Terminal software identity-assisted wcswidth()

[jquast]

VS-15 narrows width

• VS-15 support is also re-introduced (Previously Variation Selector 15 (VS-15, U+FE0E) support. #120, Should wcswidth return Narrow (1) for Basic Emojis followed by VS-15, U+FE0E? #211), now that we can properly account for terminal overrides and that ghostty and kitty match our interpretation of standard regarding VS-15 as applied to terminals.

Correction Tables

• disabled by default (term_program=False),

  • although it would improve user experience for downstream dependencies like urwid or python-prompt-toolkit to set term_program=True, that they benefit when unique identifying TERM or TERM_PROGRAM
  • it is likely to cause downstream projects to inexplicably fail automatic tests,
  • and with varying results depending on the terminal used by the developer troubleshooting it.

• Use environment values of TERM_PROGRAM and TERM, and suggest that calling programs to use XTVERSION or ENQ query result, to hint to width() and similar for wcswidth(term_program='xterm.js'),

  • When set, and it matches the automatic "overrides" tables, a correction is applied.
  • This is made possible by the data gathered by https://ucs-detect.readthedocs.io/ project.
  • Only terminals that can be automatically identified by TERM_PROGRAM, XTVERSION, ENQ, or unique TERM are offered
  • Small performance impact due to extra call for terminal resolution and whether any correction should be applied.

• Correction tables for ~20+ modern terminals, ~14 unique fingerprints, shared by hashing:

  • alacritty
  • apple_terminal
  • bobcat
  • contour
  • extraterm
  • foot
  • ghostty
  • iterm2
  • kitty
  • konsole
  • mintty
  • mlterm
  • pterm
  • rio
  • st
  • terminology
  • urxvt
  • vte
  • warp
  • wezterm
  • xterm.js
  • zellij

• And their aliases: hyper, iterm.app, putty, rxvt, rxvt-unicode-256color, screen, st-256color, tabby, tmux-256color, vscode, warpterminal, xterm-ghostty, xterm-kitty.

Closes #211.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 100.00%. Comparing base (e4f76d5) to head (61b7877).
⚠️ Report is 1 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##            master      #220    +/-   ##
==========================================
  Coverage   100.00%   100.00%            
==========================================
  Files           23        27     +4     
  Lines         1570      1924   +354     
  Branches       390       456    +66     
==========================================
+ Hits          1570      1924   +354     

☔ 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.

[codspeed-hq]

Merging this PR will degrade performance by 13.5%

⚠️ 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

❌ 4 (👁 4) regressed benchmarks
✅ 84 untouched benchmarks
🆕 20 new benchmarks

Performance Changes

	Benchmark	BASE	HEAD	Efficiency
🆕	test_wcstwidth_emoji_term_program[ghostty]	N/A	1.3 ms	N/A
🆕	test_wcstwidth_emoji_term_program[xterm.js]	N/A	547 µs	N/A
🆕	test_wcstwidth_ri_term_program[ghostty]	N/A	343.4 ms	N/A
🆕	test_wcstwidth_ri_term_program[xterm.js]	N/A	344.7 ms	N/A
🆕	test_wcstwidth_term_program[ghostty]	N/A	729.1 µs	N/A
🆕	test_wcstwidth_term_program[xterm.js]	N/A	874.7 µs	N/A
🆕	test_wcstwidth_vs15_term_program[ghostty]	N/A	680.2 µs	N/A
🆕	test_wcstwidth_vs15_term_program[xterm.js]	N/A	1 ms	N/A
🆕	test_wcstwidth_vs16_term_program[ghostty]	N/A	666.8 µs	N/A
🆕	test_wcstwidth_vs16_term_program[xterm.js]	N/A	760.3 µs	N/A
🆕	test_wcstwidth_wide_term_program[ghostty]	N/A	561.3 µs	N/A
🆕	test_wcstwidth_wide_term_program[xterm.js]	N/A	1.1 ms	N/A
🆕	test_width_ri_term_program[ghostty]	N/A	343.6 ms	N/A
🆕	test_width_ri_term_program[xterm.js]	N/A	344.9 ms	N/A
🆕	test_width_term_program[ghostty]	N/A	783 µs	N/A
🆕	test_width_term_program[xterm.js]	N/A	932.8 µs	N/A
🆕	test_width_vs15_term_program[ghostty]	N/A	735.1 µs	N/A
🆕	test_width_vs15_term_program[xterm.js]	N/A	1.1 ms	N/A
🆕	test_width_vs16_term_program[ghostty]	N/A	718.8 µs	N/A
🆕	test_width_vs16_term_program[xterm.js]	N/A	812.3 µs	N/A
...	...	...	...	...

ℹ️ Only the first 20 benchmarks are displayed. Go to the app to view all benchmarks.

---

_{Comparing jq/dynawidth (61b7877) with master (e4f76d5)}

[jquast]

I forgot to include language grapheme overrides, which is arguably the most useful -- but there is a lot of difficulty in doing this correctly, it might require that wcswidth() call iter_graphemes(), which is high performing for python 3.15 but it will definitely have some performance impact for earlier python