From owner-freebsd-bugs@FreeBSD.ORG  Sun Feb 27 21:29:39 2011
Return-Path: <owner-freebsd-bugs@FreeBSD.ORG>
Delivered-To: freebsd-bugs@freebsd.org
Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34])
	by hub.freebsd.org (Postfix) with ESMTP id 34249106566C;
	Sun, 27 Feb 2011 21:29:39 +0000 (UTC) (envelope-from stb@lassitu.de)
Received: from gilb.zs64.net (gilb.zs64.net [IPv6:2001:470:1f0b:105e::1ea])
	by mx1.freebsd.org (Postfix) with ESMTP id EB5128FC14;
	Sun, 27 Feb 2011 21:29:38 +0000 (UTC)
Received: by gilb.zs64.net (Postfix, from stb@lassitu.de) id 8A691E687;
	Sun, 27 Feb 2011 22:29:37 +0100 (CET)
Mime-Version: 1.0 (Apple Message framework v1082)
Content-Type: text/plain; charset=us-ascii
From: Stefan Bethke <stb@lassitu.de>
In-Reply-To: <AANLkTimQAimn5WrKXkmG8xa_OocCnHGPmP6E8bvvYmRa@mail.gmail.com>
Date: Sun, 27 Feb 2011 22:29:36 +0100
Content-Transfer-Encoding: quoted-printable
Message-Id: <0A08CDDB-5482-4805-91F9-0DDD37AE7A7A@lassitu.de>
References: <201102241311.p1ODBajF050380@freefall.freebsd.org>
	<83C33215-1358-4606-A7A8-FC40BCC08E69@lassitu.de>
	<AANLkTimQAimn5WrKXkmG8xa_OocCnHGPmP6E8bvvYmRa@mail.gmail.com>
To: Eitan Adler <eadler@freebsd.org>
X-Mailer: Apple Mail (2.1082)
Cc: freebsd-bugs@freebsd.org
Subject: Re: gnu/137665: [patch] dialog(1) goes into tight loop on
	encountering eof
X-BeenThere: freebsd-bugs@freebsd.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Bug reports <freebsd-bugs.freebsd.org>
List-Unsubscribe: <http://lists.freebsd.org/mailman/listinfo/freebsd-bugs>,
	<mailto:freebsd-bugs-request@freebsd.org?subject=unsubscribe>
List-Archive: <http://lists.freebsd.org/pipermail/freebsd-bugs>
List-Post: <mailto:freebsd-bugs@freebsd.org>
List-Help: <mailto:freebsd-bugs-request@freebsd.org?subject=help>
List-Subscribe: <http://lists.freebsd.org/mailman/listinfo/freebsd-bugs>,
	<mailto:freebsd-bugs-request@freebsd.org?subject=subscribe>
X-List-Received-Date: Sun, 27 Feb 2011 21:29:39 -0000

Am 27.02.2011 um 21:55 schrieb Eitan Adler:

>> dialog does not go into a tight loop eating all CPU, but it still =
gets stuck.  I believe it should take some action on encountering EOF, =
instead of waiting for input forever.  My patch did that for the old =
dialog, typically taking the same path as pressing ESC would.
>=20
> I can not reproduce this using the command stated in the PR:
>=20
> $dialog --yesno foo -1 -1 </dev/null
>=20
> $dialog --version
> Version: 1.1-20100428

You are correct, I did assume that dialog was hung when in fact it is =
waiting for input.  The man page does not mention any behavior relative =
to stdin being redirected, and I expected dialog to honor a redirection. =
 Evidently, it manages to take input from the tty instead of from stdin. =
 Is this the expected behavior?

# printf "\n" | dialog --yesno foo -1 -1=20

This will wait for input from the tty, instead of returning immediately.

The reason I'm unhappy with the behavior of the old dialog, and the =
current behavior, is that if a script unintentionally invokes dialog =
(for example when running portupgrade from a script, and the port or =
package using dialog), the script stalls.  If you run the script from a =
cron job, it will never terminate.  I'd prefer a deterministic, =
terminating reaction to stdin not being a tty.

While dialog is meant to provide a user interface, there are situations =
where you want to be able to script someone elses use of (some embedded) =
dialog, and I'd rather have a chance to at least continue instead of =
being stuck at that point, or even having to break out expect again.


Stefan

--=20
Stefan Bethke <stb@lassitu.de>   Fon +49 151 14070811