I have had a look at this issue and I find that there is no problem with running sh "cmd" do end. I will need the error message to know where to start.

Dear Paul,
unfortunately I couldn't reproduce this issue. At the moment it works with the default and the modified 'puppetclass.rb'.
Tomorrow I am going to setup a new Foreman from scratch.
I'll let you know if the problem reoccurs. Otherwise I will close this ticket/bug.
It's quite possible that the cause of this issue is my system (e.g. puppet modules, settings.yaml...).

Thank you for taking note.

Kind regards,
mayak

Dear Paul,
I was finally able to "reproduce" the problem on the same server.
Well actually I am not quite sure how I managed to break my foreman again.

Since I've updated some of my Puppet Classes (more precisely the classes documentation) I run the following command:
rake puppet:import:puppet_classes RAILS_ENV=production

After this I executed the following command:
rake puppet:rdoc:generate RAILS_ENV=production --trace 2>&1 | less

Please follow this link to find the error message:
http://pastie.org/2031319

When I exchange the 'puppetclass.rb' with the "semi broken" one of Ohad the following error occurred:
http://www.pastie.org/2031411

In both cases I cleaned the respective directory:
/usr/share/foreman/public/puppet/rdoc/production/

Please let me know if you need more precise information.

Thank you for your efforts!

Cheers, mayak

-

System Information:

CentOS release 5.6 (Final)

foreman-0.3rc2-2
httpd-2.2.3-45.el5.centos.1
mysql-server-5.0.77-4.el5_6.6
puppet-2.6.6-1.el5
puppet-server-2.6.6-1.el5

libselinux-ruby-1.33.4-5.7.el5
ruby-1.8.5-5.el5_4.8
ruby-augeas-0.4.1-1.el5
ruby-devel-1.8.5-5.el5_4.8
ruby-docs-1.8.5-5.el5_4.8
ruby-irb-1.8.5-5.el5_4.8
ruby-libs-1.8.5-5.el5_4.8
ruby-mysql-2.7.3-1.el5
ruby-rdoc-1.8.5-5.el5_4.8
ruby-ri-1.8.5-5.el5_4.8
ruby-shadow-1.4.1-7.el5
rubygems-1.3.1-1.el5
rubygem-fastthread-1.0.7-1.el5
rubygem-json-1.4.3-3.el5.1
rubygem-mime-types-1.16-3.el5
rubygem-passenger-2.2.2-1
rubygem-rake-0.8.7-2.el5
rubygem-rest-client-1.6.1-2.el5
rubygem-sqlite3-ruby-1.2.4-1.el5

Mayak,
It appears that you are using the prepare phase and this is causing the problem. The prepare phase is supposed to return the root directory of the relocated manifest tree. In your case the prepare operation is returning "". This breaks the relocation phase later in the build. Could you post your script/rdoc_prepare_script.rb file?

Dear Paul,
Thank you for looking into this issue.

My script/rdoc_prepare_script.rb is actually empty:

#!/usr/bin/ruby

When I tried to configure the Puppet class browser I followed the Foreman Wiki which says:
Most sites will not require a scripts/rdoc_prepare_script.rb file.
I am sorry if I've misunderstood the instructions and did not configure this file at all.

In the meantime I have extended my Puppet with a production and testing environment.
It seems that I should reconfigure my Foreman before importing 'puppet_classes' or generating 'rdoc'.

I received the following message when running: rake puppet:import:puppet_classes RAILS_ENV=production

Proceed with these modifications? <yes|no> yes
Problems were detected during the execution phase
[...]
production is used by hostname.domain.tld
[...]
Import failed

So I will need to reconfigure my Foreman first. I will come back to you asap. Sorry.

Kind regards,
mayak

On the first point
My apologies, the wiki is not clear.
If you do not need a rdoc_prepare_script.rb then it should not exist. An empty script will cause an error. Please remove this file and try again.
Please open a new ticket with the second point as otherwise we are not able to track issues properly

Mayak, Is it OK to close this ticket now?