ROB: Prevent loop in Cloning

pypdf/_protocols.py

@@ -73,6 +73,9 @@ def get_object(self, indirect_reference: Any) -> Optional[PdfObjectProtocol]:
     def write(self, stream: Union[Path, StrByteType]) -> Tuple[bool, IO]:
         ...
 
+    def _add_object(self, obj: Any) -> Any:
+        ...
+
     @property
     def pages(self) -> List[Any]:
         ...

pypdf/generic/_base.py

@@ -277,7 +277,10 @@ def clone(
                 obj = NullObject()
                 assert isinstance(self, (IndirectObject,))
                 obj.indirect_reference = self
-            dup = obj.clone(pdf_dest, force_duplicate, ignore_fields)
+            dup = pdf_dest._add_object(
+                obj.clone(pdf_dest, force_duplicate, ignore_fields)
+            )
+        # asserts added to prevent errors in mypy
         assert dup is not None
         assert dup.indirect_reference is not None
         return dup.indirect_reference

pypdf/generic/_data_structures.py

@@ -750,7 +750,10 @@ def _clone(
             if decoded_self is None:
                 self.decoded_self = None
             else:
-                self.decoded_self = decoded_self.clone(pdf_dest, True, ignore_fields)  # type: ignore[assignment]
+                self.decoded_self = cast(
+                    "DecodedStreamObject",
+                    decoded_self.clone(pdf_dest, force_duplicate, ignore_fields),
+                )
         except Exception:
             pass
         super()._clone(src, pdf_dest, force_duplicate, ignore_fields)

tests/test_writer.py

@@ -1292,3 +1292,14 @@ def test_iss1723():
     in_pdf = PdfReader(BytesIO(get_pdf_from_url(url, name=name)))
     out_pdf = PdfWriter()
     out_pdf.append(in_pdf, (3, 5))
+
+
+@pytest.mark.enable_socket()
+def test_iss1767():
+    # test with a pdf which is buggy because the object 389,0 exists 3 times:
+    # twice to define catalog and one as an XObject inducing a loop when
+    # cloning
+    url = "https://github.com/py-pdf/pypdf/files/11138472/test.pdf"
+    name = "iss1723.pdf"
+    in_pdf = PdfReader(BytesIO(get_pdf_from_url(url, name=name)))
+    PdfWriter(clone_from=in_pdf)