From: Yukihiro Matsumoto <matz@...>
Date: 2007-11-03T00:12:59+09:00
Subject: Re: Didn't find String#subseq

Hi,

In message "Re: Didn't find String#subseq"
    on Fri, 2 Nov 2007 04:20:23 +0900, Wolfgang N�dasi-Donner <ed.odanow@wonado.de> writes:

|Here is exactly the problem possible I mentioned...
|
|class String
|   def byte_range(from, to)
|     e = self.encoding
|     r = self.force_encoding('ascii')[from..to]
|     r.force_encoding(e)
|end
|
|puts "b�h!".byte_range(3,4) # => "h!"
|puts "b�h!".byte_range(2,4) # ! nonsense !
|
|..., so I think it will be a good idea to have a check method.

We will provide a validation method, name not fixed yet.

|But, and this is very important for these kind of usage, how does Ruby 1.9 
|handle the following...
|
|s = "a�o�u�" * 5000
|29999.downto(29000){|i|puts s[i]}
|
|..., if s.encoding is utf-8? - Does Ruby calculate the character positions for 
|each access via [], or will the positions be stored for later usage after first 
|calculation?

Ruby calculate the position each time if encoding is variable
character width.  It may cause performance problem, but less often
than you've expect.

							matz.