Subject:
|
Re: Lost Instructions
|
Newsgroups:
|
lugnet.admin.general
|
Date:
|
Thu, 13 Apr 2000 17:02:22 GMT
|
Viewed:
|
798 times
|
| |
|
|
In lugnet.admin.general, Dan Boger writes:
> In lugnet.admin.general, Todd Lehman writes:
> > So something like
> >
> > m/^Re(?:\[\d+\]):\s*/i
> >
> > might be even safer, since this would be for broken clients in the first
> > place, and cc:Mail is-was about as broken as one can get. :-)
>
> just for efficiancy's sake, you shouldn't put a /s* at the end - it's not
> doing anything at all, is it? if there's anything that matches the /s class,
> it'll just take longer to complete the match, but if there's no /s at the
> end, it'll still match just the same...
True -- the \s* isn't doing anything in that case. The regex could be
considered a shorter form of
($subject_base) = ($subject_raw =~ m/^(Re(?:\[\d+\]):\s*)?(.*)$/i);
in which the \s* is actually needed. I left it on in the example way above
for the same reason that /i is there instead of an explicit [Rr][Ee] --
because it's likely the difference beween one gazillionth of a second and one
bazillionth of a second. (/i is slower than explicit character classes, but
it wouldn't make a practical speed difference here.) The trailing \s* is
just a reminder that there has to be zero or more spaces (rather than one or
more) following the "Re:" in order to recognize it as a "Re:".
--Todd
|
|
Message is in Reply To:
 | | Re: Lost Instructions
|
| (...) just for efficiancy's sake, you shouldn't put a /s* at the end - it's not doing anything at all, is it? if there's anything that matches the /s class, it'll just take longer to complete the match, but if there's no /s at the end, it'll still (...) (25 years ago, 13-Apr-00, to lugnet.admin.general)
|
10 Messages in This Thread:
  
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
