From 5f4357ab13fb04904d8601f828225fba72b52826 Mon Sep 17 00:00:00 2001
From: pubpub-zz <4083478+pubpub-zz@users.noreply.github.com>
Date: Thu, 10 Aug 2023 20:05:44 +0200
Subject: [PATCH 1/2] BUG: prevent stall when accessing image in corrupted pdf

closes #2077
---
 pypdf/_page.py     | 13 +++++++++++--
 tests/test_page.py | 10 ++++++++++
 2 files changed, 21 insertions(+), 2 deletions(-)

diff --git a/pypdf/_page.py b/pypdf/_page.py
index 01c5b0506..868e9c594 100644
--- a/pypdf/_page.py
+++ b/pypdf/_page.py
@@ -477,8 +477,17 @@ def _old_images(self) -> List[File]:  # deprecated
         return images_extracted
 
     def _get_ids_image(
-        self, obj: Optional[DictionaryObject] = None, ancest: Optional[List[str]] = None
+        self,
+        obj: Optional[DictionaryObject] = None,
+        ancest: Optional[List[str]] = None,
+        call_stack: Optional[List[Any]] = None,
     ) -> List[Union[str, List[str]]]:
+        if call_stack is None:
+            call_stack = []
+        if obj in call_stack:
+            return []
+        else:
+            call_stack.append(obj)
         if self.inline_images_keys is None:
             nb_inlines = len(
                 re.findall(
@@ -502,7 +511,7 @@ def _get_ids_image(
             if x_object[o][IA.SUBTYPE] == "/Image":
                 lst.append(o if len(ancest) == 0 else ancest + [o])
             else:  # is a form with possible images inside
-                lst.extend(self._get_ids_image(x_object[o], ancest + [o]))
+                lst.extend(self._get_ids_image(x_object[o], ancest + [o], call_stack))
         return lst + self.inline_images_keys
 
     def _get_image(
diff --git a/tests/test_page.py b/tests/test_page.py
index d1f6f3fcb..aaee6278e 100644
--- a/tests/test_page.py
+++ b/tests/test_page.py
@@ -1220,3 +1220,13 @@ def create_stamp_pdf() -> BytesIO:
     assert isinstance(
         writer._objects[contents.indirect_reference.idnum - 1], NullObject
     )
+
+
+@pytest.mark.enable_socket()
+@pytest.mark.timeout(30)
+def test_loop_in_image_keys():
+    """Cf #2077"""
+    url = "https://github.com/py-pdf/pypdf/files/12309492/example_134.pdf"
+    name = "iss2077.pdf"
+    reader = PdfReader(BytesIO(get_data_from_url(url, name=name)))
+    reader.pages[0].images.keys()

From 7230c02f79481565eb9d083b3bd043401ce47845 Mon Sep 17 00:00:00 2001
From: pubpub-zz <4083478+pubpub-zz@users.noreply.github.com>
Date: Fri, 11 Aug 2023 11:09:49 +0200
Subject: [PATCH 2/2] fix test and move it

---
 pypdf/_page.py       |  5 +++--
 tests/test_images.py | 10 ++++++++++
 tests/test_page.py   | 10 ----------
 3 files changed, 13 insertions(+), 12 deletions(-)

diff --git a/pypdf/_page.py b/pypdf/_page.py
index 868e9c594..00800a9bf 100644
--- a/pypdf/_page.py
+++ b/pypdf/_page.py
@@ -484,10 +484,11 @@ def _get_ids_image(
     ) -> List[Union[str, List[str]]]:
         if call_stack is None:
             call_stack = []
-        if obj in call_stack:
+        _i = getattr(obj, "indirect_reference", None)
+        if _i in call_stack:
             return []
         else:
-            call_stack.append(obj)
+            call_stack.append(_i)
         if self.inline_images_keys is None:
             nb_inlines = len(
                 re.findall(
diff --git a/tests/test_images.py b/tests/test_images.py
index b159af0d3..2f14c7b38 100644
--- a/tests/test_images.py
+++ b/tests/test_images.py
@@ -209,3 +209,13 @@ def test_image_extraction(src, page_index, image_key, expected):
         with open(f"page-{page_index}-{actual_image.name}", "wb") as fp:
             fp.write(actual_image.data)
     assert image_similarity(BytesIO(actual_image.data), expected) >= 0.99
+
+
+@pytest.mark.enable_socket()
+@pytest.mark.timeout(30)
+def test_loop_in_image_keys():
+    """Cf #2077"""
+    url = "https://github.com/py-pdf/pypdf/files/12309492/example_134.pdf"
+    name = "iss2077.pdf"
+    reader = PdfReader(BytesIO(get_data_from_url(url, name=name)))
+    reader.pages[0].images.keys()
diff --git a/tests/test_page.py b/tests/test_page.py
index aaee6278e..d1f6f3fcb 100644
--- a/tests/test_page.py
+++ b/tests/test_page.py
@@ -1220,13 +1220,3 @@ def create_stamp_pdf() -> BytesIO:
     assert isinstance(
         writer._objects[contents.indirect_reference.idnum - 1], NullObject
     )
-
-
-@pytest.mark.enable_socket()
-@pytest.mark.timeout(30)
-def test_loop_in_image_keys():
-    """Cf #2077"""
-    url = "https://github.com/py-pdf/pypdf/files/12309492/example_134.pdf"
-    name = "iss2077.pdf"
-    reader = PdfReader(BytesIO(get_data_from_url(url, name=name)))
-    reader.pages[0].images.keys()