regex, multibyte locales, and word boundaries

classic Classic list List threaded Threaded
1 message Options
Reply | Threaded
Open this post in threaded view
|

regex, multibyte locales, and word boundaries

yuripv
Hi,

We have the following note in the BUGS section of regcomp(3):

----------------------------------------------------------------------
Word-boundary matching does not work properly in multibyte locales.
----------------------------------------------------------------------

It was added ages ago along with multibyte support in our regex
implementation, though I can't think of any positive test case to see
that the problem is real, and eventually fix it.

I'm wondering if anyone has real life examples showing the bug?


signature.asc (499 bytes) Download Attachment