related work

flaws rev 90:

• Code does not pay attention to != vs &= vs =
• Code dose contain only one htmlentities which is used for quoting attribute values. So for HTML content no quoting is done at all!

See my statement in the README. The parser will never be compliant without a rewrite. The author seems to be gone away ? No replies or updates for a long time.

The idea was to use a proper Parser lexer and grammar. Current status: broken which the author admitted. To make it work maybe a lot of more work has to be done. So they assigned higher priorities to other projects meanwhile...

Probbaly fixing it is not that hard: You could approximate PHP expressions the way its done in this library by rewriting this recursive regex in the grammar:

// (?R) matches to most outer regexp recursively
// That's something Ruby can't do (yet?)
// You can't just parse until you hit a ","
// because contents can be  %tag = substr(2,4,8)
// so the substr will be matched by "anything else", the (...) will be
// matched by the ", separated func args"

// keys are documentation only
$items = array(
  '" str' => '"([^"\\]+|[\\].)*"',
  "' str" => ''([^'\\]+|[\\].)*'',
  ', separated func args' =>  '(((?R)(,(?R))*))',
  'recursion in ()' => '((?R))', // this catches nested ( 2 + (4 + ) ) ..
  // '{(?R)}
  ' anything else but terminal' => "[^'"(){},n$s]+"
);
$regex ='('. implode('|',$items).')+';

• Qiuck and dirty minimal implementation
• less than 150 lines of code
• provides some interesting ideas
• parser based on regex (hard to maintain?)
• only a subset is supoprted in particular
  • no quoting. You could workaround it by

    -echo htmlentities($value);
    

  • attributes are not handled the HAML way.