1 2010-04-20 Damien Diederen <dd@crosstwine.com>
3 * Normalization.cs: Really apply canonical reordering "recursively."
5 Before this, a sequence of code points with the combining
6 classes (22, 33, 11) would be reordered to (22, 11, 33) instead of
7 the correct (11, 22, 33). This is because the 'i--' would be
8 directly cancelled by the 'i++' in the for loop.
10 2010-04-20 Damien Diederen <dd@crosstwine.com>
12 * Normalization.cs: The correct "checkType" argument to
13 Decompose() is NKD or NKFD when normalizing to NKC resp. NKFC.
15 * StringTest.cs: More NFC test cases.
17 2010-04-20 Damien Diederen <dd@crosstwine.com>
19 * Normalization.cs: Implement algorithmic Hangul composition.
20 Calling Normalize(NormalizationForm.FormC) on Korean characters
21 now works properly (bnc#480152).
23 * StringTest.cs: Add test cases for Hangul composition.
25 2010-04-20 Damien Diederen <dd@crosstwine.com>
27 * Normalization.cs: Follow the spec when checking composition pairs.
29 Figure 7 in section 1.3 of http://unicode.org/reports/tr15/ shows
30 how when doing composition, one has to examine the successive
31 (starter, candidate) pairs, and combine if a matching canonical
34 The original algorithm was, instead, iterating on canonical
35 decompositions, and, for each one, trying to match a sequence
36 of (starter, non-starter, ...). This, however, does not produce
37 the same results as it is violating some implicit ordering
38 constraints in the Unicode tables.
40 E.g., when composing the following sequence of codepoints, the
41 original algorithm was picking:
49 and would stop at 1FC2 0313 as there is no decomposition matching
50 it. The new algorithm, which follows the guidance of the pretty
51 figure 7, ends up doing:
61 resulting in the correct 1F92.
63 2010-04-19 Damien Diederen <dd@crosstwine.com>
65 * Normalization.cs: Recursively apply the Unicode decomposition mapping.
67 According to http://www.unicode.org/reports/tr15/tr15-31.html,
70 "To transform a Unicode string into a given Unicode Normalization
71 Form, the first step is to fully decompose the string. [...] Full
72 decomposition involves recursive application of the
73 Decomposition_Mapping values, because in some cases a complex
74 composite character may have a Decomposition_Mapping into a
75 sequence of characters, one of which may also have its own
76 non-trivial Decomposition_Mapping value."
78 2010-02-18 Gabriel Burt <gabriel.burt@gmail.com>
80 * Normalization.cs: Implement algorithmic Hangul decomposition; Calling
81 string.Normalize on Korean characters now works properly (bnc#480152).
82 This reduces the number of errors in 'make test' from 27k to 4.8k.
84 * StringNormalizationTestSource.cs:
85 * Makefile: Use the local, working copy of Normalization etc,so as to make
86 modifying Normalization.cs and then testing your changes with 'make test'
87 possible. Also, fix building/running of tests, patch by Alexander
90 2009-09-18 Atsushi Enomoto <atsushi@ximian.com>
92 * Normalization.cs : Handle blocked characters which are not
93 immediately next to the primary composite character. This fixes
94 some Arabic string sequence normalization.
95 * Makefile : fix test build.
97 2009-09-17 Atsushi Enomoto <atsushi@ximian.com>
99 * Normalization.cs : some renaming for disambiguation.
100 * NormalizationTableUtil.cs : fix some wrong ranges in
101 mapIdxToComposite. This fixes some Arabic normalization (and more).
102 * normalization-notes.txt : added some notes on the implementation.
104 2008-06-19 Atsushi Enomoto <atsushi@ximian.com>
107 - reverted the previous index calculation change. It was correctly
108 implemented and I rather broke it.
109 - fix index calculation on combining.
110 - NFKD was incorrectly directed to combining path. It should not.
111 - Simplify quick check.
113 2008-06-15 Atsushi Enomoto <atsushi@ximian.com>
115 * Normalization.cs : For NFC and NFKC, IsNormalized() was not working
116 enough to check composed characters. It's not possible without
117 the actual composition, so just call Normalize() and compare them.
118 In Normalize() mapping helper didn't pick correct map index since
119 the table for index stores index for "uncompressed" numbers.
120 * NormalizationTableUtil.cs : updated to the latest UCD.
121 * Makefile : to build test, source file must be downloaded too.
123 2008-11-05 Atsushi Enomoto <atsushi@ximian.com>
125 * ucd.cs : Write type for *_count. Add notice to not edit
126 unicode-data.h directly.
128 2008-11-04 Atsushi Enomoto <atsushi@ximian.com>
130 * ucd.cs : new code to generate unicode table for eglib.
132 2008-07-04 Andreas Nahr <ClassDevelopment@A-SoftTech.com>
134 * SortKey: Fix parameter names, add attribute, small formatting
136 2008-06-27 Rodrigo Kumpera <rkumpera@novell.com>
138 * CodePointIndexer.cs : Make TableRange a struct instead
139 of a class so we save 2 memory ops per ToIndex loop.
141 2008-04-02 Atsushi Enomoto <atsushi@ximian.com>
143 * SortKey.cs : check null arguments. Fixed bug #376171.
145 2007-07-20 Atsushi Enomoto <atsushi@ximian.com>
147 * create-mscompat-collation-table.cs : I wonder how long its build
150 2007-03-06 Atsushi Enomoto <atsushi@ximian.com>
152 * SimpleCollator.cs : disable QuickCheckPossible(), which is
153 inaccurate and inefficient. Fixed bug #79714.
155 2007-02-15 Atsushi Enomoto <atsushi@ximian.com>
157 * SimpleCollator.cs : character filtering is needed for
158 OrdinalIgnoreCase in 2.0 profile. Fixed bug #80865.
160 2007-01-25 Atsushi Enomoto <atsushi@ximian.com>
162 * SimpleCollator.cs : GetTailContraction() was broken to pick correct
163 contraction/special sortkey out and thus LastIndexOf() failed when
164 it is involved. Fixed bug #80612.
166 2007-01-22 Atsushi Enomoto <atsushi@ximian.com>
168 * SimpleCollator.cs : for non-StringSort comparison, level5 (- and ')
169 should be still skipped after initial level5 check is done (while
170 they were simply treated as a normal character). Fixed bug #78748.
171 * SortKeyBuffer.cs : Fixed NRE in french sort.
173 2006-12-25 Atsushi Enomoto <atsushi@ximian.com>
175 * SimpleCollator.cs : added IndexOf() implementation for Ordinal
176 and OrdinalIgnoreCase, though Ordinal version is not used (since
177 it is slower than icall).
179 2006-05-30 Miguel de Icaza <miguel@novell.com>
181 * MSCompatUnicodeTable.cs: Remove the fixed loading and compute it
182 just when we actually consume it. This only fixes the
185 2006-04-14 Atsushi Enomoto <atsushi@ximian.com>
187 * README: removed obsolete info.
188 * Normalization.cs : canonical reordering should participate in the
189 decomposition step. In reordering, string append was incomplete.
190 Combining class check is required in NFD check. Icall is written
193 2005-12-07 Zoltan Varga <vargaz@gmail.com>
195 * SimpleCollator.cs: Fix a warning.
197 2005-11-30 Sebastien Pouliot <sebastien@ximian.com>
199 * SimpleCollator.cs: Fix CAS support. The static ctor/var try to get
200 the environment variable MUCH too soon (i.e. the security manager
203 2005-11-29 Atsushi Enomoto <atsushi@ximian.com>
205 * SimpleCollator.cs : direct fast-path optimization for IndexOf().
207 2005-11-29 Atsushi Enomoto <atsushi@ximian.com>
209 * SimpleCollator.cs :
210 - CompareQuick(): added immediateBreakup to avoid extraneous sortkey
212 - QuickCheckPossible(): index used for s1 was incorrect.
214 2005-11-29 Atsushi Enomoto <atsushi@ximian.com>
216 * SimpleCollator.cs : added another quick check for CompareInternal()
217 that does almost ordinal comparison for quick-checkable strings.
218 (It affects on Compare(), IndexOf(), IsSuffix() etc. as well.)
220 2005-11-14 Atsushi Enomoto <atsushi@ximian.com>
222 * MSCompatUnicodeTable.cs : (IsIgnorable) \0 is not ignorable.
225 2005-11-14 Atsushi Enomoto <atsushi@ximian.com>
227 * SimpleCollator.cs :
228 Created another struct to reduce method arguments. Created another
229 flags that keeps "once-matched" state (counterpart of
230 checkedFlags, now neverMatchFlags).
232 2005-11-14 Atsushi Enomoto <atsushi@ximian.com>
234 * SimpleCollator.cs :
235 - Added CompareOrdinalIgnoreCase() for NET_2_0 RTM.
236 - Reduced extra parameter from LastIndexOfSortKey().
237 - LastIndexOf() should use GetTailContraction for the source string.
238 And then, target could match in the middle of the possible
239 "replacement contraction" of the source string, so use
240 LastIndexOfSortKey() to catch them.
241 - Fixed GetTailContraction() that caused index out of range.
243 2005-11-11 Atsushi Enomoto <atsushi@ximian.com>
245 * Makefile : Now use MONO_DISABLE_MANAGED_COLLATION.
246 * SortKey.cs : some members are virtual.
248 2005-10-14 Atsushi Enomoto <atsushi@ximian.com>
250 * SimpleCollator.cs : modified to use stackalloc for byte array.
252 2005-09-27 Atsushi Enomoto <atsushi@ximian.com>
254 * SimpleCollator.cs : in CompareInternal(), there was a possibility of
255 infinite loop. Fixed bug #76243.
257 2005-09-20 Atsushi Enomoto <atsushi@ximian.com>
259 * SimpleCollator.cs : In IsPrefix/IsSuffix, if target is an empty string,
260 immediately return true.
262 2005-09-09 Atsushi Enomoto <atsushi@ximian.com>
264 * SimpleCollator.cs : IsSuffix() optimization logic was buggy, so just
265 use pretty simple way with LastIndexOf() (no significant perf.
268 2005-09-01 Atsushi Enomoto <atsushi@ximian.com>
270 * README, Collation-notes.txt, CollationDataStructures.txt :
271 removing obsolete info and some added some notes.
273 2005-08-10 Atsushi Enomoto <atsushi@ximian.com>
275 * Normalization.cs : remove warned code.
276 * managed-collation.patch : now it's not required anymore.
278 2005-08-10 Atsushi Enomoto <atsushi@ximian.com>
280 * MSCompatUnicodeTable.cs : added IsSortable(string).
282 2005-08-10 Atsushi Enomoto <atsushi@ximian.com>
284 * SimpleCollator.cs : Now all collator methods are thread safe.
286 All instance non-readonly fields turned into arguments of every
287 methods that use those fields.
288 (Sadly it is the end of no-memory-cost collator era. mcs bootstrap
289 now needs +100KB memory consumption.)
291 2005-08-09 Atsushi Enomoto <atsushi@ximian.com>
293 * SimpleCollator.cs : made "checkedFlags" as nullable and made it as
294 an argument of every index methods (to make it thread safe).
296 2005-08-09 Atsushi Enomoto <atsushi@ximian.com>
299 MSCompatUnicodeTable.cs :
300 - Now IsIgnorable() is aggregated to be one invokation to check
301 completely ignorable, nonspacing and symbols.
302 - Introduced "already checked" flags for IndexOf() and LastIndexOf()
303 to skip sortkey binary check on the same characters. Significant
304 perf. improvement for such case as IndexOf("AABCBABC...Z",'Z').
306 2005-08-08 Gert Driesen <drieseng@users.sourceforge.net>
308 * SortKey.cs: Marked Serializable to match MS.NET.
310 2005-08-08 Atsushi Enomoto <atsushi@ximian.com>
312 * create-mscompat-collation-table.cs,
313 Makefile : changed resources output directory.
315 2005-08-04 Atsushi Enomoto <atsushi@ximian.com>
317 * create-normalization-tests.cs,
318 StringNormalizationTestSource.cs : new files for Unicode
319 Normalization test generator.
320 * Makefile : added support for above.
322 2005-08-03 Atsushi Enomoto <atsushi@ximian.com>
324 * NormalizationTableUtil.cs : oops, it does not compile.
325 * managed-collation.patch : I guess having managed resource would be
326 better for collation. At least current code has such #define so
327 Makefile should be in sync with it.
329 2005-08-03 Atsushi Enomoto <atsushi@ximian.com>
331 * create-normalization-source.cs : Fixed CharMapComparer which
332 incorrectly returned 0 when the second arg is shorter. Reduced
333 extraneous helperIndex map. Other minor fixes and code removal.
334 * Normalization.cs : several fixes to support blocked combine handling.
335 * NormalizationTableUtil.cs : tiny member renaming.
337 2005-08-03 Atsushi Enomoto <atsushi@ximian.com>
339 * create-normalization-source.cs,
340 NormalizationTableUtil.cs,
341 Normalization.cs : several bugfixes on index miscomputation.
342 Renamed using aliases (csc will bork). Primary combine safety is now
343 computed during UnicodeData.txt parse.
344 Maximum NFKD length was 18, not 4 (U+FDFA).
346 2005-08-02 Atsushi Enomoto <atsushi@ximian.com>
348 * managed-collation.patch : added Normalization support.
349 * managed-collation-icall.patch : added, including normalization stuff.
351 BTW when will collation code checked in?
353 2005-08-02 Atsushi Enomoto <atsushi@ximian.com>
355 * create-normalization-source.cs : Unified three normalization source
356 generators, to compute IsUnsafe flag. Fixed helperIndex array type
358 * create-char-mapping-source.cs,
359 create-combining-class-source.cs : thus removed.
360 * Makefile : thus modified for the above integration.
361 * NormalizationTableUtil.cs : Extended to contain IsUnsafe flag.
362 * Normalization.cs : Several fixes to make Normalize() actually work.
364 2005-07-29 Atsushi Enomoto <atsushi@ximian.com>
366 * create-normalization-source.cs,
368 create-char-mapping-source.cs,
369 create-combining-class-source.cs,
370 Makefile : converted managed array to pointers (like collation stuff).
372 2005-07-29 Atsushi Enomoto <atsushi@ximian.com>
374 * NormalizationTableUtil.cs : further table range optimization.
375 * create-normalization-source.cs,
376 create-char-mapping-source.cs,
377 create-combining-class-source.cs : added C header output support.
379 2005-07-29 Atsushi Enomoto <atsushi@ximian.com>
381 * create-normalization-source.cs, Normalization.cs :
382 Now property size is < 256, so directly embed value in "props" array.
383 Add QuickCheck(c,checkType) and remove IsNFD/C/KD/KC and delegates.
385 2005-07-29 Atsushi Enomoto <atsushi@ximian.com>
387 * create-combining-class-source.cs,
388 create-char-mapping-source.cs,
389 create-normalization-source.cs,
390 NormalizationTableUtil.cs,
391 Normalization.cs : String.Normalize() does not handle surrogate
392 characters. mapping information in DerivedNormalizationProps.txt
393 are not used in the code (those from UnicodeData.txt is used).
394 Hangul syllables are computed instead of embedded in the tables.
395 * managed-collation.patch : removed IntPtrStream and Makefile patches.
397 2005-07-29 Atsushi Enomoto <atsushi@ximian.com>
399 * MSCompatUnicodeTable.cs : IsSortable() was broken.
401 2005-07-29 Atsushi Enomoto <atsushi@ximian.com>
403 * MSCompatUnicodeTable.cs : added helper for CompareInfo.IsSortable().
405 2005-07-28 Atsushi Enomoto <atsushi@ximian.com>
407 * create-tailoring.cfg : added for convenience of contraction check.
409 2005-07-28 Atsushi Enomoto <atsushi@ximian.com>
411 * create-normalization-source.cs,
414 create-mscompat-collation-table.cs,
415 MSCompatUnicodeTableUtil.cs,
417 create-collation-element-table.cs,
418 MSCompatUnicodeTable.cs,
420 create-combining-class-source.cs : added copyright lines.
422 2005-07-28 Atsushi Enomoto <atsushi@ximian.com>
424 MSCompatUnicodeTable.cs : removed extraneous definition.
426 2005-07-28 Atsushi Enomoto <atsushi@ximian.com>
428 * create-mscompat-collation-table.cs
429 MSCompatUnicodeTable.cs : full C header support, finally.
431 2005-07-28 Atsushi Enomoto <atsushi@ximian.com>
434 NormalizationTableUtil.cs,
435 create-char-mapping-source.cs : more aggressive data compression.
436 It now ignores characters that are >= U+10000.
438 2005-07-28 Atsushi Enomoto <atsushi@ximian.com>
441 Normalization.template,
442 Normalization.cs : renamed existing file.
444 2005-07-28 Atsushi Enomoto <atsushi@ximian.com>
446 * NormalizationTableUtil.cs,
447 Normalization.template,
448 create-combining-class-source.cs : GetCombiningClass is now
449 implemented as indexer based array.
450 * Makefile : renamed output filename.
451 * create-mscompat-collation-table.cs : removed comments that does not
453 * create-tailoring.cs : use utf-8 output (and fixed filename).
455 2005-07-27 Atsushi Enomoto <atsushi@ximian.com>
457 * create-mscompat-collation-table.cs : hacked safer IPA extensions.
458 * Collation-notes.txt : status of sortkey table.
460 2005-07-27 Atsushi Enomoto <atsushi@ximian.com>
462 * create-mscompat-collation-table.cs : some Greek mapping fix.
464 2005-07-27 Atsushi Enomoto <atsushi@ximian.com>
466 * create-mscompat-collation-table.cs : diacritical weight is not
467 treated correctly when they are picked from letter names, as flags.
469 2005-07-27 Atsushi Enomoto <atsushi@ximian.com>
471 * create-mscompat-collation-table.cs : fixed culture-dependent
472 nonspacing mark weight.
474 2005-07-27 Atsushi Enomoto <atsushi@ximian.com>
476 * create-mscompat-collation-table.cs : some Hebrew case letter fixes.
477 Some diacritical fixes on symbols.
479 2005-07-27 Atsushi Enomoto <atsushi@ximian.com>
481 * create-mscompat-collation-table.cs : Fixed level 3 weight of
482 Arabic presentation forms.
484 2005-07-27 Atsushi Enomoto <atsushi@ximian.com>
486 * create-mscompat-collation-table.cs : Fixed some diacritical weight
487 of Arabic presentation forms.
489 2005-07-27 Atsushi Enomoto <atsushi@ximian.com>
491 * SimpleCollator.cs : more status updates. It's almost complete,
492 except for sortkey values.
494 2005-07-27 Atsushi Enomoto <atsushi@ximian.com>
496 * SimpleCollator.cs : similar optimization also for LastIndexOf().
498 2005-07-27 Atsushi Enomoto <atsushi@ximian.com>
500 * SimpleCollator.cs : the previous patch was missing IgnoreNonSpace
503 2005-07-27 Atsushi Enomoto <atsushi@ximian.com>
505 * SimpleCollator.cs : reduced extra sortkey value computation in
506 MatchesForward(). It makes IndexOf() roughly 30% faster.
508 2005-07-26 Atsushi Enomoto <atsushi@ximian.com>
510 * SortKey.cs : GetHashCode() returns a value based on its byte data.
513 2005-07-26 Atsushi Enomoto <atsushi@ximian.com>
515 * SimpleCollator.cs : consider extractions in invariant culture.
517 2005-07-26 Atsushi Enomoto <atsushi@ximian.com>
519 * SimpleCollator.cs : (unsafeFlags) be compact ;-)
521 2005-07-26 Atsushi Enomoto <atsushi@ximian.com>
523 * SimpleCollator.cs : When the tail of the target does not match more
524 than 3 times, then IsSuffix() will never be true (3 is the max
525 length of an expansion; \uFB03 -> ffi). It brings significant
526 performance boost when "source" string is very long.
527 * MSCompatUnicodeTable.cs : added MaxExpansionLength constant.
528 Reordered code lines.
530 2005-07-26 Atsushi Enomoto <atsushi@ximian.com>
532 * Collation-notes.txt : updated implementation status.
534 2005-07-26 Atsushi Enomoto <atsushi@ximian.com>
536 * SimpleCollator.cs : Implemented quick codepoint comparison in
537 Compare(). Comparison became 125x faster.
538 * mono-tailoring-source.txt : added tiny comment.
540 2005-07-26 Atsushi Enomoto <atsushi@ximian.com>
542 * mono-tailoring-source.txt : Added all single sortkey remapping to
543 all cultures (still need to fill contractions and annotate possible
544 buggy mapping referencing to CLDR).
545 * SimpleCollator.cs : removed unused code.
546 * MSCompatUnicodeTable.cs : tiny cast removal.
548 2005-07-25 Atsushi Enomoto <atsushi@ximian.com>
551 create-mscompat-collation-table.cs
552 MSCompatUnicodeTableUtil.cs
553 MSCompatUnicodeTable.cs : Now CJK mapping data is stored as byte
554 arrays. Thus SimpleCollator does not need to use bitwise and shift
555 operations to get sortkey value and they could be managed resources.
557 2005-07-25 Atsushi Enomoto <atsushi@ximian.com>
559 * create-mscompat-collation-table.cs,
560 MSCompatUnicodeTable.cs,
561 MSCompatUnicodeTableUtil.cs : From the result of sortkey comparison
562 between None and IgnoreWidth, width compat table could be computed
563 in somewhat simple way. So removed that table and all related code.
564 Increased the collation resource version.
566 2005-07-25 Atsushi Enomoto <atsushi@ximian.com>
568 * create-mscompat-collation-table.cs : Added C header output support.
570 2005-07-25 Atsushi Enomoto <atsushi@ximian.com>
572 * create-mscompat-collation-table.cs : FillLetterNFKD() could also be
573 applied to Cyrillic letters. Saved some of them.
575 2005-07-24 Atsushi Enomoto <atsushi@ximian.com>
577 * MSCompatUnicodeTable.cs : oh, ok, so we already have
578 GetManifestResourceInternal() ;-)
579 * managed-collation.patch : in Assembly.cs made that method internal.
581 2005-07-24 Atsushi Enomoto <atsushi@ximian.com>
583 * MSCompatUnicodeTable.cs : the pointer based icall code could be
584 also applicable for USE_MANAGED_RESOURCE mode.
586 2005-07-23 Atsushi Enomoto <atsushi@ximian.com>
588 * MSCompatUnicodeTable.cs : added icall support code (not enabled
589 unless the first line is commented out).
591 2005-07-22 Atsushi Enomoto <atsushi@ximian.com>
593 * create-mscompat-collation-table.cs,
594 MSCompatUnicodeTableUtil.cs,
595 MSCompatUnicodeTable.cs : Added resource version output (and ignore
596 in case of version mismatch). Removed obsolete, commented out code.
598 2005-07-22 Atsushi Enomoto <atsushi@ximian.com>
601 MSCompatUnicodeTable.cs,
602 create-mscompat-collation-table.cs : Now they use unmanaged pointers
603 instead of managed arrays.
604 * managed-collation.patch : Now it contains patch for IntPtrStream.cs
605 and Assembly.cs as well.
607 2005-07-22 Atsushi Enomoto <atsushi@ximian.com>
609 * MSCompatUnicodeTable.cs,
610 SimpleCollator.cs : Moved tailoring support classes to
611 MSCompatUnicodeTable.cs and drawn out from SimpleCollator.
612 Now that cjk and tailoring support are filled inside
613 MSCompatUnicodeTable, no managed array is exposed.
615 2005-07-22 Atsushi Enomoto <atsushi@ximian.com>
617 * create-mscompat-collation-table.cs,
619 MSCompatUnicodeTable.cs : Now it's not exposing collation table
620 internals as managed arrays (to switch to unmanaged pointers).
622 2005-07-22 Atsushi Enomoto <atsushi@ximian.com>
624 * create-mscompat-collation-table.cs : tiny nonspacing mark fix.
626 2005-07-21 Atsushi Enomoto <atsushi@ximian.com>
628 * create-mscompat-collation-table.cs : Fixed most of Greek mappings.
629 * MSCompatUnicodeTable.cs : don't lock string.
631 2005-07-21 Atsushi Enomoto <atsushi@ximian.com>
633 * create-mscompat-collation-table.cs : More Cyrillic diacritical fixes.
635 2005-07-21 Atsushi Enomoto <atsushi@ximian.com>
637 * create-mscompat-collation-table.cs : More Latin diacritical fixes.
639 2005-07-21 Atsushi Enomoto <atsushi@ximian.com>
641 * create-mscompat-collation-table.cs : There were still missing
642 math symbol mappings. Added several hacky diacritical weight for
645 2005-07-21 Atsushi Enomoto <atsushi@ximian.com>
647 * create-mscompat-collation-table.cs : fixed a few diacritical weight
648 on Cyrillic characters. Fixed ParseTailoringSource() to handle
649 non-heading escape sequence (\uXXXX) as expected.
651 2005-07-21 Atsushi Enomoto <atsushi@ximian.com>
653 * create-mscompat-collation-table.cs,
654 MSCompatUnicodeTableUtil.cs,
655 MSCompatUnicodeTable.cs : added more aggressive index limits for
656 table optimization at data size, in cost of speed.
658 2005-07-20 Atsushi Enomoto <atsushi@ximian.com>
660 * create-mscompat-collation-table.cs : fixed Arabic thirtial weight.
662 2005-07-20 Atsushi Enomoto <atsushi@ximian.com>
664 * create-mscompat-collation-table.cs : Mapping for hyphens and
665 punctuation are kinda finished. Rewrote batch mapping method to
666 collect all NFKD. Required modification on mapping is done.
668 2005-07-20 Atsushi Enomoto <atsushi@ximian.com>
670 * create-mscompat-collation-table.cs : minor mapping fixes on accent
671 marks and punctuations.
673 2005-07-20 Atsushi Enomoto <atsushi@ximian.com>
675 * create-mscompat-collation-table.cs : Fixed some MathSymbol mapping
676 and Box drawing mapping.
678 2005-07-19 Atsushi Enomoto <atsushi@ximian.com>
680 * create-mscompat-collation-table.cs : Fixed almost all numbers.
682 2005-07-19 Atsushi Enomoto <atsushi@ximian.com>
684 * create-mscompat-collation-table.cs : Symbol mappings are almost done.
685 Removed hack that gave dummy mappings to blank symbols.
687 2005-07-19 Atsushi Enomoto <atsushi@ximian.com>
689 * create-mscompat-collation-table.cs : more fix on arrows. Fix on box
690 drawings. Some code refactoring to eliminate hack.
692 2005-07-19 Atsushi Enomoto <atsushi@ximian.com>
694 * create-mscompat-collation-table.cs : Fixed some secondary weight
695 in Devanagari and arrows.
697 2005-07-19 Atsushi Enomoto <atsushi@ximian.com>
699 * create-mscompat-collation-table.cs : a set of tiny mapping fixes.
701 2005-07-19 Atsushi Enomoto <atsushi@ximian.com>
703 * create-mscompat-collation-table.cs : some diacritical fixes for
704 Latin. Added batch mapping method that considers computed
705 diacritical weight (for numbers).
707 2005-07-15 Atsushi Enomoto <atsushi@ximian.com>
709 * managed-collation.patch : forgot to add System.String patch.
711 2005-07-15 Atsushi Enomoto <atsushi@ximian.com>
713 * MSCompatUnicodeTable.cs : added resource existence check (required
714 for mscorlib transient time from the one without resources to the
717 2005-07-15 Atsushi Enomoto <atsushi@ximian.com>
719 * create-mscompat-collation-table.cs : fixed punctuations and hyphen
720 (shift) primary weight.
722 2005-07-15 Atsushi Enomoto <atsushi@ximian.com>
724 * create-mscompat-collation-table.cs : more nonspacing mark fixes.
725 Some non-basic Cyrillic diacritical weight fixes.
727 2005-07-15 Atsushi Enomoto <atsushi@ximian.com>
729 * create-mscompat-collation-table.cs : some Gurmukhi fixes on level 1
730 and level 3. Tiny Hangul weight fixes.
731 * MSCompatUnicodeTable.cs : U+30F5 and U+30F6 are small Japanese.
733 2005-07-15 Atsushi Enomoto <atsushi@ximian.com>
735 * create-mscompat-collation-table.cs : some normal characters who have
736 "narrow" NFKD mapping are regarded as "wide" and thus level 3 weight
737 values were different. Handle U+30FB as category A.
738 * MSCompatUnicodeTable.cs : U+30FB does not have special weight.
740 2005-07-15 Atsushi Enomoto <atsushi@ximian.com>
742 * create-mscompat-collation-table.cs : more diacritical weight fixes.
743 Removed some unused code.
745 2005-07-15 Atsushi Enomoto <atsushi@ximian.com>
747 * create-mscompat-collation-table.cs : Fixed some Thai and Arabic
750 2005-07-15 Atsushi Enomoto <atsushi@ximian.com>
752 * create-mscompat-collation-table.cs : Fixed Syriac nonspacing marks.
754 2005-07-15 Atsushi Enomoto <atsushi@ximian.com>
756 * create-mscompat-collation-table.cs : Fixed nonspacing marks in
757 Malayalam, Thai and Lao. Removed extraneous hack.
759 2005-07-15 Atsushi Enomoto <atsushi@ximian.com>
761 * SimpleCollator.cs : rewrote LastIndexOf() to handle source extenders.
762 Some refactoring on IndexOf() code. Removed unused Matches().
763 * Collation-notes.txt : some methods needed to be reimplemented, so
764 rewrote the description.
766 2005-07-14 Atsushi Enomoto <atsushi@ximian.com>
768 * SimpleCollator.cs : rewrote IsSuffix() to use CompareInternal().
769 Thus supported extenders in IsSuffix().
771 2005-07-14 Atsushi Enomoto <atsushi@ximian.com>
773 * SimpleCollator.cs : more IsSuffix() simplification, but it will be
774 stopped here since it cannot handle extenders (implementing new
777 2005-07-14 Atsushi Enomoto <atsushi@ximian.com>
779 * SimpleCollator.cs : simplified IsSuffix() code.
781 2005-07-14 Atsushi Enomoto <atsushi@ximian.com>
783 * SimpleCollator.cs : Fixed IndexOf() and LasIndexOf() to search the
784 entire replacement string if char target was an expansion.
785 IsSuffix() was using a method for IsPrefix() which was incorrect.
786 Removed old IsPrefix() code.
788 2005-07-14 Atsushi Enomoto <atsushi@ximian.com>
790 * SimpleCollator.cs : IndexOf() was incorrectly sharing the same
791 byte[] field in different areas of code. Now extenders in both
792 source and target really work in IndexOf().
794 2005-07-14 Atsushi Enomoto <atsushi@ximian.com>
796 * create-mscompat-collation-table.cs : fixed U+FF9F diacritical weight.
797 * SimpleCollator.cs : handle U+FF9E and U+FF9F as extenders.
799 2005-07-14 Atsushi Enomoto <atsushi@ximian.com>
801 * SimpleCollator.cs : Now FilterExtender() handles all extender
802 support. IndexOf() and LastIndexOf() now supports extenders.
803 IndexOf() and LastIndexOf() did not proceed contraction source
804 length as expected. Tiny refactoring on private IsPrefix() to take
807 2005-07-13 Atsushi Enomoto <atsushi@ximian.com>
809 * SimpleCollator.cs : when restoring from expansion, go back to the
810 top of the loop (to avoid index out of range).
811 Now IsPrefix() is implemented to reuse Compare() and thus it now
812 supports extender as well.
813 * Collation-notes.txt : status update. Deleted optimization part in
814 status section (it is duplicate).
816 2005-07-13 Atsushi Enomoto <atsushi@ximian.com>
818 * SimpleCollator.cs : some code reordering.
819 * create-mscompat-collation-table.cs : it was still missing U+3094.
821 2005-07-13 Atsushi Enomoto <atsushi@ximian.com>
823 * SimpleCollator.cs : Compare() now supports extender (e.g. U+39FC).
825 2005-07-13 Atsushi Enomoto <atsushi@ximian.com>
827 * SimpleCollator.cs : In GetSortKey(), don't update previousChar when
828 it is not primary (e.g. don't "extend" diacritical mark).
830 2005-07-13 Atsushi Enomoto <atsushi@ximian.com>
832 * managed-collation.patch : CompareInfo.Compare() should consider
833 the possibilities that non-empty string might be actually empty
834 in culture-sensitive context.
836 2005-07-13 Atsushi Enomoto <atsushi@ximian.com>
838 * SimpleCollator.cs : IndexOf() and LastIndexOf() returns start when
839 target is "empty" (in culture-sensitive context).
841 2005-07-13 Atsushi Enomoto <atsushi@ximian.com>
843 * SimpleCollator.cs : In IndexOf() and LastIndexOf(), skip ignorable
844 characters in target string.
846 2005-07-13 Atsushi Enomoto <atsushi@ximian.com>
848 * SimpleCollator.cs : When IgnoreWidth is specified, all Kana
849 characters are regarded as half-width.
850 Even though IgnoreWidth is specified, it should not ignore case.
851 For special weight comparison, the default values (E4) are bigger
852 than non-default values.
853 * SortKeyBuffer.cs : It should save LCID and original string.
854 * create-mscompat-collation-table.cs : For Japanese half-width kana,
855 it should not be counted in widthCompat map since IgnoreWidth does
856 not really ignore those differences.
858 2005-07-13 Atsushi Enomoto <atsushi@ximian.com>
860 * create-mscompat-collation-table.cs : Fixed missing Japanese bits.
862 2005-07-13 Atsushi Enomoto <atsushi@ximian.com>
864 * create-mscompat-collation-table.cs :
865 tiny diacritical weight fix for U+20D0-U+20E1.
867 2005-07-13 Atsushi Enomoto <atsushi@ximian.com>
869 * create-mscompat-collation-table.cs : ja CJK ideograph got completed.
871 2005-07-13 Atsushi Enomoto <atsushi@ximian.com>
873 * create-mscompat-collation-table.cs : Fixed CJK custom Japanese
874 mapping. It (maybe as well as other CJK tables) mixes NFKD. For
875 Japanese, modified NFKD table (because of Windows lame design).
877 2005-07-13 Atsushi Enomoto <atsushi@ximian.com>
879 * Makefile : added MONO_USE_MANAGED_COLLATION=no almost everywhere.
880 * MSCompatUnicodeTable.cs : FillCJK() was not invoked. Now it is
881 invoked at any time it is required.
882 * SimpleCollator.cs : call FillCJK() above in .ctor().
883 * MSCompatUnicodeTableUtil.cs : CJK range was wider.
884 * create-mscompat-collation-table.cs : CJK binary was missing the
885 length. CJK remapping is being moved to ModifyUnidata().
886 For cjk-ja mapping, we have to consider compat characters to be
887 added to the map, besides the raw UCA table.
889 2005-07-12 Atsushi Enomoto <atsushi@ximian.com>
891 * SortKeyBuffer.cs : Fixed shift level computation to match w/ Windows.
893 2005-07-12 Atsushi Enomoto <atsushi@ximian.com>
895 * SimpleCollator.cs : fixed LastIndexOf() to handle _target's_
896 contraction as expected. Fixed Compare() to save s2's contraction
898 * TestDriver.cs :added LastIndexOf() tester w/ indexes.
900 2005-07-12 Atsushi Enomoto <atsushi@ximian.com>
902 * managed-collation.patch : Fixed IsPrefix() and IsSuffix(). They
903 incorrectly use Compare().
904 * TestDriver.cs : more moved to nunit tests.
906 2005-07-12 Atsushi Enomoto <atsushi@ximian.com>
908 * SimpleCollator.cs : several fixes on Compare().
909 - Ignorable characters are skippted at the top of the loop.
910 - IgnoreNonSpace is checked to avoid extraneous level 2 comparison.
911 - In such case that s1 index is increased while s2 contraction is
912 replaced, s1 is inconsistently proceeded (bug).
913 - IsIgnorable() now also checks IgnoreNonSpace.
914 - Fixed FilterOptions() that does not work for IgnoreWidth at all.
915 * TestDriver.cs : now some are moved to nunit tests.
916 * Collation-notes.txt : minor todo update.
918 2005-07-11 Atsushi Enomoto <atsushi@ximian.com>
920 * SimpleCollator.cs : Compare() was ignoring such case that both
921 entire strings have '-' to be compared.
922 * Collation-notes.txt : more status updates.
923 * TestDriver.cs : added '-' use cases.
925 2005-07-08 Atsushi Enomoto <atsushi@ximian.com>
927 * SimpleCollator.cs : to be same as other buggy part, it now handles
928 U+3005, U+3031 and U+3032 as buggy as Windows. It just repeats
930 Fixed GetSortKey(): if the repeater is U+3005, second weight is 5.
931 * create-mscompat-collation-table.cs : dummy values for extenders.
933 2005-07-08 Atsushi Enomoto <atsushi@ximian.com>
935 * SimpleCollator.cs : Special weight fixes on GetSortKey(). Dash type
936 should be computed from ExtenderType, and voice mark weight should
938 * MSCompatUnicodeTable.cs : added tiny comment.
940 2005-07-08 Atsushi Enomoto <atsushi@ximian.com>
942 * SortKey.cs : It borked when MONO_USE_MANAGED_COLLATION is not yes.
943 * SimpleCollator.cs : support for extender (U+309D etc.).
945 2005-07-08 Atsushi Enomoto <atsushi@ximian.com>
947 * create-mscompat-collation-table.cs : some punct/symbols fix.
948 * managed-collation.patch : new (and temporary) file to support
949 managed collation in mscorlib.
950 * README : described how to use managed collation.
952 2005-07-08 Atsushi Enomoto <atsushi@ximian.com>
954 * create-mscompat-collation-table.cs : Further Cyrillic fixes. Handle
955 U+482-4C8 (though needs diacritical fixes).
956 * MSCompatUnicodeTable.cs : tiny comment for alternative impl.
958 2005-07-08 Atsushi Enomoto <atsushi@ximian.com>
960 * create-mscompat-collation-table.cs : Reimplemented Cyrillic weight
961 computation code, since it looks like the same way as Latin letters
962 have. Thus removed all other approach (UCA, by letter name).
964 2005-07-07 Atsushi Enomoto <atsushi@ximian.com>
966 * create-mscompat-collation-table.cs : diacritical fix for "double-
967 struck". Syriac nonspacing fixes.
969 2005-07-07 Atsushi Enomoto <atsushi@ximian.com>
971 * create-mscompat-collation-table.cs : more math symbol weight fixes.
973 2005-07-07 Atsushi Enomoto <atsushi@ximian.com>
975 * create-mscompat-collation-table.cs : fixed Hebrew character sortkeys.
977 2005-07-07 Atsushi Enomoto <atsushi@ximian.com>
979 * create-mscompat-collation-table.cs : math symbols U+25A0-U+2600 are
980 implemented (no stub). Some other fixes on category 8-A.
982 2005-07-07 Atsushi Enomoto <atsushi@ximian.com>
984 * create-mscompat-collation-table.cs : some minor fixes on Arabic,
985 Korean and Japanese sortkey weights.
987 2005-07-07 Atsushi Enomoto <atsushi@ximian.com>
989 * create-mscompat-collation-table.cs : More diacritical fixes.
990 Georgian characters do not have level 2 weights but level 3.
992 2005-07-07 Atsushi Enomoto <atsushi@ximian.com>
994 * create-mscompat-collation-table.cs : Roman numeral characters
995 have diacritical weight. quick hack for control signs (U+2400..)
998 2005-07-06 Atsushi Enomoto <atsushi@ximian.com>
1000 * create-mscompat-collation-table.cs : improving Latin mappings.
1001 Setting non-ASCII Latin characters' primary weight between those
1002 ASCII characters, and setting diacritical weight (hacky).
1003 * MSCompatUnicodeTable.cs :
1004 Kanatype check: fixed (voice marks) and improved (comparison order).
1006 2005-07-06 Atsushi Enomoto <atsushi@ximian.com>
1008 * create-mscompat-collation-table.cs : more diacritical fixes.
1009 primary weight fixes on punctuations in category 07.
1011 2005-07-06 Atsushi Enomoto <atsushi@ximian.com>
1013 * create-mscompat-collation-table.cs : several diacritical fixes.
1014 * TestDriver.cs : sortkey dumper should use StringSort.
1016 2005-07-05 Atsushi Enomoto <atsushi@ximian.com>
1018 * SimpleCollator.cs : fixed incorrect indexer setup. Optimized
1019 GetContraction() call a bit.
1021 2005-07-05 Atsushi Enomoto <atsushi@ximian.com>
1023 * create-mscompat-collation-table.cs : fixed incorrect level 2
1025 * MSCompatUnicodeTable.cs : remove debug line.
1027 2005-07-05 Atsushi Enomoto <atsushi@ximian.com>
1029 * MSCompatUnicodeTableUtil.cs,
1030 MSCompatUnicodeTable.cs,
1031 CodePointIndexer.cs,
1032 create-mscompat-collation-table.cs : made some members internal and
1033 accessible from other classes. Many indexes could be 0 by default.
1034 * SimpleCollator.cs : optimizations. avoid method call.
1036 2005-07-05 Atsushi Enomoto <atsushi@ximian.com>
1038 * Collation-notes.txt : more updates.
1039 * SimpleCollator.cs : Added quick check for Ordinal comparison.
1040 Fixed special weight comparison. It cannot be customizable in the
1041 implementation (and it won't be harmful).
1042 * mono-tailoring-source.txt : thus updated comment.
1044 2005-07-05 Atsushi Enomoto <atsushi@ximian.com>
1046 * SimpleCollator.cs : Compare() was missing French sort support.
1047 * TestDriver.cs : added example case.
1049 2005-07-05 Atsushi Enomoto <atsushi@ximian.com>
1051 * Collation-notes.txt : updated status. Eliminated descriptions on
1052 "iterator" (I avoided it for performance concern). Fixed misc.
1053 incorrect descriptions.
1055 2005-07-05 Atsushi Enomoto <atsushi@ximian.com>
1057 * Collator.cs : Now that SimpleCollator became feature complete, it is
1060 2005-07-05 Atsushi Enomoto <atsushi@ximian.com>
1062 * SimpleCollator.cs : implemented decent Compare() that immediately
1063 stops at first primary difference.
1065 2005-07-04 Atsushi Enomoto <atsushi@ximian.com>
1067 * SimpleCollator.cs : indexers might return -1.
1069 2005-07-04 Atsushi Enomoto <atsushi@ximian.com>
1071 * SimpleCollator.cs : IsPrefix() and IsSuffix() optimization code was
1072 buggy (length check for source was missing).
1074 2005-07-04 Atsushi Enomoto <atsushi@ximian.com>
1076 * create-mscompat-collation-table.cs : Fixed tailoring table output
1077 to be in correct and countable order. Now if tailoring alias was not
1078 found, just stop the build.
1079 * MSCompatUnicodeTable.cs : several build fixes. Now it works to read
1081 * mono-tailoring-source.txt : commented out CJK aliases that miss
1083 * Makefile : needed further filename fixes.
1085 2005-07-04 Atsushi Enomoto <atsushi@ximian.com>
1087 * MSCompatUnicodeTable.cs : renamed from MSCompatUnicodeTable.template
1088 (now it is working as a standalone file).
1089 * Makefile : renamed generated file as MSCompatUnicodeTableGenerated.cs
1090 (the generator now creates both binary resources and C# source).
1092 2005-07-04 Atsushi Enomoto <atsushi@ximian.com>
1094 * create-mscompat-collation-table.cs : Now it generates binary
1095 resources (to parent directory).
1096 * MSCompatUnicodeTable.template : added conditional code that fills
1097 collation tables from manifest resources.
1098 * Makefile : remove collation table binaries as well on "make clean".
1099 Removed extraneous dependency.
1101 2005-07-01 Atsushi Enomoto <atsushi@ximian.com>
1103 * MSCompatUnicodeTable.template,
1104 SimpleCollator.cs : removed extraneous GetExpansion().
1106 2005-07-01 Atsushi Enomoto <atsushi@ximian.com>
1108 * SimpleCollator.cs : IsSuffix() also supports contractions.
1109 * TestDriver.cs : IsSuffix() example contraction cases.
1111 2005-07-01 Atsushi Enomoto <atsushi@ximian.com>
1113 * SimpleCollator.cs : reverted IsSuffix() to return bool (to match w/
1114 what current IsPrefix() does). For expansion of target, IsPrefix()
1115 should check the no-match case that expansion is longer than input.
1116 Some refactory on IsPrefix().
1117 Added GetContractionTal() for IsSuffix() (not used yet).
1119 2005-07-01 Atsushi Enomoto <atsushi@ximian.com>
1121 * TestDriver.cs : added IsPrefix() expansion cases.
1122 * SimpleCollator.cs : IsPrefix() now supports contractions (with much
1123 of complexity), and it now returns bool again.
1124 IndexOf() for replacement should make use of IndexOfPrimitiveChar()
1125 since expansions won't be expanded recursively.
1127 2005-07-01 Atsushi Enomoto <atsushi@ximian.com>
1129 * SimpleCollator.cs : commonized character comparison in IsPrefix()
1130 and IsSuffix(). csc compile fix.
1131 * CompareInfoImpl.cs : deleted.
1133 2005-06-30 Atsushi Enomoto <atsushi@ximian.com>
1135 * TestDriver.cs : added SimpleCollator.ctor() sanity check.
1136 Added replacement contraction example.
1137 * SimpleCollator.cs : Now IndexOf() and LastIndexOf() support
1138 contraction in source string. Extracted matching code to Matches().
1139 Replacement contraction was including extraneous '\x0'.
1141 2005-06-30 Atsushi Enomoto <atsushi@ximian.com>
1143 * Collation-notes.txt : updated status.
1144 * CollationDataStructures.txt : tiny fixes.
1145 * SimpleCollator.cs :
1146 Renamed alias Util to UUtil (MS sys.enterprisesvc has sucky global
1147 namespace Util and csc borked).
1148 GetContraction was incorrectly returning first item.
1149 Private IsPrefix() now returns int (but it might not be in real use).
1150 Extracted simple char comparison to CompareCharSimple().
1151 IndexOf() and LastIndexOf() now fully handle contractions (both
1152 binary key and string replacement) in "target" (for "s" not yet).
1153 * TestDriver.cs : be more verbose.
1154 * mono-tailoring-source.txt : added comment.
1155 * MSCompatUnicodeTable.template :
1156 Renamed alias Util to UUtil (MS sys.enterprisesvc has sucky global
1158 2005-06-30 Atsushi Enomoto <atsushi@ximian.com>
1160 * create-mscompat-collation-table.cs : compute COMBINING blah marks as
1161 well as those characters WITH blah.
1162 * TestDriver.cs : added combining sortkey cases.
1164 2005-06-30 Atsushi Enomoto <atsushi@ximian.com>
1166 * mono-tailoring-source.txt : fixed description on '*' in sortkeys.
1167 * SimpleCollator.cs : Now it fully uses tailoring info. Fixed
1168 contraction search that worked only when string is contraction.
1169 Removed commented code. Minor refactoring.
1170 * TestDriver.cs : added example that uses "ZS" in Hungarian sorting.
1172 2005-06-29 Atsushi Enomoto <atsushi@ximian.com>
1174 * create-mscompat-collation-table.cs,
1175 * mono-tailoring-source.txt : removed extraneous level 4 sortkey
1176 which cannot be supported.
1177 * SimpleCollator.cs : added GetContraction() and used in some places.
1178 Now CompareOptions is set only once. Reordered some code (e.g.
1179 ignorable check -> get compat char -> compare).
1181 2005-06-29 Atsushi Enomoto <atsushi@ximian.com>
1183 * SimpleCollator.cs : sort tailoring tables before actual usage.
1184 Support diacritical remappings (it is customized collation rule
1185 which does not exist in UCA).
1187 2005-06-29 Atsushi Enomoto <atsushi@ximian.com>
1189 * SimpleCollator.cs : build culture specific tailoring table from
1190 TailoringInfo and unified data array.
1191 * create-mscompat-collation-table.cs : Added null termination to
1192 sortkey map tailorings (mostly to save my eyes).
1193 * MSCompatUnicodeTable.template : added public TailoringValues.
1195 2005-06-29 Atsushi Enomoto <atsushi@ximian.com>
1197 * SortKeyBuffer.cs : handle special weight (category 06) characters.
1198 * Collation-notes.txt : Updated description on special weight (it was
1200 * TestDriver.cs : added special weight cases.
1202 2005-06-29 Atsushi Enomoto <atsushi@ximian.com>
1204 * MSCompatUnicodeTable.template : added GetTailoringInfo().
1205 * SimpleCollator.cs : Now tailoring information is acquired and used.
1206 (FrenchSort is supported but Compare() won't work expectedly since
1207 the table is still incomplete for those diacritical marks).
1208 * SortKeyBuffer.cs : On reversing diacritical weights, it should
1209 ignore zeros. Reset() should reset frenchSorted flag.
1211 2005-06-28 Atsushi Enomoto <atsushi@ximian.com>
1213 * create-mscompat-collation-table.cs : Further fixes on Jamo,
1214 diacritical weights by character name, and *Numbers primary weights.
1216 2005-06-28 Atsushi Enomoto <atsushi@ximian.com>
1218 * create-mscompat-collation-table.cs : More fix on Devanagari,
1219 Gujarati, Oliya, Tamil and Lao sortkeys.
1221 2005-06-28 Atsushi Enomoto <atsushi@ximian.com>
1223 * create-mscompat-collation-table.cs : Fixed Georgian, Thai, Gurmukhi
1226 2005-06-28 Atsushi Enomoto <atsushi@ximian.com>
1228 * create-mscompat-collation-table.cs : Fixed Thai character primary
1229 and secondary values. Fixed Thaana letters. Added more LAMESPEC
1230 CJK compat. Fixed some circled CJK secondary weight.
1231 Hacked some nonspacing mark sortkey value adjustment.
1233 2005-06-28 Atsushi Enomoto <atsushi@ximian.com>
1235 * create-mscompat-collation-table.cs : CP932.TXT was not parsed as
1236 expected. JIS ordering was incorrect. OtherNumbers that represents
1237 10 or more values were incorrectly computed the offset. Some Hangul
1238 compat characters has different offset.
1240 2005-06-28 Atsushi Enomoto <atsushi@ximian.com>
1242 * create-mscompat-collation-table.cs : Fixed 0x8 category characters.
1243 Added hack for need-to-be-fixed characters to fall into 0xA category.
1244 * create-collation-element-table.cs : previous checkin seem failed :(
1245 * README: updated a bit.
1247 2005-06-24 Atsushi Enomoto <atsushi@ximian.com>
1249 * CodePointIndexer.cs :
1250 removed extraneous switch (I could use empty array for that need).
1251 * CollationElementTableUtil.cs : primary weight type became ushort.
1252 * create-collation-element-table.cs : several bugfixes.
1253 collElem should be int. It was skipping most of entries because of
1254 incorrect string tokenization.
1256 2005-06-23 Atsushi Enomoto <atsushi@ximian.com>
1258 * create-mscompat-collation-table.cs : handle some Jamo NKFD.
1260 2005-06-23 Atsushi Enomoto <atsushi@ximian.com>
1262 * SimpleCollator.cs : forgot to commit in the last checkin.
1263 * create-mscompat-collation-table.cs : fixed arabic shift weight chars.
1264 * TestDriver.cs : switch table dumper and collator testing.
1265 * SortKey.cs : for now comment out internal indexes (not in use).
1267 2005-06-23 Atsushi Enomoto <atsushi@ximian.com>
1269 * MSCompatUnicodeTable.template,
1270 SimpleCollator.cs : support for culture dependent CJK table.
1272 2005-06-23 Atsushi Enomoto <atsushi@ximian.com>
1274 * create-mscompat-collation-table.cs,
1275 MSCompatUnicodeTableUtil.cs : make CJK table more compact.
1277 2005-06-22 Atsushi Enomoto <atsushi@ximian.com>
1279 * SimpleCollator.cs : Fixed stupid index search when start != 0.
1281 2005-06-21 Atsushi Enomoto <atsushi@ximian.com>
1283 * SimpleCollator.cs : fixed my misunderstanding on LastIndexOf(). It
1284 now starts from "start" and proceeds backward by "length".
1285 * TestDriver.cs : fix warning.
1287 2005-06-21 Atsushi Enomoto <atsushi@ximian.com>
1289 * TestDriver.cs : more tests.
1290 * SimpleCollator.cs : LastIndexOf() is not setting search length
1291 on iteration. Quick workaround fro String.LastIndexOf() bug (maybe).
1293 2005-06-21 Atsushi Enomoto <atsushi@ximian.com>
1295 * create-normalization-source.cs : output propValue as uint.
1297 2005-06-21 Atsushi Enomoto <atsushi@ximian.com>
1299 * SortKey.cs : Now it is System.Globalization.SortKey.
1300 To replace existing implementation, it now requires lcid and
1301 CompareOptions. Added required members.
1302 * SortKeyBuffer.cs : thus .ctor() requires LCID.
1303 * SimpleCollator.cs : made required changes above.
1305 2005-06-21 Atsushi Enomoto <atsushi@ximian.com>
1307 * CodePointIndexer.cs : added CompressArray(). Now it requires two more
1308 parameters for default index and codepoint.
1309 * CollationElementTableUtil.cs,
1310 NormalizationTableUtil.cs : required changes wrt above change.
1311 * MSCompatUnicodeTableUtil.cs : added for several codepoint indexers.
1312 * MSCompatUnicodeTable.template : Now it uses codepoint indexer.
1313 * create-mscompat-collation-table.cs : Now it outputs compressed array.
1314 * Makefile : now collation requires MSCompatUnicodeTableUtil.cs
1316 2005-06-21 Atsushi Enomoto <atsushi@ximian.com>
1318 * SimpleCollator.cs :
1319 Implemented IsSuffix() and LastIndexOf().
1320 Several fixes on index > 0 cases.
1321 * TestDriver.cs : sample IsSuffix() and LastIndexOf() usage and more.
1323 2005-06-21 Atsushi Enomoto <atsushi@ximian.com>
1325 * Collation-notes.txt : updated (status, impl. classes).
1326 * MSCompatUnicodeTable.cs : Korean Jamo are not really expansions.
1328 2005-06-21 Atsushi Enomoto <atsushi@ximian.com>
1330 * SimpleCollator.cs : implemented IndexOf(string,string,CompareOptions)
1331 and IsPrefix(). Tiny code refactory.
1332 * TestDriver.cs : sample IsPrefix() and IndexOf() usage.
1333 * MSCompatUnicodeTable.cs : tiny refactory for CodePointIndexer use.
1335 2005-06-20 Atsushi Enomoto <atsushi@ximian.com>
1337 * SimpleCollator.cs :
1338 IndexOf(string, char, CompareOptions) implementation.
1339 * TestDriver.cs : sample IndexOf() usage.
1341 2005-06-20 Atsushi Enomoto <atsushi@ximian.com>
1343 * create-mscompat-collation-table.cs : was missing most important
1344 kind of blocks - equivalent expansions (e.g. invariant mappings).
1345 More readable mappings.
1347 2005-06-20 Atsushi Enomoto <atsushi@ximian.com>
1349 * mono-tailoring-source.txt : new file. It describes tailoring
1350 information. Basically examined under .NET 1.x.
1351 * create-mscompat-collation-table.cs : consume the file above.
1352 * MSCompatUnicodeTable.template : now tailorings is not a stub.
1353 * CollationDataStructures.txt : minor fixes.
1355 SimpleCollator.cs : added FrenchSort support.
1356 * Collation-notes.txt : added description on Latin primary weights.
1357 * ldml-limited.rng : added note.
1358 * create-tailorings.cs : added note. more serialization (but won't be
1361 2005-06-17 Atsushi Enomoto <atsushi@ximian.com>
1363 * SortKeyBuffer.cs : non-primary character is added to previous
1365 * TestDriver.cs : added example case of above.
1367 2005-06-17 Atsushi Enomoto <atsushi@ximian.com>
1369 * SimpleCollator.cs : IgnoreSymbols support.
1370 * TestDriver.cs : compilation fix. IgnoreSymbols example.
1371 * create-mscompat-collation-table.cs : more Hangul fixes.
1373 2005-06-17 Atsushi Enomoto <atsushi@ximian.com>
1375 * create-mscompat-collation-table.cs : more Hangul fixes.
1376 * SortKey.cs : it will replace sys.globalization.SortKey. It has
1377 some internal members.
1378 * SortKeyBuffer.cs : now it uses SortKey instead of byte[].
1379 * SimpleCollator.cs : CompareOptions support. However I don't think
1380 it will be developed anymore since SortKey never enables IndexOf().
1381 * TestDriver.cs : a few CompareOptions cases.
1383 2005-06-16 Atsushi Enomoto <atsushi@ximian.com>
1385 * SimpleCollator.cs : simple collator implementation that just will
1386 use GetSortKey() for all its basis.
1387 * TestDriver.cs : sample code that uses this collator set.
1388 * MSCompatUnicodeTable.template : removed test driver from here.
1390 2005-06-16 Atsushi Enomoto <atsushi@ximian.com>
1392 * create-mscompat-collation-table.cs : Hangul fixes.
1393 Now less than 300 characters that does not have sortkey weights.
1394 * MSCompatUnicodeTable.template : added FIXME info for Hangul Jamo.
1396 2005-06-16 Atsushi Enomoto <atsushi@ximian.com>
1398 * create-mscompat-collation-table.cs : Added control picture mappings.
1399 Minor primary weight fixes.
1401 2005-06-16 Atsushi Enomoto <atsushi@ximian.com>
1403 * create-mscompat-collation-table.cs : Added mappings for box
1404 drawings and blocks.
1406 2005-06-16 Atsushi Enomoto <atsushi@ximian.com>
1408 * create-mscompat-collation-table.cs : Added mappings for arrows.
1410 2005-06-15 Atsushi Enomoto <atsushi@ximian.com>
1412 * create-mscompat-collation-table.cs : added support for letterlike
1413 characters and squared CJK compatibility characters, ordered by
1414 character names (0x0E category).
1415 * Collation-notes.txt : added description on that.
1417 2005-06-15 Atsushi Enomoto <atsushi@ximian.com>
1419 * MSCompatUnicodeTable.template : Now expansions are simulated.
1420 * create-mscompat-collation-table.cs : filled Korean number level2.
1421 Reordered some code blocks to fill correct diacritical differences.
1422 * Collation-notes.txt : some corrections and minor additions.
1424 2005-06-15 Atsushi Enomoto <atsushi@ximian.com>
1426 * MSCompatUnicodeTable.template :
1427 Now dumper test driver uses SortKeyBuffer for dogfooding.
1428 * create-mscompat-collation-table.cs : some diacritical level fixes
1429 (with non-working extra latin check).
1430 * SortKeyBuffer.cs : several fixes to get working as a practical code.
1431 * Collator.cs : make it compilable, leaving things as NotImplemented.
1433 2005-06-15 Atsushi Enomoto <atsushi@ximian.com>
1435 * create-mscompat-collation-table.cs : some fixes on primary category
1436 07 (miscellaneous symbols and punctuations).
1438 2005-06-14 Atsushi Enomoto <atsushi@ximian.com>
1440 * create-mscompat-collation-table.cs : more mapping fix on numbers,
1441 letters, variable weight characters, circled Japanese and CJK.
1442 * MSCompatUnicodeTable.template : fixed HasSpecialWeight() to be more
1443 inclusive. Simplified dumper code.
1445 2005-06-14 Atsushi Enomoto <atsushi@ximian.com>
1447 * create-mscompat-collation-table.cs : finished Hangul (both Jamo
1448 and Syllables). sortkey dumper diff lines became 8000 from 30000.
1450 2005-06-14 Atsushi Enomoto <atsushi@ximian.com>
1452 * create-mscompat-collation-table.cs : added some nonspacing marks in
1453 either correct or hacky way.
1455 2005-06-13 Atsushi Enomoto <atsushi@ximian.com>
1457 * create-mscompat-collation-table.cs : several improvements. Japanese
1458 Kana support, Hebrew accents, Bengali nonspacing marks, sorting of
1459 numeric characters, diacritically decorated latin alphabets. Fixed
1460 some diacritical weights detection.
1461 * MSCompatUnicodeTable.cs : tiny Japanese fix. Handle nonspacing
1462 marks' primary weight as empty.
1463 * Collation-notes.txt : some updates.
1465 2005-06-13 Atsushi Enomoto <atsushi@ximian.com>
1467 * create-mscompat-collation-table.cs : don't process nonexact NFKD
1468 mapping as equivalent, however store CJK extensions into NFKD map
1469 even if one does not strictly match.
1470 Now am going to fill Hangul into tables (unlike UCA it does not look
1471 possible to calculate sortkey value).
1472 Fixed Cyrillic and Georgian UCA based orderings.
1473 * MSCompatUnicodeTable.template : added CJK extension sortkey
1476 2005-06-10 Atsushi Enomoto <atsushi@ximian.com>
1478 * create-mscompat-collation-table.cs : Fixed latin alphabet support.
1479 Added latin with diacritical and CJK extension.
1480 * MSCompatUnicodeTable.cs : modified dumper code a bit (for my purpose).
1482 2005-06-10 Atsushi Enomoto <atsushi@ximian.com>
1484 * create-mscompat-collation-table.cs : now parses DerivedAge.txt (right
1485 now not used thouth). Filled CJK ideograph, still not perfect.
1486 Fixed number primary keys. NFKD numbers and CJK ideographs are now
1487 considered, including brackets elimination.
1488 * Makefile : now it downloads DerivedAge.txt.
1489 * MSCompatUnicodeTable.template : added dummy code dumper. It computes
1490 PrivateUse, Surrogate and Hangul Syllables.
1491 * Collation-notes.txt : Noted that Hangul Syllables need more love.
1493 2005-06-09 Atsushi Enomoto <atsushi@ximian.com>
1495 * create-tailorings.cs : added configuration support. sort them.
1496 I wonder if it is really usable. Having own format might be better.
1497 * create-mscompat-collation-table.cs : fixing some sortkey numbers,
1498 making closer to windows. Now it handles NFKD in some places.
1499 * MSCompatUnicodeTable.template : Added dummy sortkey dumper driver.
1500 * CollationDataStructures.txt : added description on tailoring
1501 fields, though they are subject to change.
1503 2005-06-07 Atsushi Enomoto <atsushi@ximian.com>
1505 * create-tailorings.cs, ldml-limited.rng : new file.
1506 * LdmlReader.cs : removed old file.
1508 2005-06-07 Atsushi Enomoto <atsushi@ximian.com>
1510 * SortKeyBuffer.cs : split from Collator.cs. Now it considers
1511 practical use, reflecting updated sortkey constant design.
1512 Especially level 4 weight is split to 4 arrays that are merged in
1513 the last stage of GetSortKey().
1514 * Collator.cs : thus SortKeyBuffer is removed from here.
1515 Additionally, removed some extraneous bits in other classes.
1516 * Collation-notes.txt : Some editorial fixes. Added information on
1517 Korean matter (how to compute Hangle Syllables / Hangul Jamo cannot
1518 be stored in simple byte arrays).
1519 * CodePointIndexer.cs,
1520 create-collation-element-table.cs,
1521 CollationElementTable.template,
1522 NormalizationTableUtil.cs : short CodePointIndexer method names.
1523 * create-mscompat-collation-table.cs : Additional info on why some
1524 meaningful characters are ignored in Windows (Unicode version
1525 difference). Removed U+070F from special check (was extraneous).
1527 2005-06-06 Atsushi Enomoto <atsushi@ximian.com>
1529 * MSCompatUnicodeTable.template:
1530 Moved body implementation to table creator and put those bool
1531 results into an array.
1532 * create-mscompat-collation-table.cs :
1533 So imported those methods. Modified array output to emit "0x"
1534 only for more than 9.
1535 * create-normalization-source.cs : ditto on "0x" output matter.
1536 * CollationDataStructures.txt : so now it holds ignorableFlags.
1538 2005-06-03 Atsushi Enomoto <atsushi@ximian.com>
1540 * Collation-notes.txt, CollationDataStructures.txt :
1541 separate document for data structure design.
1543 2005-06-03 Atsushi Enomoto <atsushi@ximian.com>
1545 * create-mscompat-collation-table.cs : added culture-dependent CJK
1546 table creation. It uses CLDR as its basis. (Culture independent CJK
1548 * Makefile : added CLDR archive downloading support.
1549 * MSCompatUnicodeTable.template : tiny renamings.
1550 * Collation-notes.txt : additional CJK info.
1552 2005-06-02 Atsushi Enomoto <atsushi@ximian.com>
1554 * Collation-notes.txt, create-mscompat-collation-table.cs :
1555 added secondary weight support for BlahNumber characters.
1557 2005-06-01 Atsushi Enomoto <atsushi@ximian.com>
1559 * downloaded : added directory. All downloaded files are stored here.
1560 * Makefile : use "downloaded" directory.
1561 Added more auto-download stuff.
1562 * create-mscompat-collation-table.cs :
1563 Added Japanese square kana support.
1565 2005-06-01 Atsushi Enomoto <atsushi@ximian.com>
1567 * Collation-notes.txt : added Estrangela (ancient Syriac) and Thaana.
1568 * create-mscompat-collation-table.cs : added support for Arabic abjad,
1569 Estrangela and Thaana.
1570 * MSCompatUnicodeTable.template : removed BOM.
1572 2005-05-31 Atsushi Enomoto <atsushi@ximian.com>
1574 * Collation-notes.txt : wrong comment cleanup and spelling fixes.
1575 * create-mscompat-collation-table.cs : added diacritic support for
1576 Latin letters (as long as covered in primary weight).
1578 2005-05-31 Atsushi Enomoto <atsushi@ximian.com>
1580 * Makefile : minor fixes. Added warning lines to generated sources.
1582 2005-05-31 Atsushi Enomoto <atsushi@ximian.com>
1584 * create-char-mapping-source.cs :
1585 Removed ToWidthInsensitive() generation.
1587 2005-05-31 Atsushi Enomoto <atsushi@ximian.com>
1589 * create-mscompat-collation-table.cs : Now it dumps level1 to 3 values.
1590 ToWidthInsensitive() is implemented here, using an array (which is
1591 to be optimized using CodePointIndexer).
1592 * MSCompatUnicodeTable.cs : renamed as MSCompatUnicodeTable.template
1593 * MSCompatUnicodeTable.template : now it is used to generate
1594 MSCompatUnicodeTable.cs which got ready to be used.
1595 * Makefile : added MSCompatUnicodeTable.cs build support. Now it
1596 supports "make normalization" and "make collation".
1598 2005-05-30 Atsushi Enomoto <atsushi@ximian.com>
1600 * Collation-notes.txt : Description on ICU is very incorrect. Now it
1601 became more rational and sane.
1602 * create-mscompat-collation-table.cs : fixed some indexes.
1603 * Makefile : added "mstablegen" target.
1604 * MSCompatUnicodeTable.cs : removed GetPrimaryWeight(). Minor fix.
1606 2005-05-26 Atsushi Enomoto <atsushi@ximian.com>
1608 * Collation-notes.txt : more analysis on "letters".
1609 * create-mscompat-collation-table.cs : more proof of concepts.
1611 2005-05-25 Atsushi Enomoto <atsushi@ximian.com>
1613 * Collation-notes.txt : more info. Started letter sortkey analysis
1614 (some of other stuff are really non-understandable right now.)
1615 * create-mscompat-collation-table.cs : table generator proof-of-
1616 concept source (not compilable).
1617 * MSCompatUnicodeTable.cs : moved some code to the new source.
1620 2005-05-20 Atsushi Enomoto <atsushi@ximian.com>
1622 * Collation-notes.txt : started level 2 weight analysis.
1624 2005-05-19 Atsushi Enomoto <atsushi@ximian.com>
1626 * Collation-notes.txt : Additional information on how to create
1628 * MSCompatUnicodeTable.cs : implemented part of GetLevel3Weight().
1630 2005-05-19 Atsushi Enomoto <atsushi@ximian.com>
1632 * Collation-notes.txt : More case weight (level 3) analysis. I'm
1633 likely to just write table generator.
1635 2005-05-18 Atsushi Enomoto <atsushi@ximian.com>
1637 * MSCompatUnicodeTable.cs : part of level 4 weight implementation.
1639 2005-05-18 Atsushi Enomoto <atsushi@ximian.com>
1641 * Collation-notes.txt :
1643 Revised comparison methods; backward iteration is possible.
1644 More on char-by-char comparison.
1645 Level 4 comparison is actually a bit more complex.
1647 * Collator.cs : some conceptual updates wrt above.
1649 2005-05-17 Atsushi Enomoto <atsushi@ximian.com>
1651 * Collation-notes.txt : Japanese voice mark is level 2, and Hangul
1652 properties are level 3.
1654 2005-05-17 Atsushi Enomoto <atsushi@ximian.com>
1656 * Collation-notes.txt : Make it more readable. More analysis on
1657 level 3 and 4 sortkey structures.
1658 * Collator.cs : some compilation fixes (not compilable yet).
1660 2005-05-16 Atsushi Enomoto <atsushi@ximian.com>
1662 * Collation-notes.txt : Analysis on variable-weighting (level 5)
1664 * Collator.cs : updated corresponding part of level 5, and more.
1666 2005-05-13 Atsushi Enomoto <atsushi@ximian.com>
1668 * Collation-notes.txt : more updates.
1669 * Collator.cs : rewrote from scratch. Some rough sketch for sortkey
1670 buffer, character iterator and collator methods. Not compiling.
1672 2005-05-13 Atsushi Enomoto <atsushi@ximian.com>
1674 * Collator.cs : Am going to replace it with new one. No need for
1675 CompareOptions-dependent Comparer.
1677 2005-05-13 Atsushi Enomoto <atsushi@ximian.com>
1679 * Collation-notes.txt : There seems a bit more complexity.
1681 2005-05-10 Atsushi Enomoto <atsushi@ximian.com>
1683 * Collation-notes.txt : more updates, being close to write sortkey
1686 2005-05-09 Atsushi Enomoto <atsushi@ximian.com>
1688 * CompareInfoImpl.cs, Collator.cs : conceptual update
1689 * Collation-notes.txt : some corrections and additions.
1690 * Makefile : added LDML input (but it won't be used at all).
1692 2005-04-28 Atsushi Enomoto <atsushi@ximian.com>
1694 * Collation-notes.txt : more updates.
1696 2005-04-26 Atsushi Enomoto <atsushi@ximian.com>
1698 * Collation-notes.txt : more updates.
1700 2005-04-26 Atsushi Enomoto <atsushi@ximian.com>
1702 * Collation-notes.txt : some updates.
1703 * create-mapping-char-source.cs : superscripts and subscripts are also
1704 ignored in IgnoreWidth comparison.
1705 * Makefile : tiny touch fix.
1707 2005-04-25 Atsushi Enomoto <atsushi@ximian.com>
1709 * CompareInfoImpl.cs, Collator.cs : conceptual stuff (not working).
1711 2005-04-25 Atsushi Enomoto <atsushi@ximian.com>
1713 * create-char-mapping-source.cs : Now it generates
1714 ToWidthInsensitive() from combining category <wide> and <narrow>.
1715 * MSCompatUnicodeTable.cs : added ToKanaTypeInsensitive() and
1716 ToWidthInsensitive() for IgnoreKanaType and IgnoreWidth.
1718 2005-04-25 Atsushi Enomoto <atsushi@ximian.com>
1720 * README, LdmlReader.cs, DataStructures.txt : new files.
1722 2005-04-25 Atsushi Enomoto <atsushi@ximian.com>
1724 * CodePointIndexer.cs,
1725 Collation-notes.txt,
1726 CollationElementTable.template,
1727 CollationElementTableUtil.cs,
1728 create-char-mapping-source.cs,
1729 create-collation-element-table.cs,
1730 create-combining-class-source.cs,
1731 create-normalization-source.cs,
1733 MSCompatUnicodeTable.cs,
1734 Normalization.template,
1735 NormalizationTableUtil.cs : initial checkin (to private branch).