On Tue, Oct 31, 2006 at 12:21:35AM +0100, Jan Gorski wrote: > On Sun, 25 Sep 2005 at 12:32:07 -0400, Thomas Dickey wrote: > > On Sun, Sep 25, 2005 at 04:50:14PM +0200, Jan Gorski wrote: > > > On Sat, 24 Sep 2005 at 19:25:52 -0400, Thomas Dickey wrote: > > > > (That would be "configure --enable-trace --enable-wide-chars --enable-luit", > > > > and the executable produces Trace-parent.out and Trace-child.out) > > > > > > Well, I've run it. Traces are in http://slimak.info/~slimak/traces.tar.bz2 if > > > you want to look at them. All hangs end with this line > > > > thanks (got a copy for reading) sorry - I should have followed up, hadn't seen much to report > For some time I was using xfce4-terminal and xfce4wm to avoid this bug (and > another one, 329916). At least month ago I gave a try to fvwm again, and it was > working --- till last upgrade :). Now I have checkd it in few configurations, > and it seems, that both bugs depend on xinerama(1)... Bug 329916 is a bit > upredictible, and sometimes it takes a moment till it appears, but this bug is > easier to reproduce, so maybe somebody can tell me what else can be done to find > out what happens? If you could reproduce it with a copy of xterm built with debug code, then it's possible that you could attach it with gdb from another window and see where it's hanging. You wouldn't need the setgid mode on the executable, since it's unlikely that the utmp code is involved. Doing that would give more insight into the problem - to see whether it's an X library problem, or whether xterm's gotten into some state that I could fix. -- Thomas E. Dickey http://invisible-island.net ftp://invisible-island.net
Attachment:
pgpm256HyHPkH.pgp
Description: PGP signature