fix!: remove MarkerSvg and uses

[rachel-fenichel]

The basics

• I validated my changes

The details

Resolves

Fix for #8962

Proposed Changes

Commit 1: Delete the actual drawer, subclasses, and uses
Commit 2: Delete residual SVG elements that were rendered through the cursor/marker flow instead of through the normal block flow.
Commits 3 and 4: lint and format.

Reason for Changes

MarkerSvg and friends represented a way to draw the cursor that was outside of the standard rendering flow. These SVG elements were children of the standard block (field, workspace, etc) SVG elements, but were updated through explicit calls to setCurNode on the cursor.

Ben's focus changes updated focus rendering to use elements that are rendered/updated through the standard rendering flow. Where possible, focus is shown using CSS rather than by creating new elements.

Test Coverage

Documentation

Additional Information

• keyboard-experimentation will need a related PR to remove explicit calls to hide().
• The old keyboard navigation plugin will not compile after this change (if it even compiled before).
• IASTNodeLocationSvg can be removed as a follow-up, if it hasn't already been removed during refactors of ASTNode.

[rachel-fenichel]

Does not fix #8968

[cpcallen]

Opinion: I would (and now have) described issue #8962 as a task rather than a bug, so I'd have titled this PR and its main commits as chore: rather than fix:.

[cpcallen]

Should render no longer call renderManagement.finishQueuedRenders() at all?

(Maybe the queued renderers will be finished automatically, and the method call is not needed if you don't need the resulting promise?)

[rachel-fenichel]

Yes, queued renders will finish automatically, and the method call is specifically to be able to get a promise and run after the renders have finished.

[cpcallen]

Could this interface not be deleted entirely, now that it is empty?

I suppose we could formally deprecate it first, but since anyone who was calling setFooSvg on an IASTNodeLocationSvg instance already has broken code I'm not sure it's worth it.

[rachel-fenichel]

Filed as #9006

[cpcallen]

For some reason GitHub hasn't run build, format or lint workflows, and I can't figure out how to kick it in to action.

[rachel-fenichel]

Workflows ran and passed.