From: Brian Candler <b.candler@...>
Date: 2010-06-04T19:47:58+09:00
Subject: Re: Unexpected length of £ (pound) character?

Anthony Stenhouse wrote:
> Hi,
> 
> Today I came across an issue with a customer custom report which was out
> by 1 char over 40 or so lines. At first thought I had incorrectly
> limited the field length, however the problem is only present where
> there is a '£' char.
> 
> For example:
>   "1234".length => 4
>   "1234£".length => 6 (Expect 5)
>   "1234£6".length => 7 (Expect 6)

In UTF-8, "£" is two bytes, and ruby 1.8 gives you the number of bytes.

If you want to capture (say) the first 6 characters of the string, try 
this:

>> a = "1234£6789"
=> "1234\302\2436789"
>> a =~ /\A(.{6})/u
=> 0
>> puts $1
1234£6
=> nil

This may be sufficient for simple wrapping functions. Or look at the 
Iconv library.

> Is this expected functionality in ruby? It does not seem right in my
> mind.

ruby 1.9 works in characters. It brings with it enormous complexity, 
pitfalls and inconsistencies. Pick your poison :-)
-- 
Posted via http://www.ruby-forum.com/.