mirror of https://github.com/apache/poi.git
Rewinded: StringUtil.java: Sync to REL_2 BRANCH
git-svn-id: https://svn.apache.org/repos/asf/jakarta/poi/trunk@353249 13f79535-47bb-0310-9956-ffa450edef68
This commit is contained in:
parent
d20e067403
commit
e29ec7063a
|
@ -2,7 +2,7 @@
|
||||||
* ====================================================================
|
* ====================================================================
|
||||||
* The Apache Software License, Version 1.1
|
* The Apache Software License, Version 1.1
|
||||||
*
|
*
|
||||||
* Copyright (c) 2003 The Apache Software Foundation. All rights
|
* Copyright (c) 2002 The Apache Software Foundation. All rights
|
||||||
* reserved.
|
* reserved.
|
||||||
*
|
*
|
||||||
* Redistribution and use in source and binary forms, with or without
|
* Redistribution and use in source and binary forms, with or without
|
||||||
|
@ -55,32 +55,38 @@
|
||||||
package org.apache.poi.util;
|
package org.apache.poi.util;
|
||||||
|
|
||||||
import java.io.UnsupportedEncodingException;
|
import java.io.UnsupportedEncodingException;
|
||||||
|
|
||||||
import java.text.NumberFormat;
|
import java.text.NumberFormat;
|
||||||
import java.text.FieldPosition;
|
import java.text.FieldPosition;
|
||||||
import java.util.Arrays;
|
|
||||||
/**
|
/**
|
||||||
* Title: String Utility Description: Collection of string handling utilities
|
* Title: String Utility Description: Collection of string handling utilities
|
||||||
*
|
*
|
||||||
|
* Now it is quite confusing: the method pairs, in which
|
||||||
|
* one of them write data and other read written data are:
|
||||||
|
* putUncompressedUnicodeHigh and getFromUnicode
|
||||||
|
* putUncompressedUnicode and getFromUnicodeHigh
|
||||||
*
|
*
|
||||||
*@author Andrew C. Oliver
|
*@author Andrew C. Oliver
|
||||||
*@author Sergei Kozello (sergeikozello at mail.ru)
|
*@author Sergei Kozello (sergeikozello at mail.ru)
|
||||||
*@author Toshiaki Kamoshida (kamoshida.toshiaki at future dot co dot jp)
|
|
||||||
*@created May 10, 2002
|
*@created May 10, 2002
|
||||||
*@version 1.0
|
*@version 1.0
|
||||||
*/
|
*/
|
||||||
|
|
||||||
public class StringUtil {
|
public class StringUtil {
|
||||||
private final static String ENCODING = "ISO-8859-1";
|
|
||||||
|
private final static String ENCODING="ISO-8859-1";
|
||||||
/**
|
/**
|
||||||
* Constructor for the StringUtil object
|
* Constructor for the StringUtil object
|
||||||
*/
|
*/
|
||||||
private StringUtil() {
|
private StringUtil() { }
|
||||||
}
|
|
||||||
|
|
||||||
/**
|
/**
|
||||||
* given a byte array of 16-bit unicode characters, compress to 8-bit and
|
* given a byte array of 16-bit unicode characters, compress to 8-bit and
|
||||||
* return a string
|
* return a string
|
||||||
*
|
*
|
||||||
* { 0x16, 0x00 } -0x16
|
* { 0x16, 0x00 } -> 0x16
|
||||||
*
|
*
|
||||||
*@param string the byte array to be converted
|
*@param string the byte array to be converted
|
||||||
*@param offset the initial offset into the
|
*@param offset the initial offset into the
|
||||||
|
@ -96,11 +102,11 @@ public class StringUtil {
|
||||||
* length)
|
* length)
|
||||||
*@len the length of the final string
|
*@len the length of the final string
|
||||||
*/
|
*/
|
||||||
public static String getFromUnicodeLE(
|
|
||||||
final byte[] string,
|
public static String getFromUnicodeHigh(final byte[] string,
|
||||||
final int offset,
|
final int offset, final int len)
|
||||||
final int len)
|
|
||||||
throws ArrayIndexOutOfBoundsException, IllegalArgumentException {
|
throws ArrayIndexOutOfBoundsException, IllegalArgumentException {
|
||||||
|
|
||||||
if ((offset < 0) || (offset >= string.length)) {
|
if ((offset < 0) || (offset >= string.length)) {
|
||||||
throw new ArrayIndexOutOfBoundsException("Illegal offset");
|
throw new ArrayIndexOutOfBoundsException("Illegal offset");
|
||||||
}
|
}
|
||||||
|
@ -108,34 +114,39 @@ public class StringUtil {
|
||||||
throw new IllegalArgumentException("Illegal length");
|
throw new IllegalArgumentException("Illegal length");
|
||||||
}
|
}
|
||||||
|
|
||||||
try {
|
char[] chars = new char[ len ];
|
||||||
return new String(string, offset, len * 2, "UTF-16LE");
|
for ( int i = 0; i < chars.length; i++ ) {
|
||||||
} catch (UnsupportedEncodingException e) {
|
chars[i] = (char)( string[ offset + ( 2*i ) ] & 0xFF |
|
||||||
throw new InternalError(); /*unreachable*/
|
( string[ offset + ( 2*i+1 ) ] << 8 ) );
|
||||||
}
|
}
|
||||||
|
|
||||||
|
return new String( chars );
|
||||||
}
|
}
|
||||||
|
|
||||||
|
|
||||||
/**
|
/**
|
||||||
* given a byte array of 16-bit unicode characters, compress to 8-bit and
|
* given a byte array of 16-bit unicode characters, compress to 8-bit and
|
||||||
* return a string
|
* return a string
|
||||||
*
|
*
|
||||||
* { 0x16, 0x00 } -0x16
|
* { 0x16, 0x00 } -> 0x16
|
||||||
*
|
*
|
||||||
*@param string the byte array to be converted
|
*@param string the byte array to be converted
|
||||||
*@return the converted string
|
*@return the converted string
|
||||||
*/
|
*/
|
||||||
public static String getFromUnicodeLE(final byte[] string) {
|
|
||||||
return getFromUnicodeLE(string, 0, string.length / 2);
|
public static String getFromUnicodeHigh( final byte[] string ) {
|
||||||
|
return getFromUnicodeHigh( string, 0, string.length / 2 );
|
||||||
}
|
}
|
||||||
|
|
||||||
|
|
||||||
/**
|
/**
|
||||||
* given a byte array of 16-bit unicode characters, compress to 8-bit and
|
* given a byte array of 16-bit unicode characters, compress to 8-bit and
|
||||||
* return a string
|
* return a string
|
||||||
*
|
*
|
||||||
* { 0x00, 0x16 } -0x16
|
* { 0x00, 0x16 } -> 0x16
|
||||||
*
|
*
|
||||||
*@param string the byte array to be converted
|
*@param string the byte array to be converted
|
||||||
**@param offset the initial offset into the
|
*@param offset the initial offset into the
|
||||||
* byte array. it is assumed that string[ offset ] and string[ offset +
|
* byte array. it is assumed that string[ offset ] and string[ offset +
|
||||||
* 1 ] contain the first 16-bit unicode character
|
* 1 ] contain the first 16-bit unicode character
|
||||||
*@param len
|
*@param len
|
||||||
|
@ -148,10 +159,9 @@ public class StringUtil {
|
||||||
* length)
|
* length)
|
||||||
*@len the length of the final string
|
*@len the length of the final string
|
||||||
*/
|
*/
|
||||||
public static String getFromUnicodeBE(
|
|
||||||
final byte[] string,
|
public static String getFromUnicode(final byte[] string,
|
||||||
final int offset,
|
final int offset, final int len)
|
||||||
final int len)
|
|
||||||
throws ArrayIndexOutOfBoundsException, IllegalArgumentException {
|
throws ArrayIndexOutOfBoundsException, IllegalArgumentException {
|
||||||
if ((offset < 0) || (offset >= string.length)) {
|
if ((offset < 0) || (offset >= string.length)) {
|
||||||
throw new ArrayIndexOutOfBoundsException("Illegal offset");
|
throw new ArrayIndexOutOfBoundsException("Illegal offset");
|
||||||
|
@ -159,42 +169,50 @@ public class StringUtil {
|
||||||
if ((len < 0) || (((string.length - offset) / 2) < len)) {
|
if ((len < 0) || (((string.length - offset) / 2) < len)) {
|
||||||
throw new IllegalArgumentException("Illegal length");
|
throw new IllegalArgumentException("Illegal length");
|
||||||
}
|
}
|
||||||
try {
|
|
||||||
return new String(string, offset, len * 2, "UTF-16BE");
|
|
||||||
} catch (UnsupportedEncodingException e) {
|
char[] chars = new char[ len ];
|
||||||
throw new InternalError(); /*unreachable*/
|
for ( int i = 0; i < chars.length; i++ ) {
|
||||||
|
chars[i] = (char)( ( string[ offset + ( 2*i ) ] << 8 ) +
|
||||||
|
string[ offset + ( 2*i+1 ) ] );
|
||||||
}
|
}
|
||||||
|
|
||||||
|
return new String( chars );
|
||||||
}
|
}
|
||||||
|
|
||||||
|
|
||||||
/**
|
/**
|
||||||
* given a byte array of 16-bit unicode characters, compress to 8-bit and
|
* given a byte array of 16-bit unicode characters, compress to 8-bit and
|
||||||
* return a string
|
* return a string
|
||||||
*
|
*
|
||||||
* { 0x00, 0x16 } -0x16
|
* { 0x00, 0x16 } -> 0x16
|
||||||
*
|
*
|
||||||
*@param string the byte array to be converted
|
*@param string the byte array to be converted
|
||||||
*@return the converted string
|
*@return the converted string
|
||||||
*/
|
*/
|
||||||
public static String getFromUnicodeBE(final byte[] string) {
|
|
||||||
return getFromUnicodeBE(string, 0, string.length / 2);
|
public static String getFromUnicode(final byte[] string) {
|
||||||
|
return getFromUnicode(string, 0, string.length / 2);
|
||||||
}
|
}
|
||||||
|
|
||||||
|
|
||||||
/**
|
/**
|
||||||
* read compressed unicode(8bit)
|
* read compressed unicode(8bit)
|
||||||
*
|
*
|
||||||
|
* @author Toshiaki Kamoshida(kamoshida.toshiaki at future dot co dot jp)
|
||||||
|
*
|
||||||
* @param string byte array to read
|
* @param string byte array to read
|
||||||
* @param offset offset to read byte array
|
* @param offset offset to read byte array
|
||||||
* @param len length to read byte array
|
* @param len length to read byte array
|
||||||
* @return String generated String instance by reading byte array
|
* @return String generated String instance by reading byte array
|
||||||
*/
|
*/
|
||||||
public static String getFromCompressedUnicode(
|
public static String getFromCompressedUnicode(final byte[] string,
|
||||||
final byte[] string,
|
final int offset, final int len){
|
||||||
final int offset,
|
try{
|
||||||
final int len) {
|
return new String(string,offset,len,"ISO-8859-1");
|
||||||
try {
|
}
|
||||||
return new String(string, offset, len, "ISO-8859-1");
|
catch(UnsupportedEncodingException e){
|
||||||
} catch (UnsupportedEncodingException e) {
|
throw new InternalError();/* unreachable */
|
||||||
throw new InternalError(); /* unreachable */
|
|
||||||
}
|
}
|
||||||
}
|
}
|
||||||
|
|
||||||
|
@ -206,15 +224,36 @@ public class StringUtil {
|
||||||
*@param offset an offset into the byte arrat at which the data is start
|
*@param offset an offset into the byte arrat at which the data is start
|
||||||
* when written
|
* when written
|
||||||
*/
|
*/
|
||||||
public static void putCompressedUnicode(
|
|
||||||
final String input,
|
public static void putCompressedUnicode(final String input,
|
||||||
final byte[] output,
|
final byte[] output,
|
||||||
final int offset) {
|
final int offset) {
|
||||||
try {
|
int strlen = input.length();
|
||||||
byte[] bytes = input.getBytes("ISO-8859-1");
|
|
||||||
System.arraycopy(bytes, 0, output, offset, bytes.length);
|
for (int k = 0; k < strlen; k++) {
|
||||||
} catch (UnsupportedEncodingException e) {
|
output[offset + k] = (byte) input.charAt(k);
|
||||||
throw new InternalError(); /*unreachable*/
|
}
|
||||||
|
}
|
||||||
|
|
||||||
|
|
||||||
|
/**
|
||||||
|
* Write uncompressed unicode
|
||||||
|
*
|
||||||
|
*@param input the String containing the unicode data to be written
|
||||||
|
*@param output the byte array to hold the uncompressed unicode
|
||||||
|
*@param offset the offset to start writing into the byte array
|
||||||
|
*/
|
||||||
|
|
||||||
|
public static void putUncompressedUnicode(final String input,
|
||||||
|
final byte[] output,
|
||||||
|
final int offset) {
|
||||||
|
int strlen = input.length();
|
||||||
|
|
||||||
|
for (int k = 0; k < strlen; k++) {
|
||||||
|
char c = input.charAt(k);
|
||||||
|
|
||||||
|
output[offset + (2 * k)] = (byte) c;
|
||||||
|
output[offset + (2 * k) + 1] = (byte) (c >> 8);
|
||||||
}
|
}
|
||||||
}
|
}
|
||||||
|
|
||||||
|
@ -225,36 +264,22 @@ public class StringUtil {
|
||||||
*@param output the byte array to hold the uncompressed unicode
|
*@param output the byte array to hold the uncompressed unicode
|
||||||
*@param offset the offset to start writing into the byte array
|
*@param offset the offset to start writing into the byte array
|
||||||
*/
|
*/
|
||||||
public static void putUnicodeLE(
|
|
||||||
final String input,
|
public static void putUncompressedUnicodeHigh(final String input,
|
||||||
final byte[] output,
|
final byte[] output,
|
||||||
final int offset) {
|
final int offset) {
|
||||||
try {
|
int strlen = input.length();
|
||||||
byte[] bytes = input.getBytes("UTF-16LE");
|
|
||||||
System.arraycopy(bytes, 0, output, offset, bytes.length);
|
for (int k = 0; k < strlen; k++) {
|
||||||
} catch (UnsupportedEncodingException e) {
|
char c = input.charAt(k);
|
||||||
throw new InternalError(); /*unreachable*/
|
|
||||||
|
output[offset + (2 * k)] = (byte) (c >> 8);
|
||||||
|
output[offset + (2 * k)] = (byte) c;
|
||||||
}
|
}
|
||||||
}
|
}
|
||||||
|
|
||||||
/**
|
|
||||||
* Write uncompressed unicode
|
|
||||||
*
|
|
||||||
*@param input the String containing the unicode data to be written
|
|
||||||
*@param output the byte array to hold the uncompressed unicode
|
|
||||||
*@param offset the offset to start writing into the byte array
|
|
||||||
*/
|
|
||||||
public static void putUnicodeBE(
|
|
||||||
final String input,
|
|
||||||
final byte[] output,
|
|
||||||
final int offset) {
|
|
||||||
try {
|
|
||||||
byte[] bytes = input.getBytes("UTF-16BE");
|
|
||||||
System.arraycopy(bytes, 0, output, offset, bytes.length);
|
|
||||||
} catch (UnsupportedEncodingException e) {
|
|
||||||
throw new InternalError(); /*unreachable*/
|
|
||||||
}
|
|
||||||
}
|
|
||||||
|
|
||||||
/**
|
/**
|
||||||
* Description of the Method
|
* Description of the Method
|
||||||
|
@ -266,25 +291,21 @@ public class StringUtil {
|
||||||
public static String format(String message, Object[] params) {
|
public static String format(String message, Object[] params) {
|
||||||
int currentParamNumber = 0;
|
int currentParamNumber = 0;
|
||||||
StringBuffer formattedMessage = new StringBuffer();
|
StringBuffer formattedMessage = new StringBuffer();
|
||||||
|
|
||||||
for (int i = 0; i < message.length(); i++) {
|
for (int i = 0; i < message.length(); i++) {
|
||||||
if (message.charAt(i) == '%') {
|
if (message.charAt(i) == '%') {
|
||||||
if (currentParamNumber >= params.length) {
|
if (currentParamNumber >= params.length) {
|
||||||
formattedMessage.append("?missing data?");
|
formattedMessage.append("?missing data?");
|
||||||
} else if (
|
} else if ((params[currentParamNumber] instanceof Number)
|
||||||
(params[currentParamNumber] instanceof Number)
|
|
||||||
&& (i + 1 < message.length())) {
|
&& (i + 1 < message.length())) {
|
||||||
i
|
i += matchOptionalFormatting(
|
||||||
+= matchOptionalFormatting(
|
|
||||||
(Number) params[currentParamNumber++],
|
(Number) params[currentParamNumber++],
|
||||||
message.substring(i + 1),
|
message.substring(i + 1), formattedMessage);
|
||||||
formattedMessage);
|
|
||||||
} else {
|
} else {
|
||||||
formattedMessage.append(
|
formattedMessage.append(params[currentParamNumber++].toString());
|
||||||
params[currentParamNumber++].toString());
|
|
||||||
}
|
}
|
||||||
} else {
|
} else {
|
||||||
if ((message.charAt(i) == '\\')
|
if ((message.charAt(i) == '\\') && (i + 1 < message.length())
|
||||||
&& (i + 1 < message.length())
|
|
||||||
&& (message.charAt(i + 1) == '%')) {
|
&& (message.charAt(i + 1) == '%')) {
|
||||||
formattedMessage.append('%');
|
formattedMessage.append('%');
|
||||||
i++;
|
i++;
|
||||||
|
@ -296,6 +317,7 @@ public class StringUtil {
|
||||||
return formattedMessage.toString();
|
return formattedMessage.toString();
|
||||||
}
|
}
|
||||||
|
|
||||||
|
|
||||||
/**
|
/**
|
||||||
* Description of the Method
|
* Description of the Method
|
||||||
*
|
*
|
||||||
|
@ -304,31 +326,26 @@ public class StringUtil {
|
||||||
*@param outputTo Description of the Parameter
|
*@param outputTo Description of the Parameter
|
||||||
*@return Description of the Return Value
|
*@return Description of the Return Value
|
||||||
*/
|
*/
|
||||||
private static int matchOptionalFormatting(
|
private static int matchOptionalFormatting(Number number,
|
||||||
Number number,
|
|
||||||
String formatting,
|
String formatting,
|
||||||
StringBuffer outputTo) {
|
StringBuffer outputTo) {
|
||||||
NumberFormat numberFormat = NumberFormat.getInstance();
|
NumberFormat numberFormat = NumberFormat.getInstance();
|
||||||
|
|
||||||
if ((0 < formatting.length())
|
if ((0 < formatting.length())
|
||||||
&& Character.isDigit(formatting.charAt(0))) {
|
&& Character.isDigit(formatting.charAt(0))) {
|
||||||
numberFormat.setMinimumIntegerDigits(
|
numberFormat.setMinimumIntegerDigits(Integer.parseInt(formatting.charAt(0) + ""));
|
||||||
Integer.parseInt(formatting.charAt(0) + ""));
|
if ((2 < formatting.length()) && (formatting.charAt(1) == '.')
|
||||||
if ((2 < formatting.length())
|
|
||||||
&& (formatting.charAt(1) == '.')
|
|
||||||
&& Character.isDigit(formatting.charAt(2))) {
|
&& Character.isDigit(formatting.charAt(2))) {
|
||||||
numberFormat.setMaximumFractionDigits(
|
numberFormat.setMaximumFractionDigits(Integer.parseInt(formatting.charAt(2) + ""));
|
||||||
Integer.parseInt(formatting.charAt(2) + ""));
|
|
||||||
numberFormat.format(number, outputTo, new FieldPosition(0));
|
numberFormat.format(number, outputTo, new FieldPosition(0));
|
||||||
return 3;
|
return 3;
|
||||||
}
|
}
|
||||||
numberFormat.format(number, outputTo, new FieldPosition(0));
|
numberFormat.format(number, outputTo, new FieldPosition(0));
|
||||||
return 1;
|
return 1;
|
||||||
} else if (
|
} else if ((0 < formatting.length()) && (formatting.charAt(0) == '.')) {
|
||||||
(0 < formatting.length()) && (formatting.charAt(0) == '.')) {
|
|
||||||
if ((1 < formatting.length())
|
if ((1 < formatting.length())
|
||||||
&& Character.isDigit(formatting.charAt(1))) {
|
&& Character.isDigit(formatting.charAt(1))) {
|
||||||
numberFormat.setMaximumFractionDigits(
|
numberFormat.setMaximumFractionDigits(Integer.parseInt(formatting.charAt(1) + ""));
|
||||||
Integer.parseInt(formatting.charAt(1) + ""));
|
|
||||||
numberFormat.format(number, outputTo, new FieldPosition(0));
|
numberFormat.format(number, outputTo, new FieldPosition(0));
|
||||||
return 2;
|
return 2;
|
||||||
}
|
}
|
||||||
|
|
Loading…
Reference in New Issue