ROB: Silently ignore Adobe Ascii85 whitespace for suffix detection (#3528)

Based on the PDF standards "the ASCII85Decode filter shall ignore all white-space characters".

---------

Co-authored-by: mbierma <[email protected]>

Author: mbierma

pypdf/filters.py

@@ -482,6 +482,8 @@ def decode(
         if isinstance(data, str):
             data = data.encode()
         data = data.strip(WHITESPACES_AS_BYTES)
+        if len(data) > 2 and data.endswith(b">"):
+            data = data[:-1].rstrip(WHITESPACES_AS_BYTES) + data[-1:]
         try:
             return a85decode(data, adobe=True, ignorechars=WHITESPACES_AS_BYTES)
         except ValueError as error:

tests/test_filters.py

@@ -649,6 +649,13 @@ def test_ascii85decode__non_recoverable(caplog):
     assert caplog.text == ""
 
 
+def test_ascii85decode__ignore_whitespaces(caplog):
+    """Whitespace characters must be silently ignored"""
+    data = b"Cqa;:3k~\n>"
+    result = ASCII85Decode.decode(data)
+    assert result == b"l\xbe`\x8d:"
+
+
 @pytest.mark.enable_socket
 def test_ccitt_fax_decode__black_is_1():
     url = "https://github.com/user-attachments/files/19288881/imagemagick-CCITTFaxDecode_BlackIs1-true.pdf"