Revision history for Perl module Unicode::Collate.
0.33 Sat Dec 13 14:07:27 2003
- documentation improvement: in "entry", "overrideHangul", etc.
0.32 Wed Dec 3 23:38:18 2003
- A matching part from index(), match() etc. will include illegal
code points (as well as ignorable characters) following a grapheme.
- Contraction with illegal code point will be invalid.
- Added some tests in illegal.t; added view.t.
- Some tests are separated from test.t into altern.t and rearrang.t.
- modified XSUB internals.
0.31 Sun Nov 16 15:40:15 2003
- Illegal code points (surrogate and noncharacter; they are definitely
ignorable) will be distinguished from NULL ("\0");
but porting is not successful in the case of ((Pure Perl) and
(Perl 5.7.3 or before)). If perl 5.6.X is used, XSUB may help it
in place of broken CORE::unpack('U*') in older perl.
- added illegal.t and illegalp.t.
- added XSUB (EXPERIMENTAL!) where some functions are implemented
in XSUB. Pure Perl is also supported.
0.30 Mon Oct 13 21:26:37 2003
- fix: Completely ignorable in table should be able to be overrided
by non-ignorable in entry.
- fix: Maximum length for contraction must not be shortened
by a shorter contraction following in table and/or entry.
- added normal.t.
- some doc fixes
0.29 Mon Oct 13 12:18:23 2003
- now UCA Version 11 (but no functionality is different from Version 9).
- supported hangul_terminator.
- fix: Base_Unicode_Version falsely returns Perl's Unicode version.
C4 in UTS #10 requires UTS's Unicode version.
- For variable weighting, 'variable' is recommended
and 'alternate' is deprecated.
- added version() method.
- added hangtype.t, trailwt.t, variable.t, and version.t.
0.28 Sat Sep 06 20:16:01 2003
- Fixed another inconsistency under (normalization => undef):
Non-contiguous contraction is always neglected.
- Fixed: according to S2.1 in UTS #10, a blocked combining character
should not be contracted. One test in test.t was wrong, then removed.
- Added contract.t.
- (normalization => "prenormalized") is able to be used.
0.27 Sun Aug 31 22:23:17 2003
some improvements:
- The maximum length of contracted CE was not checked (v0.22 to v0.26).
Collation of a large string including a first letter of a contraction
that is not a part of that contraction (say, 'c' of 'ca'
where 'ch' is defined) was too slow, inefficient.
- A form name for 'normalization', no longer restricted to
/^(?:NF)?K?[CD]\z/, will be allowed as long as
Unicode::Normalize::normalize() accepts it, since Unicode::Normalize
or UAX #15 may be changed/enhanced in future.
- When Hangul syllables are decomposed under <normalization => undef>,
contraction among jamo (LV, VT, LVT) derived from the same
Hangul syllable is allowed. Added hangul.t.
0.26 Sun Aug 03 22:23:17 2003
- fix: an expansion in which a CE is level 3 ignorable and others are not
was wrongly made level 3 ignorable as a whole entry.
(In DUCET, some precomposites in Musical Symbols are so)
0.25 Mon Jun 06 23:20:17 2003
- fix Makefile.PL.
- internal tweak (again): pack_U() and unpack_U().
0.24 Thu Apr 02 23:12:54 2003
- internal tweak for (?un)pack 'U'.
0.23 Wed Sep 04 19:25:20 2002
- fix: scalar match() no longer returns an lvalue substr ref.
- fix: "Ignorable after variable" should be made level 3 ignorable
even if alternate => 'blanked'.
- Now a grapheme may contain trailing level 2, level 3,
and completely ignorable characters.
0.22 Mon Sep 02 23:15:14 2002
- New File: index.t.
(The new test.t excludes tests for index.)
- tweak on index(). POSITION is supported.
- add match, gmatch, subst, gsubst methods.
- fix: ignorable after variable in 'shift'-variable weight.
0.21 Sat Aug 03 10:24:00 2002
- upgrade keys.txt and t/test.t for UCA Version 9.
0.20 Fri Jul 26 02:15:25 2002
- now UCA Version 9.
- U+FDD0..U+FDEF are new non-characters.
- fix: whitespace characters before @backwards etc. in a table file.
- now values for 'alternate', 'backwards', etc.,
which are explicitly specified via new(),
are preferred to those specified in a table file.
0.12 Sun May 05 09:43:10 2002
- add new methods, ->UCA_Version and ->Base_Unicode_Version.
- test fix: removed the needless requirement of Unicode::Normalize.
[reported by David Hand]
0.11 Fri May 03 02:28:10 2002
- fix: now derived collation elements can be used for Hangul Jamo
when their weights are not defined.
[reported by Andreas J. Koenig]
- fix: rearrangements had not worked.
- mentioned pleblem on index() in BUGS.
- more documents, more tests.
- tag names for 'alternate' are case-insensitive (i.e. 'SHIFTed' etc.).
- The <undef> value for the keys "overrideCJK", "overrideHangul",
"rearrange" has a special behavior (different from default).
0.10 Tue Dec 11 23:26:42 2001
- now you are allowed to use no table file.
- fix: fetching CE with two or more combining characters.
0.09 Sun Nov 11 17:02:40:18 2001
- add the following methods: eq, ne, lt, le, gt, le.
- relies on &Unicode::Normalize::getCombinClass()
in place of %Unicode::Normalize::Combin
(the hash is not defined in the XS version of Unicode::Normalize).
then you should install Unicode::Normalize 0.10 or later.
- now independent of Lingua::KO::Hangul::Util
(this module does decomposition of Hangul syllables for itself)
0.08 Mon Aug 20 22:40:18 2001
- add the index method.
0.07 Thu Aug 16 23:42:02 2001
- rename the module name to Unicode::Collate.
0.06 Thu Aug 16 23:18:36 2001
- add description of the getSortKey method.
0.05 Mon Aug 13 22:23:11 2001
- bug fix: on the things of 4.2.1, UTR #10
- getSortKey returns a string, but not an arrayref.
0.04 Mon Aug 13 22:23:11 2001
- some bugs are fixed.
- some tailoring parameters are added.
0.03 Mon Aug 06 06:26:35 2001
- modify README
0.02 Sun Aug 05 20:20:01 2001
- some fix
0.01 Sun Jul 29 16:16:15 2001
- original version; created by h2xs 1.21
with options -A -X -n Sort::UCA