Re: Unicode (Re: Reverse composition)

[George Russell]

Marcin 'Qrczak' Kowalczyk wrote:
[snip]
> But when Unicode finally comes... How should Haskell's textfile IO
> work? 
I don't think the current standard functions for textfile IO would
have too many problems.  You can do hSeek in Haskell, but 
"The offset is given in terms of 8-bit bytes" (library standard
section 11.5.2).  So if the file is UTF8-encoded the offset may be 
illegal (leading to a possible IO exception now or later) but at 
least the run-time system can find it.