refactor: centralize anchored graphic placement

[VladaHarbour]

• Added a shared resolveGraphicPlacement() in layout-engine to centralize anchored graphic placement for images, drawings, and tables.
• The resolver now produces one placement result for paint coordinates, text-wrap exclusion coordinates, layer data (behindDoc, zIndex), and measurement impact flags.
• Refactored paragraph anchoring, pre-registered/page-relative anchors, paragraphless anchored fallback, and float registration to consume the same resolved placement instead of recomputing X/Y separately.
• Removed duplicated table anchor X logic from the floating object manager.

[linear-code]

SD-3420

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 91054858eb

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".

[chatgpt-codex-connector]

Preserve wrap-only behindDoc z-index in cells

When a table-cell anchored block uses the legacy/wrap-only behind-doc form (wrap.type === 'None' with wrap.behindDoc, but no anchor.behindDoc or zIndex), the behindDoc variable above is true and wrap exclusions are skipped, but getFragmentZIndex() only consults anchor.behindDoc/block.zIndex, so the wrapper falls back to z-index 1 and is appended after the text instead of behind it. Please feed the computed behindDoc state into the z-index fallback (or preserve the old behind-doc fallback) so these supported blocks still render behind cell content.

Useful? React with 👍 / 👎.