AC gauges are jumpy on Brief and Overview pages

[blammit]

There's something jumpy about the AC gauges in general, as reported in slack. Needs investigation.

[blammit]

Revisit this when issues concerning the gauge calculations in #532 are addressed.

Also, some relevant points:

• The max seen value (used as the max value for the AC input gauges) refers to the maximum seen for any AC input, but it should keep a per-input maximum value instead, as the max for one input may be quite different to that for another.
• It's quite possible that Global.acInputs.model is not needed at all, as the Overview and Brief pages only ever use the currently active AC input, and the model is not used in the Device List either.

[chriadam]

Might also be an animation thing, see e.g. https://github.com/victronenergy/gui-v2/blob/main/components/controls/Slider.qml#L67

I'm currently working on a shader-based progress arc implementation which should solve a lot of problems, but let's see...

[chriadam]

#602 still need to reimplement the ProgressArcs and properly test performance on device etc.

[chriadam]

Also related to setting appropriate min/max values for gauges, avoiding dynamic scaling.

[chriadam]

Marking as low priority, as I don't think this is a showstopper for 1.0.0.

[kwindrem]

It's pretty twitchy on a system with minimal load. The AC input gauge swings from max to min with only a 1 or 2 watt change. I'd say something needs to be done for 1.0.0.

It's quite possible that a system with more load produces more stable gauges but have not tried it. Still, it's very distracting with minimal load.

[chriadam]

Ah, so this is because of the min/max values not being set properly for the gauges (#532)? Thus, because it's using the "previously seen min / max" values as the rails, it is jumping large amounts visually in reaction to small actual changes? Or am I misunderstanding?

[kwindrem]

Yes, that's my interpretation of what's going on.

[jpetrell]

I am tempted to close this one and file specific issues later if the rewritten autoranging + manual min/max logic planned in #532 is not enough? Or at least drop this from 1.0.0 Milestone.

[blammit]

Yep, let's close this, since #532 addresses these issues. This was originally filed before all of that work was discussed in further detail.