AuroraMiddleware/scuffed-code
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

ICU-1533 incorporate Mark's review comments; remove special Jamo handling

Browse Source 
X-SVN-Rev: 7057
This commit is contained in:
Alan Liu 2001-11-21 20:57:08 +00:00
parent 3912ae3f89
commit f61acd7ed7
2 changed files with 18 additions and 86 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

52
icu4j/src/com/ibm/icu/text/NormalizationTransliterator.java
View File

@ -13,7 +13,7 @@ import java.util.*;
/**
* @author Alan Liu
* @version $RCSfile: NormalizationTransliterator.java,v $ $Revision: 1.9 $ $Date: 2001/11/17 20:56:13 $
* @version $RCSfile: NormalizationTransliterator.java,v $ $Revision: 1.10 $ $Date: 2001/11/21 20:57:08 $
*/
class NormalizationTransliterator extends Transliterator {
@ -33,6 +33,8 @@ class NormalizationTransliterator extends Transliterator {
* effectively consider these to be cc!=0, for our purposes.
*
* From http://www.macchiato.com/utc/NFUnsafeStart-3.1.1dX.txt
*
* TODO Update this to 4 separate sets, one for each norm. form.
*/
static final UnicodeSet UNSAFE_START = new UnicodeSet("[\u09BE\u09D7\u0B3E\u0B56\u0B57\u0BBE\u0BD7\u0CC2\u0CD5-\u0CD6\u0D3E\u0D57\u0DCF\u0DDF\u0F73\u0F75\u0F81\u102E\u1161-\u1175\u11A8-\u11C2\u3133\u3135-\u3136\u313A-\u313F\u314F-\u3163\uFF9E-\uFF9F\uFFA3\uFFA5-\uFFA6\uFFAA-\uFFAF\uFFC2-\uFFC7\uFFCA-\uFFCF\uFFD2-\uFFD7\uFFDA-\uFFDC]", false);
@ -111,51 +113,15 @@ class NormalizationTransliterator extends Transliterator {
// character b, and pass everything from the start up to the
// character before b to normalizer.
if (isIncremental) {
// Wrinkle: Jamo has a combining class of zero, but we
// don't want to normalize individual Jamo one at a time
// if we're composing incrementally. If we are composing
// in incremental mode then we collect up trailing jamo
// and save them for next time.
boolean doStandardBackup = true;
if (mode.compose()) {
// As a minor optimization, if there are three or more
// trailing jamo, we let the first three through --
// these should be handled correctly.
char c;
while (limit > offsets.start &&
(c=text.charAt(limit-1)) >= 0x1100 &&
c < 0x1200) {
--limit;
}
// Characters in [limit, offsets.limit) are jamo.
// If we have at least 3 jamo, then allow them
// to be transliterated. If we have zero jamo,
// then proceed as usual.
if (limit < offsets.limit) {
if ((offsets.limit - limit) >= 3) {
limit += 3;
}
doStandardBackup = false;
}
}
// A standard backup consists of finding the last
// character with cc==0 and passing everything from the
// start up to (but not including) that character to
// normalizer. We use the UNSAFE_START set to detect
// characters with cc==0 that we want to treat as if they
// have cc!=0 (see above).
if (doStandardBackup) {
--limit;
char c;
while (limit > start &&
(UCharacter.getCombiningClass(c=text.charAt(limit)) != 0 ||
UNSAFE_START.contains(c))) {
--limit;
char c;
while (limit > start &&
(UCharacter.getCombiningClass(c=text.charAt(limit)) != 0 ||
UNSAFE_START.contains(c))) {
--limit;
}
}
}
if (limit > start) {
char chars[] = new char[limit - start];
text.getChars(start, limit, chars, 0);

52
icu4j/src/com/ibm/text/NormalizationTransliterator.java
View File

@ -13,7 +13,7 @@ import java.util.*;
/**
* @author Alan Liu
* @version $RCSfile: NormalizationTransliterator.java,v $ $Revision: 1.9 $ $Date: 2001/11/17 20:56:13 $
* @version $RCSfile: NormalizationTransliterator.java,v $ $Revision: 1.10 $ $Date: 2001/11/21 20:57:08 $
*/
class NormalizationTransliterator extends Transliterator {
@ -33,6 +33,8 @@ class NormalizationTransliterator extends Transliterator {
* effectively consider these to be cc!=0, for our purposes.
*
* From http://www.macchiato.com/utc/NFUnsafeStart-3.1.1dX.txt
*
* TODO Update this to 4 separate sets, one for each norm. form.
*/
static final UnicodeSet UNSAFE_START = new UnicodeSet("[\u09BE\u09D7\u0B3E\u0B56\u0B57\u0BBE\u0BD7\u0CC2\u0CD5-\u0CD6\u0D3E\u0D57\u0DCF\u0DDF\u0F73\u0F75\u0F81\u102E\u1161-\u1175\u11A8-\u11C2\u3133\u3135-\u3136\u313A-\u313F\u314F-\u3163\uFF9E-\uFF9F\uFFA3\uFFA5-\uFFA6\uFFAA-\uFFAF\uFFC2-\uFFC7\uFFCA-\uFFCF\uFFD2-\uFFD7\uFFDA-\uFFDC]", false);
@ -111,51 +113,15 @@ class NormalizationTransliterator extends Transliterator {
// character b, and pass everything from the start up to the
// character before b to normalizer.
if (isIncremental) {
// Wrinkle: Jamo has a combining class of zero, but we
// don't want to normalize individual Jamo one at a time
// if we're composing incrementally. If we are composing
// in incremental mode then we collect up trailing jamo
// and save them for next time.
boolean doStandardBackup = true;
if (mode.compose()) {
// As a minor optimization, if there are three or more
// trailing jamo, we let the first three through --
// these should be handled correctly.
char c;
while (limit > offsets.start &&
(c=text.charAt(limit-1)) >= 0x1100 &&
c < 0x1200) {
--limit;
}
// Characters in [limit, offsets.limit) are jamo.
// If we have at least 3 jamo, then allow them
// to be transliterated. If we have zero jamo,
// then proceed as usual.
if (limit < offsets.limit) {
if ((offsets.limit - limit) >= 3) {
limit += 3;
}
doStandardBackup = false;
}
}
// A standard backup consists of finding the last
// character with cc==0 and passing everything from the
// start up to (but not including) that character to
// normalizer. We use the UNSAFE_START set to detect
// characters with cc==0 that we want to treat as if they
// have cc!=0 (see above).
if (doStandardBackup) {
--limit;
char c;
while (limit > start &&
(UCharacter.getCombiningClass(c=text.charAt(limit)) != 0 ||
UNSAFE_START.contains(c))) {
--limit;
char c;
while (limit > start &&
(UCharacter.getCombiningClass(c=text.charAt(limit)) != 0 ||
UNSAFE_START.contains(c))) {
--limit;
}
}
}
if (limit > start) {
char chars[] = new char[limit - start];
text.getChars(start, limit, chars, 0);