fix: improve vml watermarks rendering

[VladaHarbour]

Improve parsing watermark opacity and font.
Word/Before comparison:

Fix applied:

[linear-code]

SD-3265

[github-actions]

Status: PASS

The changes are all within the VML watermark importer (ECMA-376 Part 4) and don't introduce any spec violations. A quick rundown of what I checked:

• parseVmlOpacity correctly handles the three forms VML allows for v:fill/@opacity: a plain fraction, a percentage (50%), and 16.16 fixed-point with f suffix (32768f → 32768/65536 = 0.5). That matches the VML ST_Fraction-style serialization. https://ooxml.dev/spec?q=fill
• decodeXmlEntities on the v:textpath/@style value is benign — VML's style attribute carries CSS in an XML attribute, so "-escaped quotes around font-family are normal XML attribute escaping. Decoding them before the CSS parse is fine.
• normalizeVmlColor only maps the standard VML named colors (black/blue/gray/green/lime/red/silver/white/yellow) to hex — those are all valid VML color names. https://ooxml.dev/spec?q=fillcolor
• The other changes (textLength, lengthAdjust, fill-opacity, font fallback) are SVG output details, outside the OOXML surface.

Nothing here reads a non-existent VML attribute or assumes a wrong default for an OOXML element.

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: def351e0ca

ℹ️ About Codex in GitHub

Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".

[codecov-commenter]

Codecov Report

✅ All modified and coverable lines are covered by tests.

📢 Thoughts on this report? Let us know!