修复URLUtil.decode无法解码UTF-16问题

This commit is contained in:
Looly 2023-04-17 23:49:31 +08:00
parent ea78c853b2
commit 923aa1cf21
3 changed files with 85 additions and 4 deletions

View File

@ -2,11 +2,12 @@
# 🚀Changelog
-------------------------------------------------------------------------------------------------------------
# 5.8.19.M1 (2023-04-16)
# 5.8.19.M1 (2023-04-17)
### 🐣新特性
### 🐞Bug修复
* 【core 】 修复URLUtil.decode无法解码UTF-16问题issue#3063@Github
-------------------------------------------------------------------------------------------------------------
# 5.8.18 (2023-04-16)

View File

@ -1,6 +1,7 @@
package cn.hutool.core.net;
import cn.hutool.core.util.CharUtil;
import cn.hutool.core.util.CharsetUtil;
import cn.hutool.core.util.StrUtil;
import java.io.ByteArrayOutputStream;
@ -71,10 +72,41 @@ public class URLDecoder implements Serializable {
* @return 解码后的字符串
*/
public static String decode(String str, Charset charset, boolean isPlusToSpace) {
if(null == charset){
return str;
if (null == str) {
return null;
}
return StrUtil.str(decode(StrUtil.bytes(str, charset), isPlusToSpace), charset);
final int length = str.length();
if(0 == length){
return StrUtil.EMPTY;
}
final StringBuilder result = new StringBuilder(length / 3);
int begin = 0;
char c;
for (int i = 0; i < length; i++) {
c = str.charAt(i);
if(ESCAPE_CHAR == c || CharUtil.isHexChar(c)){
continue;
}
// 遇到非需要处理的字符跳过
// 处理之前的hex字符
if(i > begin){
result.append(decodeSub(str, begin, i, charset, isPlusToSpace));
}
// 非Hex字符忽略本字符
result.append(c);
begin = i + 1;
}
// 处理剩余字符
if(begin < length){
result.append(decodeSub(str, begin, length, charset, isPlusToSpace));
}
return result.toString();
}
/**
@ -135,4 +167,22 @@ public class URLDecoder implements Serializable {
}
return buffer.toByteArray();
}
/**
* 解码子串
*
* @param str 字符串
* @param begin 开始位置包含
* @param end 结束位置不包含
* @param charset 编码
* @param isPlusToSpace 是否+转换为空格
* @return 解码后的字符串
*/
private static String decodeSub(final String str, final int begin, final int end,
final Charset charset, final boolean isPlusToSpace){
return new String(decode(
// 截取需要decode的部分
str.substring(begin, end).getBytes(CharsetUtil.CHARSET_ISO_8859_1), isPlusToSpace
), charset);
}
}

View File

@ -1,12 +1,42 @@
package cn.hutool.core.net;
import cn.hutool.core.util.CharsetUtil;
import cn.hutool.core.util.URLUtil;
import org.junit.Assert;
import org.junit.Test;
import java.io.UnsupportedEncodingException;
import java.nio.charset.StandardCharsets;
public class UrlDecoderTest {
@Test
public void decodeForPathTest(){
Assert.assertEquals("+", URLDecoder.decodeForPath("+", CharsetUtil.CHARSET_UTF_8));
}
@Test
public void issue3063Test() throws UnsupportedEncodingException {
// https://github.com/dromara/hutool/issues/3063
final String s = "测试";
final String expectedDecode = "%FE%FF%6D%4B%8B%D5";
final String s1 = URLUtil.encode(s, StandardCharsets.UTF_16);
Assert.assertEquals(expectedDecode, s1);
final String s2 = java.net.URLEncoder.encode(s, "UTF-16");
Assert.assertEquals(expectedDecode, s2);
final String decode = URLDecoder.decode(s1, StandardCharsets.UTF_16);
Assert.assertEquals(s, decode);
// 测试编码字符串和非编码字符串混合
final String mixDecoded = expectedDecode + "你好";
final String decode2 = URLDecoder.decode(mixDecoded, StandardCharsets.UTF_16);
Assert.assertEquals("测试你好", decode2);
Assert.assertEquals(
java.net.URLDecoder.decode(mixDecoded, "UTF-16"),
URLDecoder.decode(mixDecoded, StandardCharsets.UTF_16)
);
}
}