Nestable highlighters

[ethercrow]

Original author: [email protected] (March 16, 2009 14:56:52)

Example of usage:

• Haddock markup in haskell comments
• Literate Haskell should just be some composition of Haskell and LaTeX
• String literals could be better highlighted (escaping characters) if it
  were a sub-lexer.
• Quasiquotations one day we should be able to nicely display using a
  customizable way things [$sql| SELECT * FROM people |] or
  [$regex|(a|[bc])|] or [$html| <b>Boo</b> |]...

Original issue: http://code.google.com/p/yi-editor/issues/detail?id=252

[ethercrow]

From [email protected] on March 16, 2009 16:03:00

• Quasiquotations one day we should be able to nicely display using a
  customizable way things [$sql| SELECT * FROM people |] or
  [$regex|(a|[bc])|] or [$html| Boo |]...

I don't really follow this example. Are you trying to give an example of a Haskell
source file which Yi would highlight as Haskell, but within particular fragments (as
specified by the Haskell parser), yi would highlight as a different language?

[ethercrow]

From [email protected] on March 23, 2009 11:55:11
@gwern0
exactly

[noughtmare]

I was just reading "Combinator Parsing: A Short Tutorial" by Doaitse Swierstra and found this:

Usually one is satisfied to have just a single way of scanning the
input, but sometimes one may want to use a parser for one language as sub-
parser in the parser for another language. An example of this is when one has a
Haskell parser and wants to recognise a String value. Of course one could offload
the recognition of string values to the tokeniser, but wouldn’t it be nice if we
could just call the parser for strings as a sub-parser, which uses single characters
as its token type? A second example arises when one extend a language like
Java with a sub-language like AspectJ, which again has Java as a sub-language.
Normally this creates all kind of problems with the scanning process, but if we
are able to switch from scanner type, many problems disappear.

I am now trying to understand the parsers mentioned in that paper and adapting them to make them work with Yi.