Sorry I don't know what that means. °

Can you output each service independently or do you need to know all the data before writing?

Update

My approach would be using a flip-flop operator to read each service and to output at the end.

Each line would be processed by a dedicated subroutine based on the first word.

Like service::description() and so on.

This would make your code far more readable and easier to maintain.

°) I can help you with Perl, but not with Cisco or Juniper, sorry.

does it mean

• it's a continuation of the previous line indicated by the extra indentation
• it's nested information, and this cisco format has to be parsed like a tree

 
service instance 20 ethernet
  description Internal site
  encapsulation dot1q 17
  rewrite ingress tag pop 1 symmetric
  xconnect 192.168.51.20 838969735 encapsulation mpls
   mtu 9202
 !
[download]

The above is exactly the configuration for each service instance. The extra indentation for the MTU is because the MTU is being applied to the configuration of the xconnect (xconnect being nested under service instance, mtu nested under xconnect). Hope that helps. edit: forgot a /

Yes, the xconnect is nested under the service instance(directly configured under the service instance), and the MTU is nested under the xconnect(directly configured under the xconnect).

I don't think I know enough about those tools to help you, and you don't understand my questions.

Your code is also very hard to read for me.

A general solution is to parse your tree data into a nested array of hashes and to traverse that to output your new format.

But I don't know if that's really necessary because I don't understand the data.

I hope someone with deeper knowledge of your domain will pass by and can help you. :)

update

see also

Re: Cisco to Juniper - parser help

Cheers Rolf
_{(addicted to the Perl Programming Language :)
Wikisyntax for the Monastery}