Upstream contributions to jfx.incubator.richtext

Gaps in the incubator RichTextArea that html-to-node currently works around (split from #2; details and workaround pointers there). Each landed OpenJFX feature deletes a workaround here:

1. Character-level link attribute with activation handling (currently: custom HREF attribute + getTextPosition on click)
2. Inline node/image embedding as first-class model feature (currently: addNodeSegment, view-only model only)
3. Baseline-shift / vertical-align character attribute (currently: sub/sup lose their shift, font-size only)
4. Numeric font weights (currently: CSL's font-weight: 100 collapses to normal/bold)
5. Styleable selection/caret colors on the control (e.g. -fx-selection-fill, -fx-caret-color) instead of Modena rules on skin-internal sub-nodes — see the analysis in #2
6. A lean read-only viewer mode without caret/edit overhead

🤖 Generated with Claude Code

Start with issue #2 and its workaround pointers, then review how html-to-node currently handles the six listed RichTextArea gaps. Done means the relevant OpenJFX features are available and the corresponding workarounds can be removed, covering links, inline nodes/images, vertical alignment, numeric weights, control colors, and read-only mode.