thumbor / libthumbor / 39154363403

100%

push

github

Return None from parse_decrypted on oversized numbers

Url.parse_decrypted matches dimensions and crop coordinates with \d+ and
converts them with int(). Since Python 3.11 int() refuses strings longer
than 4300 digits and raises ValueError, so a path such as "9"*5000 + "x10/"
escaped the method with an exception that no caller expects, while every
other malformed path returns None. The conversion now runs inside a
try/except ValueError and returns None, the same result the method already
gave for every other path it cannot parse, and a docstring now states that
contract, which until now only the tests recorded.

The regex itself stays unchanged so Url.regex() keeps producing the same
pattern for callers that compile it themselves. The _int_or_0 helper moves
to module level and the dictionary construction to a classmethod to keep
the except block small.

9 of 9 new or added lines in 1 file covered. (100.0%)

272 of 272 relevant lines covered (100.0%)

5.0 hits per line