Re: match-string or match-beginning incorrect

[Dave Love]

Stefan <address@hidden> writes:

>> The doc for `match-string' says it returns nil if the subexpression
>> didn't match.
>
> Right.  AFAIK that's indeed what it does.
> Can you explain more clearly in which situation it doesn't do that?

Now I find I can't...  I should have supplied an example, but it
seemed to be a general effect and I thought there had been a
deliberate change.  (I did try to make sure it wasn't anything
spurious from interactive commands or font-lock happening.)  Sorry for
the noise.

> Maybe you're thinking of the case where the whole regexp didn't
> match?

No.  (It arose when trying to use a match string for a display
property in a font-lock clause.)