From: Robert Klemme <shortcutter@...>
Date: 2009-01-07T15:55:46+09:00
Subject: Re: Simple Regexp help

On 07.01.2009 01:17, Tim Greer wrote:
> Joe Blow wrote:
> 
>> How can I test a word to make sure it ONLY contains certain
>> characters?
>>
>> Say i have an expression like /[A-Z]/i
>>
>> How could i have "Testing" pass but "Testing123" fail?
>>
>> PS. I can not dynamically create the expression so it looks like this
>> /[A-Z{10}]/
> 
> If you only want a-z, set the character class to start and end the
> string, otherwise it'll match anything with a character that's in the
> alphabet, regardless of what follows it.  I.e., /^[a-z]+$/i will only
> match a-z characters from start to end.

Note, this is not totally safe:

irb(main):001:0> s = "foo\n"
=> "foo\n"
irb(main):002:0> /^[a-z]+$/i =~ s
=> 0

As you can see, even a string with a newline at the end passes.  This 
version is safer because anchors do actually use start and end of the 
string:

irb(main):003:0> /\A[a-z]+\z/i =~ s
=> nil

>  You can then change the
> character class to whatever you wish.  ^ is the start of the string and
> $ is the end of the string.

Actually ^ is a line start and $ is a line end (see above).  For proper 
anchoring string start and end you need \A and \z.

> + is one or more characters, so ^[a-z]+$ is
> one of more characters in the character class [], being a-z, and noting
> else. 

Note also that, depending on definition of legal string the expression 
should probably contain * instead of + because also the empty string 
does not contain any illegal characters.

Kind regards

	robert

-- 
remember.guy do |as, often| as.you_can - without end