dragging and dropping images with the same name replaces a previously uploaded image

when dragging and dropping an image with the same name into a document, all occurrences of an image with that name get overwritten. i think it would be more intuitive if internally those images would be handled as separate images.

Start by reproducing the drag-and-drop flow with two images sharing a filename, then trace how dropped images are identified and stored in the authoring document. The work is done when same-named images remain separate and existing occurrences are no longer overwritten.

Written by the indexing model from the issue text.