jweirich / one.net wrote:
>
> >>>>> "Ben" == Ben Tilly <ben_tilly / hotmail.com> writes:
>
>     >> The "slow" method is O(n*n) as expected.  Both the "strscan" and 
>the
>     >> "scan" methods are O(n).
>
>     Ben> Both of the scan methods are incompatible with
>     Ben> context-sensitive tokenizing.  Guess what I want to do? :-)
>
>And StringScanner is incompatible because ... it only matches from the
>front of the string?

/me reads..
/me goes "d'oh"!
/me realizes that lack of sleep is a problem for him...

The scan method there is right.  I hadn't realized that
$' and $` were populated by default on every match using
index.  Of course in Perl I never use those because of
the performance hit.  But I guess that Ruby has no obvious
way around them.

Perhaps when I have more than a couple of days with Ruby
I will have a better sense.  (Or perhaps after I get some
sleep.)

Thanks,
Ben
_____________________________________________________________________________________
Get more from the Web.  FREE MSN Explorer download : http://explorer.msn.com