Re: [Haskell-cafe] ANNOUNCE: A ReadP style parser for ByteStrings

2007-12-12 Thread Don Stewart
gracjanpolak:
 
 I'm happy to announce a ReadP style parser for ByteStrings,
 Text.ParserCombinators.ReadP.ByteString.
 
 http://hackage.haskell.org/cgi-bin/hackage-scripts/package/bytestringreadp
 
 Text.ParserCombinators.ReadP.ByteString is an adaptation of
 Text.ParserCombinators.ReadP to work over Data.ByteString as input
 stream representation. This gives enormous improvements in terms of
 parsing speed but most significantly in memory usage.
 
 Features:
 
  * ReadP style parser over ByteString input
  * Drop-in replacement for Text.ParserCombinators.ReadP
  * Fast
  * Good memory usage
 
 The algorithm is slightly modified to exploit ByteString as random access
 data input structure. Unlike original ReadP, that stressed garbage collection 
 very much by creating a lot of conses (:), this parser has very good memory
 allocation behaviour.
 
 Package works out of the box with GHC 6.8.1, with slight (cabal) modifications
 also with GHC 6.6.1.
 
 Thanks to everyone for their support! Happy hacking!


Great work, Gracjan. I hope to try it out soon. I think this is the
first of the many bytestring parser projects started to actually get
released onto Hackage! Well done!

-- Don
___
Haskell-Cafe mailing list
Haskell-Cafe@haskell.org
http://www.haskell.org/mailman/listinfo/haskell-cafe


[Haskell-cafe] ANNOUNCE: A ReadP style parser for ByteStrings

2007-12-11 Thread Gracjan Polak

I'm happy to announce a ReadP style parser for ByteStrings,
Text.ParserCombinators.ReadP.ByteString.

http://hackage.haskell.org/cgi-bin/hackage-scripts/package/bytestringreadp

Text.ParserCombinators.ReadP.ByteString is an adaptation of
Text.ParserCombinators.ReadP to work over Data.ByteString as input
stream representation. This gives enormous improvements in terms of
parsing speed but most significantly in memory usage.

Features:

 * ReadP style parser over ByteString input
 * Drop-in replacement for Text.ParserCombinators.ReadP
 * Fast
 * Good memory usage

The algorithm is slightly modified to exploit ByteString as random access
data input structure. Unlike original ReadP, that stressed garbage collection 
very much by creating a lot of conses (:), this parser has very good memory
allocation behaviour.

Package works out of the box with GHC 6.8.1, with slight (cabal) modifications
also with GHC 6.6.1.

Thanks to everyone for their support! Happy hacking!

--
Gracjan


___
Haskell-Cafe mailing list
Haskell-Cafe@haskell.org
http://www.haskell.org/mailman/listinfo/haskell-cafe