[issue] Crash at startup

[Stephen J. Turnbull]

jean-louis.a.leroy at fortis.com writes:

 > Oh, so you didn't get yesterday's message. I must have goofed after
 > having my mails rejected a couple of times because the attachment
 > was too big.

Please don't send core files and executables to the mailing list or to
developers unless requested.  They're almost never useful, except on
the machine they are intended for.

 > #0  0x00000000 in ?? ()
 > #1  0xfef620fc in XRegisterIMInstantiateCallback () from /usr/lib/libX11.so.4
 > #2  0x00163e40 in x_init_device (d=0x928b00, props=-33312648) at device-x.c:767

As usual with crashes in Xlib, this makes no sense whatsoever.  We
call XRegisterIMInstantiateCallback(dpy,NULL,NULL,NULL,cb,data) where
dpy is the X display, cb is the address of the callback being
registered, and data is private data which will be passed to the
callback.  The three NULLs have to do with resources that are private
to particular input methods, and are explicitly allowed to be NULL in
the man page for the two Xlib implementation I have access to.  If cb
were being called, we should see it in the backtrace.

If I can get the man page for XRegisterIMInstantiateCallback on the
sparc_sun_solaris2_10 platform, and it says something different, maybe
we can do something about it.  However I suspect that this is an Xlib
bug on the Sun.