Reuben Thomas wrote: > I like the balance between having a documented "quote task" data > structure that can be constructed manually, while having constructor > functions for the POSIX, PCRE and GNU cases, so as to make casual use > risk-free.
Thanks for the review. > The only thing I find confusing is the "multibyte" flag, which is zero > for UTF-8 text (which is a multibyte format). Did you really mean > "wide character" rather than "multibyte"? I meant "multibyte character in the sense of the locale", that is, what mbtowc() and mblen() recognize. I've added a comment about it and pushed the result. Bruno -- In memoriam Amar Singh Chamkila <http://en.wikipedia.org/wiki/Amar_Singh_Chamkila>
