On Mon, Aug 18, 2008 at 07:59:28PM +0200, Harald Dunkel wrote:
> I've got a problem with fvwm on Linux (current snapshot as well
> as fvwm-2.5.24):
>
> run "xine" (without arguments)
> right click on the movie window
> select Settings --> Setup
> select the "video" tab
> select the "video driver to use" drop down menu
> scroll to the end and select the last entry
>
> After that fvwm used up a complete CPU (100%). A backtrace showed
> where it got stuck:
* Which xine-Version?
* When I open the "video driver to use" list, I get a decorated
list window that appears below the headline of the box first but
jumps to the middle of the screen.
* How do you scroll to the end of the list? By pulling the
scrollbar?
* What is the name of the list entry you select?
* How do you select it? By clicking on it with the mouse?
For me, *xine* eats up 50% to 100% of one cpu all the time. Can
you please send me your fvwm and xine configuration files?
> #0 0x00007f16cc0f4a25 in XCheckIfEvent () from /usr/lib/libX11.so.6
> #1 0x000000000049e8d4 in FCheckPeekIfEvent (display=0x1007810,
> event_return=0x7fffd480cec0, predicate=<value optimized out>,
> arg=<value optimized out>) at FEvent.c:349
> #2 0x000000000042bd87 in __handle_configure_request (cre=
> {type = 23, serial = 320163, send_event = 0, display = 0x1007810,
> parent = 2097886, window = 10489443, x = 479, y = 187, width = 260,
> height = 114, border_width = 0, above = 0, detail = 0, value_mask =
> 3}, ea=0x7fffd480d200, fw=0x10a2a50, force=0, force_gravity=0) at
> events.c:858
> #3 0x000000000042d073 in HandleConfigureRequest (ea=0x7fffd480d200) at
> events.c:1882
> #4 0x0000000000427464 in dispatch_event (e=0x7fffd480d230) at events.c:4042
> #5 0x0000000000427bc5 in HandleEvents () at events.c:4086
> #6 0x0000000000449d0d in main (argc=<value optimized out>, argv=<value
> optimized out>) at fvwm.c:2563
Can you attach a debugger to fvwm and single step through it to
see where it's looping? By starting fvwm on a second display or
by running gdb from the console).
> events.c is something fragile, AFAICS,
Why do you think so?
> so I tried it with
> #undef EXPERIMENTAL_ANTI_RACE_CONDITION_CODE
And did it loop? THe backtrace below shows that you were
navigating an fvwm menu at that time.
> Now the bt showed:
>
> #0 0x00007f829ca53b87 in XCheckMaskEvent () from /usr/lib/libX11.so.6
> #1 0x000000000049e821 in FCheckMaskEvent (display=0x2436250, event_mask=0,
> event_return=0x7fffa516ce40) at FEvent.c:326
> #2 0x0000000000412484 in __menu_loop (pmp=0x7fffa516d580,
> pmret=0x7fffa516d610, pdkp=0x7fffa516d4c0) at menus.c:4131
> #3 0x0000000000410a7c in do_menu (pmp=0x7fffa516d580,
> pmret=0x7fffa516d610) at menus.c:6102
> #4 0x0000000000480b41 in menu_func (cond_rc=0x7fffa516d930, exc=0x7f28300,
> action=0x7f28270 "Nop", fStaysUp=<value optimized out>) at menucmd.c:122
> #5 0x0000000000466585 in __execute_function (cond_rc=0x0, exc=0x79d40d0,
> action=<value optimized out>, exec_flags=0 '\0', args=<value optimized
> out>, has_ref_window_moved=0) at functions.c:645
> #6 0x000000000042b6bb in HandleButtonPress (ea=<value optimized out>) at
> events.c:1566
> #7 0x0000000000427464 in dispatch_event (e=0x7fffa516db90) at events.c:4042
> #8 0x0000000000427bc5 in HandleEvents () at events.c:4086
> #9 0x0000000000449b5d in main (argc=<value optimized out>, argv=<value
> optimized out>) at fvwm.c:2563
>
>
> This was generated using the snapshot of today (20080818). Distro
> is Debian (amd64) on a Core2 Duo.
Ciao
Dominik ^_^ ^_^
--
Dominik Vogt
