Parser question + Markdown "wysiwyg"

[david parsons]

In article <20100420094250339725.51497217 at yahoo.co.uk>,
Albert Skye  <markdown-discuss at six.pairlist.net> wrote:

>http://eigenclass.org/R2/writings/fast-extensible-simplified-markdown-in-ocaml

>http://eigenclass.org/R2/writings/markdown-redux-html-generation


    The ocaml markdown clone (it's explicitly not markdown by design, and
    would take a little bit of editing to override the writer's intent) is
    quite fast and small, but that's not really relevant to the original
    question, is it?

    In regards to the actual question, I do have code in discount that
    creates a block-level parse tree, but I don't build a parse tree for
    the content inside paragraphs.   And I don't keep track of byte
    offsets into the source.    It wouldn't be difficult to add the
    offset to the start of each paragraph into the data structures;
    you'd need to keep track of byte & line offsets when lineilating the
    input into the library, and then each paragraph object could easily
    get to that information for its first line.

    If you wanted to do the same for span-level markup, you'd be
    completely on your own.   I'm sorry about that, but I never
    considered using discount as an interpreter instead of a compiler,
    and it would probably take a fairly extensive redesign to get it
    to work nicely without suffering a staggering performance penalty.

    -david parsons