Re: string-match: \` and \= do not match when START is non-zero

Re: string-match: \` and \= do not match when START is non-zero

[Kevin Rodgers]

I wrote:

> \` or \= should match the empty string at index START in STRING.
> 
> (string-match "\\`bar\\'" "foobar" 3) and (string-match "\\=bar\\'" 
> "foobar" 3)
> return nil.  I think at least one of them should succeed and return 3, 
> just like
> (string-match "bar\\'" "foobar" 3) does.  Otherwise, how can you anchor the
> REGEXP at index START in STRING?
> 
> [time passes]
> 
> (equal (string-match REGEXP STRING START) START)
> 
> [never mind]

Since START can be negative, that should be:

(equal (string-match REGEXP STRING START)
        (if (minusp START)
	   (+ (length STRING) START)
	 START))

I'm studying regex.c to see if I can implement \= matching in strings...

-- 
<a href="mailto:&lt;kevin.rodgers&#64;ihs.com&gt;">Kevin Rodgers</a>

string-match: \` and \= do not match when START is non-zero

[Kevin Rodgers]

\` or \= should match the empty string at index START in STRING.

(string-match "\\`bar\\'" "foobar" 3) and (string-match "\\=bar\\'" "foobar" 3)
return nil.  I think at least one of them should succeed and return 3, just like
(string-match "bar\\'" "foobar" 3) does.  Otherwise, how can you anchor the
REGEXP at index START in STRING?

[time passes]

(equal (string-match REGEXP STRING START) START)

[never mind]
-- 
<a href="mailto:&lt;kevin.rodgers&#64;ihs.com&gt;">Kevin Rodgers</a>

Re: string-match: \` and \= do not match when START is non-zero

[Richard Stallman]

    \` or \= should match the empty string at index START in STRING.


\' certainly should only match at the beginning of the string,
just as it only matches at the beginning of the buffer.

For \= I am not sure.  I am not sure whether the START value in
string-match is really analogous to point in a buffer.  On the other hand,
currently \= is meaningless in a string, and it might do no harm to
give it this meaning.

I won't try to implement it myself.