Last "segment" was often missing.
The function doing escaping was sadly quadratic in the length of successive character that are converted to htmlentities (now it's linear). Avoid to construct intermediary list for block of non escaped and escaped chars Avoid to construct list of escaped chars Speep-up the most frequent path by a factor 3 (block of non escaped chars) CHANGELOG optimisation of html escaping (x3 speed and use less memory)
…utf-8 page Provide more readable html source and unlimited speed improvment on complex utf-8 character (e.g. chinese) CHANGELOG Keep multibyte characters "as is" during html escaping when the target resource is utf-8
CHANGELOG Fix url decoding on multibyte utf8 characters
…d positional update
…ed by syntactic integer)