部分gbk文件无法识别，加入新库

kekingcn · Dec 26, 2023 · be2e522 · be2e522
1 parent 8ac8cd8
commit be2e522
Show file tree

Hide file tree

Showing 3 changed files with 16 additions and 1 deletion.
diff --git a/pom.xml b/pom.xml
@@ -35,6 +35,7 @@
         <aspose-cad.version>23.7</aspose-cad.version>
         <bcprov-jdk15on.version>1.70</bcprov-jdk15on.version>
         <juniversalchardet.version>1.0.3</juniversalchardet.version>
+        <icu4j.version>70.1</icu4j.version>
 
         <commons-cli.version>1.5.0</commons-cli.version>
         <commons-net.version>3.9.0</commons-net.version>

diff --git a/server/pom.xml b/server/pom.xml
@@ -129,6 +129,11 @@
             <artifactId>juniversalchardet</artifactId>
             <version>${juniversalchardet.version}</version>
         </dependency>
+        <dependency>
+            <groupId>com.ibm.icu</groupId>
+            <artifactId>icu4j</artifactId>
+            <version>${icu4j.version}</version>
+        </dependency>
 
         <!-- 解压(rar)-->
         <dependency>

diff --git a/server/src/main/java/cn/keking/utils/EncodingDetects.java b/server/src/main/java/cn/keking/utils/EncodingDetects.java
@@ -1,5 +1,7 @@
 package cn.keking.utils;
 
+import com.ibm.icu.text.CharsetDetector;
+import com.ibm.icu.text.CharsetMatch;
 import org.mozilla.universalchardet.UniversalDetector;
 import org.slf4j.Logger;
 import org.slf4j.LoggerFactory;
@@ -44,7 +46,14 @@ public static String getJavaEncode(byte[] content) {
         detector.dataEnd();
         String charsetName = detector.getDetectedCharset();
         if (charsetName == null) {
-            charsetName = Charset.defaultCharset().name();
+            CharsetDetector cd = new CharsetDetector();
+            cd.setText(content);
+            CharsetMatch cm = cd.detect();
+            if (cm != null) {
+                charsetName = cm.getName();
+            } else {
+                charsetName = Charset.defaultCharset().name();
+            }
         }
         return charsetName;
     }