[rahul05ranjan]

Fixes #1074.

Summary

from_format crashed with a raw AttributeError: 'NoneType' object has no attribute 'values' for every ordinal token:

input	before	after
from_format("7 Juni 1982", "Do MMMM YYYY", locale="de") (no ordinals in de)	AttributeError	1982-06-07
from_format("2nd 1982", "Qo YYYY", locale="en")	AttributeError (any locale)	1982-04-01
from_format("6th 1982", "Mo YYYY")	AttributeError	1982-06-01
from_format("158th 1982", "DDDo YYYY")	AttributeError	1982-06-07
from_format(..., "do/e/eo/Wo/wo ...")	AttributeError	ValueError: Unsupported token: ...

Meanwhile format() happily emits all of these (format("Qo YYYY") → '2nd 1982'), so no ordinal token could ever be round-tripped.

Root cause

src/pendulum/formatting/formatter.py:

1. The Do entry in _LOCALIZABLE_TOKENS calls locale.get("custom.ordinal").values() — None for locales without ordinals (bg, de, ru, pl, uk, da, sv, ...). Locale.ordinalize() (the format() side) already falls back to the plain number there, so the parse pattern must accept plain digits too.
2. Qo/Mo/DDDo/do/e/eo/Wo/wo are None in _LOCALIZABLE_TOKENS; _replace_tokens passes that to locale.translation(None).values() and crashes before reaching its own ValueError("Unsupported token: ...") guard.

Fix

• _ordinal_candidates(locale): builds \d+<suffix> patterns from custom.ordinal, falling back to \d+ for locales without ordinals — mirroring Locale.ordinalize(), so format()/from_format round-trip in every locale.
• Do, Mo, Qo, DDDo now parse to their fields (day/month/quarter/day_of_year), all of which _check_parsed already consumes.
• The deliberately unparseable do, e, eo, Wo, wo now raise the regular ValueError("Unsupported token: X") instead of AttributeError.
• The bracket-literal / escaped-character protection issue is out of scope here (it is being addressed in Fix from_format crash on multi-character literal blocks #972).

Verification

• 10 new regression tests in tests/datetime/test_from_format.py (Do with/without locale ordinals incl. de, Mo, Qo, DDDo, and the clean ValueError for the unsupported tokens).
• The pre-existing failure set of tests/datetime/test_from_format.py is byte-identical before/after (those 49 failures are the time-machine-gated travel tests; time_machine is unavailable in this environment — diff of failing test names verified with git stash).
• tests/formatting + tests/parsing: identical failure set before/after (all previously-passing tests still pass).
• Python 3.13, Linux, system tzdata + dateutil from source (no optional deps installed).