public final Class CollationData

Array of CE32 values. At index 0 there must be CE32(U+0000) to support U+0000's special-tag for NUL-termination handling.

Array of CE values for expansions and OFFSET_TAG.

Array of prefix and contraction-suffix matching data.

Header portion of the fastLatinTable. In C++, these are one array, and the header is skipped for mapping characters. In Java, two arrays work better.

See Also

jamoCE32s

Simple array of JAMO_CE32S_LENGTH=19+21+27 CE32s, one per canonical Jamo L/V/T. They are normally simple CE32s, rarely expansions. For fast handling of HANGUL_TAG.

The single-byte primary weight (xx000000) for numeric collation.

Data for scripts and reordering groups. Uses include building a reordering permutation table and providing script boundaries to AlphabeticIndex.

The length of scriptsIndex is numScripts+16. It maps from a UScriptCode or a special reorder code to an entry in scriptStarts. 16 special reorder codes (not all used) are mapped starting at numScripts. Up to MAX_NUM_SPECIAL_REORDER_CODES are codes for special groups like space/punct/digit. There are special codes at the end for reorder-reserved primary ranges.

Multiple scripts may share a range and index, for example Hira & Kana.

Start primary weight (top 16 bits only) for a group/script/reserved range indexed by scriptsIndex. The first range (separators & terminators) and the last range (trailing weights) are not reorderable, and no scriptsIndex entry points to them.

Main lookup trie.

Set of code points that are unsafe for starting string comparison after an identical prefix, or in backwards CE iteration.

Returns the CE32 from two contexts words. Access to the defaultCE32 for contraction and prefix matching.

Computes a CE from c's ce32 which has the OFFSET_TAG.

Returns the FCD16 value for code point c. c must be >= 0.

Returns the CE32 for an indirect special CE32 (e.g., with DIGIT_TAG), if ce32 is special.

Returns the first primary for the script's reordering group.

Returns:long

the primary with only the first primary lead byte of the group (not necessarily an actual root collator primary weight), or 0 if the script is unknown

Returns the CE32 for an indirect special CE32 (e.g., with DIGIT_TAG). Requires that ce32 is special.

Returns the single CE that c maps to. Throws UnsupportedOperationException if c does not map to a single CE.

Writes the permutation of primary-weight ranges for the given reordering of scripts and groups. The caller checks for illegal arguments and takes care of [DEFAULT] and memory allocation.

Each list element will be a (limit, offset) pair as described for the CollationSettings.reorderRanges. The list will be empty if no ranges are reordered.