test: consolidate discovery coverage and replace copied oracles (#32211)

Co-authored-by: OrcaWin <293788423+OrcaWin@users.noreply.github.com>

src/shared/command-code-output-status.test.ts

@@ -1,8 +1,6 @@
 import { afterEach, describe, expect, it, vi } from 'vitest'
-import {
-  createCommandCodeOutputStatusDetector,
-  stripTerminalControl
-} from './command-code-output-status'
+import { createCommandCodeOutputStatusDetector } from './command-code-output-status'
+import { stripTerminalControl } from './terminal-control-stripping'
 
 afterEach(() => {
   vi.restoreAllMocks()
@@ -389,44 +387,6 @@ function maxStringContextLength(contexts: unknown[]): number {
 }
 
 describe('terminal control stripping', () => {
-  const esc = String.fromCharCode(0x1b)
-  const bel = String.fromCharCode(0x07)
-  const ansiEscape = new RegExp(
-    `${esc}(?:[@-Z\\\\-_]|\\[[0-?]*[ -/]*[@-~]|\\][^${bel}]*(?:${bel}|${esc}\\\\))`,
-    'g'
-  )
-  const incompleteAnsiEscape = new RegExp(
-    `${esc}(?:\\[[0-?]*[ -/]*|\\][^${bel}${esc}]*|\\S?)?$`,
-    'g'
-  )
-
-  function legacyStripTerminalControl(data: string): string {
-    const withoutAnsi = data.replace(ansiEscape, '').replace(incompleteAnsiEscape, '')
-    let output = ''
-    for (let index = 0; index < withoutAnsi.length; index += 1) {
-      const code = withoutAnsi.charCodeAt(index)
-      if ((code <= 0x1f && code !== 0x0a && code !== 0x0d) || (code >= 0x7f && code <= 0x9f)) {
-        continue
-      }
-      output += withoutAnsi[index]
-    }
-    return output
-  }
-
-  function makeRandom(seed: number): () => number {
-    let state = seed >>> 0
-    return () => {
-      state ^= state << 13
-      state ^= state >>> 17
-      state ^= state << 5
-      return (state >>> 0) / 0x1_0000_0000
-    }
-  }
-
-  function expectLegacyEquivalent(data: string): void {
-    expect(stripTerminalControl(data)).toBe(legacyStripTerminalControl(data))
-  }
-
   function promptFrom(raw: string): string | null {
     let captured: string | null = null
     const detector = createCommandCodeOutputStatusDetector({
@@ -458,77 +418,26 @@ describe('terminal control stripping', () => {
     )
   })
 
-  it('matches the legacy filter across ANSI, C0/C1, line breaks, and Unicode', () => {
-    for (const data of [
-      '',
-      'plain text',
-      '\x1b[35mstyled\x1b[0m',
-      '\x1b[unterminated',
-      '\x00leading',
-      'trailing\x9f',
-      '\x01\x02adjacent\x7f\x80',
-      'keep\r\nline breaks',
-      '日本語 \u{1f389} \ud83d \ude00',
-      '\x1b]0;title\x07prompt',
-      '\x1b]0;title\x1b\\prompt'
-    ]) {
-      expectLegacyEquivalent(data)
-    }
-  })
-
-  it('matches legacy output at density thresholds and block resets', () => {
-    const control = '\x01'
-    const fixtures = [
-      `${control.repeat(31)}${'a'.repeat(33)}`,
-      `${control.repeat(32)}${'a'.repeat(32)}`,
-      `${'a'.repeat(32)}${control.repeat(31)}a`,
-      `${control.repeat(31)}${'a'.repeat(33)}${control.repeat(31)}z`,
-      `${control.repeat(31)}${'a'.repeat(33)}${control.repeat(32)}z`,
-      `${'a'.repeat(64 * 3)}${control.repeat(32)}tail`,
-      `${'a\x01'.repeat(2048)}tail`
-    ]
-    for (const data of fixtures) {
-      expectLegacyEquivalent(data)
-    }
-  })
-
-  it('exhaustively matches short strings over control and Unicode code units', () => {
-    const alphabet = ['a', '\r', '\n', '\x01', '\x7f', '\x80', '\u20ac', '\ud83d']
-    for (let encoded = 0; encoded < alphabet.length ** 4; encoded += 1) {
-      let cursor = encoded
-      let data = ''
-      for (let position = 0; position < 4; position += 1) {
-        data += alphabet[cursor % alphabet.length]
-        cursor = Math.floor(cursor / alphabet.length)
-      }
-      expectLegacyEquivalent(data)
-    }
-  })
-
-  it('matches seeded random terminal text', () => {
-    const random = makeRandom(0xc0de_0727)
-    const alphabet = [
-      'a',
-      'Z',
-      '\r',
-      '\n',
-      '\x00',
-      '\x1f',
-      '\x7f',
-      '\x9f',
-      '\u00e9',
-      '\u20ac',
-      '\u{1f389}',
-      '\x1b[35m',
-      '\x1b[0m'
-    ]
-    for (let trial = 0; trial < 2_000; trial += 1) {
-      let data = ''
-      const parts = Math.floor(random() * 256)
-      for (let index = 0; index < parts; index += 1) {
-        data += alphabet[Math.floor(random() * alphabet.length)]
-      }
-      expectLegacyEquivalent(data)
-    }
+  it.each([
+    { data: '', expected: '' },
+    { data: 'plain text', expected: 'plain text' },
+    { data: '\x1b[35mstyled\x1b[0m', expected: 'styled' },
+    { data: 'prompt\x1b[35', expected: 'prompt' },
+    { data: '\x00leading', expected: 'leading' },
+    { data: 'trailing\x9f', expected: 'trailing' },
+    { data: '\x01\x02adjacent\x7f\x80', expected: 'adjacent' },
+    { data: 'keep\r\nline breaks', expected: 'keep\r\nline breaks' },
+    { data: '日本語 🎉 \ud83d \ude00', expected: '日本語 🎉 \ud83d \ude00' },
+    { data: '\x01'.repeat(31) + 'a'.repeat(33), expected: 'a'.repeat(33) },
+    { data: '\x01'.repeat(32) + 'a'.repeat(32), expected: 'a'.repeat(32) },
+    { data: `${'a'.repeat(32) + '\x01'.repeat(31)}a`, expected: 'a'.repeat(33) },
+    {
+      data: `${'\x01'.repeat(31) + 'a'.repeat(33) + '\x01'.repeat(32)}z`,
+      expected: `${'a'.repeat(33)}z`
+    },
+    { data: `${'a'.repeat(192) + '\x01'.repeat(32)}tail`, expected: `${'a'.repeat(192)}tail` },
+    { data: `${'a\x01'.repeat(2048)}tail`, expected: `${'a'.repeat(2048)}tail` }
+  ])('strips controls while preserving printable text (case %#)', ({ data, expected }) => {
+    expect(stripTerminalControl(data)).toBe(expected)
   })
 })