(This is a snapshot of my old weblog. New posts and selected republished essays can be found at raganwald.com.)
Friday, February 22, 2008
The recursive implementation of /bin/true
Here is the source code to /bin/true in Solaris (via The Daily WTF):
/bin/true simply exits successfully, and Solaris implements this as a shell script. There’s a lot to laugh about on a Friday afternoon, but a thought struck me: What exactly is the copyright notice copyrighting? Not any implementation code: This “implementation” of /bin/true doesn’t do anything, so it has no executable code. Not the idea of implementing /bin/true as a shell script without implementation code: you’d need a patent to protect that idea.
I conclude that the copyright notice is protecting… The copyright notice! There is a delicious self-reference there in the style of Gödel, Escher, Bach
…
Here is the source code to /bin/true in Solaris (via The Daily WTF):
#!/usr/bin/sh
# Copyright (c) 1984, 1986, 1987, 1988, 1989 AT&T
# All Rights Reserved
# THIS IS UNPUBLISHED PROPRIETARY SOURCE CODE OF AT&T
# The copyright notice above does not evidence any
# actual or intended publication of such source code.
#ident "@(#)true.sh 1.6 93/01/11 SMI" /* SVr4.0 1.4 *//bin/true simply exits successfully, and Solaris implements this as a shell script. There’s a lot to laugh about on a Friday afternoon, but a thought struck me: What exactly is the copyright notice copyrighting? Not any implementation code: This “implementation” of /bin/true doesn’t do anything, so it has no executable code. Not the idea of implementing /bin/true as a shell script without implementation code: you’d need a patent to protect that idea.
I conclude that the copyright notice is protecting… The copyright notice! There is a delicious self-reference there in the style of Gödel, Escher, Bach
…
Comments on “The recursive implementation of /bin/true”:
<< Home
As stated at the Daily WTF, the other notable quality of this implementation is that it's up to version 1.6. They'd like to see the change log. I'd like to see the bug reports.
Oddly, if I remember correctly, the first line, the one starting with #! is a directive to the shell to start a specific program, in this case sh. The rest is actual comments, but I think what this is doing is calling sh without a terminal or command, so it starts briefly and then exits. If that's true, then there is actually a line of code in this file :-)
Given it probably forks and execs, its a fairly expensive way of assigning a true value. A shell variable would be a lot cheaper :-)
Just a guess...
Paul.
http://theprogrammersparadox.blogspot.com
Given it probably forks and execs, its a fairly expensive way of assigning a true value. A shell variable would be a lot cheaper :-)
Just a guess...
Paul.
http://theprogrammersparadox.blogspot.com
Paul, fortunately speed isn't really an issue here, as most often this command isn't actually called. As GNU's true (which responds to --help and --version, by the way) says:
NOTE: your shell may have its own version of true, which usually supersedes the version described here. Please refer to your shell's documentation for details about the options it supports.
And indeed it does, at least for bash. Calling "true --help" here does nothing, while "`which true` --help" says a few things, including that note I just cited. For the record, GNU true is a 12 KB binary, and claims here (Ubuntu 7.10) to be on version 5.97.
NOTE: your shell may have its own version of true, which usually supersedes the version described here. Please refer to your shell's documentation for details about the options it supports.
And indeed it does, at least for bash. Calling "true --help" here does nothing, while "`which true` --help" says a few things, including that note I just cited. For the record, GNU true is a 12 KB binary, and claims here (Ubuntu 7.10) to be on version 5.97.
That is the best code i have seen in years (Reference: Coding Horror: The Best Code is No Code At All)
<< Home
