Ragel SIP parser¶

A SIP parser based on Ragel State Machine Compiler. It's 100% strict and implements BNF grammar in RFC 3261 and 3966.

Installation¶

Install Ragel¶

The first step is installing the last stable version of Ragel.

Download Ragel from http://www.complang.org/ragel/ragel-6.6.tar.gz.

Install it by running ./configure, make, make install.

Check the installacion:

~$ ragel -v
Ragel State Machine Compiler version X.X NNNNN YYYY
Copyright (c) 2001-2009 by Adrian Thurston

Install ABNF Generator¶

ABFN Generator parses ABNF definitions from source file and generates output as state machine definition for Ragel.

Download the version 0.2 from http://www.2p.cz/files/2p.cz/downloads/abnfgen-src.0.2.tgz.

Compile it by running make. You will get an executable abnfgen.

Copy/move/link the executable abnfgen to a directory into your PATH environment variable (ie: /usr/local/bin). Make sure it has execution permissions for your user.

Check the installation:

~$ abnfgen -V
abnfgen - ABNF generator, v0.1

Install Ruby 1.8 and RubyGems¶

Make sure you have Ruby 1.8 installed in your computer:

~$ ruby -v
ruby 1.8.6 (2007-09-24 patchlevel 111) [i486-linux]

Download RubyGems from http://rubyforge.org/frs/download.php/45905/rubygems-1.3.1.tgz.

Install it by running ruby setup.rb as root.

Check the installation:
```
~$ gem1.8 -v
1.3.1
```

Install term-ansicolor gem (as root):
```
~# gem1.8 install term-ansicolor
```

Download Ragel SIP Parser¶

Get it using Subversion:

~$ svn co svn://dev.sipdoc.net/Ragel-SIP-Parser Ragel-SIP-Parser

Enter into the directory Ragel-SIP-Parser and make sure that sip_parser.sh has running permissions. If not add them by running:
```
~$ chmod +x sip_parser.sh
```

Usage¶

~$ ./sip_parser.sh KEYWORD 'TEXT'

KEYWORD : Any BNF keyword in RFC 3261 or RFC 3966
TEXT : Text to match enclosed with ' or "

~$ ./sip_parser.sh KEYWORD -f FILE

KEYWORD : Any BNF keyword in RFC 3261 or RFC 3966
FILE : File containing the text to match, don't add newline (<CRLF> string is replaced with "\r\n")

After running the above command, the script will compile the Ragel machine according to the KEYWORD parameter and show a colorized output. In case the text is invalid, the first wrong character will be pointed by a red arrow.

Examples¶

Checking a valid 'Request-Line':

~$ ./sip_parser.sh Request-Line 'INVITE sip:ibc@aliax.net;transport=TcP SIP/2.0'

Compiling 'sip_parser.rl' ...
OK

Parsing 'Request_Line':  INVITE sip:ibc@aliax.net;transport=TcP SIP/2.0 ...
  --------------------------------
  Finished:  cs: 387  p: 47  pe: 47
  Time:      0.000954
  --------------------------------
OK

Checking an invalid 'SIP-URI' (host part contains illegal character '_'):

~$ ./sip_parser.sh SIP-URI 'SIp:alice:mypassword@invalid_hostpart.com:5070;param=xxx'

Compiling 'sip_parser.rl' ...
OK

Parsing 'SIP_URI':  SIp:alice:mypassword@invalid_hostpart.com:5070;param=xxx ...
  --------------------------------
  Finished:  cs: 0  p: 28  pe: 57
  Error parsing the data in character 28:

    SIp:alice:mypassword@invalid_hostpart.com:5070;param=xxx
                                ^
  Time:      0.001084
  --------------------------------
ERROR

Checking a valid 'From' header:

./sip_parser.sh From 'f: "Iñaki B.C." <sips:ibc@aliax.net;cool=yes>;hparam1=on;hparam2;hparam3=xxx'

Compiling 'sip_parser.rl' ...
OK

Parsing 'From':  f: "Iñaki B.C." <sips:ibc@aliax.net;cool=yes>;hparam1=on;hparam2;hparam3=xxx ...
  --------------------------------
  Finished:  cs: 858  p: 78  pe: 78
  Time:      0.001584
  --------------------------------
OK

Checking an invalid 'Contact' multi-value header (the third URI is in 'addr-spec' form containing an invalid question mark '?' symbol):

~$ ./sip_parser.sh Contact 'ConTACT:  Alice <sip:alice?VALID-QUESTION-MARK@domain.org>;hparam=1 ,  tel:+1234;param=asd,  sips:user?INVALID-QUESTION-MARK@mydomain.org'

Compiling 'sip_parser.rl' ...
OK

Parsing 'Contact':  ConTACT:  Alice <sip:alice?VALID-QUESTION-MARK@domain.org>;hparam=1 ,  tel:+1234;param=asd,  sips:user?INVALID-QUESTION-MARK@mydomain.org ...
  --------------------------------
  Finished:  cs: 0  p: 102  pe: 138
  Error parsing the data in character 102:

    ConTACT:  Alice <sip:alice?VALID-QUESTION-MARK@domain.org>;hparam=1 ,  tel:+1234;param=asd,  sips:user?INVALID-QUESTION-MARK@mydomain.org
                                                                                                          ^
  Time:      0.001813
  --------------------------------
ERROR

Checking an invalid 'telephone-uri' (it's a 'local number' so it needs the param phone-context):

~$ ./sip_parser.sh telephone-uri 'tel:1234'

Compiling 'sip_parser.rl' ...
OK

Parsing 'telephone_uri':  tel:1234 ...
  --------------------------------
  Finished:  cs: 0  p: 8  pe: 9
  Error parsing the data in character 8:

    tel:1234
            ^
  Time:      0.000826
  --------------------------------
ERROR

Checking a valid IPv6:

~$ ./sip_parser.sh IPv6address '2001:0db8:85a3:08d3:1319:8a2e:0370:7334'

Compiling 'sip_parser.rl' ...
OK

Parsing 'IPv6address':  2001:0db8:85a3:08d3:1319:8a2e:0370:7334 ...
  --------------------------------
  Finished:  cs: 152  p: 40  pe: 40
  Time:      0.00083
  --------------------------------
OK

Examples using grammar written in a file¶

Some SIP grammars cannot be written as Bash command line argument since they include invalid characters as ! or '. We can write the grammar in a file and test it. Also we can simulate a CRLF by writting "<CRLF>" in the file.

For example let check a valid 'Request-Line' appearing in RFC 4475 (SIP Torture Messages):

First create a file called 'FILE' (note there must not be a new line at the end):

!interesting-Method0123456789_*+`.%indeed'~ sip:1_unusual.URI~(to-be!sure)&isn't+it$/crazy?,/;;*:&it+has=1,weird!*pas$wo~d_too.(doesn't-it)@example.com SIP/2.0

Check it:

~$ ./sip_parser.sh Request-Line -f FILE

Compiling 'sip_parser.rl' ...
OK

Parsing 'Request_Line':  !interesting-Method0123456789_*+`.0ndeed'~ sip:1_unusual.URI~(to-be!sure)&isn't+it$/crazy?,/;;*:&it+has=1,weird!*pas$wo~d_too.(doesn't-it)@example.com SIP/2.0 ...
  --------------------------------
  Finished:  cs: 387  p: 160  pe: 160
  Time:      0.001961
  --------------------------------
OK

Now, let insert a valid CRLF into a 'Via' header:

Create a file called 'FILE2':

Via: SIP / 2.0 / UDP<CRLF> first.example.com: 4000;ttl=16;maddr=224.2.0.1 ;branch=z9hG4bKa7c6a8dlze.1

Let's check it:

$ ./sip_parser.sh Via -f FILE

Compiling 'sip_parser.rl' ...
OK

Parsing 'Via':  Via: SIP / 2.0 / UDP
 first.example.com: 4000;ttl=16;maddr=224.2.0.1 ;branch=z9hG4bKa7c6a8dlze.1 ...
  --------------------------------
  Finished:  cs: 624  p: 98  pe: 98
  Time:      0.001065
  --------------------------------
OK

Issues¶

svn update¶

Some files are generated or modified each time 'sip_parser.sh' is executed (for example 'main.rl'). But these files are also included in the Subversion repository so when doing a svn update you will get some conflicts, ie:

~$/Ragel-SIP-Parser$ svn update
A    FILE
C    main.rl
U    sip_parser.sh
C    sip_parser.rb
Updated to revision 47.

In order to avoid it, delete the files in conflict and perform the svn update (this is a really dirty workaround, perhaps it's possible to force an update of the trunk version replacing our local modifed files instead of creating "conflict" files).

An included command called svn_update.sh will perform these steps for us:

It deletes 'main.rl' and 'sip_parser.rb' files.
It executes svn update.

Example:

~$ ./svn_update.sh

Deleting 'main.rl' and 'sip_parser.rb' ...

svn update ...
Se restituyó 'main.rl'
Se restituyó 'sip_parser.rb'
In revision 48.

OK

UTF-8 in Ruby output¶

UTF-8 symbols are wrongly displayed in the Ruby output so the error pointer (^) could point to incorrect position.

Also available in: HTML TXT

Ragel SIP parser

Wiki