On Fri, Nov 27, 2009 at 4:04 AM, John Lato <jwl...@gmail.com> wrote: > My apologies for not replying; I have been traveling and am only now > working through my email. > > Oleg's response is much better than anything I would have written. > I'd like to add one point. > > stream2list is very inefficient as he mentioned, however only for > large values of 'n'. For small n it should be fine. Assuming you're > using Word8 elements, "small" means "< 4096". This is because the > default chunk size reading from a file is 2048 elements, so for any n > < 4096 you have at most two concatenations in producing the > stream2list. > > Sincerely, > John Lato > > PS for one example of a binary data parser, please see > http://inmachina.net/~jwlato/haskell/iter-audio/ > > This is similar to the audio codec included with iteratee, but much > more efficient. In particular, the functions "convFunc" and > "unroller" in Sound.Iteratee.Codecs.Common are pretty highly > optimized.
Wonderful! Sample code is very helpful to get familiar with iteratees. Thank you, John. Thanks to both of you. -- vvv _______________________________________________ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
