[Haskell-cafe] Re: Draft MissingH Reorg Plan

[Malcolm Wallace]

John Goerzen <jgoerzen at complete.org> wrote:

> > About Text.ParserCombinators.Parsec.RFC2234 and the like.  ...
> >        ...             you do not in
> > fact define any new combinators for parsing.  What you have there is
> > a parser
> 
> This module provides parsers for the grammar defined in RFC2234,
> 
> alpha :: CharParser st Char
> bit :: CharParser st Char
> character :: CharParser st Char
> cr :: CharParser st Char

Yes, I do see the point that these are rather generic and basic parsers.
(But they are not combinators as such, just alternative "primitives" of
the domain-specific parsing language.)  I guess they could easily live
with Parsec after all, unless there is an obviously better place for
them in the hierarchy.

> Does the above suggest a single good place for them to you?

You mean, for the constructed parsers, as well as the basics?  I suppose
I would go for much the same suggestions as before, apart from RFC2234:

    Text.ParserCombinators.Parsec.RFC2234
    Network.SMTP.RFC2821.Parse
    Network.Email.RFC2822.Parse

Conceivably, you might want to drop the RFC part of the latter two names,
and merely refer users to the relevant RFC in the Haddock docs?

    Network.SMTP.Parse
    Network.Email.Parse

Regards,
    Malcolm