isEmpty

Meh. Why did that pass D10.1? In the test setting the $field->value directly doesn't cause the calculation as setValue(). Maybe it should be really.

I'd say this is blocked on 🐛 Computed values Active

thanks @ekes for your updates. Going to re-jump and review your last comment asap ...

Soo @ekes ...
I better inspected all this, and I ended up with the conviction that is still nice to define a GeofieldItem::isEmpty() method that not only make sure that the geofield primitive/row (WKT) input is not empty but also that it generates a valid /Geometry.
That is correct, and we should keep it, and also the Test code base that assumes it ...

What we should rather limit, as you correctly mentioned, is the redundant generation of the same /Geometry in the same GeofieldItem class, to limit the time/memory script overhead and better apply the DRY principle.

That's why in my latest commit a preferred to keep basically unchanged the GeofieldItem::isEmpty() method and rather to define a private \Geometry $geometry property that could be generated once and reused also in the GeofieldItem::populateComputedValues() method.

Tests are still passing nicely with this change and $geo_php_wrapper->load($value) is only called once in the class.

I would commit this that looks a nice improvement to me.
Let me know if you disagree, if you see some issues or potential regressions, or evident mistake with this solution of mine.

Well ... I didn't read any feedback to my latest update.
I am feeling my last update / commit is pretty solid, hence I am going to merge this into 8.x-1.x dev, for being part of new upcoming release.

Automatically closed - issue fixed for 2 weeks with no activity.