Dave Whipp wrote:

If I want to parse a language that is sensitive to whitespace indentation (e.g. Python, Haskell), how do I do it using P6 rules/grammars?

The way I'd usually handle it is to have a lexer that examines leading whitespace and converts it into "indent" and "unindent" tokens. The grammer can then use these tokens in the same way that it would any other block-delimiter.

This requires a stateful lexer, because to work out the number of "unindent" tokens on a line, it needs to know what the indentation positions are. How would I write a P6 rule that defines <indent> and <unindent> tokens? Alternatively (if a different approach is needed) how would I use P6 to parse such a language?

In this context, I thought readers of this list might be interested in the following extract from mediawiki.lmn, a ruleset for generating html pages from a subset of mediawiki markup. These rules are written in lmn, the metalanguage of the language machine, and the extract deals with unordered and ordered lists, where entries are prefixed by '*' and '#' characters, and repeated prefix characters indicate nesting.

Not that the source text of lmn rules is written using a subset of the mediawiki markup, with preformatted text (lines that start with at least one space) treated as actual source with no markup and everything else treated as annotation:

----------------- start of extract from mediawiki.lmn ------------------
== bulleted and numbered lists ==
Unordered and ordered lists are a bit tricky - essentially they are like indented blocks in Python, but a little more complex because of the way ordered and unordered lists can be combined with each other. The solution is that at each level, the prefix pattern of '#' and '*' characters is known, and the level continues while that pattern is recognised. This can be done by matching the value of a variable which holds the pattern for the current level.

   '*'                                  <- unit - ulist :'*';
   '#'                                  <- unit - olist :'#';
   ulist :A item :X repeat more item :Y <- unit ul :{X each Y} eom;
   olist :A item :X repeat more item :Y <- unit ol :{X each Y} eom;

   '*'                                  <- item - ulist :{A'*'};
   '#'                                  <- item - olist :{A'#'};
   ulist :A item :X repeat more item :Y <- item :{ ul :{X each Y}};
   olist :A item :X repeat more item :Y <- item :{ ol :{X each Y}};
   - wikitext :X                        <- item :{ li :X };

The following rule permits a level to continue as long as the input matches the current prefix. We recurse for each level before getting here, so we will always try to match the innermost levels first - they have the longest prefix strings, and so there is no danger of a premature match

   - A                                  <- more ;
-----------------  end of extract from mediawiki.lmn  ------------------

The complete ruleset can be seen at:
http://languagemachine.sourceforge.net/website.html    - summary
http://languagemachine.sourceforge.net/mediawiki.html  - markup
http://languagemachine.sourceforge.net/sitehtml.html   - wrappings

I have fairly recently published the language machine under Gnu GPL at sourceforge. It consists of a minimal main program, a shared library written in D using the gdc frontend to gnu gcc, and several flavours of an lmn metalanguage compiler - these are all written in lmn and share a common frontend.

The metalanguage compiler sources are on the website (with many other examples) as web pages that have been generated directly from lmn source text by applying the markup-to-html translation rules.

The language machine in previous incarnations has a long history, but it is not much like any other language toolkit that I know of. This is a page that relates it to received wisdom about language and language implementations:


There is an extremely useful diagram which shows what happens when unrestricted grammatical substitution rules are applied to an input stream - this is explained here in relation to a couple of trivially simple examples:


My intention in creating this implementation has been to make something that can be combined with other free languages and toolchains, and I have recently asked the grants-secretary at the Perl Foundation for feedback on a draft proposal to create a language machine extension for perl.

I would be very interested to hear what you think.

Peri Hankey

http://languagemachine.sourceforge.net - The language machine

Reply via email to