From: Paul Brannan <paul@...>
Date: 2001-12-05T05:36:01+09:00
Subject: [ruby-talk:27480] Re: Shebangikowksy

On Wed, Dec 05, 2001 at 04:09:26AM +0900, Martin Weber wrote:
> If I write software I want people to be able to use it easily, all they
> should need to know is what the script (which sometimes turns out to be
> a big bad huge program, not a "dirty little script") will do, not which
> interpreter is used or where to find this one. I've had enough "bug-reports"
> from a couple of people who work in a department as chip-designers, using
> unix as their platform as of the availability of certain software for it,
> that the program I've wrote for them wouldn't work, because dumb me hard-
> coded the path in the first place.

Personally, if I write software, I want to have fun writing it.

If I want to make sure that software I write works on as many systems as
possible, then I'm going to have to do things that may not be fun.  For
example, the /usr/bin/env solution doesn't work in the case that /usr is
not mounted or /usr is simply not present (a condition that I have run
into before).  So to make sure that the script works in this condition:

#!/bin/sh
case "$PATH" in
"") PATH=$PATH:/usr/bin:/usr/local/bin ;; esac
ruby -w -e"eval((['']*5+File.readlines(\"$0\")[5..-1]).join(\"\n\"),nil,\"$0\")"
exit $?
---    Ruby script follows ---
puts "foo!"

This also solves the -w problem (/usr/bin/env won't let you pass
any arguments to the program it calls, because it receives everything in
ARGV[1]).

Another solution is to use:
#!/bin/nice ruby

which solves the missing /usr problem but not the -w problem (and runs
the script with a modified priority).

> In the end I have to admit, though, that there is no 'perfect' way. Sigh :(

Yep.  I think this is a problem that can't be solved without changing
Unix.  Apart from that, every solution we come up with is going to be a
tradeoff of some form or another.

Personally, I'll leave my scripts -x and tell the user to use ruby -S.

Paul