datetime error message is different between _pydatetime.py and _datetimemodule.c

[mattip]

Bug report

Bug description:

import _datetime, _pydatetime
_datetime.date(1, 1, 50)
Traceback (most recent call last):
 File "<stdin>", line 1, in <module>
ValueError: day is out of range for month```

_pydatetime.date(1, 1, 50)
Traceback (most recent call last):
 File "<stdin>", line 1, in <module>
 File "/usr/lib/python3.12/_pydatetime.py", line 960, in __new__
 year, month, day = _check_date_fields(year, month, day)
 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
 File "/usr/lib/python3.12/_pydatetime.py", line 535, in _check_date_fields
 raise ValueError('day must be in 1..%d' % dim, day)
ValueError: ('day must be in 1..31', 50)

The error message differs between the two implementations of datetime. This came up when testing PyPy, which uses the pure-python datetime implementation. xref conda-forge/rtoml-feedstock#1 (comment)

CPython versions tested on:

3.12

Operating systems tested on:

Linux

Linked PRs

• gh-109798: Normalize _datetime and datetime error messages #127345

[donbarbos]

@mattip So, you suggest to leave the option as in the _datetime implementation?

I can send a PR

[mattip]

I don't really have an opinion which error message is better, only that they be the same.

[donbarbos]

Ok, I will leave error message as in the c implementation

Firstly, it is better to output a string (not tuple), and in the python implementation, the second value in tuple is redundant.
Secondly, it seems to me that it is the main implementation.

[donbarbos]

I also add more inconsistency in the error messages that I found in the code

1. Errors about date and time have inconsistent messages, for example problem with year field may raise next errors:

ValueError: ('year must be in 1..9999', {year})
ValueError: Year is out of range: {year}
ValueError: year {year} is out of range
ValueError: Year {year} is out of range

We need to make same message template for the fields year, month, day, hour, minute, second, microsecond and fold.

2. Method date.fromisocalendar may raise next errors depends on implementation:

ValueError: Invalid weekday: {day} (range is [1, 7])
ValueError: Invalid day: {day} (range is [1, 7])

3. Error with offset argument in timezone may raise next errors:

ValueError: offset must be a timedelta strictly between -timedelta(hours=24) and timedelta(hours=24), not datetime.timedelta(days=21). # with dot
ValueError: offset must be a timedelta strictly between -timedelta(hours=24) and timedelta(hours=24), not datetime.timedelta(days=21) # without dot
ValueError: offset must be a timedelta strictly between -timedelta(hours=24) and timedelta(hours=24).
ValueError: {name}()={offset}, must be a timedelta strictly between -timedelta(hours=24) and timedelta(hours=24)

4. Errors about type of arugment

for example if i call _pydatetime.datetime() and _datetime.datetime() with tz i get:

TypeError: tzinfo argument must be None or of a tzinfo subclass # in _pydatetime
TypeError: tzinfo argument must be None or of a tzinfo subclass, not type 'int' # in _datetime

[StanFromIreland]

PR was merged, issue is complete :-)

It seems there was no plan to back port which is for the best for backwards compatibility.