First, my heart felt thanks for taking the issue seriously!
: I'll have to reread it to understand it more fully, but it sounds to me
: like you're looking for some kind of lookbehind assertion. Have you
: thought about the possibility of using a lookbehind assertion with a \p
: in it? A lookbehind feature was recently added.
The last table of the last page of the paper probably has the real meat of
what should be happening. I did try a number of permutations with look
behinds and \p (there should be a footnote on the last page for one such test).
A look behind should be sufficient here as I understand them through the man
pages (?<=\p{MyP}) seems straight forward enough but my intuition and
experience with the extended regex escapes is not well developed.
I have prepared now a mini testing suite for Ethiopic character classes in
Perl. The look behind test and others are found in the suite:
ftp://ftp.cs.indiana.edu/pub/fidel/perl-unicode/TestEthiopic.pl
But it's no fun without Ethiopic fonts. Linux users can download:
ftp://ftp.cs.indiana.edu/pub/fidel/perl-unicode/EthiopicDemo.tar.gz
there is a setup that will switch on an Ethiopic font in your console (since
Linux 2.0 you should have 9 Ethiopic fonts already, but the package includes
one plus the loader just in case). The file "TestEthiopic.pl" is included
and it is assumed that you have downloaded and install the class files:
ftp://ftp.cs.indiana.edu/pub/fidel/perl-unicode/perl-unicode.tar.gz
which were updated today as well.
:: The utf8 support feels like a new lease on life. I love it! Is there any
:: intention to extend it beyond regex sequences? I would like also to use
:: non-ASCII terms for string, array, and subroutine names.
: It should already work for that, as long as you've set up the properties
: to reflect which characters should be considered letters.
I have this working now :-)
thanks,
/Daniel
