PPI Manual - The ( unfinished ) manual for PPI


About this Document

This is the PPI ( Parse::Perl::Isolated ) user manual. It describes PPI, it's structure, it's use, an overview of it's API, and provides implementation samples.


The ability to read, and understand perl ( programmatically ) outside of the perl executable is one that has caused difficulty for a very long time.

The root cause of this problem is perl's dynamic grammer. Although there are typically not huge differences in the grammer, some things cause large problems.

An example of these are function signatures, as demonstrated by the following.

  @result = (&dothis $foo, $bar);

  # Which of the following is it equivalent to?
  @result = (&dothis($foo), $bar);
  @result = &dothis($foo, $bar);

This code can be interpreted in two different ways, depending on whether the &dothis function is expecting one argument, or two, or several.

To restate this, a parser would need context information that could not be found in the immediate vicinity. In fact, this information might not even be in the same file. It might also not be able to determine this without the prior execution of a BEGIN {} block. In other words, to parse perl, you must also execute it, or if not it, everything that it depends on for it's grammer.

This, while possibly feasable in some circumstances, is not a valid solution ( at least, so far as this module is concerned ). Imagine trying to parse some code that had a dependency on the Win32::* modules from a Unix machine, or trying to parse some code with a dependency on another module that had not even been written yet...

For more information on why it isn't possible to parse perl, see:

Why "Isolated"?

In aknowledgement that someone may some day come up with a valid solution for this problem, we leave the Parse::Perl namespace free. The namespace of Parse::Perl::Isolated ( shortened to PPI ), has been chosen because our purpose here is only to "parse" perl code that is isolated from other resources. That is, we assume that there is no possiblity of accessing other code on which we have a dependency, or of running an instance of perl alongside the parser ( a possible solution for Parse::Perl that is investigated from time to time ).

Goals of PPI

Once we have recognised that an attempt to parse perl well enough to execute it either impossible or beyond an immediate solution, it is important we establish just what uses we want to put a Parse::Perl type module to.

Presentation / Colouring etc

Modify, improve, syntax colour etc the presentation of code.


Analyze the contents of code to automatically generate documentation, parrellel to, or as a replacement for, POD documentation.

Structural Analysis

Determine quality or other metrics across a body of code, and identify situations relating to particular phrases, techniques or locations.


Make structural, syntax, or other changes to code in an automated manner, independantly, or in assistance to an editor.


This includes techniques such as tidying ( ala perltidy ), obfuscation, or to implement formatting preferences or policies.

As long as the above tasks can be achieved, without damaging code, then PPI can be considered to be a success.

Good Enough(TM)

With the above tasks in mind, PPI seeks to be good enough to achieve the above tasks within a single framework, and provide a sufficient good API to allow others to implement code in these and related areas.

However, some limits are applied. Because PPI cannot adapt to changing grammers, any code written using code filters should not assume to be parsable. This includes anything munged by Acme::Bleach, as well as ( arguably ) more common cases like and We do not assume to be able to parse code using these modules, although someone may be able to extend PPI to handle them.

Our goals for success are to be able to successfully parse 99% of all perl source files contained in CPAN. This means the entire file in each case. In pratical tests, we actually do much better. Aside from very intentionally nasty things, the only code which we currently fail to parse is the "selfgol" obscured code entry. We are in fact able to parse most obscured code, but selfgol uses the sub definition sub'new, an apparently special case which is equivalent to sub ::new and not sub::new.


General Layout

PPI has two major parsing componants, PPI::Tokenizer and PPI::Lexer, upon which everything else is based.

The Tokenizer

The Tokenizer takes source code and converts it into a series of tokens. It does this using a slow but thorough character by character manual process, rather than using complex regexs. Currently, this implementation means that PPI is not of use for highly interactive tasks, such as an editor which checks and formats code on the fly.

How slow? As an example, tokenizing, a 7112 line, 40,000 token file takes about 5 seconds on my little Duron 800 test server. So you should expect the tokenizer to work at a rate of about 1700 lines of code per gigacycle. The code is currently being optomised, and there is plenty of scope remaining for speed improvements, but it is fairly slow work. The target rate is about 5000 lines per gigacycle.

The Lexer

The Lexer is currently structurally complete, but lacks a fair percentage of it's logic code, especially relating to code blocks.

At the top level of all lex trees is the PPI::Document. This will contain non-significant tokens ( whitespace, comments, pod etc ), as well as statements and structure. A PPI::Structure is any series of tokens terminated by matching braces ( with a few exceptions ). This includes things like code blocks, conditionals, function argument braces, anonymous array square brackets et al. A PPI::Statement is any series of tokens that is included in a single contigious statement.

The following example shows how Statements and Structures nest.

  PPI::Statement      "Hello World!"
  PPI::Structure      ( "Hello World!" )
  PPI::Statement      print( "Hello World!" );
  PPI::Document       #!/usr/bin/perl
                      print( "Hello World!" );

As you can see, a statement can have as little as one token. The actual nested structure for some code would look something like this.


  print( "Hello World!" );


    PPI::Token::Comment               "#!/usr/bin/perl\n"
    PPI::Token::Whitespace                  "\n"
      PPI::Token::Bareword            'print'
        PPI::Structure                '(' ... ')'
          PPI::Token::Whitespace            ' '
            PPI::Token::Quote::Double '"Hello World!"'
          PPI::Token::Whitespace            ' '
      PPI::Token::Structure           ';'
    PPI::Token::Whitespace                  "\n"
    PPI::Token::Whitespace                  "\n"
      PPI::Token::Bareword            'exit'
      PPI::Structure                  '(' ... ')'
      PPI::Token::Structure           ';'
    PPI::Token::Whitespace                  "\n"

As you can see, the tree can get fairly deep at time, especially when every isolated token in a bracket becomes it's own statement. This is needed to allow anything inside the tree the ability to grow. It also makes the search and analysis algorithms simpler.

Please note that the lexer is partly completed, but not enough to use it on anything "real". It needs major work before it can be considered to be "working".