Fix: Cursor Positioning After Line Wrap

[MazenSamehR]

Summary

Implemented a character-based ANSI-safe wrapper directly in the breakLines function to force deterministic wrapping across all terminals:

• Implemented internal wrapper: The breakLines function now uses character-based wrapping instead of word-based wrapping

Fixes: #1818

Testing

All existing tests pass. Run:

yarn pretest && yarn test

demo

https://vimeo.com/1147364471?fl=pl&fe=sh

[codecov]

Codecov Report

❌ Patch coverage is 89.28571% with 3 lines in your changes missing coverage. Please review.
⚠️ Please upload report for BASE (main@67c31c1). Learn more about missing BASE report.

Files with missing lines	Patch %	Lines
packages/core/src/lib/utils.ts	89.28%	3 Missing ⚠️

Additional details and impacted files

@@           Coverage Diff           @@
##             main    #1943   +/-   ##
=======================================
  Coverage        ?   94.75%           
=======================================
  Files           ?       37           
  Lines           ?     1562           
  Branches        ?      456           
=======================================
  Hits            ?     1480           
  Misses          ?       80           
  Partials        ?        2           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[SBoudrias]

One thing here that isn't clear anymore about this approach;

We were talking of how the bug reported seemed to be an issue with how different terminal auto-wrap content and how that conflicted with the cursor position and the way wrap-ansi would do the wrapping.

Given that, it's unclear why a different wrapping algorithm fix this issue. This would fix it for terminals doing simple wraps, but won't it now break for those who're maybe implementing word wrapping instead?

[MazenSamehR]

By implementing character-based wrapping directly in the code rather than relying on wrap-ansi we Make wrapping deterministic, so the exact same wrapping happens on all terminals. So since we control the wrapping, we can accurately calculate cursor positions.

Now a terminal's auto-wrap behavior becomes irrelevant because we've already wrapped the content at exactly width characters, and the terminal receives pre-wrapped lines that fit within its display width. So even if a terminal has word-wrapping enabled, it only applies to content that exceeds the width, which ours won't

Example

Terminal A (word-wrap): "This is a long message"
Terminal B (char-wrap): "This is a long message"

using wrap-ansi:
- Terminal A might wrap: "This is a long" / "message"
- Terminal B might wrap: "This is a lo" / "ng message"
- Cursor calculation broken because we don't know which one we're on!

character-based:
- All terminals receive: "This is a lo" / "ng message"
- Both terminals see it's already wrapped
- Cursor calculation is accurate because we control the wrapping!

[SBoudrias]

Hey @MazenSamehR coming back from some time off, I'm looking at this PR and the source code of wrap-ansi, and I see we'll be dropping wrapping of many things the current library handles.

Given how widely Inquirer is, I wouldn't be surprised this will break things for many users. I'm wondering if you talk to anyone who maintains wrap-ansi about fixing the hard wrap issues on their side instead?