I think the accuracy has already been accurately stated but I will add my 2 cents anyway. Metal moves with temperature but not enough to matter at temperatures our body can tolerate. I looked up a thermal expansion coefficient (googled it) of 13% per degree F but that is 10 to the negative 6 for aluminum. In other words .00000013 times the length. Steel is 6.7% - about half. So if you really care, use steel measurement devices not aluminum. But neither moves enough to matter.

A square built of the same materials will not become un-square due to thermal expansion. All the material will move exactly the same amount. If built of dissimilar materials you could do the math but I am pretty sure it also won't matter.

This is physics, not thermo dynamics - but that also doesn't matter.