Roamy here. NASA had a great deal of egg on its face after finding out that an error during mirror polishing messed up the Hubble Space Telescope. Hubble did go on to become a national treasure in astronomy and cosmology. (minus 1,000 points if you think cosmology has anything to do with mascara and eyeliner.)
One of the lessons learned was to have a full test of any mirror, through the full thermal cycle, on the ground before flight, to catch any alignment problems. Rule of thumb for thermal cycling in space is +250 °F to -250 °F, but there are worse extremes, and I’ll let the thermal analysts handle that. Imagine going from a fairly hot oven into near liquid nitrogen temperature ( which is actually −321 °F for the pedants out there). Coefficient of thermal expansion really has an effect, especially in assemblies where it’s mismatched. And this is my personal view, but one test is worth 100 analyses.
NASA was determined not to have any mirror alignment problems with the Chandra X-ray Observatory, and they modified an existing facility at Marshall just for that. Originally called the X-Ray Calibration Facility, now it’s called the X-ray and Cryogenic Facility after it was upgraded for much colder temperatures. Either way, it’s still the XRCF. There were 550 heater panels processed for the XRCF, and I could bore you silly with the details.
X-ray source on the upper left, vacuum tube, vacuum chamber and clean room where you put your mirror assembly on the right. The mirrors for the James Webb telescope are being tested now. Here’s a pretty good article about it.
Test what you fly, fly what you test.