diff options
| author | Karl Williamson <khw@cpan.org> | 2017-12-23 12:39:16 -0700 |
|---|---|---|
| committer | Karl Williamson <khw@cpan.org> | 2017-12-24 17:12:03 -0700 |
| commit | 226b74db842dc6466710b278cae98a80b0f77458 (patch) | |
| tree | e10b233b09d56bbebc11e8fe39ced47499f9e253 /charclass_invlists.h | |
| parent | 39c4defe660921fe7b772ebc0f18b2473a6c5a4c (diff) | |
| download | perl-226b74db842dc6466710b278cae98a80b0f77458.tar.gz | |
regen/mk_invlists.pl: Simplify handling of early Unicode releases
Several Unicode properties are used internally in regex matching, for
example \b{wb} and \X. The code is supposed to compile and work as best
it can on all Unicode releases. But the rules for these have been
expanded by Unicode over the years, including new types of boundary
conditions. Each such type is given an enum value in an enum type for
the given boundary. For early Unicodes, this enum isn't present.
This commit simplifies how this generating code copes with this,
removing the hard-coded lists in favor of using a different mechanism
already in place for handling slightly different circumstances.
Diffstat (limited to 'charclass_invlists.h')
| -rw-r--r-- | charclass_invlists.h | 2 |
1 files changed, 1 insertions, 1 deletions
diff --git a/charclass_invlists.h b/charclass_invlists.h index a60a94fa9b..1d20548e95 100644 --- a/charclass_invlists.h +++ b/charclass_invlists.h @@ -97457,5 +97457,5 @@ static const U8 WB_table[24][24] = { * 4d296f9555c2df469eb9629c7cd53b3bf5d10331baa88087d8451966f1563865 lib/unicore/mktables * 21653d2744fdd071f9ef138c805393901bb9547cf3e777ebf50215a191f986ea lib/unicore/version * 913d2f93f3cb6cdf1664db888bf840bc4eb074eef824e082fceda24a9445e60c regen/charset_translations.pl - * 11f3bbdf78d86d8fe9c353130c57217c722260e17c1a67b15c38eaebf2b7da6b regen/mk_invlists.pl + * 5aef86190c343205c29243efaacec3fb208bdecea8f8eb69f8027d021c389499 regen/mk_invlists.pl * ex: set ro: */ |