[Grida Canvas] Dropping (from external) insertion position does not handle the current transform accurately

[softmarshmallow]

[Grida Canvas] When dropping items (externally, e.g. an image) the position of the dropped item is not accurately handled. I think the zoom part in transform is not handled, with the translate.

When dropping, the dropped item, if handled, must be placed visually accuratelu in the dropped position. (center if the size of the dropping item be can calculated, otherwise top-left (0,0) of the drop position)

[k08200]

Opened #700 with the drop-position fix. It converts the drop point to canvas space before subtracting known image/SVG half-size, centers image/SVG insertions, and applies the same behavior to inline image data-transfer drops.