chore: support all GB* charsets

soulteary · Dec 19, 2022 · fba16e8 · fba16e8
1 parent 490bc0e
commit fba16e8
Show file tree

Hide file tree

Showing 2 changed files with 7 additions and 6 deletions.
diff --git a/internal/define/charset.go b/internal/define/charset.go
@@ -1,7 +1,10 @@
 package define
 
 const (
-	CHARSET_GBK    = "gbk"
-	CHARSET_GB2312 = "gb2312"
-	CHARSET_UTF8   = "utf-8"
+	CHARSET_GBK     = "gbk"
+	CHARSET_GB2312  = "gb2312"
+	CHARSET_GB18030 = "gb18030"
+	CHARSET_BIG5    = "big5"
+
+	CHARSET_UTF8 = "utf-8"
 )
diff --git a/internal/fn/charset.go b/internal/fn/charset.go
@@ -26,9 +26,7 @@ func DecodeHTMLBody(body io.Reader, charset string) (io.Reader, error) {
 	if charset == "" {
 		charset = DetectContentEncoding(body)
 	} else {
-		// use UTF-8 as fallback
-		// TODO maybe support more preset types
-		if !(charset == define.CHARSET_UTF8 || charset == define.CHARSET_GB2312 || charset == define.CHARSET_GBK) {
+		if !(charset == define.CHARSET_UTF8 || charset == define.CHARSET_GB2312 || charset == define.CHARSET_GBK || charset == define.CHARSET_GB18030) {
 			charset = define.DEFAULT_DOCUMENT_CHARSET
 		}
 	}