From: Patrick Hurley <phurley@...>
Date: 2006-06-28T11:21:39+09:00
Subject: Re: string manipulation

On 6/27/06, Jeff Pritchard <jp@jeffpritchard.com> wrote:
> My question is, is String#scan(/./) the best way to convert a string
> into an array so that one could then use Array#each or
> Array#each_with_index to mess with the characters?  It seems to work in
> irb, but I'm guessing there are some corner cases where the regexp would
> falter.  Also seems more computationally  intensive that should be
> necessary to just split it up into characters.

Try String#each_byte and then .chr the numeric byte value.

> Also, my first thought on a solution for this was to simply walk along
> the string and insert dashes where appropriate. Why isn't there a
> "String#insert(where,str)" method?  Inserting substrings into strings
> seems like such a common thing to do.

You mean like String#insert :-) it is there, of course there is no
insert!, so you have to remember to save the result.

> And finally, a question about Alex's gsub.  I don't quite follow the
> "\\0-" syntax.  It appears related to the group matching discussed on
> page 326 of pickaxe, but not quite.  Clearly in the "\\0-" sequence we
> have the first slash escaping the second one, so it is effectively \0
> plus the '-', in other words replace the four characters matched with
> the four characters plus a dash.  What I don't get is how \0 outside of
> the pattern is giving us anything useful.  Seems like it should be $1
> instead, since it is outside of the regular expression being matched.
> But then, there aren't any parenthesis, so there is no group, so perhaps
> $1 doesn't apply either.  Can someone offer some clarification?

This is as documented and an artifact of Ruby's history (i.e. a
Perlism), which in turn was a sedism, which was brought in from ed...

pth