A list of known limitations in gl2d graphs in comparison with svg2d feature set:
mode: 'text' (for text charts) is not availabletonexty, tonexty, and toself are not avaliable (tozerox and tozeroy work) :white_check_mark: Fixed with regl.marker.symbol is pretty good since v1.28.0 but not on-par with svg :white_check_mark: Fixed with regl.line.dash are not availableline.shape: 'spline' update: 'hv', 'vh', 'hvh', 'vhv' added in https://github.com/plotly/plotly.js/pull/3087\n (most common in date axes)autorange: 'reversed') ranges. :white_check_mark: Fixed with regl.x and y are not available (toggling hover on/off in modebar works). :white_check_mark: Fixed with regl.fixedrange (would need to tweak gl-select-box). :white_check_mark: Fixed with regl.plotly_hover, plotly_unhover and plotly_click are not emitted [done in #994]hoverinfo to 'none' should disable the hover spikes. :white_check_mark: Fixed with regl.Plotly.plot callcan you also include ADD d3 events to plotly_click plotly_hover plotly_unhover events? Same as non webgl version?
Also, no support for tags as to Scattergl.
For example,
layout = go.Layout(xaxis=dict(title='<b>Title</b>'))
Title will not appear if enclosed in the tags.
can you also include ADD d3 events to plotly_click plotly_hover plotly_unhover events? Same as non webgl version?
Also, no support for tags as to Scattergl.
For example,
layout = go.Layout(xaxis=dict(title='Title'))
Title will not appear if enclosed in the tags.
Both will be fixed by @dfcreative 's regl-based scattergl.
@dfcreative quick question: is your implementation of line.dash on-par with SVG scatter?
... if so, the only remaining items in the list above are:
line.shape: 'spline' now in -> https://github.com/plotly/plotly.js/issues/3229pretty good :tada:
@etpinard dashes are not perfectly the same px-wise, but cover all SVG modes.
@etpinard dash patterns are covered by https://github.com/plotly/plotly.js/commit/dd120ee541475b833cf8c8dd447ebc6f312181d1:

A few more things that scattergl doesn't do quite right as of https://github.com/plotly/plotly.js/pull/2258:
hoveron: 'fills' now in -> https://github.com/plotly/plotly.js/issues/2887axis.layercliponaxismarker.colorscale now in -> https://github.com/plotly/plotly.js/issues/2953Writing down a couple more thoughts on future gl2d development. Some of these items may benefit from Github issue of their own down the road.
scatter and scattergl use the same set of rules e.g. we could say: _for graphs with more than 1e5 don't consider marker.size in auto-range computation._ DONE in https://github.com/plotly/plotly.js/pull/2404pointcloud to using regl. Better yet, merge pointcloud fast data pass into a scattergl _mode_ to allows fast updates.heatmapgl and contourgl? Canvas-based heatmap is currently out-performing heatmapgl in the latest Chrome version. contourgl is full of bugs and offers little to no benefits compare to svg contour. Perhaps we should just deprecate them.@dfcreative 's new regl-based scattergl was merged in #2258 and released in v1.33.0. We'll leave this issue open as a starting point to discuss future gl2d development, but we'll close off the On-par gl2d.
Not sure if y'all group scattermapbox under this category, but I would to see that support hoveron='fill' as well!
Also rangeslider:
https://codepen.io/cpsievert/pen/JLwQyx
Edit: in -> https://github.com/plotly/plotly.js/issues/2627
Closing. Looks like all the remaining items have their own ticket.
If I missed something, please open a new ticket.
Most helpful comment
can you also include ADD d3 events to
plotly_clickplotly_hoverplotly_unhoverevents? Same as non webgl version?