[adimalkar]

This PR makes scattergl, scatterpolargl and splom use marker size when they pick the hovered point, as scatter does.

Closes #8082.

Why this is correct

• scatter scores a point by its distance to the marker edge. Inside a marker, a smaller marker wins.
• scattergl and splom measured only the distance to the marker center. So the marker drawn on top always won.
• This PR moves the existing scatter math into getAxisDistance and getPointDistance in scatter/hover.js, with no change in behavior. scattergl and splom now call them.
• Each radius is half of the sizes/size that convert already computes for the WebGL markers.
• For large traces, the spatial index search box grows by the largest marker radius.
• scatterpolargl gets the fix through scattergl.

Behavior change

The WebGL traces now give the same result as their SVG counterparts. I checked this with real pointer events in Chrome:

Case	SVG	WebGL before	WebGL after
Issue snippet (A size 6, B size 20, same point)	A	B	A
One trace, sizes [6, 20] at the same point	small	big	small
Big marker drawn first, pointer 8px from center	B	A	B
splom, sizes [6, 20]	—	big	small
scatterpolargl vs scatterpolar, sizes [6, 20]	small	big	small
1e5 points, pointer 30px inside an 80px marker	big	no hover	big

Changes that affect every WebGL plot with markers:

• A point is in range when the pointer is within hoverdistance of the marker edge.
• spikedistance and the label order in x hovermode use the scatter scoring.
• When the pointer is inside several markers of the same size, the point drawn last wins. Before, WebGL picked the nearest center.

The last item changes five expectations in scatterpolargl_test.js. The default pointer is 3.7px from point 2 and 5.8px from point 26 of 'Trial 3'. Both markers have a 6px radius. scatterpolar picks point 26 for the same figure.

The hover label still anchors at the marker center in WebGL. I left that out of this PR.

Tests

• New tests in hover_test.js (hover on points at the same position):
  • the smaller marker wins for scatter, scattergl and splom;
  • a marker hovers when its edge, but not its center, is within hoverdistance;
  • a 1e5-point scattergl trace finds a marker whose center is outside the search box.
• Suites run locally (Chrome 157):
  • hover 237/237, scattergl 48/48, splom 51/51, gl2d_click 12/12;
  • hover_spikeline 45/45, scatterpolargl 32/32, gl2d_plot_interact 16/16, scattergl_select 20/20.
• select: 115/117. The two failures are the ohlc and candlestick lasso tests, which fail the same way on main (90f954a) on this machine.
• npm run lint and npm run test-syntax pass.
• No image baselines change. Hover does not draw in the image tests. I did not run regl-codegen, because no shader code changed.

This change was developed with AI assistance (Claude Code). A human reviewed the code and ran the jasmine suites.

[Copilot]

Copilot was unable to review this pull request because the user who requested the review has insufficient remaining quota. Try requesting a review at a lower effort level.

An organization owner can choose the repository organization as the billing source for reviews requested by members with a paid Copilot license. Learn about Copilot code review billing. Review the Copilot code review billing settings.