ROB: Consider root objects without catalog type as fallback (#3175)

Closes #3164.

Author: stefan6419846

pypdf/_reader.py

@@ -229,7 +229,16 @@ def root_object(self) -> DictionaryObject:
                     self._validated_root = o
                     logger_warning(f"Root found at {o.indirect_reference!r}", __name__)
                     break
-            if self._validated_root is None:
+        if self._validated_root is None:
+            if not is_null_or_none(root) and "/Pages" in cast(DictionaryObject, cast(PdfObject, root).get_object()):
+                logger_warning(
+                    f"Possible root found at {cast(PdfObject, root).indirect_reference!r}, but missing /Catalog key",
+                    __name__
+                )
+                self._validated_root = cast(
+                    DictionaryObject, cast(PdfObject, root).get_object()
+                )
+            else:
                 raise PdfReadError("Cannot find Root object in pdf")
         return self._validated_root
 

tests/test_reader.py

@@ -1760,7 +1760,7 @@ def test_repair_root(caplog):
     caplog.clear()
     reader = PdfReader(
         BytesIO(
-            b.replace(b"/Root 1 0 R", b"/Root 2 0 R").replace(b"/Catalog", b"/Catalo ")
+            b.replace(b"/Root 1 0 R", b"/Root 2 0 R").replace(b"/Catalog/Pages 3 0 R", b"/Catalo ")
         )
     )
     with pytest.raises(PdfReadError):
@@ -1775,9 +1775,9 @@ def test_repair_root(caplog):
 
     # Invalid /Root Entry + error in get_object
     caplog.clear()
-    b = b.replace(b"/Root 1 0 R", b"/Root 2 0 R").replace(b"/Catalog", b"/Catalo ")
-    b = b[:5124] + b"A" + b[5125:]
-    reader = PdfReader(BytesIO(b))
+    data = b.replace(b"/Root 1 0 R", b"/Root 2 0 R").replace(b"/Catalog/Pages 3 0 R", b"/Catalo ")
+    data = data[:5124] + b"A" + data[5125:]
+    reader = PdfReader(BytesIO(data))
     with pytest.raises(PdfReadError):
         len(reader.pages)
     assert all(
@@ -1788,6 +1788,23 @@ def test_repair_root(caplog):
         )
     )
 
+    # Invalid /Root Entry without /Type, but /Pages.
+    caplog.clear()
+    reader = PdfReader(
+        BytesIO(
+            b.replace(b"/Root 1 0 R", b"/Root 2 0 R").replace(b"/Catalog", b"/Catalo ")
+        )
+    )
+    assert len(reader.pages) == 1
+    assert all(
+        msg in caplog.text
+        for msg in (
+            "Invalid Root object in trailer",
+            'Searching object with "/Catalog" key',
+            f"Possible root found at IndirectObject(2, 0, {id(reader)}), but missing /Catalog key"
+        )
+    )
+
 
 @pytest.mark.enable_socket
 def test_issue3151(caplog):