Problems with shared autobins on single-sample-point histogram2d

[nicolaskruchten]

One of these facets has misaligned bins: https://codepen.io/nicolaskruchten/pen/MdVdGw

[nicolaskruchten]

Note also: that misaligned one has x:NaN, y:NaN in the hover for some reason

[etpinard]

They are misaligned, but are they incompatible? It's not obvious to me by looking at it quickly. I'll take deep dive tomorrow.

[etpinard]

Thanks for reporting. Something is up with histogram2d traces that have 1 x and y sample point. I can't replicate the problem with other samples with 2 or more points (example adding one pt to gd.data[2] https://codepen.io/etpinard/pen/byjbbr?editors=1010)

Is this blocking for 1.48.0?

[etpinard]

Similar to #3881

[nicolaskruchten]

agreed that it is "compatible" and that it shouldn't block 1.48, but it does seem like a bug to me.

[etpinard]

but it does seem like a bug to me.

Yep, no doubt it's a bug.

[etpinard]

In fact, even non-shared single-sample-point histogram2d traces do not respect xbins / ybins settings:

https://codepen.io/etpinard/pen/eajory?editors=1010

[etpinard]

Ha, looks like things were working ok back in 1.42.x: https://codepen.io/etpinard/pen/eajory?editors=1010

Commit e0116cb#diff-98b2c1750afe8f855914e1647f72dbfe didn't work for histogram2d traces