Add Terminal.text_sized() kitty text sizing

[jquast]

From bin/text_sizing_protocol.py:

From bin/scroller.py:

kitty-text-sizing-.2.mp4

[codspeed-hq]

Merging this PR will degrade performance by 20.7%

⚠️ Different runtime environments detected

Some benchmarks with significant performance changes were compared across different runtime environments,
which may affect the accuracy of the results.

Open the report in CodSpeed to investigate

⚡ 2 improved benchmarks
❌ 3 (👁 3) regressed benchmarks
✅ 23 untouched benchmarks

Performance Changes

	Benchmark	BASE	HEAD	Efficiency
⚡	test_truncate_cjk	7,034.3 µs	485.9 µs	×14
⚡	test_truncate_emoji_zwj	1,609.3 µs	937.9 µs	+71.59%
👁	test_rjust_ansi	363.2 µs	458 µs	-20.7%
👁	test_ljust_ansi	363.4 µs	458 µs	-20.65%
👁	test_center_ansi	366.2 µs	460.8 µs	-20.54%

---

_{Comparing jq/text-sizing (58821f2) with master (ab1b0be)}

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 97.46%. Comparing base (ab1b0be) to head (58821f2).
⚠️ Report is 1 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #368      +/-   ##
==========================================
+ Coverage   97.44%   97.46%   +0.02%     
==========================================
  Files          13       13              
  Lines        3642     3673      +31     
  Branches      631      640       +9     
==========================================
+ Hits         3549     3580      +31     
  Misses         71       71              
  Partials       22       22              

☔ View full report in Codecov by Sentry.
📢 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.

[jquast]

The 3 regressions in ljust, center, and, rjust is that we now use wcwidth.width() with argumentcontrol_codes=parse, instead of control_codes=ignore as before.

"ignoring" control codes worked pretty well, but the more complex ones like OSC 66 kitty text sizing, and basic backspacing and cursor left "overtyping" all new in wcwidth 0.7.0 release require 'parse' to parse them.

So this regression is acceptable, we get better results by using 'parse' by default,

>>> wcwidth.ljust('hello\b\b\b\bworld', 10, control_codes='ignore', fillchar='.')
'hello\x08\x08\x08\x08world'
>>> wcwidth.ljust('hello\b\b\b\bworld', 10, control_codes='parse', fillchar='.')
'hello\x08\x08\x08\x08world....'

❌ test_ljust_ansi 363.4 µs 461.7 µs -21.29%
❌ test_center_ansi 366.2 µs 464.2 µs -21.11%
❌ test_rjust_ansi 363.2 µs 461.2 µs -21.24%

I have included new documentation in docs/measuring.txt about the new ability to parse cursor movement sequences.