Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!watmath!clyde!rutgers!mit-eddie!bacchus!rlk
From: rlk@bacchus.UUCP
Newsgroups: comp.unix.wizards
Subject: Re: Make is Buggy (was \"special\" shells a security hole?)
Message-ID: <226@bacchus.MIT.EDU>
Date: Mon, 16-Feb-87 15:09:15 EST
Article-I.D.: bacchus.226
Posted: Mon Feb 16 15:09:15 1987
Date-Received: Tue, 17-Feb-87 06:16:28 EST
Sender: daemon@bacchus.MIT.EDU
Reply-To: rlk@athena.MIT.EDU
Organization: MIT Project Athena
Lines: 25

In article <263@its63b.ed.ac.uk> simon@its63b.ed.ac.uk writes:
]
]In article <13252@sun.uucp> guy@sun.UUCP (Guy Harris) writes:
]>Furthermore, it is a botch to *import* SHELL from the environment;
]>this causes Makefiles to work differently for people with different
]>login shells!  "make" should NOT import SHELL.
]
]And it *certainly* should not attempt to import things like IFS!
]This can lead to extremely confusing bugs in complicated makefiles.
]The problem is that it strips out initial blank-space when assigning
]a make-variable (even from an environ-string), so IFS=" \t\n" gets
]changed to IFS="" (double-quotes inserted here by me for clarity), which
]can cause shell-scripts executed as subprocesses by make to go totally
]haywire!

It's bad enough that make imports PATH.  I was building something from
the source, and a couple of subdirectories failed.  Turned out it was
because the line in the makefile used a program in the working
directory, not prefixing it with ./ .  Since I don't have . in my path
for various reasons, the make blew out.

Of course, this is still better than some of the other things that
could have happened...but it was still messy.

Robert^Z