从外壳调用红宝石

Question

When I have an executable ruby script foo starting with a shebang

foo

#!/usr/bin/env ruby

and call that ruby script from within a bash script bar as executable (ie, not calling ruby foo , but directly foo ), how can I get the full path of bar from within the ruby script foo ?

---

Edit

If this is not possible, then is it possible if I have a bash script baz in between so that:

bar (bash) calls baz (bash) which calls foo (ruby)

where bar calls baz without any explicit argument and baz figures out the path of its caller bar , and passes that as an argument when calling foo ?

For my purpose, it is okay whether or not foo needs to receive the path information as an argument as long as the original bash script bar does not need to pass that explicitly.

Answer 1

First, note that the question is not well defined. If bar is not a unique link to the executable, then there is no unique path. Assuming you don't care about that issue and you just want to know how bar was accessed and if you are running on Linux that information is available in /proc/pid-of-bar/cmdline . If foo's parent is the process running bar (it should be, unless you've daemonized or foo is not a direct descendant), the bar's pid is available to foo in the environment at PPID (although ruby almost certainly provides a better way to access the parent's pid.) So, get your parent's pid and read /proc/parent-pid/cmdline. If bar was invoked as a shell script, the first string of cmdline will be the interpreter (terminated by a null), and the second string will be the path you care about.

Answer 2

If I understand your question $0 is all you need.

Edit: and of course pass this into your ruby script via the bar script.