Hello Johan, Friday, February 1, 2008, 11:56:09 AM, you wrote:
I think the Haskell Way of interpreting bytes as Latin-1 - while unfortunate in today's multi-everything environment - is something we just have to live with. Too much code expects this behavior, and too many tasks require just reading ASCII to be burdened with complications of locales and character sets.
Why do we have to live with it? I understand why we ended up with the situation we have today. Most languages have/had the same problem.
the problem is just over-complication of existing Handle implementation - it's about 2kloc of high-messed code. taking into account that it lacks a lot of other required features, it's quite meaningless to spend a lot of time fixing it for just one particular problem i've implemented 2 versions of Streams library which includes everything i know about i/o (i.e. no networking, overlapped i/o or iconv) and started to implement its 3rd version which includes bytestring i/o but then switched to more interesting project. Streams is still the most complete haskell i/o library and quite useful. its main drawback is lack of docs, especially for second version, but to some degree you can understand it just by analogy with System.IO interfaces -- Best regards, Bulat mailto:Bulat.Ziganshin@gmail.com