https://bugs.documentfoundation.org/show_bug.cgi?id=173967

--- Comment #2 from [email protected] ---
Thank you for reviewing this. Although Windows participates in font fallback,
the evidence suggests an application-specific issue.

On the affected Windows 11 computer, Notepad, Microsoft Edge, and Microsoft
Office displayed emoji in color through normal fallback, while LibreOffice
displayed monochrome glyphs.

Two computers had the same Windows build and LibreOffice version, identical
Segoe font binaries, and matching font-linking and substitution settings. Their
enumeration orders differed:

Affected PC: Windows listed Segoe UI Symbol at 115 and Segoe UI Emoji at 247.
LibreOffice listed them at 204 and 277. Emoji fallback was monochrome.

Working PC: Windows listed Segoe UI Emoji at 107 and Segoe UI Symbol at 116.
LibreOffice listed them at 212 and 221. Emoji fallback was colored.

Importantly, IDWriteFontFallback::MapCharacters selected Segoe UI Emoji for
U+1F600 on BOTH computers, including the one where LibreOffice displayed
monochrome.

The intervention: On the affected PC, I deleted and immediately recreated only
the Segoe UI Symbol registration value in the Windows Fonts registry,
preserving its original data and type. No font files were modified.

After the registry intervention, the relative order of Segoe UI Emoji and Segoe
UI Symbol changed in the Windows Fonts registry enumeration. LibreOffice's font
descriptor enumeration reflected the same change.

Other Windows applications correctly rendered color emojis through automatic
font fallback both before and after the intervention. LibreOffice only did so
after the registry enumeration order changed without a font replacement rule.

This does not establish the precise LibreOffice code path, but strongly
connects its behavior to enumeration order. Could the Windows VCL font
collection or fallback selection depend on candidate discovery order? I am not
suggesting LibreOffice change Windows registry settings, only that it
investigate why its result differed from DirectWrite and other Windows
applications.

Please reconsider the NOB conclusion, or suggest a diagnostic to identify where
the final fallback choice occurs.

-- 
You are receiving this mail because:
You are the assignee for the bug.

Reply via email to