Should we add a *strict* (or *debug*) mode that runs Plotly.validate on every newPlot and update call before attempting to draw?

[etpinard]

which would essentially be a more general way to resolve #1736

I'm thinking this strict/debug mode would be set as a config argument. We could either log the attribute errors (as Plotly.validate currently does) or throw an error.

[rreusser]

Random question: does a bundle with stripped metadata affect at all what or how it's able to report errors? (Otherwise I like the idea of this as a config argument and of reporting as much/helpful information as we can.)

[etpinard]

does a bundle with stripped metadata affect at all what or how it's able to report errors

In its current implementation, Plotly.validate doesn't log the attribute descriptions, so no, it wouldn't matter.

[etpinard]

cc @phxnsharp if you weren't aware of this.

[phxnsharp]

This sounds like exactly what we would prefer. Is there a sense of what type of performance hit this would involve?

[etpinard]

Is there a sense of what type of performance hit this would involve?

At the moment, Plotly.validate is fairly fast. It clocks in at no worse than 50ms. Plotly.validate's execution time, in its current state, scales with the number of "data" and "layout" attributes set.

That said, if we choose to implement #1998, Plotly.validate will likely slow down a little bit. It we only check for data arrays lengths, the performance hit will be small. But if we choose to check for invalid data items (e.g. null in scatter marker traces), this will make Plotly.validate's execution time scale with the length of the data array - which will slow it down considerably for graphs with many data points.

[etpinard]

cross-referencing #167

[gvwilson]

Hi - this issue has been sitting for a while, so as part of our effort to tidy up our public repositories I'm going to close it. If it's still a concern, we'd be grateful if you could open a new issue (with a short reproducible example if appropriate) so that we can add it to our stack. Cheers - @gvwilson