Re: error in regex handling of \n?

On Saturday, August 18, 2012 4:14:40 PM UTC-5, Tim Chase wrote:
> On 08/17/12 15:59, Chris Jones wrote:
>
> > On Fri, Aug 17, 2012 at 02:01:53PM EDT, Tim Chase wrote:
>
> >> On 08/17/12 12:51, Chris Jones wrote:
>
> >
>
> >>> On Fri, Aug 17, 2012 at 11:09:21AM EDT, Tim Chase wrote:
>
> >>>> /r[^\n]
>
> >>>> /r\_[^\n]
>
> >>>>
>
> >>>> to no avail.
>
> >>>
>
> >>> /r[^\r]
>
> >>
>
> >> Well, bust my buttons. So "\n" is a newline on the search side, and
>
> >> "\r" is a newline on the replacement side, but it's also a newline
>
> >> on the search side if it's in a negated character class, but not in
>
> >> a non-negated character class of a search?
>
> >>
>
> >> Confusing.
>
> >
>
> > Vim fun..?
>
> >
>
> > Maybe it's just a matter of putting 2 & 2 together.. even (as Ben noted)
>
> > when they don't add up..?
>
> >
>
> > All the more interesting, since in keeping with my usual non-technical
>
> > approach.. it was a post of yours on these aspects some two months ago
>
> > that made me think of this course of action.. ;-)
>
>
>
> Sorry I can't claim I remember the post :-)
>
>
>
> So playing around with this, I've tried multiple cases and must say
>
> I'm somewhat confused by what happens:
>
>
>
> 1) /e$
>
> 2) /e[\n]
>
> 3) /e[\r]
>
> 4) /e\_[\n]
>
> 5) /e\_[\r]
>
> 6) /e[^\n]
>
> 7) /e[^\r]
>
> 8) /e\_[^\n]
>
> 9) /e\_[^\r]
>
>
>
> 1) finds "e" as the last character
>
> 2) works like #1
>
> 3) does not find "e" as the last character (but will find one if
>
> there's a stray ^M after an "e" in the file)
>
> 4) works like #1
>
> 5) works like #1, but also finds "e" followed by ^M
>
> 6) finds all "e" characters, including those at the EOL
>
> 7) finds all "e" characters except those at the EOL
>
> 8) same as #7
>
> 9) finds all "e" characters except those followed by ^M
>
>
>
> I guess it's the disparity between #2 (and #4) and #6 (and
>
> correspondingly, #8) that I find most disconcerting: that the
>
> negation ("^") of the set doesn't alter it from matching.
>
>
>
> -tim

As I mentioned, I think it has something to do with the behavior at :help /[\n] which says it acts just like /_[], and it goes on to say that /_[] includes the newline even if the class is negated by starting with ^. Really /[^\n] is pointless anyway because [] doesn't match the \n character already.

It's still surprising, and still probably wrong, but that help section might indicate WHY it's wrong.

--
You received this message from the "vim_use" maillist.
Do not top-post! Type your reply below the text you are replying to.
For more information, visit http://www.vim.org/maillist.php