LoadICUBreakIterator

Breaking encapsulation of ICU BreakIterator

Because of Issue 84467 (duplicate of the Issue 81519 ) we are using RuleBasedBreakIterator() constructor and then we want to setBreakType() there.

There is a fix to this that removes the patch to ICU by creating a subclass of RuleBasedBreakIterator which can access the protected setBreakType() member. The bug is here: Issue 88411

ICU code:

• BreakIterator reference
• RuleBasedBreakIterator reference

OpenOffice.org code:

• BreakIterator_Unicode::loadICUBreakIterator function

Mailing list discussions:

• Discussion about ICULanguageBreakFactory
• ports/121787 FreeBSD problem report
• Debian bug 448745
• icu-support

Example reasons to use custom rules:

• Issue 72868 Writer/Impress: line does not break after Chinese punctuation and before Latin letters
• Issue 80891 character in the forbidden list sometimes appears at the home of line
• Issue 83229 wrong hyphenation when word does contain a hyphen
• Issue 83649 Line break should be between typographical quote and left bracket
• Issue 83464 line brake between letter and $
• Issue 81448 slash and backslash make non-braking spaces of preceding spaces

Use cases of loadICUBreakIterator

Questions:

• Why does wordRule need to be static and preserved across the calls?
• Is rulestring word used at all? Other WordTypes?

public method	loadICU call	resulting rule text
nextCharacters(Text, nStartPos, rLocale, SKIPCELL, sal_Int32 nCount, nDone) prevCharacters(Text, nStartPos, rLocale, SKIPCELL, sal_Int32 nCount, nDone)	loadICUBreakIterator(rLocale, LOAD_CHARACTER_BREAKITERATOR, 0, "char", Text)	char
nextWord( const OUString& Text, sal_Int32 nStartPos, rLocale, ANYWORD_IGNOREWHITESPACES) previousWord(const OUString& Text, sal_Int32 nStartPos, rLocale, ANYWORD_IGNOREWHITESPACES) getWordBoundary( const OUString& Text, sal_Int32 nPos, rLocale, ANYWORD_IGNOREWHITESPACES, sal_Bool bDirection)	loadICUBreakIterator(rLocale, LOAD_WORD_BREAKITERATOR, ANYWORD_IGNOREWHITESPACES, NULL, Text)	edit_word
nextWord( const OUString& Text, sal_Int32 nStartPos, rLocale, DICTIONARY_WORD) previousWord(const OUString& Text, sal_Int32 nStartPos, rLocale, DICTIONARY_WORD) getWordBoundary( const OUString& Text, sal_Int32 nPos, rLocale, DICTIONARY_WORD, sal_Bool bDirection)	loadICUBreakIterator(rLocale, LOAD_WORD_BREAKITERATOR, DICTIONARY_WORD, NULL, Text)	dict_word
nextWord( const OUString& Text, sal_Int32 nStartPos, rLocale, WORD_COUNT) previousWord(const OUString& Text, sal_Int32 nStartPos, rLocale, WORD_COUNT) getWordBoundary( const OUString& Text, sal_Int32 nPos, rLocale, WORD_COUNT, sal_Bool bDirection)	loadICUBreakIterator(rLocale, LOAD_WORD_BREAKITERATOR, WORD_COUNT, NULL, Text)	count_word
nextWord( const OUString& Text, sal_Int32 nStartPos, rLocale, another_word_type) previousWord(const OUString& Text, sal_Int32 nStartPos, rLocale, another_word_type) getWordBoundary( const OUString& Text, sal_Int32 nPos, rLocale, another_word_type, sal_Bool bDirection)	loadICUBreakIterator(rLocale, LOAD_WORD_BREAKITERATOR, another_word_type NULL, Text)	word (???)
beginOfSentence( const OUString& Text, sal_Int32 nStartPos, rLocale) endOfSentence( const OUString& Text, sal_Int32 nStartPos,rLocale)	loadICUBreakIterator(rLocale, LOAD_SENTENCE_BREAKITERATOR, 0, NULL, Text);	NULL
getLineBreak( const OUString& Text, sal_Int32 nStartPos, const lang::Locale& rLocale, sal_Int32 nMinBreakPos, const LineBreakHyphenationOptions& hOptions, const LineBreakUserOptions& /*rOptions*/ )	loadICUBreakIterator(rLocale, LOAD_LINE_BREAKITERATOR, 0, "line", Text);	line

1. Figure out if locale BreakIteratorRules ({edit_word, dict_word, count_word, char, line}) gives something for the requested locale
2. If not, try to load rule+_ + lang string anyway.