From ward@wright.lcs.mit.edu  Sun Jan  5 06:21:05 1997
X-VM-v5-Data: ([nil nil nil nil t nil nil nil nil]
	[nil "Sun" " 5" "January" "1997" "09:21:03" "-0500" "Steve Ward" "ward@wright.lcs.mit.edu" "<199701051421.JAA25026@wright.lcs.mit.edu>" "31" "ntemacs call-process hang???" "^From:" nil nil "1" nil nil nil nil]
	nil)
Received: from wright.lcs.mit.edu (wright.lcs.mit.edu [18.111.0.157]) by june.cs.washington.edu (8.8.3+CSE/7.2ju) with ESMTP id GAA19838 for <voelker@cs.washington.edu>; Sun, 5 Jan 1997 06:21:04 -0800
Received: (from ward@localhost) by wright.lcs.mit.edu (8.7.5/8.7.3) id JAA25026; Sun, 5 Jan 1997 09:21:03 -0500 (EST)
Message-Id: <199701051421.JAA25026@wright.lcs.mit.edu>
Reply-to: ward@mit.edu
From: Steve Ward <ward@wright.lcs.mit.edu>
To: voelker@cs.washington.edu
cc: patl@wright.lcs.mit.edu
Subject: ntemacs call-process hang???
Date: Sun, 5 Jan 1997 09:21:03 -0500 (EST)


Geoff -

I've installed and am happily using your emacs 19.34 port on several systems
running W95  and NT4.  The sole problem that I am currently exasperated by
involves an unexplained, intermittant, long delay in executing a command
in a subprocess.

Normally, typing something like "N-! echo foo" responds immediately.  However,
for mysterious reasons, sometimes (often!) the system gets into a state where
a long (15-20 second) period of apparent autistic behavior -- no disk or net
activity, no response to keyboard input -- precedes the correct behavior.
This mode persists despite system reboots, re-installation of emacs, etc.
Then, mysteriously, normal (delay-free) behavior is unexpectedly restored,
often during an editing session and with no apparent cause.

With some primitive elisp-level debugging, I've determined that the delay
is entirely with a call to call-process-region, with reasonable argument
values.  I haven't taken the trouble to look at the c code for this builtin
to understand where the time could possibly be going.

I believe I've observed this problem on NT4 as well, but can't duplicate
it at the moment.

Any ideas?  Is this a known anomoly?

Thanks for your help... and, btw, for an outstanding contribution to the
community.  The care and support you provide for ntemacs should shame
most commercial software vendors.

-Steve