Condense the 2.1 release notes and give them an intro

[derek73]

Summary

The 2.1.0 section ran 4,908 words across 26 entries (188 average, longest 635). That is a specification rather than release notes, and it was largely a second copy of the docs: usage.rst already carries the glued/spaced honorific peel, the comma and 间隔号 rules, the reach limits, the period rule, and even the vocabulary rationale for 殿/양/博士.

Now 2,137 words in the same 26 entries — a summary and a link each. Nothing is dropped; where content had no docs home it moved there instead.

Changes

An intro, in the shape 2.0.0 uses: what the release is for, that most of it is default-on and why, and the one-liner that restores 2.0's reading.

Entries condensed to a summary plus a cross-reference. Every :ref:/:doc: target was verified to exist before use.

Em-dashes: 75 → 0. At one every 65 words they had stopped marking anything.

A new Decomposed text section in usage.rst. This was the one topic with genuinely zero docs coverage (grep for NFD/NFC/decomposed across docs/ returned nothing), and it is a real concern for anyone reading names off macOS. Two other uncovered topics, the 〆 shime mark and the Ukrainian й, stay release-log-only on purpose: they are "we now recognize X" facts rather than how-to material.

One correction worth flagging

The release log implied this gotcha:

parse(decomposed).family == "김"      # False, same name

For an unsplit decomposed name that is False for the wrong reason: family holds the whole string 김민준, so it differs in value, not just in encoding. The spaced form is the one that demonstrates the actual point, and the docs now use it:

decomposed = unicodedata.normalize("NFD", "김 민준")
parse(decomposed).family == "김"                          # False
unicodedata.normalize("NFC", parse(decomposed).family)    # '김'

Both lines verified against the parser.

Verification

• sphinx-build -b html — exit 0, 0 warnings
• sphinx-build -b doctest — exit 0
• python -m doctest README.rst — exit 0
• uv run pytest -q — 3060 passed

No code changed.

🤖 Generated with Claude Code

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 98.48%. Comparing base (bc8f8a5) to head (95bc27b).
⚠️ Report is 4 commits behind head on master.

Additional details and impacted files

@@           Coverage Diff           @@
##           master     #336   +/-   ##
=======================================
  Coverage   98.48%   98.48%           
=======================================
  Files          41       41           
  Lines        2845     2845           
=======================================
  Hits         2802     2802           
  Misses         43       43           

☔ 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.