Unify TUI input editing in Tui::InputField

[tobi]

Stacked on #133.

Steal from tobi/try-cli: one real input widget instead of four copy-pasted editors. try-cli's tui_push/tui_pop is an ANSI style stack, not a view stack; Ruby's Screen/Section/Line layout is already better, so this only ports the input/terminal layer.

• Tui::InputField#handle_key used by search, rename, graduate, and the YES prompt
• Left/Right move the cursor; Ctrl-U kill-to-start; Delete-forward; Ctrl-H backspace
• Reverse-video cursor #to_s unchanged (existing screenshots still match)
• TryEntry#method_missing replaced with explicit accessors
• at_exit restores alt-screen / cooked / cursor if we crash in raw mode
• Drain leftover stdin on TUI exit; consume unknown CSI/mouse so trackpad junk doesn't land in the filter

MRI and native: 399/399 (original 397 plus two keyboard tests). runner_and_compare.sh 13/13.

Not in this PR: C style stack, C fuzzy scoring, ghost-placeholder rename UX, collapsing dialogs into a single-loop mode enum.

Demo

Same key story on both runtimes (cursor into spin, insert x, Ctrl-A/E, Ctrl-U, rename):

MRI

Native / Spinel

Playable: MRI asciinema · native asciinema

Benchmark (hyperfine)

MRI ruby try.rb vs native dist/try, 2026-08-13 16:04 EDT. hyperfine 1.19.0, --shell=none -w 3 -r 20. Ruby 3.3.8, Spinel 79b1aba8bdef (PR 3906). Native at measurement was ~1.0 MiB unstripped; current dist/try after -Os+strip is 705 KB. Native won every case.

Case	MRI	native	Speedup
--help	69.4 ± 4.8 ms	0.956 ± 0.096 ms	72.6×
--version	70.8 ± 6.3 ms	0.959 ± 0.125 ms	73.8×
no args (exit 2)	68.7 ± 4.3 ms	0.982 ± 0.068 ms	69.9×
init	96.5 ± 9.4 ms	21.2 ± 1.3 ms	4.5×
TUI empty	74.0 ± 8.4 ms	1.28 ± 0.14 ms	57.8×
TUI 5 dirs	72.0 ± 6.0 ms	1.38 ± 0.08 ms	52.1×
TUI --and-keys=beta	72.4 ± 5.9 ms	2.70 ± 0.35 ms	26.8×
TUI exec beta	74.6 ± 8.1 ms	1.49 ± 0.16 ms	50.2×
clone script	70.3 ± 6.2 ms	0.959 ± 0.103 ms	73.3×
worktree script	72.9 ± 8.6 ms	1.01 ± 0.09 ms	71.9×
scan 1000 dirs	80.3 ± 7.3 ms	8.57 ± 0.69 ms	9.4×
filter 1000 dirs	93.9 ± 5.5 ms	40.9 ± 4.8 ms	2.3×

init still forks ps to detect fish vs bash (~20 ms on native). The 1000-entry filter is whole-process (scan + four keystrokes + re-renders), not an isolated matcher.

[tobi]

TUI demos

Same key story on MRI and the Spinel native binary: type spin, Left into the word, insert x → sxpin, Ctrl-A/E, Backspace, Ctrl-U, filter tui, open rename, Esc.

MRI (ruby try.rb)

Native / Spinel (dist/try)

Playable asciinema (unclaimed uploads expire in 7 days):

• MRI: https://asciinema.org/a/EVjOgQHXacrCG3KN
• Native: https://asciinema.org/a/qcE8h41nHAF9n23A

.cast files are in demo/ (asciinema play demo/try-mri.cast).

[tobi]

Benchmark (hyperfine, MRI vs native)

Measured 2026-08-13 16:04 EDT on this box. hyperfine 1.19.0, --shell=none -w 3 -r 20 (native is <5 ms; a shell would swamp it). Ruby 3.3.8, Spinel 79b1aba8bdef (PR 3906). Native at measurement was the unstripped ~1.0 MiB binary; current dist/try after -Os+strip is 705 KB.

Speedup is MRI mean / native mean. Native won every case.

| Case | MRI mean ± σ | native mean ± σ | Speedup |
|------|----------------|-------------------|
| --help | 69.4 ± 4.8 ms | 0.956 ± 0.096 ms | 72.6× |
| --version | 70.8 ± 6.3 ms | 0.959 ± 0.125 ms | 73.8× |
| no args (exit 2) | 68.7 ± 4.3 ms | 0.982 ± 0.068 ms | 69.9× |
| init | 96.5 ± 9.4 ms | 21.2 ± 1.3 ms | 4.5× |
| TUI empty + --and-exit | 74.0 ± 8.4 ms | 1.28 ± 0.14 ms | 57.8× |
| TUI 5 dirs | 72.0 ± 6.0 ms | 1.38 ± 0.08 ms | 52.1× |
| TUI filter --and-keys=beta | 72.4 ± 5.9 ms | 2.70 ± 0.35 ms | 26.8× |
| TUI exec beta | 74.6 ± 8.1 ms | 1.49 ± 0.16 ms | 50.2× |
| clone script only | 70.3 ± 6.2 ms | 0.959 ± 0.103 ms | 73.3× |
| worktree script only | 72.9 ± 8.6 ms | 1.01 ± 0.09 ms | 71.9× |
| TUI scan 1000 dirs | 80.3 ± 7.3 ms | 8.57 ± 0.69 ms | 9.4× |
| TUI filter 1000 dirs | 93.9 ± 5.5 ms | 40.9 ± 4.8 ms | 2.3× |

init is the outlier because native still forks ps to detect fish vs bash (~20 ms). The 1000-entry filter is whole-process (scan + four keystrokes + re-renders), not an isolated matcher microbench.

[tobi]

Moved off the Spinel stack. The TUI InputField work now lives on main as #135 (no TryCompat, 399/399 MRI).

This stacked PR mixed Spinel + TUI + native demos; closing so #130 → #131 → #133 stays Spinel-only.