From 5a5a411e1857cd3f87017f4ca83e046fb39bbc5e Mon Sep 17 00:00:00 2001
From: Neil <4138956+nwparker@users.noreply.github.com>
Date: Mon, 10 Aug 2026 00:29:26 -0700
Subject: [PATCH] perf(main): skip impossible advertised URL scans (#13514)

(cherry picked from commit c5023fa0ab07a9a9800f0885b6d6d7b23aec40bf)
---
 .../advertised-url-watcher-benchmark.mjs      | 255 ++++++++++++++++++
 src/main/ports/advertised-url-watcher.test.ts |  36 +++
 src/main/ports/advertised-url-watcher.ts      |  23 +-
 3 files changed, 310 insertions(+), 4 deletions(-)
 create mode 100644 config/scripts/advertised-url-watcher-benchmark.mjs

diff --git a/config/scripts/advertised-url-watcher-benchmark.mjs b/config/scripts/advertised-url-watcher-benchmark.mjs
new file mode 100644
index 00000000000..cae7195d024
--- /dev/null
+++ b/config/scripts/advertised-url-watcher-benchmark.mjs
@@ -0,0 +1,255 @@
+#!/usr/bin/env node
+// Measures the no-URL PTY path that every terminal chunk enters.
+import { spawnSync } from 'node:child_process'
+import { existsSync, readFileSync } from 'node:fs'
+import nodeModule from 'node:module'
+import { performance } from 'node:perf_hooks'
+import { fileURLToPath } from 'node:url'
+
+if (!process.execArgv.includes('--experimental-transform-types')) {
+  const result = spawnSync(
+    process.execPath,
+    ['--experimental-transform-types', '--no-warnings', import.meta.filename],
+    { stdio: 'inherit' }
+  )
+  process.exit(result.status ?? 1)
+}
+
+nodeModule.registerHooks({
+  resolve(specifier, context, nextResolve) {
+    if (specifier.startsWith('.') && !/\.[cm]?[jt]s$/.test(specifier) && context.parentURL) {
+      const candidate = new URL(`${specifier}.ts`, context.parentURL)
+      if (existsSync(fileURLToPath(candidate))) {
+        return { url: candidate.href, shortCircuit: true }
+      }
+    }
+    return nextResolve(specifier, context)
+  }
+})
+
+const source = readFileSync(
+  new URL('../../src/main/ports/advertised-url-watcher.ts', import.meta.url),
+  'utf8'
+)
+for (const marker of [
+  "return mayContainHttpUrl(finalized) ? stripTerminalControls(finalized) : ''",
+  'if (chunk.length >= PER_PTY_BUFFER_LIMIT)',
+  'this.raw.length + chunk.length > PER_PTY_BUFFER_LIMIT'
+]) {
+  if (!source.includes(marker)) {
+    throw new Error(`advertised-url-watcher.ts no longer contains \`${marker}\``)
+  }
+}
+
+const { AdvertisedUrlWatcher, extractUrlCandidates, stripTerminalControls } = await import(
+  new URL('../../src/main/ports/advertised-url-watcher.ts', import.meta.url).href
+)
+
+const BUFFER_LIMIT = 4096
+const ITERATIONS = Number(process.env.ORCA_ADVERTISED_URL_BENCH_ITERATIONS ?? '10000')
+const ROUNDS = Number(process.env.ORCA_ADVERTISED_URL_BENCH_ROUNDS ?? '12')
+const WARMUP = Number(process.env.ORCA_ADVERTISED_URL_BENCH_WARMUP ?? '1000')
+
+for (const [name, value] of [
+  ['ORCA_ADVERTISED_URL_BENCH_ITERATIONS', ITERATIONS],
+  ['ORCA_ADVERTISED_URL_BENCH_ROUNDS', ROUNDS],
+  ['ORCA_ADVERTISED_URL_BENCH_WARMUP', WARMUP]
+]) {
+  if (!Number.isSafeInteger(value) || value <= 0) {
+    throw new Error(`${name} must be a positive integer, received ${value}`)
+  }
+}
+if (ROUNDS % 2 !== 0) {
+  throw new Error('ORCA_ADVERTISED_URL_BENCH_ROUNDS must be even')
+}
+
+class BeforePtyBuffer {
+  raw = ''
+
+  ingest(chunk) {
+    const chunkHasLineBreak = chunk.includes('\n') || chunk.includes('\r')
+    this.raw += chunk
+    if (this.raw.length > BUFFER_LIMIT) {
+      this.raw = this.raw.slice(-BUFFER_LIMIT)
+    }
+    if (!chunkHasLineBreak) {
+      return ''
+    }
+    const lastNewline = lastLineBreak(this.raw)
+    if (lastNewline === -1) {
+      return ''
+    }
+    const finalized = this.raw.slice(0, lastNewline + 1)
+    this.raw = this.raw.slice(lastNewline + 1)
+    return stripTerminalControls(finalized)
+  }
+}
+
+class BeforeWatcher {
+  buffers = new Map()
+  ptyToWorktree = new Map()
+
+  bindPty(ptyId, worktreeId) {
+    this.ptyToWorktree.set(ptyId, worktreeId)
+  }
+
+  ingest(ptyId, chunk) {
+    if (!this.ptyToWorktree.has(ptyId)) {
+      return 0
+    }
+    let buffer = this.buffers.get(ptyId)
+    if (!buffer) {
+      buffer = new BeforePtyBuffer()
+      this.buffers.set(ptyId, buffer)
+    }
+    const finalized = buffer.ingest(chunk)
+    return finalized ? extractUrlCandidates(finalized).length : 0
+  }
+}
+
+function lastLineBreak(text) {
+  for (let index = text.length - 1; index >= 0; index -= 1) {
+    const code = text.charCodeAt(index)
+    if (code === 0x0a || code === 0x0d) {
+      return index
+    }
+  }
+  return -1
+}
+
+function repeatToLine(text, length, sample) {
+  const unit = `${text}${sample}\n`
+  return `${unit.repeat(Math.ceil(length / unit.length)).slice(0, length - 1)}\n`
+}
+
+function repeatWithoutLineBreak(text, length, sample) {
+  const unit = `${text}${sample}`
+  return unit.repeat(Math.ceil(length / unit.length)).slice(0, length)
+}
+
+const fixtures = [
+  {
+    label: 'plain 32B line',
+    chunks: Array.from({ length: 32 }, (_, index) => `compiled module ${index} in 183ms\n`)
+  },
+  {
+    label: 'ANSI 32B line',
+    chunks: Array.from({ length: 32 }, (_, index) => `\x1b[2K\x1b[1GThinking ${index}...\r\n`)
+  },
+  {
+    label: 'guard fallthrough line',
+    chunks: Array.from(
+      { length: 32 },
+      (_, index) => `path/to/http_server: compiled module ${index}\n`
+    )
+  },
+  {
+    label: 'plain 4KiB line',
+    chunks: Array.from({ length: 32 }, (_, index) =>
+      repeatToLine('INFO build completed module=renderer duration=12ms ', BUFFER_LIMIT, index)
+    )
+  },
+  {
+    label: 'ANSI 4KiB line',
+    chunks: Array.from({ length: 32 }, (_, index) =>
+      repeatToLine(
+        '\x1b[2K\x1b[1Gthinking about compiler output and writing patches ',
+        BUFFER_LIMIT,
+        index
+      )
+    )
+  },
+  {
+    label: 'guard fallthrough 4KiB',
+    chunks: Array.from({ length: 32 }, (_, index) =>
+      repeatToLine(
+        'path/to/http_server: compiled without an advertised endpoint ',
+        BUFFER_LIMIT,
+        index
+      )
+    )
+  },
+  {
+    label: '64KiB partial chunk',
+    chunks: Array.from({ length: 32 }, (_, index) =>
+      repeatWithoutLineBreak('compiler output without a line break ', 64 * 1024, index)
+    )
+  },
+  {
+    label: '256KiB partial chunk',
+    chunks: Array.from({ length: 32 }, (_, index) =>
+      repeatWithoutLineBreak('compiler output without a line break ', 256 * 1024, index)
+    )
+  }
+]
+
+let checksum = 0
+
+function runBefore(chunks, iterations) {
+  const watcher = new BeforeWatcher()
+  watcher.bindPty('pty', 'repo::/bench')
+  let found = 0
+  for (let index = 0; index < iterations; index += 1) {
+    found += watcher.ingest('pty', chunks[index % chunks.length])
+  }
+  checksum = Math.imul(checksum ^ found ^ watcher.buffers.get('pty').raw.length, 16777619) >>> 0
+}
+
+function runAfter(chunks, iterations) {
+  const watcher = new AdvertisedUrlWatcher()
+  watcher.bindPty('pty', 'repo::/bench')
+  for (let index = 0; index < iterations; index += 1) {
+    watcher.ingest('pty', chunks[index % chunks.length], index)
+  }
+  const buffer = watcher.buffers.get('pty')
+  checksum = Math.imul(checksum ^ (buffer?.raw.length ?? 0), 16777619) >>> 0
+}
+
+function median(values) {
+  const sorted = [...values].sort((left, right) => left - right)
+  const middle = Math.floor(sorted.length / 2)
+  return sorted.length % 2 === 0 ? (sorted[middle - 1] + sorted[middle]) / 2 : sorted[middle]
+}
+
+function measure(fixture) {
+  runBefore(fixture.chunks, WARMUP)
+  runAfter(fixture.chunks, WARMUP)
+  const before = []
+  const after = []
+  for (let round = 0; round < ROUNDS; round += 1) {
+    const measureBefore = () => {
+      const startedAt = performance.now()
+      runBefore(fixture.chunks, ITERATIONS)
+      before.push(((performance.now() - startedAt) * 1000) / ITERATIONS)
+    }
+    const measureAfter = () => {
+      const startedAt = performance.now()
+      runAfter(fixture.chunks, ITERATIONS)
+      after.push(((performance.now() - startedAt) * 1000) / ITERATIONS)
+    }
+    if (round % 2 === 0) {
+      measureBefore()
+      measureAfter()
+    } else {
+      measureAfter()
+      measureBefore()
+    }
+  }
+  return { before: median(before), after: median(after) }
+}
+
+console.log('Advertised URL watcher no-match cost in microseconds/chunk. Lower is better.')
+console.log(`iterations=${ITERATIONS}, rounds=${ROUNDS}, warmup=${WARMUP}`)
+console.log('fixture                     before      after    speedup')
+for (const fixture of fixtures) {
+  const result = measure(fixture)
+  console.log(
+    `${fixture.label.padEnd(26)} ${result.before.toFixed(3).padStart(8)} ${result.after
+      .toFixed(3)
+      .padStart(10)} ${(result.before / result.after).toFixed(2).padStart(9)}x`
+  )
+}
+console.log(`checksum=${checksum}`)
+console.log(
+  'Limitations: synthetic in-process scan; explicit timestamps omit the production clock read and surrounding PTY dispatch.'
+)
diff --git a/src/main/ports/advertised-url-watcher.test.ts b/src/main/ports/advertised-url-watcher.test.ts
index d46b8268fab..17fcc5803bb 100644
--- a/src/main/ports/advertised-url-watcher.test.ts
+++ b/src/main/ports/advertised-url-watcher.test.ts
@@ -147,6 +147,13 @@ describe('AdvertisedUrlWatcher.ingest', () => {
     expect(watcher.lookup(WORKTREE, 3001)?.origin).toBe('https://local.getmontecarlo.com:3001')
   })
 
+  it('preserves a split URL inside the bounded suffix', () => {
+    const watcher = bindFresh()
+    watcher.ingest(PTY, `${'x'.repeat(4080)} https://local.`)
+    watcher.ingest(PTY, 'example.com:3003/\n')
+    expect(watcher.lookup(WORKTREE, 3003)?.origin).toBe('https://local.example.com:3003')
+  })
+
   it('does not scan buffered PTY text until a line break arrives', () => {
     const watcher = bindFresh()
     const charCodeAtSpy = vi.spyOn(String.prototype, 'charCodeAt')
@@ -168,6 +175,35 @@ describe('AdvertisedUrlWatcher.ingest', () => {
     expect(watcher.lookup(WORKTREE, 3001)?.origin).toBe('https://example.com:3001')
   })
 
+  it('keeps URLs whose scheme is split by terminal controls', () => {
+    const watcher = bindFresh()
+    watcher.ingest(PTY, 'H\x1b[32mtT\x1b[0mP://example.com:3001/\n')
+    expect(watcher.lookup(WORKTREE, 3001)?.origin).toBe('http://example.com:3001')
+  })
+
+  it('rejects finalized lines missing any character required by an HTTP URL', () => {
+    const watcher = bindFresh()
+    for (const line of [
+      'ttp://example.com:3001/\n',
+      'hhp://example.com:3001/\n',
+      'htt://example.com:3001/\n',
+      'http//example.com/3001\n',
+      'http:example.com:3001\n'
+    ]) {
+      watcher.ingest(PTY, line)
+    }
+    expect(watcher.lookup(WORKTREE, 3001)).toBeUndefined()
+  })
+
+  it('retains exactly the latest bounded tail across oversized chunks', () => {
+    const watcher = bindFresh()
+    watcher.ingest(PTY, `https://stale.example.com:3001/${'x'.repeat(4096)}`)
+    watcher.ingest(PTY, ' https://fresh.example.com:3002/\n')
+
+    expect(watcher.lookup(WORKTREE, 3001)).toBeUndefined()
+    expect(watcher.lookup(WORKTREE, 3002)?.host).toBe('fresh.example.com')
+  })
+
   it('sanitizes to origin (drops path, query, fragment, userinfo)', () => {
     const watcher = bindFresh()
     watcher.ingest(PTY, 'go to https://user:pass@example.com:3001/callback?token=secret#x\n')
diff --git a/src/main/ports/advertised-url-watcher.ts b/src/main/ports/advertised-url-watcher.ts
index befc11ab94f..3f8ca60bcce 100644
--- a/src/main/ports/advertised-url-watcher.ts
+++ b/src/main/ports/advertised-url-watcher.ts
@@ -66,9 +66,13 @@ class PtyBuffer {
   /** Append a chunk; return cleaned text up to the last newline. The tail stays buffered so a URL or ANSI sequence split across chunks survives. */
   ingest(chunk: string): string {
     const chunkHasLineBreak = chunk.includes('\n') || chunk.includes('\r')
-    this.raw += chunk
-    if (this.raw.length > PER_PTY_BUFFER_LIMIT) {
-      this.raw = this.raw.slice(-PER_PTY_BUFFER_LIMIT)
+    // Keep the suffix directly so oversized chunks never materialize a throwaway full concatenation.
+    if (chunk.length >= PER_PTY_BUFFER_LIMIT) {
+      this.raw = chunk.slice(-PER_PTY_BUFFER_LIMIT)
+    } else if (this.raw.length + chunk.length > PER_PTY_BUFFER_LIMIT) {
+      this.raw = `${this.raw.slice(-(PER_PTY_BUFFER_LIMIT - chunk.length))}${chunk}`
+    } else {
+      this.raw += chunk
     }
     if (!chunkHasLineBreak) {
       return ''
@@ -79,10 +83,21 @@ class PtyBuffer {
     }
     const finalized = this.raw.slice(0, lastNewline + 1)
     this.raw = this.raw.slice(lastNewline + 1)
-    return stripTerminalControls(finalized)
+    return mayContainHttpUrl(finalized) ? stripTerminalControls(finalized) : ''
   }
 }
 
+function mayContainHttpUrl(text: string): boolean {
+  // Control stripping cannot create any character required by an HTTP scheme.
+  return (
+    (text.includes('h') || text.includes('H')) &&
+    (text.includes('t') || text.includes('T')) &&
+    (text.includes('p') || text.includes('P')) &&
+    text.includes(':') &&
+    text.includes('/')
+  )
+}
+
 function lastLineBreak(text: string): number {
   // Accept either \n or \r as a finalize point (\r\n is normalized later in stripTerminalControls).
   for (let i = text.length - 1; i >= 0; i--) {