For what reason? That's the least likely culprit.
Every single IO pin has a voltage drop going towards one IO pin, which is the pin that takes in the row input for the function key. Flipping the leads results in an open circuit, which means that there is at least one diode in between this IO pin and all of the other IO pins. Unfortunately with switches soldered in I can only see half of the board, but there's no obvious defects that I can see, especially one that would cause this pin to get connected to every single other pin. I didn't completely mangle my board, which should be obvious from the picture, and there is no circuitry that I can find outside of the chip that could produce the results I see from my multimeter. I can think of reasons why this circuitry would exist inside of the chip however, and if the internal diode went bad and is now allowing bidirectional current that would explain everything.
The only reason I'm hesitating is because I have very limited experience with smt soldering (my Iris was the first time I had ever done it).
Well, that's annoying. I would offer to replace the 32u4 for you but shipping two ways may make it less appealing than just getting that new PCB that was offered.
You were right that it wasn't the chip...apparently one of the switch pads is incredibly easy to bridge to ground, to the point were you can bridge it, desolder the switch and the bridge will still remain
Pictures of the evil pad in question to come tomorrow when I'm not falling asleep
EDIT: yup, that was the culprit, a single leg shorted to ground...now I need to figure out how to fix the two LEDs that I murdered in the process of fixing the switches
EDIT 2: how do you desolder one of the rgb leds with a conventional soldering iron? I removed all of the visible solder and it's still stuck...
Here's the bad pad in question. You can't really tell from this, but the ground is exposed a fraction of a millimeter away from the pad...