Re: patch for string library

["Asakawa, Kazuaki" <k_asakawa@...>]

> As Rici pointed out, zero-length matches are tricky. Conceptually, 
> there
> are infinite empty strings at any point of any string.  So, the 
> specific
> behavior of a given implementation seems more a device to avoid an
> infinite loop than actually a feature (or a bug). As such, programmers
> should not count on that behavior.

OK, I agree.
I withdraw the second patch.

--
asakawa