gh-124529: Fix _strptime to make %c/%x accept a year with fewer digits

[zuo]

_strptime.TimeRE has been modified, regarding the %c and %x format
codes, to make datetime.datetime.strptime(), datetime.date.strptime()
and time.strptime() accept inputs that contain year numbers consisting
of fewer digits than usual, i.e., not zero-padded to the usual width
of 2 or 4 (which is appropriate for the format code and current locale).

Thanks to that, now the desired behavior described in gh-124529 should be
consistently observed; certain strftime/strptime round trips (involving
%c/%x and small year numbers) no longer raise ValueError for some
platforms/locales (in particular, this was the case on Linux, for various
locales, including C/C.UTF-8).

[zuo]

PS Note that the changes concern only the %c/%x-parsing-related behavior, and nothing changes when it comes to %y, %Y and %G (and any non-standard year-representation-related format codes).

[zuo]

PPS I added (obviously) the necessary tests (directly related to the changes), but also some indirectly-related ones (especially confirming the unchanged behavior for %y/%Y/%G).

PPPS A suitable mention in the docs has also been added.

[zuo]

I hope it's ready. [EDIT: not yet]

[zuo]

OK, I hope now it's ready. :)

[zuo]

TODO: if this PR's approach is accepted at all, it would be good to adjust it to the changes from PR #124946 (after it is merged).

[donbarbos]

looks like good idea and important fix :)
@zuo are you ready to continue?

[zuo]

looks like good idea and important fix :) @zuo are you ready to continue?

Yes, I am :)

Yet I'm not sure what is the next step...

[StanFromIreland]

This has a lot of conflicts, and needs updates.

[github-actions]

This PR is stale because it has been open for 30 days with no activity.

[zuo]

For the record, I'd be happy to continue with this, but I guess first the discussion in the related issue (#124529) needs to be resolved -- to decide whether this path is OK or not.