[net.emacs] GNU Emacs Hanging on Suns

pwl@fluke.UUCP (10/20/86)

There have been a number of articles about GNU emacs hanging when run
from a window on a Sun workstation.  It turns out that this problem is
caused by a problem with Sun's socket code.  In Sun release 3.0,
out-of-band data through a socket can get lost.  It turns out that GNU
emacs does an "ioctl" to turn off XOFF/XON stalling, so that control-S
and control-Q are available as command characters.  Rlogin uses the
out-of-band mechanism to send this configuration information from the
machine emacs is running on to the machine where the window resides.
The flakey out-of-band connection causes the rlogin session to loop
endlessly waiting for out-of-band data.

Sun has a patch available for 3.0 which consists of a replacement object
module for the kernel and a total replacement for the "rlogin" program.

Sun claims that this bug is fixed in release 3.2, which was due to ship
October 15th.

Hope this helps all you out there who have finally stumbled onto this
problem.
-- 
	Paul Lutt          (206) 356-5059
	John Fluke Mfg. Co.
	P.O. Box C9090  Everett WA  98206  
	{allegra,decvax!microsof,ssc-vax,ucbvax!lbl-csam,uw-beaver}!fluke!pwl