BUG: Dates conversion not working with Z00'00' (#1946)

Closes #1943
py-pdf · Jul 6, 2023 · c239073 · c239073
1 parent 29c79fc
commit c239073
Show file tree

Hide file tree

Showing 2 changed files with 29 additions and 5 deletions.
diff --git a/pypdf/_reader.py b/pypdf/_reader.py
@@ -242,14 +242,16 @@ def creation_date(self) -> Optional[datetime]:
  text = self._get_text(DI.CREATION_DATE)
  if text is None:
  return None
- return datetime.strptime(text.replace("'", ""), "D:%Y%m%d%H%M%S%z")
+ return datetime.strptime(
+ text.replace("Z", "+").replace("'", ""), "D:%Y%m%d%H%M%S%z"
+ )
 
  @property
  def creation_date_raw(self) -> Optional[str]:
  """
  The "raw" version of creation date; can return a ``ByteStringObject``.
 
- Typically in the format ``D:YYYYMMDDhhmmss[+-]hh'mm`` where the suffix
+ Typically in the format ``D:YYYYMMDDhhmmss[+Z-]hh'mm`` where the suffix
  is the offset from UTC.
  """
  return self.get(DI.CREATION_DATE)
@@ -264,15 +266,17 @@ def modification_date(self) -> Optional[datetime]:
  text = self._get_text(DI.MOD_DATE)
  if text is None:
  return None
- return datetime.strptime(text.replace("'", ""), "D:%Y%m%d%H%M%S%z")
+ return datetime.strptime(
+ text.replace("Z", "+").replace("'", ""), "D:%Y%m%d%H%M%S%z"
+ )
 
  @property
  def modification_date_raw(self) -> Optional[str]:
  """
  The "raw" version of modification date; can return a
  ``ByteStringObject``.
 
- Typically in the format ``D:YYYYMMDDhhmmss[+-]hh'mm`` where the suffix
+ Typically in the format ``D:YYYYMMDDhhmmss[+Z-]hh'mm`` where the suffix
  is the offset from UTC.
  """
  return self.get(DI.MOD_DATE)
@@ -644,7 +648,10 @@ def _build_field(
  if s not in states:
  states.append(s)
  retval[key][NameObject("/_States_")] = ArrayObject(states)
- if obj.get(FA.Ff, 0) & FA.FfBits.NoToggleToOff != 0 and "/Off" in retval[key]["/_States_"]:
+ if (
+ obj.get(FA.Ff, 0) & FA.FfBits.NoToggleToOff != 0
+ and "/Off" in retval[key]["/_States_"]
+ ):
  del retval[key]["/_States_"][retval[key]["/_States_"].index("/Off")]
 
  def _check_kids(

diff --git a/tests/test_reader.py b/tests/test_reader.py
@@ -112,6 +112,23 @@ def test_read_metadata(pdf_path, expected):
  assert metadict["/Title"] == docinfo.title
 
 
+def test_iss1943():
+ reader = PdfReader(RESOURCE_ROOT / "crazyones.pdf")
+ docinfo = reader.metadata
+ docinfo.update(
+ {
+ NameObject("/CreationDate"): TextStringObject("D:20230705005151Z00'00'"),
+ NameObject("/ModDate"): TextStringObject("D:20230705005151Z00'00'"),
+ }
+ )
+ docinfo.creation_date
+ docinfo.creation_date_raw
+ docinfo.modification_date
+ docinfo.modification_date_raw
+ docinfo.update({NameObject("/CreationDate"): NumberObject(1)})
+ assert docinfo.creation_date is None
+
+
 @pytest.mark.samples()
 @pytest.mark.parametrize(
  "pdf_path", [SAMPLE_ROOT / "017-unreadable-meta-data/unreadablemetadata.pdf"]