Path: utzoo!attcan!uunet!husc6!rutgers!ucsd!ucbvax!decwrl!pyramid!prls!philabs!ttidca!mb
From: mb@ttidca.TTI.COM (Michael Bloom)
Newsgroups: comp.mail.elm
Subject: Re: Ignoring Hangups
Message-ID: <2636@ttidca.TTI.COM>
Date: 5 Jun 88 17:54:01 GMT
References: <5615@bloom-beacon.MIT.EDU>
Reply-To: mb@ttidca.tti.com (The Mb of all Evil)
Organization: Citicorp/TTI, Santa Monica
Lines: 14

In article <5615@bloom-beacon.MIT.EDU> xev@hstbme.mit.edu (Xev Gittler) writes:

> ... Could someone tell me the reason that elm ignores HUP's at some
>points, and mre importantly, is it bad to change it not to ignore
>them?

I think it may be "bad" NOT to change it, considering that on many
systems terminal opens can be invalidated (after which programs that
do ignore HUPS may subsequently loop on apparent EOF's).

While there may be critical sections to protect, it would be better to
hold the HUP rather than to ignore it.  You can do this for systems
without sighold or sigblock by using a dummy handler that just notes
that the HUP occurred and deal with it later.