But nevertheless, I think the problem you have now is that you've painted yourself into the corner of the fallacy of argument ad ignorantium, the argument from ignorance - because it has not been definitively proven true, it can be assumed to be false. Or, more precisely, because you can't imagine a path, there is no path. But, as I alluded to earlier, the limits of the universe are not determined by the limits of our imaginations.
What can I say? Examining the differences between eyes of various members of the animal kingdom, combined with what we know from the fossil record about when they arose on earth, can lead us to plausible pathways for the development of mammalian eyes.
So, here's one possibility, complete with real, live examples of each step. Simple photoreceptive cells, a la ticks -> eye spots consisting of multiple photoreceptive cells; e.g., flukes -> compound eyes composed of multiple photoreceptive cells differentiated into discrete units, viz arthropods -> simple pinhole eyes, with multiple photoreceptive cells differentiated into discrete units forming a retina behind a pinhole aperture; e.g., nautilus -> simple lens with variable focal length sitting in front of multiple photoreceptive cells differentiated into discrete units which form a retina, viz fishes and snakes -> fixed focal-length eyes with variable density lenses sitting in front of multiple photoreceptive cells differentiated into discrete units which form a retina, a la reptiles and birds -> further differentiation of discrete units of multiple photoreceptive cells which form a retina, such that discrete units exhibit variable responses to external stimuli in order to enhance image reception in low light and provide differential responses to varying wavelengths of light (i.e., color vision), A/K/A mammalian eyes.
;-)