dateutil parser bug for years < 100

[Holger Joukl]

dateutil parser erroneously ignores significant leading year digits for 4-digit years < 100:

(Python 2.7.3)
>>> import datetime
>>> import dateutil
>>> import dateutil.parser
>>> dateutil.__version__
'1.5'
>>> dt = datetime.datetime(3, 2, 1)
>>> print dt.isoformat()
0003-02-01T00:00:00
>>> dateutil.parser.parse(dt.isoformat(), yearfirst=True, dayfirst=False)datetime.datetime(2003, 2, 1, 0, 0)
>>> dateutil.parser.parse(dt.isoformat(), yearfirst=True, dayfirst=False).isoformat()
'2003-02-01T00:00:00'
>>>
>>> dt = datetime.datetime(99, 2, 1)
>>> print dt.isoformat()
0099-02-01T00:00:00
>>> dateutil.parser.parse(dt.isoformat(), yearfirst=True, dayfirst=False).isoformat()
'1999-02-01T00:00:00'

Looks like the significant leading digits get ignored, triggering the 2-digit-year rules.

Years >= 100 work as expected:
>>> dt = datetime.datetime(100, 2, 1)
>>> dateutil.parser.parse(dt.isoformat(), yearfirst=True, dayfirst=False).isoformat()
'0100-02-01T00:00:00'
>>>

[Holger Joukl (jholg)]

A note for posterity: This has long(?) been fixed - modern dateutil versions (e.g. 2.8.1) do not suffer from this bug anymore.

(dateutil issue tracking has moved to GitHub years ago)