
Yes, but because they could previously be individually selected, I gave the users of my application the option of setting bk_parens and/or bk_braces, separately from choosing POSIX or Perl semantics. It seems that syntax and semantics are no longer orthogonal. I don't have strong feelings about that, but just wanted to check that it was intentional, before hacking the application.
OK understood, but if you use Basic-style syntax with Perl like extensions you can get into trouble particularly if you try and match non-greedy repeats POSIX style. So, I guess what I'm saying is it avoid a whole lot of confusion all round if you stick to one of the standard syntaxes or their variations: basic, extended, perl, emacs etc. My guess is your users would find it less confusing as well - most people (all of them?) would want regexes that follow whatever form they've used before, rather than inventing their own new syntax. You can still choose whether to use Perl-style matching or POSIX-style leftmost longest matching independently of the syntax type, but I'd recomend that you don't, or at least that you don't try and use Perl syntax with leftmost-longest rules. The other way around however: POSIX syntax with Perl matching rules is OK (and is more or less what emacs does). John.