Re: [Haskell-cafe] Alex Lexer Performance Issues

2011-06-26 Thread Malcolm Wallace
On 22 Jun 2011, at 15:53, Tristan Ravitch wrote: On Wed, Jun 22, 2011 at 07:48:40AM +0100, Stephen Tetley wrote: How fast is good old String rather than ByteString? For lexing, String is a good fit (cheap deconstruction at the head / front). For your particular case, maybe it loses due to

Re: [Haskell-cafe] Alex Lexer Performance Issues

2011-06-26 Thread Richard O'Keefe
On 26/06/2011, at 10:19 PM, Malcolm Wallace wrote: There is an old folklore that lexing is usually the most expensive phase of any compiler-like traversal. 50% of time and space expended on lexing was pretty common twenty years ago. Indeed it is old, but no, it isn't folklore, you'll find

Re: [Haskell-cafe] Alex Lexer Performance Issues

2011-06-22 Thread Stephen Tetley
How fast is good old String rather than ByteString? For lexing, String is a good fit (cheap deconstruction at the head / front). For your particular case, maybe it loses due to the large file size, maybe it doesn't... ___ Haskell-Cafe mailing list

Re: [Haskell-cafe] Alex Lexer Performance Issues

2011-06-22 Thread Tristan Ravitch
On Wed, Jun 22, 2011 at 07:48:40AM +0100, Stephen Tetley wrote: How fast is good old String rather than ByteString? For lexing, String is a good fit (cheap deconstruction at the head / front). For your particular case, maybe it loses due to the large file size, maybe it doesn't... I gave it