[iamsaugatpandey]

date.fromisoformat() accepts strings of length 7, 8 or 10, but neither parser checks that it read the whole string. A 10-character string without a - after the year goes down the basic-format path, which reads YYYYMMDD (or YYYYWwwD) and returns, ignoring the last two characters. So a truncated or mistyped timestamp turns into a plausible but wrong date:

>>> from datetime import date
>>> date.fromisoformat('2020010112')    # e.g. a YYYYMMDDHH stamp
datetime.date(2020, 1, 1)
>>> date.fromisoformat('2020W011xx')
datetime.date(2019, 12, 30)
>>> date.fromisoformat('20200101é')     # C only: 9 characters, but 10 bytes in UTF-8
datetime.date(2020, 1, 1)

With this change, all three raise ValueError: Invalid isoformat string, as datetime.fromisoformat() already does for the same input.

The fix checks that the parser stopped exactly at the end of the date, in both parse_isoformat_date() in C and _parse_isoformat_date() in _pydatetime. datetime.fromisoformat() uses the same function for its date part, but there the date always ends at the separator for valid strings. I compared the C module before and after on ~300k fuzzed strings: only date.fromisoformat() results changed (all from accepted to rejected), and datetime.fromisoformat() results did not change at all.

While adding the tests I found that the pure-Python datetime.fromisoformat() has a similar problem. It skipped time parsing whenever the time part was empty, so a date followed by any single character was accepted:

>>> import _pydatetime
>>> _pydatetime.datetime.fromisoformat('2020-01-01T')
datetime.datetime(2020, 1, 1, 0, 0)
>>> _pydatetime.datetime.fromisoformat('2021-01-071')
datetime.datetime(2021, 1, 7, 0, 0)

The C implementation already rejects these. It now parses the time whenever something follows the date, the same as C. It's the same "didn't consume the whole input" bug, so I kept it here, but I'm happy to split it out if you'd prefer.

Tests are added to test_fromisoformat_fails and test_fromisoformat_fails_datetime, so they run against both implementations.

• Issue: date.fromisoformat() silently ignores the last two characters of a 10-character basic-format string #158500

[Copilot]

Copilot was unable to review this pull request because the user who requested the review has insufficient remaining quota.

[Copilot]

Copilot was unable to review this pull request because the user who requested the review has insufficient remaining quota.