Add "Remove" transliterator
X-SVN-Rev: 4442
This commit is contained in:
Alan Liu
parent
73d42ad240
commit
9f8d255013
@ -5,8 +5,8 @@
|
||||
*******************************************************************************
|
||||
*
|
||||
* $Source: /xsrl/Nsvn/icu/icu4j/src/com/ibm/icu/dev/test/translit/TransliteratorTest.java,v $
|
||||
* $Date: 2001/04/03 18:21:57 $
|
||||
* $Revision: 1.33 $
|
||||
* $Date: 2001/04/04 18:07:08 $
|
||||
* $Revision: 1.34 $
|
||||
*
|
||||
*****************************************************************************************
|
||||
*/
|
||||
@ -742,6 +742,11 @@ public class TransliteratorTest extends TestFmwk {
|
||||
"Null[abc]",
|
||||
"xyz",
|
||||
"xyz",
|
||||
|
||||
"Remove[abc]",
|
||||
"Remove[abc]",
|
||||
"axbycz",
|
||||
"xyz",
|
||||
};
|
||||
|
||||
for (int i=0; i<DATA.length; i+=4) {
|
||||
@ -767,6 +772,15 @@ public class TransliteratorTest extends TestFmwk {
|
||||
}
|
||||
}
|
||||
|
||||
/**
|
||||
* Test the "Remove" transliterator.
|
||||
*/
|
||||
public void TestRemove() {
|
||||
Transliterator t = Transliterator.getInstance("Remove[aeiou]");
|
||||
expect(t, "The quick brown fox.",
|
||||
"Th qck brwn fx.");
|
||||
}
|
||||
|
||||
//======================================================================
|
||||
// Support methods
|
||||
//======================================================================
|
||||
|
||||
74
icu4j/src/com/ibm/icu/text/RemoveTransliterator.java
Executable file
74
icu4j/src/com/ibm/icu/text/RemoveTransliterator.java
Executable file
@ -0,0 +1,74 @@
|
||||
/*
|
||||
*******************************************************************************
|
||||
* Copyright (C) 1996-2000, International Business Machines Corporation and *
|
||||
* others. All Rights Reserved. *
|
||||
*******************************************************************************
|
||||
*
|
||||
* $Source: /xsrl/Nsvn/icu/icu4j/src/com/ibm/icu/text/RemoveTransliterator.java,v $
|
||||
* $Date: 2001/04/04 18:06:53 $
|
||||
* $Revision: 1.1 $
|
||||
*
|
||||
*****************************************************************************************
|
||||
*/
|
||||
package com.ibm.text;
|
||||
import java.util.*;
|
||||
|
||||
/**
|
||||
* A transliterator that removes characters. This is useful in conjunction
|
||||
* with a filter.
|
||||
*/
|
||||
public class RemoveTransliterator extends Transliterator {
|
||||
|
||||
/**
|
||||
* Package accessible ID for this transliterator.
|
||||
*/
|
||||
static String _ID = "Remove";
|
||||
|
||||
/**
|
||||
* Constructs a transliterator.
|
||||
*/
|
||||
public RemoveTransliterator() {
|
||||
super(_ID, null);
|
||||
}
|
||||
|
||||
/**
|
||||
* Implements {@link Transliterator#handleTransliterate}.
|
||||
*/
|
||||
protected void handleTransliterate(Replaceable text,
|
||||
Position offsets, boolean incremental) {
|
||||
// Find runs of unfiltered characters and replace them with the
|
||||
// empty string. This loop has been optimized to what is probably
|
||||
// an unnecessary degree.
|
||||
String empty = "";
|
||||
int start = offsets.start;
|
||||
for (;;) {
|
||||
// Find first unfiltered character, if any
|
||||
while (start < offsets.limit &&
|
||||
filteredCharAt(text, start) == '\uFFFE') {
|
||||
++start;
|
||||
}
|
||||
if (start >= offsets.limit) {
|
||||
break;
|
||||
}
|
||||
|
||||
// assert(start < offsets.limit &&
|
||||
// filteredCharAt(text, start) != 0xFFFE);
|
||||
|
||||
// Find last unfiltered character
|
||||
int limit = start+1; // sic: +1
|
||||
while (limit < offsets.limit &&
|
||||
filteredCharAt(text, limit) != '\uFFFE') {
|
||||
++limit;
|
||||
}
|
||||
|
||||
// assert(start < limit);
|
||||
|
||||
// Remove characters
|
||||
text.replace(start, limit, empty);
|
||||
limit -= start; // limit <= deleted length
|
||||
offsets.contextLimit -= limit;
|
||||
offsets.limit -= limit;
|
||||
}
|
||||
offsets.start = start;
|
||||
}
|
||||
}
|
||||
@ -5,8 +5,8 @@
|
||||
*******************************************************************************
|
||||
*
|
||||
* $Source: /xsrl/Nsvn/icu/icu4j/src/com/ibm/icu/text/Transliterator.java,v $
|
||||
* $Date: 2001/03/31 01:31:13 $
|
||||
* $Revision: 1.30 $
|
||||
* $Date: 2001/04/04 18:06:25 $
|
||||
* $Revision: 1.31 $
|
||||
*
|
||||
*****************************************************************************************
|
||||
*/
|
||||
@ -240,7 +240,7 @@ import com.ibm.text.resources.ResourceReader;
|
||||
* <p>Copyright © IBM Corporation 1999. All rights reserved.
|
||||
*
|
||||
* @author Alan Liu
|
||||
* @version $RCSfile: Transliterator.java,v $ $Revision: 1.30 $ $Date: 2001/03/31 01:31:13 $
|
||||
* @version $RCSfile: Transliterator.java,v $ $Revision: 1.31 $ $Date: 2001/04/04 18:06:25 $
|
||||
*/
|
||||
public abstract class Transliterator {
|
||||
/**
|
||||
@ -794,7 +794,8 @@ public abstract class Transliterator {
|
||||
if (direction == REVERSE) {
|
||||
int i = id.indexOf('-');
|
||||
if (i < 0) {
|
||||
if (!id.equals(NullTransliterator._ID)) {
|
||||
if (!id.equals(NullTransliterator._ID) &&
|
||||
!id.equals(RemoveTransliterator._ID)) {
|
||||
throw new IllegalArgumentException("No inverse for: "
|
||||
+ id);
|
||||
}
|
||||
@ -1083,5 +1084,7 @@ public abstract class Transliterator {
|
||||
UnicodeToHexTransliterator.class, null);
|
||||
registerClass(NullTransliterator._ID,
|
||||
NullTransliterator.class, null);
|
||||
registerClass(RemoveTransliterator._ID,
|
||||
RemoveTransliterator.class, null);
|
||||
}
|
||||
}
|
||||
|
||||
@ -5,8 +5,8 @@
|
||||
*******************************************************************************
|
||||
*
|
||||
* $Source: /xsrl/Nsvn/icu/icu4j/src/com/ibm/test/translit/Attic/TransliteratorTest.java,v $
|
||||
* $Date: 2001/04/03 18:21:57 $
|
||||
* $Revision: 1.33 $
|
||||
* $Date: 2001/04/04 18:07:08 $
|
||||
* $Revision: 1.34 $
|
||||
*
|
||||
*****************************************************************************************
|
||||
*/
|
||||
@ -742,6 +742,11 @@ public class TransliteratorTest extends TestFmwk {
|
||||
"Null[abc]",
|
||||
"xyz",
|
||||
"xyz",
|
||||
|
||||
"Remove[abc]",
|
||||
"Remove[abc]",
|
||||
"axbycz",
|
||||
"xyz",
|
||||
};
|
||||
|
||||
for (int i=0; i<DATA.length; i+=4) {
|
||||
@ -767,6 +772,15 @@ public class TransliteratorTest extends TestFmwk {
|
||||
}
|
||||
}
|
||||
|
||||
/**
|
||||
* Test the "Remove" transliterator.
|
||||
*/
|
||||
public void TestRemove() {
|
||||
Transliterator t = Transliterator.getInstance("Remove[aeiou]");
|
||||
expect(t, "The quick brown fox.",
|
||||
"Th qck brwn fx.");
|
||||
}
|
||||
|
||||
//======================================================================
|
||||
// Support methods
|
||||
//======================================================================
|
||||
|
||||
74
icu4j/src/com/ibm/text/RemoveTransliterator.java
Executable file
74
icu4j/src/com/ibm/text/RemoveTransliterator.java
Executable file
@ -0,0 +1,74 @@
|
||||
/*
|
||||
*******************************************************************************
|
||||
* Copyright (C) 1996-2000, International Business Machines Corporation and *
|
||||
* others. All Rights Reserved. *
|
||||
*******************************************************************************
|
||||
*
|
||||
* $Source: /xsrl/Nsvn/icu/icu4j/src/com/ibm/text/Attic/RemoveTransliterator.java,v $
|
||||
* $Date: 2001/04/04 18:06:53 $
|
||||
* $Revision: 1.1 $
|
||||
*
|
||||
*****************************************************************************************
|
||||
*/
|
||||
package com.ibm.text;
|
||||
import java.util.*;
|
||||
|
||||
/**
|
||||
* A transliterator that removes characters. This is useful in conjunction
|
||||
* with a filter.
|
||||
*/
|
||||
public class RemoveTransliterator extends Transliterator {
|
||||
|
||||
/**
|
||||
* Package accessible ID for this transliterator.
|
||||
*/
|
||||
static String _ID = "Remove";
|
||||
|
||||
/**
|
||||
* Constructs a transliterator.
|
||||
*/
|
||||
public RemoveTransliterator() {
|
||||
super(_ID, null);
|
||||
}
|
||||
|
||||
/**
|
||||
* Implements {@link Transliterator#handleTransliterate}.
|
||||
*/
|
||||
protected void handleTransliterate(Replaceable text,
|
||||
Position offsets, boolean incremental) {
|
||||
// Find runs of unfiltered characters and replace them with the
|
||||
// empty string. This loop has been optimized to what is probably
|
||||
// an unnecessary degree.
|
||||
String empty = "";
|
||||
int start = offsets.start;
|
||||
for (;;) {
|
||||
// Find first unfiltered character, if any
|
||||
while (start < offsets.limit &&
|
||||
filteredCharAt(text, start) == '\uFFFE') {
|
||||
++start;
|
||||
}
|
||||
if (start >= offsets.limit) {
|
||||
break;
|
||||
}
|
||||
|
||||
// assert(start < offsets.limit &&
|
||||
// filteredCharAt(text, start) != 0xFFFE);
|
||||
|
||||
// Find last unfiltered character
|
||||
int limit = start+1; // sic: +1
|
||||
while (limit < offsets.limit &&
|
||||
filteredCharAt(text, limit) != '\uFFFE') {
|
||||
++limit;
|
||||
}
|
||||
|
||||
// assert(start < limit);
|
||||
|
||||
// Remove characters
|
||||
text.replace(start, limit, empty);
|
||||
limit -= start; // limit <= deleted length
|
||||
offsets.contextLimit -= limit;
|
||||
offsets.limit -= limit;
|
||||
}
|
||||
offsets.start = start;
|
||||
}
|
||||
}
|
||||
@ -5,8 +5,8 @@
|
||||
*******************************************************************************
|
||||
*
|
||||
* $Source: /xsrl/Nsvn/icu/icu4j/src/com/ibm/text/Attic/Transliterator.java,v $
|
||||
* $Date: 2001/03/31 01:31:13 $
|
||||
* $Revision: 1.30 $
|
||||
* $Date: 2001/04/04 18:06:25 $
|
||||
* $Revision: 1.31 $
|
||||
*
|
||||
*****************************************************************************************
|
||||
*/
|
||||
@ -240,7 +240,7 @@ import com.ibm.text.resources.ResourceReader;
|
||||
* <p>Copyright © IBM Corporation 1999. All rights reserved.
|
||||
*
|
||||
* @author Alan Liu
|
||||
* @version $RCSfile: Transliterator.java,v $ $Revision: 1.30 $ $Date: 2001/03/31 01:31:13 $
|
||||
* @version $RCSfile: Transliterator.java,v $ $Revision: 1.31 $ $Date: 2001/04/04 18:06:25 $
|
||||
*/
|
||||
public abstract class Transliterator {
|
||||
/**
|
||||
@ -794,7 +794,8 @@ public abstract class Transliterator {
|
||||
if (direction == REVERSE) {
|
||||
int i = id.indexOf('-');
|
||||
if (i < 0) {
|
||||
if (!id.equals(NullTransliterator._ID)) {
|
||||
if (!id.equals(NullTransliterator._ID) &&
|
||||
!id.equals(RemoveTransliterator._ID)) {
|
||||
throw new IllegalArgumentException("No inverse for: "
|
||||
+ id);
|
||||
}
|
||||
@ -1083,5 +1084,7 @@ public abstract class Transliterator {
|
||||
UnicodeToHexTransliterator.class, null);
|
||||
registerClass(NullTransliterator._ID,
|
||||
NullTransliterator.class, null);
|
||||
registerClass(RemoveTransliterator._ID,
|
||||
RemoveTransliterator.class, null);
|
||||
}
|
||||
}
|
||||
|
||||
Loading…
Reference in New Issue
Block a user