fix(offset): support the military timezone letter j (local time)

[0xSoftBoi]

GNU date accepts every letter a–y as a military time zone. All of them are already in our table except one: j.

It is missing for a reason. j is the single letter that does not denote a fixed offset — it means local time, so it cannot be expressed as an Offset at all.

What GNU actually does

Verified against GNU coreutils 9.4, in TZ=America/New_York (EDT, UTC−4):

input	GNU result	meaning
8a	08:00 +01:00	UTC+1
8n	08:00 −01:00	UTC−1
8z	04:00 −0400	UTC
8j	08:00 −0400	local — identical to bare 8
2026-01-01 j	00:00 −0500	local, DST-aware
8foo	invalid date	unknown words are errors

The 2026-01-01 j row is the important one: it resolves to −05:00 in January and −04:00 in August. It follows the zone's DST rules, so no fixed offset can represent it.

The change

j is parsed as its own item and recorded on the builder as a flag rather than as an offset, so the local rule zone is used instead of a fixed one.

• 8j, 8 j, j 8 and 8J all work.
• It is still a time zone item, so 8j utc, 8 utc j and 8 j j are rejected as a repeated zone, matching GNU.
• The whole alphabetic word is consumed before comparing, so 8jan still parses as a date and jj is rejected rather than matching a j prefix.
• Unknown words stay errors — 8foo still fails, as it does in GNU.

Verification

I compared every letter a–z against GNU coreutils 9.4, comparing resolved Unix timestamps in a DST zone:

8a gnu=1787641200  pd=1787641200  ok
...
8j gnu=1787659200  pd=1787659200  ok
...
8z gnu=1787644800  pd=1787644800  ok
=== mismatches: 0 ===

All 26 letters now resolve to the same instant as GNU. 403 tests pass, fmt and clippy clean.

Relationship to #286

This supersedes #286, and confirms @sylvestre's review there was correct.

#286 assumed GNU silently discards unrecognized trailing alphabetic tokens after a number. It does not — date -d '8foo' is an invalid date. The only such token GNU actually accepts is j, and for a reason #286 did not model: it is a time zone, not noise to be skipped.

#286 has also drifted five months behind main and its diff would revert several later fixes, so this is a fresh branch from current main rather than an update to it.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 99.34%. Comparing base (1c7fdea) to head (180746e).

Additional details and impacted files

@@           Coverage Diff           @@
##             main     #320   +/-   ##
=======================================
  Coverage   99.33%   99.34%           
=======================================
  Files          21       21           
  Lines        4075     4125   +50     
  Branches      129      136    +7     
=======================================
+ Hits         4048     4098   +50     
  Misses         26       26           
  Partials        1        1           

Flag	Coverage Δ
macos_latest	99.34% <100.00%> (?)
ubuntu_latest	99.34% <100.00%> (+<0.01%)	⬆️
windows_latest	13.72% <3.70%> (-0.20%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.