Re: strange behavior of customize widget buttons

[Jason Rumney]

Eli Zaretskii <address@hidden> writes:

> On Mon, 27 Nov 2000, Jason Rumney wrote:
> 
> > > Do you still have that backtrace, perhaps?
> > 
> > [1] _kill(0x0, 0xb, 0x0, 0x0, 0xef323138, 0xa3434), at 0xef308244
> > [2] fatal_error_signal(0xb, 0x0, 0xefffa190, 0x0, 0x0, 0x0), at 0xa3434
> > ---- called from signal handler with signal 11 (SIGSEGV) ------
> > [3] show_busy_cursor(), at 0x9981c
> > [4] alarm_signal_handler(0xe, 0x0, 0xefffa648, 0x0, 0x0, 0x0), at 0x1407fc
> > ---- called from signal handler with signal 14 (SIGALRM) ------
> > [5] _memcmp(), at 0xef2719a4
> > [6] Fstring_equal(0xe86af4, 0x302957b4, 0xe86af4, 0x30e86af4, 0x126f030, 
> > 0x0), at 0x10b790
> 
> SIGALRM during a library function call?  Some non-reentrancy in the 
> library, perhaps?

The SIGALRM is a timer for displaying the busy cursor. The SEGV is
what causes the core dump. This is clearly in show_busy_cursor, but
dbx is not co-operating by telling me where in that function the error
occurred (does anyone know how to force ddd/dbx to believe that it is
in the correct source file? I'd use gdb, but it is not installed on
the machine, and I don't have root access).

I wonder whether the fact that we are in the middle of popping up a
menu here affects this?

-- 
Jason Rumney <address@hidden>