BUG: Check version of crypt provider (#2115)

Closes #2107

Author: MartinThoma

pypdf/_crypt_providers/__init__.py

@@ -39,6 +39,12 @@
         rc4_decrypt,
         rc4_encrypt,
     )
+    from pypdf._utils import Version
+
+    if Version(crypt_provider[1]) <= Version("3.0"):
+        # This is due to the backend parameter being required back then:
+        # https://cryptography.io/en/latest/changelog/#v3-1
+        raise ImportError("cryptography<=3.0 is not supported")  # pragma: no cover
 except ImportError:
     try:
         from pypdf._crypt_providers._pycryptodome import (  # type: ignore

pypdf/_utils.py

@@ -31,6 +31,7 @@
 
 import functools
 import logging
+import re
 import warnings
 from dataclasses import dataclass
 from datetime import datetime, timezone
@@ -41,6 +42,7 @@
     Any,
     Callable,
     Dict,
+    List,
     Optional,
     Pattern,
     Tuple,
@@ -588,3 +590,52 @@ def replace(self, new_image: Any, **kwargs: Any) -> None:
         self.name = self.name[: self.name.rfind(".")] + extension
         self.data = byte_stream
         self.image = img
+
+
+@functools.total_ordering
+class Version:
+    COMPONENT_PATTERN = re.compile(r"^(\d+)(.*)$")
+
+    def __init__(self, version_str: str) -> None:
+        self.version_str = version_str
+        self.components = self._parse_version(version_str)
+
+    def _parse_version(self, version_str: str) -> List[Tuple[int, str]]:
+        components = version_str.split(".")
+        parsed_components = []
+        for component in components:
+            match = Version.COMPONENT_PATTERN.match(component)
+            if not match:
+                parsed_components.append((0, component))
+                continue
+            integer_prefix = match.group(1)
+            suffix = match.group(2)
+            if integer_prefix is None:
+                integer_prefix = 0
+            parsed_components.append((int(integer_prefix), suffix))
+        return parsed_components
+
+    def __eq__(self, other: object) -> bool:
+        if not isinstance(other, Version):
+            return False
+        return self.components == other.components
+
+    def __lt__(self, other: Any) -> bool:
+        if not isinstance(other, Version):
+            raise ValueError(f"Version cannot be compared against {type(other)}")
+        min_len = min(len(self.components), len(other.components))
+        for i in range(min_len):
+            self_value, self_suffix = self.components[i]
+            other_value, other_suffix = other.components[i]
+
+            if self_value < other_value:
+                return True
+            elif self_value > other_value:
+                return False
+
+            if self_suffix < other_suffix:
+                return True
+            elif self_suffix > other_suffix:
+                return False
+
+        return len(self.components) < len(other.components)

tests/test_utils.py

@@ -7,6 +7,7 @@
 import pypdf._utils
 from pypdf._utils import (
     File,
+    Version,
     _get_max_pdf_version_header,
     _human_readable_bytes,
     deprecate_with_replacement,
@@ -351,3 +352,51 @@ def test_parse_datetime_err():
         parse_iso8824_date("D:20210408T054711Z")
     assert ex.value.args[0] == "Can not convert date: D:20210408T054711Z"
     assert parse_iso8824_date("D:20210408054711").tzinfo is None
+
+
+@pytest.mark.parametrize(
+    ("left", "right", "is_less_than"),
+    [
+        ("1", "2", True),
+        ("2", "1", False),
+        ("1", "1", False),
+        ("1.0", "1.1", True),
+        ("1", "1.1", True),
+        # suffix left
+        ("1a", "2", True),
+        ("2a", "1", False),
+        ("1a", "1", False),
+        ("1.0a", "1.1", True),
+        # I'm not sure about that, but seems special enoguht that it
+        # probably doesn't matter:
+        ("1a", "1.1", False),
+        # suffix right
+        ("1", "2a", True),
+        ("2", "1a", False),
+        ("1", "1a", True),
+        ("1.0", "1.1a", True),
+        ("1", "1.1a", True),
+        ("", "0.0.0", True),
+        # just suffix matters ... hm, I think this is actually wrong:
+        ("1.0a", "1.0", False),
+        ("1.0", "1.0a", True),
+    ],
+)
+def test_version_compare(left, right, is_less_than):
+    assert (Version(left) < Version(right)) is is_less_than
+
+
+def test_version_compare_equal_str():
+    a = Version("1.0")
+    assert (a == "1.0") is False
+
+
+def test_version_compare_lt_str():
+    a = Version("1.0")
+    with pytest.raises(ValueError) as exc:
+        a < "1.0"  # noqa
+    assert exc.value.args[0] == "Version cannot be compared against <class 'str'>"
+
+
+def test_bad_version():
+    assert Version("a").components == [(0, "a")]