fix: avoid splitting surrogate pairs in trimEnd

[Alexandre Kohler (kwy404)]

Root cause: trimEnd in src/common/stringUtils.ts cuts the string at maxLength - 1 UTF-16 code units. When that cut lands between the two halves of a surrogate pair (an emoji or any other astral character), the result keeps a lone high surrogate before the ellipsis. Truncated object previews, keys, descriptions and console.table rows then show a replacement character. For example trimEnd('ab\u{1F600}cd', 4) returns 'ab\uD83D…'. trimMiddle, right below it, already guards against this case, but trimEnd did not.

Fix: if the last kept code unit starts a surrogate pair, drop it too, using the same codePointAt(...) >= 0x10000 check that trimMiddle uses.

Test: added trimEnd does not split surrogate pairs to src/common/stringUtils.test.ts. It fails before the change (expected 'ab\uD83D…' to equal 'ab…') and passes after. The rest of the unit suite is unchanged.

[Copilot]

Copilot review overview

🟢 Approval recommended

The implementation matches the established trimMiddle approach and the regression test covers both boundary cases.

Review effort: Balanced
Findings: None

What changed in this PR

Fixes trimEnd so truncation does not split UTF-16 surrogate pairs.

Changes:

• Adjusts truncation when the boundary intersects an astral character.
• Adds regression coverage for safe and unsafe truncation boundaries.

File	Description
src/​common/​stringUtils.ts	Prevents retaining a lone high surrogate.
src/​common/​stringUtils.test.ts	Tests emoji truncation boundaries.

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

[Connor Peet (connor4312)]

thanks!