🐛 fix: filter bad chars in comment content

Author: SigureMo

packages/biliass/rust/src/reader/xml.rs

@@ -152,6 +152,8 @@ fn filter_bad_chars(string: &str) -> String {
             if ('\u{00}'..='\u{08}').contains(&c)
                 || c == '\u{0b}'
                 || c == '\u{0c}'
+                || c == '\u{2028}'
+                || c == '\u{2029}'
                 || ('\u{0e}'..='\u{1f}').contains(&c)
             {
                 '\u{fffd}'

packages/biliass/src/biliass/biliass.py

@@ -85,7 +85,7 @@ def read_comments_bilibili_protobuf(protobuf: bytes | str, fontsize: float) -> G
                     elem.progress / 1000,  # 视频内出现的时间
                     elem.ctime,  # 弹幕的发送时间（时间戳）
                     i,
-                    c,
+                    filter_bad_chars(c),
                     {1: 0, 4: 2, 5: 1, 6: 3}[elem.mode],
                     elem.color,
                     size,
@@ -98,7 +98,7 @@ def read_comments_bilibili_protobuf(protobuf: bytes | str, fontsize: float) -> G
                     elem.progress / 1000,
                     elem.ctime,
                     i,
-                    c,
+                    filter_bad_chars(c),
                     "bilipos",
                     elem.color,
                     elem.fontsize,
@@ -589,6 +589,10 @@ def convert_type2(row, height, bottom_reserved):
     return height - bottom_reserved - row
 
 
+def filter_bad_chars(string: str) -> str:
+    return re.sub("[\\x00-\\x08\\x0b\\x0c\\x0e-\\x1f\u2028\u2029]", "\ufffd", string)
+
+
 class safe_list(list):
     def get(self, index, default=None):
         def is_empty(value):