The documentation doesn't say the function gives warnings. Other string functions do not give warnings when $length=0. substr_count should not give warnings.

But the change is not documented - not in the substr_count docs (eg, in the Changelog) or in the 7.1 migration guide's "Changed functions" list.

To add to this, the reason is that substr_count() now also accepts a negative length parameter with the usual meaning. We could still explicitly exclude the case length=0, but, as @requinix mentions, it doesn't really make sense. We allow length=0 elsewhere and it is useful if $length is not a hardcoded value.

Do you think this is normal behavior, 0 function takes long and looking over the entire length? Here is an example:
https://3v4l.org/3B2Ct
In PHP 5.6 this behavior in the case of other

<https://3v4l.org/bAu27> looks like a bug in substr_count() to me.
Shouldn't the respective if condition[1] actually be `length < 0`?

[1] <https://github.com/php/php-src/blob/php-7.1.0RC6/ext/standard/string.c#L5220>

For the record: the actual bug has been fixed in the meantime (see
bug #74041).

This bug has been fixed in the documentation's XML sources. Since the
online and downloadable versions of the documentation need some time
to get updated, we would like to ask you to be a bit patient.

Thank you for the report, and for helping us make our documentation better.