fix: resolve 15-second SSH delay with fish shell prompt detection

[jasperan]

Summary

Fixes #1597. The previous fix (#1615) addressed a different issue (focus-stealing, #1467) and didn't touch prompt detection.

• Root cause: waitForShellPrompt tested each PTY data chunk independently against the prompt regex. Fish shell renders its prompt in multiple small writes over SSH (colored segments, escape sequences, prompt character). When > arrived in its own TCP segment without preceding non-whitespace context, the regex failed and fell back to the 15-second timeout.
• Secondary cause: Fish's cursor save/restore escapes (\e7/\e8) and OSC-ST terminators weren't stripped, leaving non-whitespace bytes after > that broke the \s*$ anchor.

Fix (2 files, 3 changes):

• Accumulate PTY chunks into a rolling buffer (4KB cap) so prompts split across TCP segments match against the full accumulated output
• Pass additional strip options (stripOtherEscapes, stripOscSt, stripCarriageReturn) to handle fish's escape sequences
• Expand stripOtherEscapes regex to cover Fe/Fp/Fs single-byte escapes (was only matching [A-Za-z], now matches all except CSI/OSC openers)

Test plan

• 7 new test cases: fish prompt, TCP-split prompt, bare > rejection, Fe escapes, OSC-ST
• All 27 waitForShellPrompt tests pass
• All 9 stripAnsi tests pass (no regressions)
• Full suite: 78 files, 782 tests pass
• Type-check clean, format clean, lint clean on changed files
• Manual verification: connect to SSH remote with fish shell, confirm no 15s delay

Summary by CodeRabbit

• Bug Fixes

  • Improved shell prompt detection for fish-style prompts (e.g., "user@host ~>"), including prompts split across multiple terminal/chunked inputs.
  • Expanded removal of noisy terminal escape bytes and OSC terminators to reduce false positives while preserving valid control sequences.

• Tests

  • Added tests covering split-chunk prompts, escape-sequence noise resilience, and a negative case ensuring a bare "> " alone does not trigger detection.

[vercel]

@jasperan is attempting to deploy a commit to the General Action Team on Vercel.

A member of the Team first needs to authorize it.

[greptile-apps]

Your free trial has ended. If you'd like to continue receiving code reviews, you can add a payment method here.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro

Run ID: 621054e8-df78-4fec-98c4-a34fcf49454e

📥 Commits

Reviewing files that changed from the base of the PR and between 10d88f2 and ad446aa.

📒 Files selected for processing (3)

• src/main/utils/__tests__/waitForShellPrompt.test.ts
• src/main/utils/waitForShellPrompt.ts
• src/shared/text/stripAnsi.ts

✅ Files skipped from review due to trivial changes (2)

• src/main/utils/waitForShellPrompt.ts
• src/main/utils/tests/waitForShellPrompt.test.ts

🚧 Files skipped from review as they are similar to previous changes (1)

• src/shared/text/stripAnsi.ts

---

📝 Walkthrough

Walkthrough

Appends incoming PTY chunks to a capped rolling buffer, expands escape-stripping to remove OSC/other single-byte escapes and CRs, and runs prompt regexes against the cleaned buffer; adds tests covering fish-style prompts (including prompts split across chunks and noisy escape sequences).

Changes

Cohort / File(s)	Summary
Test Coverage src/main/utils/__tests__/waitForShellPrompt.test.ts	Added tests validating fish-style prompt detection (user@host ~>), prompts split across PTY chunks, delayed > arrivals, handling of cursor save/restore and OSC-ST noise, and rejecting bare > without prior context.
Prompt Detection Logic src/main/utils/waitForShellPrompt.ts	Now sanitizes each incoming chunk with expanded stripAnsi options (strip other escapes, OSC, CRs), appends to a rolling prompt buffer (trimmed to max), and continues matching the combined cleaned buffer against prompt regexes. No exported signatures changed.
Escape Sequence Stripping src/shared/text/stripAnsi.ts	When stripOtherEscapes is enabled, broadened final escape removal from /\x1b[A-Za-z]/g to /\x1b[^[\]]/g, removing ESC + any single byte except [ or ] to cover more single-byte ESC sequences while preserving CSI/OSC openers.

Sequence Diagram(s)

sequenceDiagram
  participant RemotePTY as Remote PTY
  participant Buffer as Rolling Buffer
  participant Stripper as stripAnsi
  participant Matcher as Prompt Regex
  participant Caller as waitForShellPrompt caller

  RemotePTY->>Buffer: emit chunk (may be partial)
  Buffer->>Buffer: append & trim (max buffer size)
  Buffer->>Stripper: sanitize(buffer, options)
  Stripper->>Matcher: cleaned text
  alt regex matches prompt
    Matcher->>Caller: finish() — prompt detected
  else no match yet
    Matcher-->>RemotePTY: await further chunks
  end

Loading

Estimated code review effort

🎯 3 (Moderate) | ⏱️ ~25 minutes

Possibly related PRs

• fix(ssh): detect fish prompts without timeout fallback #1622: Similar changes to waitForShellPrompt and tests to handle fish-style prompts split across PTY chunks and sanitize escape sequences.

Poem

🐰 I hop through bytes and clear the haze,
Stitching chunks in patient ways,
Fishy prompts no longer hide,
I scrub the noise and find the guide. 🥕

🚥 Pre-merge checks | ✅ 4 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Docstring Coverage	⚠️ Warning	Docstring coverage is 50.00% which is insufficient. The required threshold is 80.00%.	Write docstrings for the functions missing them to satisfy the coverage threshold.

✅ Passed checks (4 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title accurately summarizes the main change: fixing a 15-second SSH delay caused by fish shell prompt detection issues.
Linked Issues check	✅ Passed	The pull request directly addresses all coding objectives from issue #1597: buffering PTY chunks, stripping fish-specific escapes, expanding stripAnsi regex, and adding comprehensive test coverage.
Out of Scope Changes check	✅ Passed	All changes are scoped to the prompt detection fix: modifications to waitForShellPrompt, stripAnsi, and corresponding tests; no unrelated changes detected.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests

---

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[coderabbitai]

Actionable comments posted: 1

🤖 Prompt for all review comments with AI agents

Verify each finding against the current code and only fix it if needed.

Inline comments:
In `@src/main/utils/__tests__/waitForShellPrompt.test.ts`:
- Around line 350-363: Test title says "detects prompt when > arrives alone
after accumulated context" but the test emits '$ ' instead of '> '; update the
test so the description matches the emitted token: either rename the it(...)
string to reference '$' (e.g. mention "$ arrives alone") or change the emitted
prompt in the body from pty.emit('$ ') to pty.emit('> '); ensure references to
waitForShellPrompt, createMockPty, and pty.emit are updated accordingly so the
assertion remains valid.

🪄 Autofix (Beta)

Fix all unresolved CodeRabbit comments on this PR:

• Push a commit to this branch (recommended)
• Create a new PR with the fixes

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro

Run ID: 46b76c10-a1ac-4bfe-9612-8d7cd30c6d1c

📥 Commits

Reviewing files that changed from the base of the PR and between e78b862 and 12549b9.

📒 Files selected for processing (3)

• src/main/utils/__tests__/waitForShellPrompt.test.ts
• src/main/utils/waitForShellPrompt.ts
• src/shared/text/stripAnsi.ts

[arnestrickmann]

@jasperan thank you for taking this on again!
Would be amazing if you could resolve the merge conflicts in waitForShellPrompt.ts - happy to merge then!

[jasperan]

done @arnestrickmann

[arnestrickmann]

Thanks!