# Shell \<-\> Ruby question
**URL:** https://rubytalk.org/t/shell-ruby-question/2506
**Category:** ruby-talk
**Created:** [27 October 2002 02:27 UTC](https://rubytalk.org/t/shell-ruby-question/2506 "2002-10-27T02:27:38Z")
**Posts on this page:** 4
**Page:** 1
### Author:  [@Christian\_Szegedy1](https://rubytalk.org/u/Christian_Szegedy1)
#### Post date: [27 October 2002 02:27 UTC](https://rubytalk.org/t/shell-ruby-question/2506/1 "2002-10-27T02:27:38Z")
This is not a strictly Ruby question, but perhaps
you have some idea which could help me.
Recently I have written some some concise ruby scripts
which help me while working in the (bash) shell. (E.g.
x extracts any archive etc.)
I also wanted to write a more intelligent cd
(change sirectory) script in Ruby. Of course there is
an obvious problem I can’t solve: the bash opens
a new shell for the script, so the working directory
remains unchanged in the shell.
I also tried to run ruby with
… /usr/bin/ruby
I thought, it would run ruby in the current environment
(like ksh), but it simpy gives the error:
bash: ELF: command not found
Does someone have an idea, how could I change the
working directory by a Ruby script?
Regards, Christian
---
### Author:  [@Nobuyoshi\_Nakada](https://rubytalk.org/u/Nobuyoshi_Nakada)
#### Post date: [27 October 2002 02:46 UTC](https://rubytalk.org/t/shell-ruby-question/2506/2 "2002-10-27T02:46:02Z")
Hi,
> Does someone have an idea, how could I change the
> working directory by a Ruby script?
Fundamently, child processes can never modify directly parent
process’es environment. Unix process model prohibits it. You
can just receive child’s output and cd by the shell process
itself.
foo() {eval `ruby somescript`}
# somescript
puts “cd #{directroy}”
> **···**
>
> At Sun, 27 Oct 2002 11:27:38 +0900, Christian Szegedy wrote:
>
> –
> Nobu Nakada
---
### Author:  [@YANAGAWA\_Kazuhisa](https://rubytalk.org/u/YANAGAWA_Kazuhisa)
#### Post date: [27 October 2002 03:38 UTC](https://rubytalk.org/t/shell-ruby-question/2506/3 "2002-10-27T03:38:36Z")
> I also wanted to write a more intelligent cd
> (change sirectory) script in Ruby. Of course there is
> an obvious problem I can’t solve: the bash opens
> a new shell for the script, so the working directory
> remains unchanged in the shell.
How about ruby script only make a path to cd and do actual cd by shell
function or alias? Namely:
```
~> function mcd () {cd `ruby /tmp/foo.rb`}
~> cat > /tmp/foo.rb
puts "/tmp"
^D
~> mcd
/tmp>
```
Of course a script may do more useful thing, and a function/alias must
be more safe; at least a function/alias should check exit status of
the script and the script signals failure on such as no directory to
cd being found.
> . /usr/bin/ruby
>
> I thought, it would run ruby in the current environment
No, probably ksh also doesn’t. `.' in sh-derived shells means `run
the _script_ in the current context. So `. /usr/bin/ruby’ evaluates
ruby itself as shell script.
> **···**
>
> In message [apfia9$aq1$05$1@news.t-online.com](mailto:apfia9$aq1$05$1@news.t-online.com) [szegedy@t-online.de](mailto:szegedy@t-online.de) writes:
>
> –
> kjana@dm4lab.to October 27, 2002
> Behind the clouds is the sun still shining.
---
### Author:  [@Christian\_Szegedy1](https://rubytalk.org/u/Christian_Szegedy1)
#### Post date: [27 October 2002 13:28 UTC](https://rubytalk.org/t/shell-ruby-question/2506/4 "2002-10-27T13:28:11Z")
YANAGAWA Kazuhisa wrote:
How about ruby script only make a path to cd and do actual cd by shell
function or alias? Namely:
```
~> function mcd () {cd `ruby /tmp/foo.rb`}
~> cat > /tmp/foo.rb
```
Thanks! This was a great idea! It solves my problem.
Thank you for the help!
Regards, Christian