So for a car with a 12 gallon tank, let’s suppose it’s got two gallons of fuel and ten gallons of air. That works out to about 0.011 pounds of air. Suppose that air is at 45F, 50% humidity. The water vapor content is 22 grains of water per pound of air.
Now cool that tank down to 20F. at 100% relative humidity, the moisture content is 15 grains/lb, so some of that water vapor has condensed: 7 grains of condensate per pound of air, times 0.011 pounds of air, equals 0.077 grains of water. The density of liquid water is about 253 grains per cubic inch, so now your tank has roughly 0.0003 cubic inches of ice at the bottom of the tank. That’s a cube of water about 1/16" on a side, assuming it all coalesces into a single bolus.
The tank will breathe due to daily temperature cycling, but with not much volume change, since the daily temperature swing won’t be very big; swinging between 20 and 40F each day, it’ll exhale/inhale about a half-gallon of outside air every day. If the ambient humidity of the air is greater than about 34%, then some portion of that humidity will condense inside the tank the next night when the temp gets back down to 20F. But it won’t be much, since you’ve only brought in a half-gallon of humid air instead of the 10 gallons of humid air that the tank started with.
If the diurnal temperature cycle is entirely below freezing, the frost that evolves inside the tank won’t coalesce into a single solid icebergling; you’ll just have something like frazil ice, a gasoline slushee.
Large bits of ice should get caught on the gravel screen at the pump’s inlet; anything small enough to get through is not going to damage or stall the fuel pump, and once it gets to the fuel injectors, it should melt and get squirted into the engine without incident.
If your engine wouldn’t turn over but your battery was OK, then maybe your starter wasn’t kicking its pinion gear out to engage with the flywheel’s ring gear. If it was on a cold winter day, this may have been because the mechanism was iced up, and hammering cracked the ice up enough to allow free movement again.
Here’s what that’s supposed to look like:
If the little pinion gear can’t come out and play with the flywheel’s big ring gear, it can’t turn the engine.
Am I right in guessing that over-wing GA planes don’t use a fuel pump? If so, it’s not hard to imagine ice crystals accumulating in (and eventually blocking) the line from the wing to the carb, since you wouldn’t have a fuel pump chewing up ice and cramming fuel past minor obstructions.