diff options
Diffstat (limited to 'libjava/gnu/gcj/text')
| -rw-r--r-- | libjava/gnu/gcj/text/BaseBreakIterator.java | 82 | ||||
| -rw-r--r-- | libjava/gnu/gcj/text/CharacterBreakIterator.java | 188 | ||||
| -rw-r--r-- | libjava/gnu/gcj/text/LineBreakIterator.java | 168 | ||||
| -rw-r--r-- | libjava/gnu/gcj/text/LocaleData_en.java | 75 | ||||
| -rw-r--r-- | libjava/gnu/gcj/text/LocaleData_en_US.java | 71 | ||||
| -rw-r--r-- | libjava/gnu/gcj/text/SentenceBreakIterator.java | 226 | ||||
| -rw-r--r-- | libjava/gnu/gcj/text/WordBreakIterator.java | 224 |
7 files changed, 1034 insertions, 0 deletions
diff --git a/libjava/gnu/gcj/text/BaseBreakIterator.java b/libjava/gnu/gcj/text/BaseBreakIterator.java new file mode 100644 index 0000000..794fb2e --- /dev/null +++ b/libjava/gnu/gcj/text/BaseBreakIterator.java @@ -0,0 +1,82 @@ +// Base class for default BreakIterators. + +/* Copyright (C) 1999 Cygnus Solutions + + This file is part of libgcj. + +This software is copyrighted work licensed under the terms of the +Libgcj License. Please consult the file "LIBGCJ_LICENSE" for +details. */ + +package gnu.gcj.text; + +import java.text.BreakIterator; +import java.text.CharacterIterator; + +/** + * @author Tom Tromey <tromey@cygnus.com> + * @date March 22, 1999 + */ + +public abstract class BaseBreakIterator extends BreakIterator +{ + public int current () + { + return iter.getIndex(); + } + + public int first () + { + iter.first(); + return iter.getBeginIndex(); + } + + public int following (int pos) + { + int save = iter.getIndex(); + iter.setIndex(pos); + int r = next (); + iter.setIndex(save); + return r; + } + + public CharacterIterator getText () + { + return iter; + } + + public int last () + { + iter.last(); + return iter.getEndIndex(); + } + + public int next (int n) + { + int r = iter.getIndex (); + if (n > 0) + { + while (n > 0 && r != DONE) + { + r = next (); + --n; + } + } + else if (n < 0) + { + while (n < 0 && r != DONE) + { + r = previous (); + ++n; + } + } + return r; + } + + public void setText (CharacterIterator newText) + { + iter = newText; + } + + protected CharacterIterator iter; +} diff --git a/libjava/gnu/gcj/text/CharacterBreakIterator.java b/libjava/gnu/gcj/text/CharacterBreakIterator.java new file mode 100644 index 0000000..e1777fd --- /dev/null +++ b/libjava/gnu/gcj/text/CharacterBreakIterator.java @@ -0,0 +1,188 @@ +// Default character BreakIterator. + +/* Copyright (C) 1999 Cygnus Solutions + + This file is part of libgcj. + +This software is copyrighted work licensed under the terms of the +Libgcj License. Please consult the file "LIBGCJ_LICENSE" for +details. */ + +package gnu.gcj.text; + +import java.text.BreakIterator; +import java.text.CharacterIterator; + +/** + * @author Tom Tromey <tromey@cygnus.com> + * @date March 19, 1999 + * Written using The Unicode Standard, Version 2.0. + */ + +public class CharacterBreakIterator extends BaseBreakIterator +{ + // Hangul Jamo constants from Unicode book. + private static final int LBase = 0x1100; + private static final int VBase = 0x1161; + private static final int TBase = 0x11a7; + private static final int LCount = 19; + private static final int VCount = 21; + private static final int TCount = 28; + + // Information about surrogates. + private static final int highSurrogateStart = 0xD800; + private static final int highSurrogateEnd = 0xDBFF; + private static final int lowSurrogateStart = 0xDC00; + private static final int lowSurrogateEnd = 0xDFFF; + + public Object clone () + { + return new CharacterBreakIterator (this); + } + + public CharacterBreakIterator () + { + iter = null; // FIXME? + } + + private CharacterBreakIterator (CharacterBreakIterator other) + { + iter = (CharacterIterator) other.iter.clone(); + } + + // Some methods to tell us different properties of characters. + private final boolean isL (char c) + { + return c >= LBase && c <= LBase + LCount; + } + private final boolean isV (char c) + { + return c >= VBase && c <= VBase + VCount; + } + private final boolean isT (char c) + { + return c >= TBase && c <= TBase + TCount; + } + private final boolean isLVT (char c) + { + return isL (c) || isV (c) || isT (c); + } + private final boolean isHighSurrogate (char c) + { + return c >= highSurrogateStart && c <= highSurrogateEnd; + } + private final boolean isLowSurrogate (char c) + { + return c >= lowSurrogateStart && c <= lowSurrogateEnd; + } + + public int next () + { + int end = iter.getEndIndex(); + if (iter.getIndex() == end) + return DONE; + + char c; + for (char prev = CharacterIterator.DONE; iter.getIndex() < end; prev = c) + { + c = iter.next(); + if (c == CharacterIterator.DONE) + break; + int type = Character.getType(c); + + // Break after paragraph separators. + if (type == Character.PARAGRAPH_SEPARATOR) + break; + + // Now we need some lookahead. + char ahead = iter.next(); + iter.previous(); + if (ahead == CharacterIterator.DONE) + break; + int aheadType = Character.getType(ahead); + + if (aheadType != Character.NON_SPACING_MARK + && ! isLowSurrogate (ahead) + && ! isLVT (ahead)) + break; + if (! isLVT (c) && isLVT (ahead)) + break; + if (isL (c) && ! isLVT (ahead) + && aheadType != Character.NON_SPACING_MARK) + break; + if (isV (c) && ! isV (ahead) && !isT (ahead) + && aheadType != Character.NON_SPACING_MARK) + break; + if (isT (c) && ! isT (ahead) + && aheadType != Character.NON_SPACING_MARK) + break; + + if (! isHighSurrogate (c) && isLowSurrogate (ahead)) + break; + if (isHighSurrogate (c) && ! isLowSurrogate (ahead)) + break; + if (! isHighSurrogate (prev) && isLowSurrogate (c)) + break; + } + + return iter.getIndex(); + } + + public int previous () + { + if (iter.getIndex() == iter.getBeginIndex()) + return DONE; + + int start = iter.getBeginIndex(); + while (iter.getIndex() >= iter.getBeginIndex()) + { + char c = iter.previous(); + if (c == CharacterIterator.DONE) + break; + int type = Character.getType(c); + + if (type != Character.NON_SPACING_MARK + && ! isLowSurrogate (c) + && ! isLVT (c)) + break; + + // Now we need some lookahead. + char ahead = iter.previous(); + if (ahead == CharacterIterator.DONE) + { + iter.next(); + break; + } + char ahead2 = iter.previous(); + iter.next(); + iter.next(); + if (ahead2 == CharacterIterator.DONE) + break; + int aheadType = Character.getType(ahead); + + if (aheadType == Character.PARAGRAPH_SEPARATOR) + break; + + if (isLVT (c) && ! isLVT (ahead)) + break; + if (! isLVT (c) && type != Character.NON_SPACING_MARK + && isL (ahead)) + break; + if (! isV (c) && ! isT (c) && type != Character.NON_SPACING_MARK + && isV (ahead)) + break; + if (! isT (c) && type != Character.NON_SPACING_MARK + && isT (ahead)) + break; + + if (isLowSurrogate (c) && ! isHighSurrogate (ahead)) + break; + if (! isLowSurrogate (c) && isHighSurrogate (ahead)) + break; + if (isLowSurrogate (ahead) && ! isHighSurrogate (ahead2)) + break; + } + + return iter.getIndex(); + } +} diff --git a/libjava/gnu/gcj/text/LineBreakIterator.java b/libjava/gnu/gcj/text/LineBreakIterator.java new file mode 100644 index 0000000..b42881c --- /dev/null +++ b/libjava/gnu/gcj/text/LineBreakIterator.java @@ -0,0 +1,168 @@ +// Default word BreakIterator. + +/* Copyright (C) 1999 Cygnus Solutions + + This file is part of libgcj. + +This software is copyrighted work licensed under the terms of the +Libgcj License. Please consult the file "LIBGCJ_LICENSE" for +details. */ + +package gnu.gcj.text; + +import java.text.BreakIterator; +import java.text.CharacterIterator; + +/** + * @author Tom Tromey <tromey@cygnus.com> + * @date March 22, 1999 + * Written using The Unicode Standard, Version 2.0. + */ + +public class LineBreakIterator extends BaseBreakIterator +{ + public Object clone () + { + return new LineBreakIterator (this); + } + + public LineBreakIterator () + { + iter = null; + } + + private LineBreakIterator (LineBreakIterator other) + { + iter = (CharacterIterator) other.iter.clone(); + } + + // Some methods to tell us different properties of characters. + private final boolean isNb (char c) + { + return (c == 0x00a0 // NO-BREAK SPACE + || c == 0x2011 // NON-BREAKING HYPHEN + || c == 0xfeff); // ZERO WITH NO-BREAK SPACE + } + private final boolean isClose (int type) + { + return (type == Character.END_PUNCTUATION + // Unicode book says "comma, period, ...", which I take to + // mean "Po" class. + || type == Character.OTHER_PUNCTUATION); + } + private final boolean isIdeo (char c) + { + return (c >= 0x3040 && c <= 0x309f // Hiragana + || c >= 0x30a0 && c <= 0x30ff // Katakana + || c >= 0x4e00 && c <= 0x9fff // Han + || c >= 0x3100 && c <= 0x312f); // Bopomofo + } + + public int next () + { + int end = iter.getEndIndex(); + if (iter.getIndex() == end) + return DONE; + + while (iter.getIndex() < end) + { + char c = iter.current(); + int type = Character.getType(c); + + char n = iter.next(); + + if (n == CharacterIterator.DONE + || type == Character.PARAGRAPH_SEPARATOR + || type == Character.LINE_SEPARATOR) + break; + + // Handle two cases where we must scan for non-spacing marks. + int start = iter.getIndex(); + if (type == Character.SPACE_SEPARATOR + || type == Character.START_PUNCTUATION + || isIdeo (c)) + { + while (n != CharacterIterator.DONE + && Character.getType(n) == Character.NON_SPACING_MARK) + n = iter.next(); + if (n == CharacterIterator.DONE) + break; + + if (type == Character.SPACE_SEPARATOR) + { + int nt = Character.getType(n); + if (nt != Character.NON_SPACING_MARK + && nt != Character.SPACE_SEPARATOR + && ! isNb (n)) + break; + } + else if (type == Character.START_PUNCTUATION) + { + if (isIdeo (n)) + { + // Open punctuation followed by non spacing marks + // and then ideograph does not have a break in + // it. So skip all this. + start = iter.getIndex(); + } + } + else + { + // Ideograph preceded this character. + if (isClose (Character.getType(n))) + break; + } + } + iter.setIndex(start); + } + + return iter.getIndex(); + } + + public int previous () + { + int start = iter.getBeginIndex(); + if (iter.getIndex() == start) + return DONE; + + while (iter.getIndex() >= start) + { + char c = iter.previous(); + if (c == CharacterIterator.DONE) + break; + int type = Character.getType(c); + + char n = iter.previous(); + if (n == CharacterIterator.DONE) + break; + iter.next(); + + int nt = Character.getType(n); + // Break after paragraph separators. + if (nt == Character.PARAGRAPH_SEPARATOR + || nt == Character.LINE_SEPARATOR) + break; + + // Skip non-spacing marks. + int init = iter.getIndex(); + while (n != CharacterIterator.DONE && nt == Character.NON_SPACING_MARK) + { + n = iter.previous(); + nt = Character.getType(n); + } + + if (nt == Character.SPACE_SEPARATOR + && type != Character.SPACE_SEPARATOR + && type != Character.NON_SPACING_MARK + && ! isNb (c)) + break; + if (! isClose (type) && isIdeo (n)) + break; + if (isIdeo (c) && nt != Character.START_PUNCTUATION) + break; + iter.setIndex(init); + } + + return iter.getIndex(); + } +} diff --git a/libjava/gnu/gcj/text/LocaleData_en.java b/libjava/gnu/gcj/text/LocaleData_en.java new file mode 100644 index 0000000..df5a0de --- /dev/null +++ b/libjava/gnu/gcj/text/LocaleData_en.java @@ -0,0 +1,75 @@ +// Generic English locale data for java.text. + +/* Copyright (C) 1999 Cygnus Solutions + + This file is part of libgcj. + +This software is copyrighted work licensed under the terms of the +Libgcj License. Please consult the file "LIBGCJ_LICENSE" for +details. */ + +package gnu.gcj.text; + +import java.util.ListResourceBundle; + +/** + * @author Tom Tromey <tromey@cygnus.com> + * @date March 4, 1999 + */ + +public final class LocaleData_en extends ListResourceBundle +{ + // These are for DateFormatSymbols. + static final String[] ampmsDefault = {"AM", "PM" }; + static final String[] erasDefault = {"BC", "AD" }; + static final String localPatternCharsDefault = "GyMdkHmsSEDFwWahKz"; + static final String[] monthsDefault = { + "January", "February", "March", "April", "May", "June", + "July", "August", "September", "October", "November", "December", "" + }; + static final String[] shortMonthsDefault = { + "Jan", "Feb", "Mar", "Apr", "May", "Jun", + "Jul", "Aug", "Sep", "Oct", "Nov", "Dec", "" + }; + static final String[] shortWeekdaysDefault = { + "", "Sun", "Mon", "Tue", "Wed", "Thu", "Fri", "Sat" + }; + static final String[] weekdaysDefault = { + "", "Sunday", "Monday", "Tuesday", + "Wednesday", "Thursday", "Friday", "Saturday" + }; + + private static final Object[][] contents = + { + // These are for DecimalFormatSymbols. + { "decimalSeparator", "." }, + { "digit", "#" }, + { "exponential", "E" }, + { "groupingSeparator", "," }, + { "infinity", "\u221e" }, + { "minusSign", "-" }, + { "NaN", "\ufffd" }, + { "patternSeparator", ";" }, + { "percent", "%" }, + { "perMill", "\u2030" }, + { "zeroDigit", "0" }, + + // These are for NumberFormat. + { "numberFormat", "#,##0.###" }, + { "percentFormat", "#,##0%" }, + + // These are for DateFormatSymbols. + { "ampm", ampmsDefault }, + { "eras", erasDefault }, + { "datePatternChars", localPatternCharsDefault }, + { "months", monthsDefault }, + { "shortMonths", shortMonthsDefault }, + { "shortWeekdays", shortWeekdaysDefault }, + { "weekdays", weekdaysDefault } + }; + + protected Object[][] getContents () + { + return contents; + } +} diff --git a/libjava/gnu/gcj/text/LocaleData_en_US.java b/libjava/gnu/gcj/text/LocaleData_en_US.java new file mode 100644 index 0000000..a0f0209 --- /dev/null +++ b/libjava/gnu/gcj/text/LocaleData_en_US.java @@ -0,0 +1,71 @@ +// US English locale data for java.text. + +/* Copyright (C) 1999 Cygnus Solutions + + This file is part of libgcj. + +This software is copyrighted work licensed under the terms of the +Libgcj License. Please consult the file "LIBGCJ_LICENSE" for +details. */ + +package gnu.gcj.text; + +import java.util.ListResourceBundle; + +/** + * @author Tom Tromey <tromey@cygnus.com> + * @date March 4, 1999 + */ + +public final class LocaleData_en_US extends ListResourceBundle +{ + // These are for DateFormatSymbols. + static String[][] zoneStringsDefault = { + { "PST", "Pacific Standard Time", "PST", + /**/ "Pacific Daylight Time", "PDT", "San Francisco" }, + { "MST", "Mountain Standard Time", "MST", + /**/ "Mountain Daylight Time", "MDT", "Denver" }, + { "PNT", "Mountain Standard Time", "MST", + /**/ "Mountain Standard Time", "MST", "Phoenix" }, + { "CST", "Central Standard Time", "CST", + /**/ "Central Daylight Time", "CDT", "Chicago" }, + { "EST", "Eastern Standard Time", "EST", + /**/ "Eastern Daylight Time", "EDT", "Boston" }, + { "IET", "Eastern Standard Time", "EST", + /**/ "Eastern Standard Time", "EST", "Indianapolis" }, + { "PRT", "Atlantic Standard Time", "AST", + /**/ "Atlantic Daylight Time", "ADT", "Halifax" }, + { "HST", "Hawaii Standard Time", "HST", + /**/ "Hawaii Daylight Time", "HDT", "Honolulu" }, + { "AST", "Alaska Standard Time", "AST", + /**/ "Alaska Daylight Time", "ADT", "Anchorage" } + }; + + private static final Object[][] contents = + { + // These are for DecimalFormatSymbols. + { "currency", "$" }, + { "intlCurrencySymbol", "$" }, // FIXME? + + // These are for NumberFormat. + { "currencyFormat", "$#,##0.00;($#,##0.00)" }, + + // These are for DateFormatSymbols. + { "zoneStrings", zoneStringsDefault }, + + // These are for DateFormat. + { "shortDateFormat", "M/d/yy" }, // Java's Y2K bug. + { "mediumDateFormat", "d-MMM-yy" }, + { "longDateFormat", "MMMM d, yyyy" }, + { "fullDateFormat", "EEEE MMMM d, yyyy G" }, + { "shortTimeFormat", "h:mm a" }, + { "mediumTimeFormat", "h:mm:ss a" }, + { "longTimeFormat", "h:mm:ss a z" }, + { "fullTimeFormat", "h:mm:ss;S 'o''clock' a z" } + }; + + protected Object[][] getContents () + { + return contents; + } +} diff --git a/libjava/gnu/gcj/text/SentenceBreakIterator.java b/libjava/gnu/gcj/text/SentenceBreakIterator.java new file mode 100644 index 0000000..8668087 --- /dev/null +++ b/libjava/gnu/gcj/text/SentenceBreakIterator.java @@ -0,0 +1,226 @@ +// Default sentence BreakIterator. + +/* Copyright (C) 1999 Cygnus Solutions + + This file is part of libgcj. + +This software is copyrighted work licensed under the terms of the +Libgcj License. Please consult the file "LIBGCJ_LICENSE" for +details. */ + +package gnu.gcj.text; + +import java.text.BreakIterator; +import java.text.CharacterIterator; + +/** + * @author Tom Tromey <tromey@cygnus.com> + * @date March 23, 1999 + * Written using The Unicode Standard, Version 2.0. + */ + +public class SentenceBreakIterator extends BaseBreakIterator +{ + public Object clone () + { + return new SentenceBreakIterator (this); + } + + public SentenceBreakIterator () + { + iter = null; + } + + private SentenceBreakIterator (SentenceBreakIterator other) + { + iter = (CharacterIterator) other.iter.clone(); + } + + public int next () + { + int end = iter.getEndIndex(); + if (iter.getIndex() == end) + return DONE; + + while (iter.getIndex() < end) + { + char c = iter.current(); + if (c == CharacterIterator.DONE) + break; + int type = Character.getType(c); + + char n = iter.next(); + if (n == CharacterIterator.DONE) + break; + + // Always break after paragraph separator. + if (type == Character.PARAGRAPH_SEPARATOR) + break; + + if (c == '!' || c == '?') + { + // Skip close punctuation. + while (n != CharacterIterator.DONE + && Character.getType(n) == Character.END_PUNCTUATION) + n = iter.next(); + // Skip spaces. + while (n != CharacterIterator.DONE + && Character.getType(n) == Character.SPACE_SEPARATOR) + n = iter.next(); + // Skip optional paragraph separator. + if (n != CharacterIterator.DONE + && Character.getType(n) == Character.PARAGRAPH_SEPARATOR) + n = iter.next(); + + // There's always a break somewhere after `!' or `?'. + break; + } + + if (c == '.') + { + int save = iter.getIndex(); + // Skip close punctuation. + while (n != CharacterIterator.DONE + && Character.getType(n) == Character.END_PUNCTUATION) + n = iter.next(); + // Skip spaces. We keep count because we need at least + // one for this period to represent a terminator. + int spcount = 0; + while (n != CharacterIterator.DONE + && Character.getType(n) == Character.SPACE_SEPARATOR) + { + n = iter.next(); + ++spcount; + } + if (spcount > 0) + { + int save2 = iter.getIndex(); + // Skip over open puncutation. + while (n != CharacterIterator.DONE + && Character.getType(n) == Character.START_PUNCTUATION) + n = iter.next(); + // Next character must not be lower case. + if (n == CharacterIterator.DONE + || ! Character.isLowerCase(n)) + { + iter.setIndex(save2); + break; + } + } + iter.setIndex(save); + } + } + + return iter.getIndex(); + } + + private final int previous_internal () + { + int start = iter.getBeginIndex(); + if (iter.getIndex() == start) + return DONE; + + while (iter.getIndex() >= start) + { + char c = iter.previous(); + if (c == CharacterIterator.DONE) + break; + + char n = iter.previous(); + if (n == CharacterIterator.DONE) + break; + iter.next(); + int nt = Character.getType(n); + + if (! Character.isLowerCase(c) + && (nt == Character.START_PUNCTUATION + || nt == Character.SPACE_SEPARATOR)) + { + int save = iter.getIndex(); + int save_nt = nt; + char save_n = n; + // Skip open punctuation. + while (n != CharacterIterator.DONE + && Character.getType(n) == Character.START_PUNCTUATION) + n = iter.previous(); + if (n == CharacterIterator.DONE) + break; + if (Character.getType(n) == Character.SPACE_SEPARATOR) + { + // Must have at least once space after the `.'. + int save2 = iter.getIndex(); + while (n != CharacterIterator.DONE + && Character.getType(n) == Character.SPACE_SEPARATOR) + n = iter.previous(); + // Skip close punctuation. + while (n != CharacterIterator.DONE + && Character.getType(n) == Character.END_PUNCTUATION) + n = iter.previous(); + if (n == CharacterIterator.DONE || n == '.') + { + // Communicate location of actual end. + period = iter.getIndex(); + iter.setIndex(save2); + break; + } + } + iter.setIndex(save); + nt = save_nt; + n = save_n; + } + + if (nt == Character.PARAGRAPH_SEPARATOR) + { + // Communicate location of actual end. + period = iter.getIndex(); + break; + } + else if (nt == Character.SPACE_SEPARATOR + || nt == Character.END_PUNCTUATION) + { + int save = iter.getIndex(); + // Skip spaces. + while (n != CharacterIterator.DONE + && Character.getType(n) == Character.SPACE_SEPARATOR) + n = iter.previous(); + // Skip close punctuation. + while (n != CharacterIterator.DONE + && Character.getType(n) == Character.END_PUNCTUATION) + n = iter.previous(); + int here = iter.getIndex(); + iter.setIndex(save); + if (n == CharacterIterator.DONE || n == '!' || n == '?') + { + // Communicate location of actual end. + period = here; + break; + } + } + else if (n == '!' || n == '?') + { + // Communicate location of actual end. + period = iter.getIndex(); + break; + } + } + + return iter.getIndex(); + } + + public int previous () + { + // We want to skip over the first sentence end to the second one. + // However, at the end of the string we want the first end. + int here = iter.getIndex(); + period = here; + int first = previous_internal (); + if (here == iter.getEndIndex() || first == DONE) + return first; + iter.setIndex(period); + return previous_internal (); + } + + // This is used for communication between previous and + // previous_internal. + private int period; +} diff --git a/libjava/gnu/gcj/text/WordBreakIterator.java b/libjava/gnu/gcj/text/WordBreakIterator.java new file mode 100644 index 0000000..b2fcb97 --- /dev/null +++ b/libjava/gnu/gcj/text/WordBreakIterator.java @@ -0,0 +1,224 @@ +// Default word BreakIterator. + +/* Copyright (C) 1999 Cygnus Solutions + + This file is part of libgcj. + +This software is copyrighted work licensed under the terms of the +Libgcj License. Please consult the file "LIBGCJ_LICENSE" for +details. */ + +package gnu.gcj.text; + +import java.text.BreakIterator; +import java.text.CharacterIterator; + +/** + * @author Tom Tromey <tromey@cygnus.com> + * @date March 22, 1999 + * Written using The Unicode Standard, Version 2.0. + */ + +public class WordBreakIterator extends BaseBreakIterator +{ + public Object clone () + { + return new WordBreakIterator (this); + } + + public WordBreakIterator () + { + iter = null; + } + + private WordBreakIterator (WordBreakIterator other) + { + iter = (CharacterIterator) other.iter.clone(); + } + + // Some methods to tell us different properties of characters. + private final boolean isHira (char c) + { + return c >= 0x3040 && c <= 0x309f; + } + private final boolean isKata (char c) + { + return c >= 0x30a0 && c <= 0x30ff; + } + private final boolean isHan (char c) + { + return c >= 0x4e00 && c <= 0x9fff; + } + + public int next () + { + int end = iter.getEndIndex(); + if (iter.getIndex() == end) + return DONE; + + while (iter.getIndex() < end) + { + char c = iter.current(); + if (c == CharacterIterator.DONE) + break; + int type = Character.getType(c); + + char n = iter.next(); + if (n == CharacterIterator.DONE) + break; + + // Break after paragraph separators. + if (type == Character.PARAGRAPH_SEPARATOR + || type == Character.LINE_SEPARATOR) + break; + + // Break between letters and non-letters. + // FIXME: we treat apostrophe as part of a word. This + // is an English-ism. + boolean is_letter = Character.isLetter(c); + if (c != '\'' && ! is_letter && type != Character.NON_SPACING_MARK + && Character.isLetter(n)) + break; + + // Always break after certain symbols, such as punctuation. + // This heuristic is derived from hints in the JCL book and is + // not part of Unicode. It seems to be right, however. + // FIXME: we treat apostrophe as part of a word. This + // is an English-ism. + if (c != '\'' + && (type == Character.DASH_PUNCTUATION + || type == Character.START_PUNCTUATION + || type == Character.END_PUNCTUATION + || type == Character.CONNECTOR_PUNCTUATION + || type == Character.OTHER_PUNCTUATION + || type == Character.MATH_SYMBOL + || type == Character.CURRENCY_SYMBOL + || type == Character.MODIFIER_SYMBOL + || type == Character.OTHER_SYMBOL + || type == Character.FORMAT + || type == Character.CONTROL)) + break; + + boolean is_hira = isHira (c); + boolean is_kata = isKata (c); + boolean is_han = isHan (c); + + // Special case Japanese. + if (! is_hira && ! is_kata && ! is_han + && type != Character.NON_SPACING_MARK + && (isHira (n) || isKata (n) || isHan (n))) + break; + + if (is_hira || is_kata || is_han || is_letter) + { + // Now we need to do some lookahead. We might need to do + // quite a bit of lookahead, so we save our position and + // restore it later. + int save = iter.getIndex(); + // Skip string of non spacing marks. + while (n != CharacterIterator.DONE + && Character.getType(n) == Character.NON_SPACING_MARK) + n = iter.next(); + if (n == CharacterIterator.DONE) + break; + if ((is_hira && ! isHira (n)) + || (is_kata && ! isHira (n) && ! isKata (n)) + || (is_han && ! isHira (n) && ! isHan (n)) + // FIXME: we treat apostrophe as part of a word. This + // is an English-ism. + || (is_letter && ! Character.isLetter(n) && n != '\'')) + break; + iter.setIndex(save); + } + } + + return iter.getIndex(); + } + + public int previous () + { + int start = iter.getBeginIndex(); + if (iter.getIndex() == start) + return DONE; + + while (iter.getIndex() >= start) + { + char c = iter.previous(); + if (c == CharacterIterator.DONE) + break; + + boolean is_hira = isHira (c); + boolean is_kata = isKata (c); + boolean is_han = isHan (c); + boolean is_letter = Character.isLetter(c); + + char n = iter.previous(); + if (n == CharacterIterator.DONE) + break; + iter.next(); + int type = Character.getType(n); + // Break after paragraph separators. + if (type == Character.PARAGRAPH_SEPARATOR + || type == Character.LINE_SEPARATOR) + break; + + // Break between letters and non-letters. + // FIXME: we treat apostrophe as part of a word. This + // is an English-ism. + if (n != '\'' && ! Character.isLetter(n) + && type != Character.NON_SPACING_MARK + && is_letter) + break; + + // Always break after certain symbols, such as punctuation. + // This heuristic is derived from hints in the JCL book and is + // not part of Unicode. It seems to be right, however. + // FIXME: we treat apostrophe as part of a word. This + // is an English-ism. + if (n != '\'' + && (type == Character.DASH_PUNCTUATION + || type == Character.START_PUNCTUATION + || type == Character.END_PUNCTUATION + || type == Character.CONNECTOR_PUNCTUATION + || type == Character.OTHER_PUNCTUATION + || type == Character.MATH_SYMBOL + || type == Character.CURRENCY_SYMBOL + || type == Character.MODIFIER_SYMBOL + || type == Character.OTHER_SYMBOL + || type == Character.FORMAT + || type == Character.CONTROL)) + break; + + // Special case Japanese. + if ((is_hira || is_kata || is_han) + && ! isHira (n) && ! isKata (n) && ! isHan (n) + && type != Character.NON_SPACING_MARK) + break; + + // We might have to skip over non spacing marks to see what's + // on the other side. + if (! is_hira || (! is_letter && c != '\'')) + { + int save = iter.getIndex(); + while (n != CharacterIterator.DONE + && Character.getType(n) == Character.NON_SPACING_MARK) + n = iter.previous(); + iter.setIndex(save); + // This is a strange case: a bunch of non-spacing marks at + // the beginning. We treat the current location as a word + // break. + if (n == CharacterIterator.DONE) + break; + if ((isHira (n) && ! is_hira) + || (isKata (n) && ! is_hira && ! is_kata) + || (isHan (n) && ! is_hira && ! is_han) + // FIXME: we treat apostrophe as part of a word. This + // is an English-ism. + || (! is_letter && c != '\'' && Character.isLetter(n))) + break; + } + } + + return iter.getIndex(); + } +} |