Comments on: Parsing a simple config file in Haskell

By: links for 2009-04-14 | Perpetually Curious

links for 2009-04-14 | Perpetually Curious — Wed, 15 Apr 2009 04:30:41 +0000

[...] teideal glic deisbhéalach » Blog Archive » Parsing a simple config file in Haskell Yet another parsec article. This one is about parsing a simple configuration file. (tags: haskell parsec parser tutorial) [...]

By: Bryan O'Sullivan

Bryan O'Sullivan — Fri, 02 Feb 2007 00:20:52 +0000

Tristram -

I wrote listToMap because I overlooked the existence of Map.fromList.

I have no intuition as to whether readConfig should be more or less efficient than readConf2, as Haskell compiler technology has advanced a long way since the last time I looked at it in any depth.

By: Tristram Brelstaff

Tristram Brelstaff — Thu, 01 Feb 2007 12:44:12 +0000

A nice article. It was discovering how to do parsing (as explained by J.S.Rohl in his “An Introduction to Compiler Writing”) that first got me into Pascal programming back in 1980. I think your article has done the same for me with Haskell. I have couple of questions though:

1. Is there any reason that you used listToMap rather than Map.fromList in readconf2?

2. Would you expect the “point-free” readConfig to be more efficient than the “beginner’s Haskell” readCon2?

By: Chris Eidhof

Chris Eidhof — Thu, 01 Feb 2007 11:39:24 +0000

Nice tutorial! Most of the times, when building more advanced parser, you’d want to make use of a Lexer. Parsec also provides very good support for that. A lexer recognizes separate tokens, and the parser parses those tokens and makes something useful of it.

For example, a very common and useful task of the Lexer is to deal with the whitespace. Parsec also provides good helpers for that, for example the function “lexeme”. When you use it (for example like: lexeme identifier), it’ll eat all the trailing whitespace, which is quite useful in practice!