Path: utzoo!attcan!utgpu!jarvis.csri.toronto.edu!rutgers!tut.cis.ohio-state.edu!ucbvax!ulysses!andys
From: andys@ulysses.homer.nj.att.com (Andy Sherman)
Newsgroups: comp.bugs.4bsd
Subject: Here Documents not closed in at
Message-ID: <11703@ulysses.homer.nj.att.com>
Date: 26 Jun 89 17:24:27 GMT
Sender: netnews@ulysses.homer.nj.att.com
Reply-To: andys@ulysses.homer.nj.att.com (Andy Sherman)
Organization: AT&T Bell Laboratories - Murray Hill, NJ
Lines: 28


The 4.2 and 4.3 BSD versions of at will fail if you use a shell that
strictly follows the POSIX standard for handling here documents.  Note
that previous documentation of the shell indicated that a here
document could be closed by either the key line or by an end of file.
According to POSIX, an end of file while processing a here document is
a syntax error.

The shell script generated by the stock 4.3 version of at contains the
construct:

$SHELL << '...the rest of the file is shell input'
  ...
Your commands submitted to at
  ...
END OF FILE

With a POSIX conforming shell, all your at jobs will terminate with
syntax errors rather than executed jobs.  The fix to at.c is easy.
Output the line:
...the rest of the file is shell input

just before closing the at queue file.
-- 
Andy Sherman/AT&T Bell Laboratories/Murray Hill, NJ           *NEW ADDRESS*
AUDIBLE:  (201) 582-5928                                      *NEW PHONE*
READABLE: andys@ulysses.ATT.COM  or att!ulysses!andys         *NEW EMAIL*
The views and opinions are my own.  Who else would want them? *OLD DISCLAIMER*