My regex-ordering module (Regexp-Compare) is based on a view of regex internals described in http://perl.plover.com/Rx/paper/, which is nice, but getting old and limited in what it describes. Specifically, it doesn't mention Unicode support, which apparently uses some flags to instruct the regex matcher to look outside the basic compiled expression someplace I can't find... Is there any documentation of Unicode-using compiled regex format? I've tried to read regexec.c couple of times, but never got very far...