CommonMark Formal Grammar

12

+++ imsky [Sep 03 14 21:12 ]:

Can the re2c specifications be made portable so that the two
implementations (and, in the future, additional implementations) could
rely on one grammar?

What do you mean, “portable”?

Yes, and that is exactly what both the C and javascript
implementations
of stmd do.

Can the C and JS implementations of stmd output the serialized
in-memory representation somehow? In the JS case, this could be a JSON
object with tokens and line numbers, perhaps even with nesting (i.e.
H2s are nested under H1s).

You can use the --ast flag with the C implementation, and it will
give you a representation of the syntax tree (not in any
machine-readable format, but that could be added).

With the js implementation, you can easily get it to output the parse
tree. Apply the following patch to js/markdown:

--- js/markdown 2014-08-14 11:00:36.000000000 -0700
+++ js/markdownast      2014-09-03 14:25:30.000000000 -0700
@@ -10,6 +10,5 @@
     return console.log(err);
   }
   var parser   = new stmd.DocParser();
-  var renderer = new stmd.HtmlRenderer();
-  console.log(renderer.render(parser.parse(data)));
+  console.log(util.inspect(parser.parse(data), { depth: null }));
 });
3 Likes