diff --git a/CHANGES.md b/CHANGES.md
index cd7f4c9..f6dcafe 100644
--- a/CHANGES.md
+++ b/CHANGES.md
@@ -2,6 +2,10 @@
## Unreleased
+- Fixed: `markfind import` crashed with a traceback on a property list whose
+ XML declaration named an unknown encoding. It is now reported like any other
+ unreadable file. Found by fuzzing the parsers (#78).
+
- Fixed: a bookmark pointing at a page that only redirects, through
``, was indexed as the stub. One real example
indexed 28 characters of "Click here to be redirected." instead of the
diff --git a/src/markfind/importers.py b/src/markfind/importers.py
index c0a05c7..0705612 100644
--- a/src/markfind/importers.py
+++ b/src/markfind/importers.py
@@ -324,7 +324,14 @@ def parse_safari_plist(data: bytes) -> list[Bookmark]:
"""Parse Safari's ``Bookmarks.plist`` (binary or XML property list)."""
try:
root = plistlib.loads(data)
- except (plistlib.InvalidFileException, ExpatError, ValueError, OverflowError) as exc:
+ # LookupError: the XML declaration names an encoding Python doesn't have.
+ except (
+ plistlib.InvalidFileException,
+ ExpatError,
+ ValueError,
+ OverflowError,
+ LookupError,
+ ) as exc:
raise ValueError(f"not a readable Safari bookmarks file: {exc}") from exc
if not isinstance(root, dict):
raise ValueError("not a readable Safari bookmarks file: unexpected contents")
diff --git a/tests/test_safari.py b/tests/test_safari.py
index ade3d63..64e93eb 100644
--- a/tests/test_safari.py
+++ b/tests/test_safari.py
@@ -83,6 +83,8 @@ def test_load_file_detects_plists(tmp_path, fmt):
[
b"bplist00 this is not really a binary plist",
b'x',
+ # An encoding Python doesn't have: plistlib raises LookupError, not ValueError.
+ b'',
plistlib.dumps(["a list, not a dict"]),
],
)
@@ -101,6 +103,13 @@ def test_binary_plist_written_by_apples_own_tool(tmp_path):
assert load_file(path) == EXPECTED
+def test_a_corrupt_plist_file_is_reported_not_raised(tmp_path, capsys):
+ path = tmp_path / "Bookmarks.plist"
+ path.write_bytes(b'')
+ assert cli.main(["--db", str(tmp_path / "db.sqlite"), "import", str(path)]) == 1
+ assert "could not read" in capsys.readouterr().err
+
+
def test_other_formats_still_detected(tmp_path):
html = tmp_path / "export.html"
html.write_text('A', encoding="utf-8")