BUG: Ambiguous translated references

pypdf/_protocols.py

@@ -3,6 +3,7 @@
 from abc import abstractmethod
 from pathlib import Path
 from typing import IO, Any, Dict, List, Optional, Tuple, Union
+from weakref import WeakKeyDictionary
 
 try:
     # Python 3.8+: https://peps.python.org/pep-0586
@@ -78,7 +79,7 @@ def trailer(self) -> Dict[str, Any]:
 
 class PdfWriterProtocol(PdfCommonDocProtocol, Protocol):
     _objects: List[Any]
-    _id_translated: Dict[int, Dict[int, int]]
+    _id_translated: "WeakKeyDictionary[PdfReaderProtocol, Dict[int, int]]"
 
     @abstractmethod
     def write(self, stream: Union[Path, StrByteType]) -> Tuple[bool, IO[Any]]:

pypdf/_writer.py

@@ -52,7 +52,9 @@
     Union,
     cast,
 )
+from weakref import WeakKeyDictionary
 
+from ._protocols import PdfReaderProtocol
 from ._cmap import build_char_map_from_dict
 from ._doc_common import PdfDocCommon
 from ._encryption import EncryptAlgorithm, Encryption
@@ -178,7 +180,7 @@ def __init__(
         self._idnum_hash: Dict[bytes, IndirectObject] = {}
         """Maps hash values of indirect objects to their IndirectObject instances."""
 
-        self._id_translated: Dict[int, Dict[int, int]] = {}
+        self._id_translated: "WeakKeyDictionary[PdfReaderProtocol, dict[int, int]]" = WeakKeyDictionary()
 
         # The root of our page tree node.
         pages = DictionaryObject()
@@ -389,7 +391,7 @@ def _add_page(
         # page, we need to create a new dictionary for the page, however the
         # objects below (including content) are not duplicated:
         try:  # delete an already existing page
-            del self._id_translated[id(page_org.indirect_reference.pdf)][  # type: ignore
+            del self._id_translated[page_org.indirect_reference.pdf][  # type: ignore
                 page_org.indirect_reference.idnum  # type: ignore
             ]
         except Exception:
@@ -2451,7 +2453,7 @@ def merge(
                     ArrayObject,
                     cast(DictionaryObject, self._root_object["/AcroForm"])["/Fields"],
                 )
-            trslat = self._id_translated[id(reader)]
+            trslat = self._id_translated[reader]
             try:
                 for f in reader.root_object["/AcroForm"]["/Fields"]:  # type: ignore
                     try:
@@ -2564,7 +2566,7 @@ def add_filtered_articles(
                 else:
                     thr = thr.get_object()
                 if thr.indirect_reference.idnum not in self._id_translated[
-                    id(reader)
+                    reader
                 ] and fltr.search((thr["/I"] if "/I" in thr else {}).get("/Title", "")):
                     self._add_articles_thread(thr, pages, reader)
 
@@ -2780,15 +2782,15 @@ def reset_translation(
                 if set to None or omitted, all tables will be reset.
         """
         if reader is None:
-            self._id_translated = {}
+            self._id_translated = WeakKeyDictionary()
         elif isinstance(reader, PdfReader):
             try:
-                del self._id_translated[id(reader)]
+                del self._id_translated[reader]
             except Exception:
                 pass
         elif isinstance(reader, IndirectObject):
             try:
-                del self._id_translated[id(reader.pdf)]
+                del self._id_translated[reader.pdf]
             except Exception:
                 pass
         else:

pypdf/generic/_base.py

@@ -124,19 +124,18 @@ def _reference_clone(
             return clone
         i = len(pdf_dest._objects) + 1
         if ind is not None:
-            if id(ind.pdf) not in pdf_dest._id_translated:
-                pdf_dest._id_translated[id(ind.pdf)] = {}
-                pdf_dest._id_translated[id(ind.pdf)]["PreventGC"] = ind.pdf  # type: ignore
+            if ind.pdf not in pdf_dest._id_translated:
+                pdf_dest._id_translated[ind.pdf] = {}
             if (
                 not force_duplicate
-                and ind.idnum in pdf_dest._id_translated[id(ind.pdf)]
+                and ind.idnum in pdf_dest._id_translated[ind.pdf]
             ):
                 obj = pdf_dest.get_object(
-                    pdf_dest._id_translated[id(ind.pdf)][ind.idnum]
+                    pdf_dest._id_translated[ind.pdf][ind.idnum]
                 )
                 assert obj is not None
                 return obj
-            pdf_dest._id_translated[id(ind.pdf)][ind.idnum] = i
+            pdf_dest._id_translated[ind.pdf][ind.idnum] = i
         pdf_dest._objects.append(clone)
         clone.indirect_reference = IndirectObject(i, 0, pdf_dest)
         return clone
@@ -250,11 +249,11 @@ def clone(
         if self.pdf == pdf_dest and not force_duplicate:
             # Already duplicated and no extra duplication required
             return self
-        if id(self.pdf) not in pdf_dest._id_translated:
-            pdf_dest._id_translated[id(self.pdf)] = {}
+        if self.pdf not in pdf_dest._id_translated:
+            pdf_dest._id_translated[self.pdf] = {}
 
-        if self.idnum in pdf_dest._id_translated[id(self.pdf)]:
-            dup = pdf_dest.get_object(pdf_dest._id_translated[id(self.pdf)][self.idnum])
+        if self.idnum in pdf_dest._id_translated[self.pdf]:
+            dup = pdf_dest.get_object(pdf_dest._id_translated[self.pdf][self.idnum])
             if force_duplicate:
                 assert dup is not None
                 assert dup.indirect_reference is not None