Negative Lookahead Equivalent in emacs

Negative Lookahead Equivalent in emacs

[luishenriquezperez]

Hi,

I'm trying to write a regex that matches the last character of a sequence of non-whitespace characters '[^\n\r\t\f ]', or an empty line matching ^$. 

Thus: 
Hello World! --> "o" and "!" would be matched

In non-elisp regex languages I know the code for this is: \S(?!\S) 
I know that \S is equivalent too [^ /n/r/t/f]. 
But I'm unsure of what the elisp equivalent (if any) of the negative lookahead (?!).

I saw on this forum a post  "gnu.emacs.help › regex nirvana - near miss"
Where Drew Adams said: "Typically, what you want to do for this in Emacs Lisp is to combine 
the use of a regexp for positive matching with other code that takes 
care of the non-matching (negation) need. "

However, I'm not sure how to go about doing this.

Re: Negative Lookahead Equivalent in emacs

[Emanuel Berg]

> I'm trying to write a regex that matches the
> last character of a sequence of
> non-whitespace characters '[^\n\r\t\f ]', or
> an empty line matching ^$.
>
> Thus: Hello World! --> "o" and "!" would be
> matched

I would probably use syntax classes [1] for
this, with numbered groups, as in this Elisp.
Regexps are hairy tho so some fiddling and
fine-tuning is expected:

    (when (re-search-forward
           "\\([[:graph:]]\\)[ \n]" (point-max) t)
      (goto-char (match-beginning 0))
      (message (match-string-no-properties 1)) )

Hello World!

[1] https://www.emacswiki.org/emacs/RegularExpression

-- 
underground experts united
http://user.it.uu.se/~embe8573

Re: Negative Lookahead Equivalent in emacs

[hector]

On Mon, May 08, 2017 at 10:15:42PM -0700, luishenriquezperez@gmail.com wrote:
> Hi,
> 
> I'm trying to write a regex that matches the last character of a sequence of non-whitespace characters '[^\n\r\t\f ]', or an empty line matching ^$. 
> 
> Thus: 
> Hello World! --> "o" and "!" would be matched
> 
> In non-elisp regex languages I know the code for this is: \S(?!\S) 
> I know that \S is equivalent too [^ /n/r/t/f]. 
> But I'm unsure of what the elisp equivalent (if any) of the negative lookahead (?!).
> 
> I saw on this forum a post  "gnu.emacs.help › regex nirvana - near miss"
> Where Drew Adams said: "Typically, what you want to do for this in Emacs Lisp is to combine 
> the use of a regexp for positive matching with other code that takes 
> care of the non-matching (negation) need. "
> 
> However, I'm not sure how to go about doing this.

Perhaps

\\S-\\s-

or

[^[:space:]][[:space:]]

not including the ^$. I guess you just had to add "\\|^$" to the preceding expressions.
I'm not sure if there is any difference between the two syntaxes.

Note that these expressions depend on the syntax table. So the definition of
"[[:space:]]" can change depending on syntax table. It is not necessarily
"[\n\t\f\r\v ]"

Re: Negative Lookahead Equivalent in emacs

[Wasell]

On Mon, 8 May 2017 22:15:42 -0700 (PDT), in article <49e1dd7d-4be5-4b03-b9e2-
e26b15b0a6cb@googlegroups.com>, luishenriquezperez@gmail.com wrote:
> 
> Hi,
> 
> I'm trying to write a regex that matches the last character of a sequence of non-whitespace characters '[^\n\r\t\f ]', or an empty line matching ^$. 
> 
> Thus: 
> Hello World! --> "o" and "!" would be matched
> 
> In non-elisp regex languages I know the code for this is: \S(?!\S) 
> I know that \S is equivalent too [^ /n/r/t/f]. 
> But I'm unsure of what the elisp equivalent (if any) of the negative lookahead (?!).
> 
> I saw on this forum a post  "gnu.emacs.help ? regex nirvana - near miss"
> Where Drew Adams said: "Typically, what you want to do for this in Emacs Lisp is to combine 
> the use of a regexp for positive matching with other code that takes 
> care of the non-matching (negation) need. "
> 
> However, I'm not sure how to go about doing this.


If I'm not miss-understanding you completely, you want to match a non-
whitespace, followed either a whitespace or an end-of-line. That would be:

	[^[:space:]]\(?:[[:space:]]\|$\)

Am I missing something?

Re: Negative Lookahead Equivalent in emacs

[luishenriquezperez]

On Tuesday, May 9, 2017 at 5:54:29 AM UTC-4, Wasell wrote:
> On Mon, 8 May 2017 22:15:42 -0700 (PDT), in article <49e1dd7d-4be5-4b03-b9e2-
> e26b15b0a6cb@googlegroups.com>, luishenriquezperez@gmail.com wrote:
> > 
> > Hi,
> > 
> > I'm trying to write a regex that matches the last character of a sequence of non-whitespace characters '[^\n\r\t\f ]', or an empty line matching ^$. 
> > 
> > Thus: 
> > Hello World! --> "o" and "!" would be matched
> > 
> > In non-elisp regex languages I know the code for this is: \S(?!\S) 
> > I know that \S is equivalent too [^ /n/r/t/f]. 
> > But I'm unsure of what the elisp equivalent (if any) of the negative lookahead (?!).
> > 
> > I saw on this forum a post  "gnu.emacs.help ? regex nirvana - near miss"
> > Where Drew Adams said: "Typically, what you want to do for this in Emacs Lisp is to combine 
> > the use of a regexp for positive matching with other code that takes 
> > care of the non-matching (negation) need. "
> > 
> > However, I'm not sure how to go about doing this.
> 
> 
> If I'm not miss-understanding you completely, you want to match a non-
> whitespace, followed either a whitespace or an end-of-line. That would be:
> 
> 	[^[:space:]]\(?:[[:space:]]\|$\)
> 
> Am I missing something?

No you're not missing anything. This one is a lot simpler :). 
Thank you.

Re: Negative Lookahead Equivalent in emacs

[Felix Dietrich]

luishenriquezperez@gmail.com writes:

> I'm trying to write a regex that matches the last character of a
> sequence of non-whitespace characters '[^\n\r\t\f ]', or an empty line
> matching ^$.
>
> Thus: 
> Hello World! --> "o" and "!" would be matched
>
> In non-elisp regex languages I know the code for this is: \S(?!\S) 
> I know that \S is equivalent too [^ /n/r/t/f].

For your described behaviour something like the following might come
close to what you had in mind and does not require negative lookahead
support:

    \\(?:[^\n\r\t\f ]*\\([^\n\r\t\f ]\\)\\)\\|^\n

The first shy group "\(?:\)" matches zero or more non-whitespace
characters and one more non-whitespace character (everything following
must be a whitespace character); the one non-whitespace character will
be available as the first group of the resultant match.  Alternatively
"\|" it matches a line beginning with and (depending possibly on the
newline convention) therefore containing only a newline character,
i.e. an empty line.

Don't get confused by the plethora of backslashes: they require escaping
in an Emacs Lisp string to reach the regular expression functions as
proper backslashes; if they were not escaped they themself would escape
the following character.

> But I'm unsure of what the elisp equivalent (if any) of the negative
> lookahead (?!).

I do not know of the existence of an equivalent for the negative
lookahead feature of other regular expression engines in Emacs Lisp
regular expressions.

--
Felix Dietrich

Re: Negative Lookahead Equivalent in emacs

[luishenriquezperez]

On Tuesday, May 9, 2017 at 1:15:46 AM UTC-4, luishenri...@gmail.com wrote:
> Hi,
> 
> I'm trying to write a regex that matches the last character of a sequence of non-whitespace characters '[^\n\r\t\f ]', or an empty line matching ^$. 
> 
> Thus: 
> Hello World! --> "o" and "!" would be matched
> 
> In non-elisp regex languages I know the code for this is: \S(?!\S) 
> I know that \S is equivalent too [^ /n/r/t/f]. 
> But I'm unsure of what the elisp equivalent (if any) of the negative lookahead (?!).
> 
> I saw on this forum a post  "gnu.emacs.help › regex nirvana - near miss"
> Where Drew Adams said: "Typically, what you want to do for this in Emacs Lisp is to combine 
> the use of a regexp for positive matching with other code that takes 
> care of the non-matching (negation) need. "
> 
> However, I'm not sure how to go about doing this.

Thank you this worked out.

Re: Negative Lookahead Equivalent in emacs

[Emanuel Berg]

> Thank you this worked out.

How?

-- 
underground experts united
http://user.it.uu.se/~embe8573

Re: Negative Lookahead Equivalent in emacs

[luishenriquezperez]

On Tuesday, May 9, 2017 at 1:15:46 AM UTC-4, luishenri...@gmail.com wrote:
> Hi,
> 
> I'm trying to write a regex that matches the last character of a sequence of non-whitespace characters '[^\n\r\t\f ]', or an empty line matching ^$. 
> 
> Thus: 
> Hello World! --> "o" and "!" would be matched
> 
> In non-elisp regex languages I know the code for this is: \S(?!\S) 
> I know that \S is equivalent too [^ /n/r/t/f]. 
> But I'm unsure of what the elisp equivalent (if any) of the negative lookahead (?!).
> 
> I saw on this forum a post  "gnu.emacs.help › regex nirvana - near miss"
> Where Drew Adams said: "Typically, what you want to do for this in Emacs Lisp is to combine 
> the use of a regexp for positive matching with other code that takes 
> care of the non-matching (negation) need. "
> 
> However, I'm not sure how to go about doing this.

I edited the solution slightly so no new lines are included.

\[^[:space:]\n\]\\(?:\[[:space:]\]\\|$\\)

There are a lot of backslashes because I used regex-builder to test it.