Re: latin encodings, mail-lib, and posix regexp classes

Thanks for looking at this. Be aware that this stuff is not easy, and the code is crufty. Some references are the Dragon Book's (Aho, Sethi, Ullman _Compilers_) chapter on lexical analysis and Sedgewick's _Algorithms_ (get the C version). PB> i believe the perl regular expression engine supports PB> [:alpha:] constructs, maybe we could have a look at how they PB> do it to get ideas on how XEmacs can support those nifty PB> constructs I don't think this is terribly hard, just a little tedious. PB> i think implementing POSIX bracket-expression's in XEmacs will PB> enhance its multi-language capabilities Hard to say. The problem is that POSIX bracket-expressions depend on the POSIX locale, which we pretty much ignore elsewhere. That means we are likely to get weird interactions between people's expectations of how XEmacs "should" behave and what it does in regexps. PB> the other idea i had was wether pcre could be used in XEmacs PB> (pcre is a library that does perl compatible regular PB> expressions for those who dont know) No, at least not as a _replacement_ for the current stuff (backward compatibility, remember? lots and lots of Lisp that would have to be ported). Also, we have extensions (the \sX syntax matcher) and documented but unimplemented extensions (generic character categories using \cX) that are not in POSIX that we really should implement. What has been suggested in the past is (1) synch to GNU Emacs's regexp engine (Stefan Monnier did a _lot_ of work on it---he is extremely busy now but might be able to spare a clue or two; Jan Vroonhof, also very busy, has looked at this in the past) (2) implementing something like Python's raw-strings which would allow regexps to be written without all those doubled and quadrupled backslashes. -- Institute of Policy and Planning Sciences http://turnbull.sk.tsukuba.ac.jp University of Tsukuba Tennodai 1-1-1 Tsukuba 305-8573 JAPAN My nostalgia for Icon makes me forget about any of the bad things. I don't have much nostalgia for Perl, so its faults I remember. Scott Gilbert c.l.py