
3 May
2005
3 May
'05
7:37 a.m.
On Fri, Apr 29, 2005 at 07:38:28PM +0100, Krasimir Angelov wrote:
In Haddock there is another good Haskell Lexer/Parser. It should be good to have one reusable parser suitable for all projects.
That's yet another fork of the same code. The problem is that LALR parsers aren't usually parameterizable. You can turn stuff on and off by hiding symbols in the lexer or testing after parsing, but you have to have thought of it in advance. There's the danger of bloat -- do we want Haddock's doc comments in the shared parser? And then as soon as someone wants to add some extra bit of syntax, they'll fork again.