Merge pull request #23 from jepler/no-tracemalloc

Improve coverage & create InvalidContentError

.coveragerc

@@ -1,7 +1,15 @@
-# SPDX-FileCopyrightText: 2021 Jeff Epler
+# SPDX-FileCopyrightText: 2021-2024 Jeff Epler
 #
 # SPDX-License-Identifier: GPL-3.0-only
-[run]
-omit =
-    */site-packages/*
-    test*.py
+[report]
+exclude_also =
+    def __repr__
+    if self.debug:
+    if settings.DEBUG
+    raise AssertionError
+    raise NotImplementedError
+    if 0:
+    if __name__ == .__main__.:
+    if TYPE_CHECKING:
+    class .*\bProtocol\):
+    @(abc\.)?abstractmethod

.github/workflows/test.yml

@@ -67,7 +67,7 @@ jobs:
       run: make mypy
 
     - name: Test
-      run: python -X tracemalloc=3 -mcoverage run --branch -m unittest testleapseconddata.py && python -mcoverage report --fail-under=100 && python -mcoverage xml
+      run: python -mcoverage run --branch -m unittest testleapseconddata.py && python -mcoverage report --fail-under=100 && python -mcoverage xml
 
   pre-commit:
     runs-on: ubuntu-latest

leapseconddata/__init__.py

@@ -25,12 +25,16 @@
 import datetime
 import hashlib
 import io
+import itertools
 import logging
 import pathlib
 import re
 import urllib.request
 from dataclasses import dataclass, field
-from typing import BinaryIO, ClassVar
+from typing import TYPE_CHECKING, BinaryIO, ClassVar
+
+if TYPE_CHECKING:  # pragma no cover
+    from collections.abc import Sequence
 
 tai = datetime.timezone(datetime.timedelta(0), "TAI")
 
@@ -58,6 +62,10 @@ class InvalidHashError(ValueError):
     """The file hash could not be verified"""
 
 
+class InvalidContentError(ValueError):
+    """A line in the file was not valid"""
+
+
 def _from_ntp_epoch(value: int) -> datetime.datetime:
     return NTP_EPOCH + datetime.timedelta(seconds=value)
 
@@ -237,6 +245,7 @@ def from_standard_source(
         when: datetime.datetime | None = None,
         *,
         check_hash: bool = True,
+        custom_sources: Sequence[str] = (),
     ) -> LeapSecondData:
         """Get the list of leap seconds from a standard source.
 
@@ -246,22 +255,28 @@ def from_standard_source(
         Using a list of standard sources, including network sources, find a
         leap-second.list data valid for the given timestamp, or the current
         time (if unspecified)
+
+        If ``custom_sources`` is specified, this list of URLs is checked before
+        the hard-coded sources.
         """
-        for location in cls.standard_file_sources + cls.standard_network_sources:
+        for location in itertools.chain(custom_sources, cls.standard_file_sources, cls.standard_network_sources):
             logging.debug("Trying leap second data from %s", location)
             try:
                 candidate = cls.from_url(location, check_hash=check_hash)
-            except InvalidHashError:  # pragma no cover
+            except InvalidHashError:
                 logging.warning("Invalid hash while reading %s", location)
                 continue
-            if candidate is None:  # pragma no cover
+            except InvalidContentError as e:
+                logging.warning("Invalid content while reading %s: %s", location, e)
                 continue
-            if candidate.valid(when):  # pragma no branch
+            if candidate is None:
+                continue
+            if candidate.valid(when):
                 logging.info("Using leap second data from %s", location)
                 return candidate
-            logging.warning("Validity expired for %s", location)  # pragma no cover
+            logging.warning("Validity expired for %s", location)
 
-        raise ValidityError("No valid leap-second.list file could be found")  # pragma no cover
+        raise ValidityError("No valid leap-second.list file could be found")
 
     @classmethod
     def from_file(
@@ -338,36 +353,39 @@ def from_open_file(cls, open_file: BinaryIO, *, check_hash: bool = True) -> Leap
 
         hasher = hashlib.sha1()
 
-        for row in open_file:
-            row = row.strip()  # noqa: PLW2901
-            if row.startswith(b"#h"):
-                content_hash = cls._parse_content_hash(row)
-                continue
+        for row_ws in open_file:
+            row = row_ws.strip()
+            try:
+                if row.startswith(b"#h"):
+                    content_hash = cls._parse_content_hash(row)
+                    continue
 
-            if row.startswith(b"#@"):
-                parts = row.split()
-                hasher.update(parts[1])
-                valid_until = _from_ntp_epoch(int(parts[1]))
-                continue
+                if row.startswith(b"#@"):
+                    parts = row.split()
+                    hasher.update(parts[1])
+                    valid_until = _from_ntp_epoch(int(parts[1]))
+                    continue
 
-            if row.startswith(b"#$"):
-                parts = row.split()
-                hasher.update(parts[1])
-                last_updated = _from_ntp_epoch(int(parts[1]))
-                continue
+                if row.startswith(b"#$"):
+                    parts = row.split()
+                    hasher.update(parts[1])
+                    last_updated = _from_ntp_epoch(int(parts[1]))
+                    continue
 
-            row = row.split(b"#")[0].strip()  # noqa: PLW2901
-            content_to_hash.extend(re.findall(rb"\d+", row))
+                row = row.split(b"#")[0].strip()
+                content_to_hash.extend(re.findall(rb"\d+", row))
 
-            parts = row.split()
-            if len(parts) != 2:  # noqa: PLR2004
-                continue
-            hasher.update(parts[0])
-            hasher.update(parts[1])
+                parts = row.split()
+                if len(parts) != 2:  # noqa: PLR2004
+                    continue
+                hasher.update(parts[0])
+                hasher.update(parts[1])
 
-            when = _from_ntp_epoch(int(parts[0]))
-            tai_offset = datetime.timedelta(seconds=int(parts[1]))
-            leap_seconds.append(LeapSecondInfo(when, tai_offset))
+                when = _from_ntp_epoch(int(parts[0]))
+                tai_offset = datetime.timedelta(seconds=int(parts[1]))
+                leap_seconds.append(LeapSecondInfo(when, tai_offset))
+            except Exception as e:
+                raise InvalidContentError(f"Failed to parse: {row!r}: {e}") from e
 
         if check_hash:
             if content_hash is None:

testleapseconddata.py

@@ -91,6 +91,18 @@ def test_empty(self) -> None:
             datetime.timedelta(seconds=0),
         )
 
+    def test_invalid2(self) -> None:
+        when = datetime.datetime(datetime.MAXYEAR, 1, 1, tzinfo=datetime.timezone.utc) - datetime.timedelta(seconds=1)
+        with self.assertRaises(leapseconddata.ValidityError):
+            leapseconddata.LeapSecondData.from_standard_source(
+                when,
+                custom_sources=[
+                    "data:text/plain;base64,SGVsbG8sIFdvcmxkIQ==",
+                    "data:text/plain,%23h%099dac5845%208acd32c0%202947d462%20daf4a943%20f58d9391%0A",
+                    "file:///doesnotexist",
+                ],
+            )
+
     def test_tz(self) -> None:
         when = datetime.datetime(1999, 1, 1, tzinfo=datetime.timezone.utc) - datetime.timedelta(seconds=1)
         when = when.replace(fold=True)