[0.81] fix: Unicode Text length Calculation

[vineethkuttan]

Description

The length of text that is being rendered is not calculated for Unicode characters, this change will fix it

Type of Change

• Bug fix (non-breaking change which fixes an issue)

Why

Text Highlight for Unicode is not working

• Resolves Nested Text components apply styles at wrong offsets when text contains multi-byte characters #15828

Screenshots

Testing

Tested in Playground

Changelog

Should this change be included in the release notes: yes

(fix: Unicode Text length Calculation)

Microsoft Reviewers: Open in CodeFlow

[sundaramramaswamy]

Looks good — correctly fixes the UTF-8 vs UTF-16 length mismatch for DirectWrite text ranges. One suggestion for a follow-up optimization:

Minor optimization: Each Utf8ToUtf16(fragment.string).length() allocates a temporary std::wstring just to get the length. Would it be worth adding a Utf8ToUtf16Length() helper to vnext/Common/unicode.h that calls MultiByteToWideChar with a null output buffer? That would avoid the allocation entirely — could matter in hot paths like RenderText and hit-testing loops.

Not blocking — the fix is correct as-is.

[vineethkuttan]

Looks good — correctly fixes the UTF-8 vs UTF-16 length mismatch for DirectWrite text ranges. One suggestion for a follow-up optimization:

Minor optimization: Each Utf8ToUtf16(fragment.string).length() allocates a temporary std::wstring just to get the length. Would it be worth adding a Utf8ToUtf16Length() helper to vnext/Common/unicode.h that calls MultiByteToWideChar with a null output buffer? That would avoid the allocation entirely — could matter in hot paths like RenderText and hit-testing loops.

Not blocking — the fix is correct as-is.

Will add this while forward porting to main and other stable versions. And for 0.81 will raise the separate PR for this.