On Thu, Sep 19, 2002 at 08:35:23PM +0300, Jarkko Hietaniemi wrote: > > I'll be happy to oblige and make a U::EAW with that, then. > If it is found to work fine, we can certainly merge that later back > into the core (maybe when Unicode 3.2.1 comes out).
Okay, I've put it together now. Bug uncovered, though. Line 88 in utf8_heavy.pl tests for user-definedness: if ($type =~ /^I[ns](\w+)$/) { But it couldn't contain 'Is' anyway, because line 53 removed it: $type =~ s/^Is(?:\s+|[-_])?//i So basically a user cannot define 'Is' properties on his/her own, because the canonized form does not match the specified $type: sub Foo { '0-0' } # this doesn't work sub IsFoo { '0-0' } # neither does this sub InFoo { '0-0' } print /\p{InFoo}/; # okay print /\p{IsFoo}/; # dies Which is kind of disappointing, because I'll be forced to use 'InEastAsianFullwidth' instead of the more natural 'IsEastAsianFullwidth' as the name. Thanks, /Autrijus/
msg01455/pgp00000.pgp
Description: PGP signature