wasn't anticipated by the original program coding. Or it might be some
other way of making the code failproof (.. "nothing can go wrong",
"nothing can.." ..).

Regards,
..jim


-- 
[email protected]
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-list

Reply via email to