datetime.datetime.fromtimestamp(0, tzlocal()) throws error

[mzhang13]

BPO	38645
Nosy	@abalkin, @ammaraskar, @pganssle, @corona10, @mzhang13, @PythonCoderAS

^{Note: these values reflect the state of the issue at the time it was migrated and might not reflect the current state.}

Show more details

GitHub fields:

assignee = None
closed_at = <Date 2019-11-01.03:11:09.212>
created_at = <Date 2019-10-30.15:07:25.514>
labels = ['3.7', 'type-crash']
title = 'datetime.datetime.fromtimestamp(0, tzlocal()) throws error'
updated_at = <Date 2019-11-01.03:11:09.206>
user = 'https://github.com/mzhang13'

bugs.python.org fields:

activity = <Date 2019-11-01.03:11:09.206>
actor = 'ammar2'
assignee = 'none'
closed = True
closed_date = <Date 2019-11-01.03:11:09.212>
closer = 'ammar2'
components = []
creation = <Date 2019-10-30.15:07:25.514>
creator = 'mzhang13'
dependencies = []
files = []
hgrepos = []
issue_num = 38645
keywords = []
message_count = 9.0
messages = ['355720', '355772', '355780', '355783', '355784', '355785', '355786', '355787', '355788']
nosy_count = 6.0
nosy_names = ['belopolsky', 'ammar2', 'p-ganssle', 'corona10', 'mzhang13', 'ThePokestarFan']
pr_nums = []
priority = 'normal'
resolution = 'third party'
stage = 'resolved'
status = 'closed'
superseder = None
type = 'crash'
url = 'https://bugs.python.org/issue38645'
versions = ['Python 3.7']

[mzhang13]

Discovered this while trying to use a function in boto3. Seems like when tzlocal() is passed with a 0 to datetime.datetime.fromtimestamp(), it throws an "[Errno 22] Invalid argument" error. Using 3.7.4 via Anaconda on Windows 10, and tzlocal().tznames's local timezone for me is EST/EDT.

Can replicate through this:

import datetime
from dateutil.tz import tzlocal

print(datetime.datetime.fromtimestamp(0, tzlocal()))

[PythonCoderAS]

For me, it seems that dateutil is not a standard Python library module. When I run import dateutil it reports that there is no module called dateutil.

>>> import dateutil
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
ImportError: No module named dateutil

[corona10]

Python 3.9.0a0 (heads/master:112f2b805b, Nov  1 2019, 10:48:20)
[GCC 7.4.0] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import datetime
>>> utc_tzinfo = datetime.timezone.utc
>>> print(datetime.datetime.fromtimestamp(0, utc_tzinfo))
1970-01-01 00:00:00+00:00

works properly on my Linux machine.
Can you gave me which can reproduce based on stdlib library?

[mzhang13]

I think it might be something Windows specific; the original comment includes the version and code to reproduce.

[PythonCoderAS]

Can you run

>>import dateutil
>>print(repr(dateutil))

[ammaraskar]

It is windows specific, but I don't think this is a dateutil bug rather than the python stdlib:

Python 3.6.5 (v3.6.5:f59c0932b4, Mar 28 2018, 16:07:46) [MSC v.1900 32 bit (Intel)] on win32
Type "help", "copyright", "credits" or "license" for more information.
>>> import datetime
>>> from dateutil.tz import tzlocal
>>> print(datetime.datetime.fromtimestamp(0, tzlocal()))
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "D:\Python365\lib\site-packages\dateutil\tz\_common.py", line 144, in fromutc
    return f(self, dt)
  File "D:\Python365\lib\site-packages\dateutil\tz\_common.py", line 258, in fromutc
    dt_wall = self._fromutc(dt)
  File "D:\Python365\lib\site-packages\dateutil\tz\_common.py", line 238, in _fromutc
    dtdst = enfold(dt, fold=1).dst()
  File "D:\Python365\lib\site-packages\dateutil\tz\tz.py", line 225, in dst
    if self._isdst(dt):
  File "D:\Python365\lib\site-packages\dateutil\tz\tz.py", line 288, in _isdst
    if self.is_ambiguous(dt):
  File "D:\Python365\lib\site-packages\dateutil\tz\tz.py", line 250, in is_ambiguous
    (naive_dst != self._naive_is_dst(dt - self._dst_saved)))
  File "D:\Python365\lib\site-packages\dateutil\tz\tz.py", line 254, in _naive_is_dst
    return time.localtime(timestamp + time.timezone).tm_isdst
OSError: [Errno 22] Invalid argument

As you can see, none of that backtrace is within Python, it's all inside dateutil code. More than likely they are calling time.localtime with a value of less than 0 which throws an error on Windows. (see also bpo-29097)

[ammaraskar]

Err...I mean I think this is a dateutil bug rather than in the Python stdlib.

[mzhang13]

Yeah, seconded with my own tracing as well.

[ammaraskar]

Closing, here's the bug in dateutil's repo: dateutil/dateutil#197