feat(desktop): use headless xterm for terminal scrollback storage#686
feat(desktop): use headless xterm for terminal scrollback storage#686
Conversation
Replace raw PTY data storage with headless xterm + serialize addon for cleaner scrollback. This properly processes escape sequences instead of storing them verbatim. Changes: - Add @xterm/headless dependency for Node.js terminal emulation - Store terminal state in HeadlessTerminal instead of raw string - Use SerializeAddon to produce clean ANSI output for scrollback - Recreate headless terminal on ED3 (clear) sequences to handle async - Keep headless terminal dimensions in sync on resize - Update tests for new architecture 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
|
Warning Rate limit exceeded@Kitenite has exceeded the limit for the number of commits that can be reviewed per hour. Please wait 4 minutes and 13 seconds before requesting another review. ⌛ How to resolve this issue?After the wait time has elapsed, a review can be triggered using the We recommend that you space out your commits to avoid hitting the rate limit. 🚦 How do rate limits work?CodeRabbit enforces hourly rate limits for each developer per organization. Our paid plans have higher rate limits than the trial, open-source and free plans. In all cases, we re-allow further reviews after a brief timeout. Please see our FAQ for further information. 📒 Files selected for processing (3)
📝 WalkthroughWalkthroughRefactors terminal scrollback to use an xterm headless terminal with the Serialize addon. Sessions now carry Changes
Sequence Diagram(s)sequenceDiagram
participant Manager
participant Session
participant Headless
participant Serializer
Manager->>Session: createSession(cols, rows, existingScrollback?)
Session->>Headless: createHeadlessTerminal(cols, rows, scrollback?)
Headless->>Serializer: attach SerializeAddon
alt existingScrollback present
Manager->>Session: recoverScrollback({ existingScrollback, headless })
Session->>Headless: write(existingScrollback)
Headless->>Serializer: serialize()
Serializer-->>Session: serializedOutput
Session-->>Manager: wasRecovered = true
end
Manager->>Session: getSerializedScrollback(session)
Session->>Serializer: serialize()
Serializer-->>Manager: stringScrollback
Manager->>Session: resize(cols, rows)
Session->>Headless: resize(cols, rows)
Manager->>Session: clearScrollback()
Session->>Headless: clear() / recreate headless
Estimated code review effort🎯 4 (Complex) | ⏱️ ~50 minutes Possibly related PRs
Poem
🚥 Pre-merge checks | ✅ 2 | ❌ 1❌ Failed checks (1 warning)
✅ Passed checks (2 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 2
🤖 Fix all issues with AI agents
In @apps/desktop/package.json:
- Around line 67-68: The two @xterm dependencies use inconsistent version
ranges; update the package.json entries for "@xterm/headless" and "@xterm/xterm"
so they use the same version strategy (e.g., both pinned to "5.5.0" or both
caret-prefixed "^5.5.0"), then run your package install to update the lockfile
and run the app/tests to verify no compatibility breakage.
In @apps/desktop/src/main/lib/terminal/session.ts:
- Around line 186-197: This block duplicates headless terminal setup; replace it
by calling the existing createHeadlessTerminal helper and reuse its returned
terminal/serializer instead of manually new-ing HeadlessTerminal and
SerializeAddon. Specifically, call createHeadlessTerminal with the session's
cols/rows (or the helper's expected args), assign its returned headless instance
to session.headless and the returned serializer to session.serializer (or if the
helper only returns the terminal, ensure it loads and exposes the SerializeAddon
so you still set session.serializer). Remove the manual new HeadlessTerminal /
new SerializeAddon / loadAddon code and use createHeadlessTerminal to keep
configuration consistent.
🧹 Nitpick comments (3)
apps/desktop/src/main/lib/terminal/manager.test.ts (2)
580-582: Consider a more deterministic synchronization approach.The 50ms delay is used to wait for the headless terminal's async write queue. While functional, fixed delays can be flaky on slow systems and wasteful on fast ones.
Explore whether xterm's headless terminal provides a write completion callback or drain event for more reliable synchronization:
// Current approach: await new Promise((resolve) => setTimeout(resolve, 50)); // Potential alternative (if available): await new Promise<void>((resolve) => { session.headless.write("", resolve); });If such an API exists, it would make tests more reliable and potentially faster.
610-612: Duplicated async delay pattern.This is the same 50ms delay pattern from line 580-582. Consider extracting to a test helper to reduce duplication and make future adjustments easier.
🔄 Extract delay helper
// At the top of the test file async function waitForHeadlessWrite() { await new Promise((resolve) => setTimeout(resolve, 50)); } // Then use it: await waitForHeadlessWrite();apps/desktop/src/main/lib/terminal/session.ts (1)
90-93: Consider includingonDatacallback in the params object.Per coding guidelines, functions with 2+ parameters should accept a single params object. The callback could be included as a named property for consistency.
♻️ Suggested signature refactor
-export async function createSession( - params: InternalCreateSessionParams, - onData: (paneId: string, data: string) => void, -): Promise<TerminalSession> { +export async function createSession( + params: InternalCreateSessionParams & { + onData: (paneId: string, data: string) => void; + }, +): Promise<TerminalSession> { + const { onData, ...sessionParams } = params;
📜 Review details
Configuration used: defaults
Review profile: CHILL
Plan: Pro
⛔ Files ignored due to path filters (1)
bun.lockis excluded by!**/*.lock
📒 Files selected for processing (6)
apps/desktop/package.jsonapps/desktop/src/main/lib/terminal/manager.test.tsapps/desktop/src/main/lib/terminal/manager.tsapps/desktop/src/main/lib/terminal/session.test.tsapps/desktop/src/main/lib/terminal/session.tsapps/desktop/src/main/lib/terminal/types.ts
🧰 Additional context used
📓 Path-based instructions (5)
apps/desktop/**/*.{ts,tsx}
📄 CodeRabbit inference engine (apps/desktop/AGENTS.md)
apps/desktop/**/*.{ts,tsx}: For Electron interprocess communication, ALWAYS use tRPC as defined insrc/lib/trpc
Use alias as defined intsconfig.jsonwhen possible
Prefer zustand for state management if it makes sense. Do not use effect unless absolutely necessary.
For tRPC subscriptions with trpc-electron, ALWAYS use the observable pattern from@trpc/server/observableinstead of async generators, as the library explicitly checksisObservable(result)and throws an error otherwise
Files:
apps/desktop/src/main/lib/terminal/manager.test.tsapps/desktop/src/main/lib/terminal/session.test.tsapps/desktop/src/main/lib/terminal/types.tsapps/desktop/src/main/lib/terminal/manager.tsapps/desktop/src/main/lib/terminal/session.ts
**/*.{ts,tsx}
📄 CodeRabbit inference engine (AGENTS.md)
**/*.{ts,tsx}: Use object parameters for functions with 2+ parameters instead of positional arguments
Functions with 2+ parameters should accept a single params object with named properties for self-documentation and extensibility
Use prefixed console logging with context pattern: [domain/operation] message
Extract magic numbers and hardcoded values to named constants at module top
Use lookup objects/maps instead of repeated if (type === ...) conditionals
Avoid usinganytype - maintain type safety in TypeScript code
Never swallow errors silently - at minimum log them with context
Import from concrete files directly when possible - avoid barrel file abuse that creates circular dependencies
Avoid deep nesting (4+ levels) - use early returns, extract functions, and invert conditions
Use named properties in options objects instead of boolean parameters to avoid boolean blindness
Files:
apps/desktop/src/main/lib/terminal/manager.test.tsapps/desktop/src/main/lib/terminal/session.test.tsapps/desktop/src/main/lib/terminal/types.tsapps/desktop/src/main/lib/terminal/manager.tsapps/desktop/src/main/lib/terminal/session.ts
**/*.test.{ts,tsx}
📄 CodeRabbit inference engine (AGENTS.md)
Co-locate tests with implementation files using .test.ts or .test.tsx suffix
Files:
apps/desktop/src/main/lib/terminal/manager.test.tsapps/desktop/src/main/lib/terminal/session.test.ts
apps/**/*.{ts,tsx}
📄 CodeRabbit inference engine (AGENTS.md)
Use Drizzle ORM for all database operations - never use raw SQL
Files:
apps/desktop/src/main/lib/terminal/manager.test.tsapps/desktop/src/main/lib/terminal/session.test.tsapps/desktop/src/main/lib/terminal/types.tsapps/desktop/src/main/lib/terminal/manager.tsapps/desktop/src/main/lib/terminal/session.ts
**/*.{ts,tsx,js,jsx}
📄 CodeRabbit inference engine (AGENTS.md)
Use Biome for formatting and linting - run at root level with
bun run lint:fixorbiome check --write
Files:
apps/desktop/src/main/lib/terminal/manager.test.tsapps/desktop/src/main/lib/terminal/session.test.tsapps/desktop/src/main/lib/terminal/types.tsapps/desktop/src/main/lib/terminal/manager.tsapps/desktop/src/main/lib/terminal/session.ts
🧠 Learnings (1)
📚 Learning: 2026-01-02T06:50:28.671Z
Learnt from: CR
Repo: superset-sh/superset PR: 0
File: AGENTS.md:0-0
Timestamp: 2026-01-02T06:50:28.671Z
Learning: Applies to apps/desktop/src/renderer/**/*.{ts,tsx} : Never import Node.js modules (fs, path, os, net) in renderer process or shared code - they are externalized for browser compatibility
Applied to files:
apps/desktop/src/main/lib/terminal/session.ts
🧬 Code graph analysis (2)
apps/desktop/src/main/lib/terminal/session.test.ts (2)
apps/desktop/src/main/lib/terminal/session.ts (3)
recoverScrollback(59-69)getSerializedScrollback(52-54)flushSession(235-238)apps/desktop/src/main/lib/terminal/types.ts (1)
TerminalSession(6-24)
apps/desktop/src/main/lib/terminal/manager.ts (1)
apps/desktop/src/main/lib/terminal/session.ts (1)
getSerializedScrollback(52-54)
⏰ Context from checks skipped due to timeout of 90000ms. You can increase the timeout in your CodeRabbit configuration to a maximum of 15 minutes (900000ms). (6)
- GitHub Check: Deploy API
- GitHub Check: Deploy Docs
- GitHub Check: Deploy Marketing
- GitHub Check: Deploy Admin
- GitHub Check: Deploy Web
- GitHub Check: Build
🔇 Additional comments (17)
apps/desktop/src/main/lib/terminal/types.ts (2)
1-2: LGTM - Type imports align with new architecture.The imports for
SerializeAddonandHeadlessTerminalare correctly sourced from the respective packages.
14-17: Good architectural separation between internal state and external API.Replacing the raw
scrollback: stringwithheadless: HeadlessTerminalandserializer: SerializeAddonenables proper escape sequence processing. The external API (SessionResult.scrollback: stringon line 41) remains stable as a computed value, which is excellent design.apps/desktop/src/main/lib/terminal/manager.ts (4)
6-11: LGTM - Proper import of serialization helper.The import of
getSerializedScrollbackcentralizes the scrollback serialization logic, promoting consistency across the codebase.
41-41: Excellent consistency in scrollback serialization.All usages of
session.scrollbackhave been systematically replaced withgetSerializedScrollback(session). This ensures:
- Consistent scrollback representation across all code paths
- Proper escape sequence processing via the headless terminal
- Single source of truth for scrollback generation
The changes cover:
- Existing session reattachment (line 41)
- New session creation with existing scrollback (line 49)
- New session return (line 100)
- Fallback shell recovery (line 130)
Also applies to: 49-49, 100-100, 130-130
194-195: Critical: Headless terminal dimensions synchronized on resize.Keeping
session.headless.resize(cols, rows)in sync with the PTY resize ensures the headless terminal accurately reflects the terminal state for proper scrollback serialization. This is essential for correctness.
261-262: LGTM - Clear scrollback now properly handled.The change from string manipulation to
session.headless.clear()aligns with the new architecture and ensures the headless terminal's buffer is properly cleared.apps/desktop/src/main/lib/terminal/session.test.ts (5)
11-26: Good test helper for consistent test setup.The
createTestHeadless()helper promotes DRY principles and ensures consistent configuration across tests. The headless terminal is properly initialized with reasonable defaults (80x24, 1000 scrollback).
30-50: Excellent async handling in tests.The test properly waits for xterm's async write to complete using the callback pattern before asserting on serialized content. This is more reliable than fixed delays.
52-63: Good negative test case coverage.Testing the null/empty scrollback path ensures
recoverScrollbackhandles both scenarios correctly.
88-113: Comprehensive resource cleanup verification.The test verifies both
dataBatcher.dispose()andheadless.dispose()are called, ensuring no resource leaks. This is critical given that headless terminals may hold significant memory.
23-23: This type assertion is a documented workaround for xterm's type definitions. The@xterm/addon-serialize@0.13.0is the correct compatible version for@xterm/headless@5.5.0, and the double cast is the standard approach used across xterm projects to load addons. No changes needed.Likely an incorrect or invalid review comment.
apps/desktop/src/main/lib/terminal/session.ts (6)
1-19: LGTM!Imports are correctly organized. Constants are well-named with appropriate documentation for
DEFAULT_SCROLLBACK.
21-47: LGTM!Well-documented helper with clear justification for the type cast workaround. The object params pattern follows coding guidelines.
49-69: LGTM!Clean helper functions. The
recoverScrollbackcorrectly uses object parameters per coding guidelines and returns a meaningful boolean for the caller to track recovery state.
71-88: LGTM!Standard PTY spawn configuration with proper object parameter usage.
198-206: LGTM!The content-after-clear handling is correct. Writing only the post-clear content ensures the fresh headless terminal starts with the right state.
235-238: LGTM!Proper resource cleanup disposing both the data batcher and headless terminal.
🧹 Preview Cleanup CompleteThe following preview resources have been cleaned up:
Thank you for your contribution! 🎉 |
There was a problem hiding this comment.
Actionable comments posted: 0
Caution
Some comments are outside the diff and can’t be posted inline due to platform limitations.
⚠️ Outside diff range comments (2)
apps/desktop/src/main/lib/terminal/manager.ts (2)
196-208: Handle partial resize failure to prevent state inconsistency.If
session.headless.resize()throws butsession.pty.resize()succeeds, the catch block will log the error but leave the terminals out of sync. Additionally,session.colsandsession.rowswon't be updated, creating a mismatch between the tracked dimensions and the actual PTY size.Consider restructuring to handle this scenario:
🔧 Proposed fix to handle partial failure
try { session.pty.resize(cols, rows); + session.cols = cols; + session.rows = rows; + // Keep headless terminal in sync for accurate scrollback serialization - session.headless.resize(cols, rows); - session.cols = cols; - session.rows = rows; + try { + session.headless.resize(cols, rows); + } catch (headlessError) { + console.error( + `[TerminalManager] Failed to resize headless terminal ${paneId} (cols=${cols}, rows=${rows}):`, + headlessError, + ); + } session.lastActive = Date.now(); } catch (error) { console.error( `[TerminalManager] Failed to resize terminal ${paneId} (cols=${cols}, rows=${rows}):`, error, ); }This ensures:
- PTY dimensions are tracked even if headless resize fails
- Headless resize errors are logged separately for better diagnostics
- Main PTY resize errors still propagate to the outer catch block
253-267: Add error handling for headless terminal clear operation.The
session.headless.clear()call lacks error handling, unlike the similarsession.headless.resize()operation in theresize()method. If the clear operation throws, it could break the user-facing "clear scrollback" action (Cmd+K) without any diagnostic logging.🛡️ Proposed fix to add error handling
// Clear the headless terminal's scrollback buffer - session.headless.clear(); + try { + session.headless.clear(); + } catch (error) { + console.error( + `[TerminalManager] Failed to clear scrollback for terminal ${paneId}:`, + error, + ); + } session.lastActive = Date.now();
📜 Review details
Configuration used: defaults
Review profile: CHILL
Plan: Pro
📒 Files selected for processing (1)
apps/desktop/src/main/lib/terminal/manager.ts
🧰 Additional context used
📓 Path-based instructions (4)
apps/desktop/**/*.{ts,tsx}
📄 CodeRabbit inference engine (apps/desktop/AGENTS.md)
apps/desktop/**/*.{ts,tsx}: For Electron interprocess communication, ALWAYS use tRPC as defined insrc/lib/trpc
Use alias as defined intsconfig.jsonwhen possible
Prefer zustand for state management if it makes sense. Do not use effect unless absolutely necessary.
For tRPC subscriptions with trpc-electron, ALWAYS use the observable pattern from@trpc/server/observableinstead of async generators, as the library explicitly checksisObservable(result)and throws an error otherwise
Files:
apps/desktop/src/main/lib/terminal/manager.ts
**/*.{ts,tsx}
📄 CodeRabbit inference engine (AGENTS.md)
**/*.{ts,tsx}: Use object parameters for functions with 2+ parameters instead of positional arguments
Functions with 2+ parameters should accept a single params object with named properties for self-documentation and extensibility
Use prefixed console logging with context pattern: [domain/operation] message
Extract magic numbers and hardcoded values to named constants at module top
Use lookup objects/maps instead of repeated if (type === ...) conditionals
Avoid usinganytype - maintain type safety in TypeScript code
Never swallow errors silently - at minimum log them with context
Import from concrete files directly when possible - avoid barrel file abuse that creates circular dependencies
Avoid deep nesting (4+ levels) - use early returns, extract functions, and invert conditions
Use named properties in options objects instead of boolean parameters to avoid boolean blindness
Files:
apps/desktop/src/main/lib/terminal/manager.ts
apps/**/*.{ts,tsx}
📄 CodeRabbit inference engine (AGENTS.md)
Use Drizzle ORM for all database operations - never use raw SQL
Files:
apps/desktop/src/main/lib/terminal/manager.ts
**/*.{ts,tsx,js,jsx}
📄 CodeRabbit inference engine (AGENTS.md)
Use Biome for formatting and linting - run at root level with
bun run lint:fixorbiome check --write
Files:
apps/desktop/src/main/lib/terminal/manager.ts
🧬 Code graph analysis (1)
apps/desktop/src/main/lib/terminal/manager.ts (1)
apps/desktop/src/main/lib/terminal/session.ts (1)
getSerializedScrollback(52-54)
⏰ Context from checks skipped due to timeout of 90000ms. You can increase the timeout in your CodeRabbit configuration to a maximum of 15 minutes (900000ms). (6)
- GitHub Check: Deploy Marketing
- GitHub Check: Deploy Web
- GitHub Check: Deploy Admin
- GitHub Check: Deploy Docs
- GitHub Check: Deploy API
- GitHub Check: Build
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
There was a problem hiding this comment.
Actionable comments posted: 0
🧹 Nitpick comments (1)
apps/desktop/src/main/lib/terminal/session.ts (1)
56-69: Potential async timing issue with scrollback recovery.
headless.write()queues data asynchronously. IfgetSerializedScrollback()is called immediately afterrecoverScrollback(), the write may not have completed, resulting in incomplete scrollback serialization.Consider whether a small delay or synchronization is needed when recovering scrollback, similar to what the tests do. The PR summary mentions tests use "small asynchronous delays to accommodate headless processing."
📜 Review details
Configuration used: defaults
Review profile: CHILL
Plan: Pro
📒 Files selected for processing (2)
apps/desktop/src/main/lib/terminal/manager.tsapps/desktop/src/main/lib/terminal/session.ts
🧰 Additional context used
📓 Path-based instructions (4)
apps/desktop/**/*.{ts,tsx}
📄 CodeRabbit inference engine (apps/desktop/AGENTS.md)
apps/desktop/**/*.{ts,tsx}: For Electron interprocess communication, ALWAYS use tRPC as defined insrc/lib/trpc
Use alias as defined intsconfig.jsonwhen possible
Prefer zustand for state management if it makes sense. Do not use effect unless absolutely necessary.
For tRPC subscriptions with trpc-electron, ALWAYS use the observable pattern from@trpc/server/observableinstead of async generators, as the library explicitly checksisObservable(result)and throws an error otherwise
Files:
apps/desktop/src/main/lib/terminal/manager.tsapps/desktop/src/main/lib/terminal/session.ts
**/*.{ts,tsx}
📄 CodeRabbit inference engine (AGENTS.md)
**/*.{ts,tsx}: Use object parameters for functions with 2+ parameters instead of positional arguments
Functions with 2+ parameters should accept a single params object with named properties for self-documentation and extensibility
Use prefixed console logging with context pattern: [domain/operation] message
Extract magic numbers and hardcoded values to named constants at module top
Use lookup objects/maps instead of repeated if (type === ...) conditionals
Avoid usinganytype - maintain type safety in TypeScript code
Never swallow errors silently - at minimum log them with context
Import from concrete files directly when possible - avoid barrel file abuse that creates circular dependencies
Avoid deep nesting (4+ levels) - use early returns, extract functions, and invert conditions
Use named properties in options objects instead of boolean parameters to avoid boolean blindness
Files:
apps/desktop/src/main/lib/terminal/manager.tsapps/desktop/src/main/lib/terminal/session.ts
apps/**/*.{ts,tsx}
📄 CodeRabbit inference engine (AGENTS.md)
Use Drizzle ORM for all database operations - never use raw SQL
Files:
apps/desktop/src/main/lib/terminal/manager.tsapps/desktop/src/main/lib/terminal/session.ts
**/*.{ts,tsx,js,jsx}
📄 CodeRabbit inference engine (AGENTS.md)
Use Biome for formatting and linting - run at root level with
bun run lint:fixorbiome check --write
Files:
apps/desktop/src/main/lib/terminal/manager.tsapps/desktop/src/main/lib/terminal/session.ts
🧠 Learnings (1)
📚 Learning: 2026-01-02T06:50:28.671Z
Learnt from: CR
Repo: superset-sh/superset PR: 0
File: AGENTS.md:0-0
Timestamp: 2026-01-02T06:50:28.671Z
Learning: Applies to apps/desktop/src/renderer/**/*.{ts,tsx} : Never import Node.js modules (fs, path, os, net) in renderer process or shared code - they are externalized for browser compatibility
Applied to files:
apps/desktop/src/main/lib/terminal/session.ts
🧬 Code graph analysis (1)
apps/desktop/src/main/lib/terminal/session.ts (2)
apps/desktop/src/main/lib/terminal/types.ts (1)
TerminalSession(6-24)apps/desktop/src/main/lib/terminal-escape-filter.ts (2)
containsClearScrollbackSequence(29-31)extractContentAfterClear(38-46)
⏰ Context from checks skipped due to timeout of 90000ms. You can increase the timeout in your CodeRabbit configuration to a maximum of 15 minutes (900000ms). (7)
- GitHub Check: Deploy Marketing
- GitHub Check: Deploy Docs
- GitHub Check: Deploy Admin
- GitHub Check: Deploy Web
- GitHub Check: Deploy API
- GitHub Check: Build
- GitHub Check: Typecheck
🔇 Additional comments (10)
apps/desktop/src/main/lib/terminal/manager.ts (4)
6-12: LGTM!The imports are well-organized and correctly bring in the new headless terminal utilities from the session module.
114-117: Good defensive ordering.Capturing scrollback before
flushSessionis correct sinceflushSessiondisposes the headless terminal. The comment clearly explains the intent.
196-209: LGTM!Keeping the headless terminal dimensions synchronized ensures accurate scrollback serialization. The error handling properly wraps both resize operations.
254-275: LGTM!Recreating the headless terminal is the correct approach given xterm's asynchronous write queue. This ensures a clean slate for scrollback.
apps/desktop/src/main/lib/terminal/session.ts (6)
49-54: LGTM!Clean abstraction that encapsulates the serialization logic.
123-163: LGTM!Session creation correctly initializes the headless terminal before spawning the PTY, and properly stores both
headlessandserializerin the session object.
178-200: LGTM!The ED3 clear sequence handling is well-implemented:
- Properly disposes the old terminal before creating a new one
- Extracts and preserves any content after the clear sequence
- The conditional write on line 194 is correct (empty string would be a no-op)
229-232: LGTM!Proper cleanup of both the data batcher and headless terminal resources.
14-19: Good constant extraction.The
DEFAULT_SCROLLBACKconstant follows the coding guideline to extract magic numbers. 10,000 lines is a reasonable default for terminal scrollback.
21-47: LGTM — type cast is justified, andallowProposedApi: trueis required.The type assertion on lines 40-44 is correct.
SerializeAddonis an experimental/proposed API in xterm.js that requiresallowProposedApi: trueto function; it cannot be instantiated otherwise. The comment correctly explains thatSerializeAddononly accesses the buffer API, which both browser and headless terminals share at runtime. The code also properly disposes the headless terminal influshSession()and handles terminal recreation when clear scrollback sequences are detected.
* feat(desktop): use headless xterm for terminal scrollback storage Replace raw PTY data storage with headless xterm + serialize addon for cleaner scrollback. This properly processes escape sequences instead of storing them verbatim. Changes: - Add @xterm/headless dependency for Node.js terminal emulation - Store terminal state in HeadlessTerminal instead of raw string - Use SerializeAddon to produce clean ANSI output for scrollback - Recreate headless terminal on ED3 (clear) sequences to handle async - Keep headless terminal dimensions in sync on resize - Update tests for new architecture
1 participant
Summary
@xterm/headless(official Node.js version of xterm.js) with@xterm/addon-serializeto produce clean ANSI outputChanges
@xterm/headless@5.5.0dependency (matches@xterm/xterm@5.5.0)scrollback: stringwithheadless: HeadlessTerminal+serializer: SerializeAddonin session typeserializer.serialize()to get clean scrollback outputHow it works
The serialize addon outputs minimal ANSI sequences that reconstruct the terminal state, rather than storing raw escape-sequence-laden PTY data.
Test plan
🤖 Generated with Claude Code
Summary by CodeRabbit
Tests
Chores
Refactor
✏️ Tip: You can customize this high-level summary in your review settings.