Spurred to curiosity by this thread, I tried a little experiment. I suspended a very thin, low-quality, tea light case above a tea light with the edge of the case in the hottest part of the flame (the tip, approx 1,000 deg C). The room was free of draughts. 1 hour later there was no sign of any melting, deformation or burning of the aluminium. It was too hot to touch, but as it hadn't melted, no part of it was above 600 deg C. Also, as there was no glow whatsoever, the temperature at the hottest point was less than 525 deg C - the starting temperature of incandescence.
In the real world the flame never touches the cup. The hottest part of the flame is about 1-1.5 cm above the melted wax pool, in free air. You can drip the melted wax onto your finger and although it is a bit painful, it doesn't do any damage. The wax pool does not normally reach the cup, although it might if the flame flickers in a draught, so there is, most of the time, a layer of insulation between the flame and the cup - candle wax. That insulator is itself a heat absorber since its ignition depends on it vapourising, so it absorbs the latent heat of vapourisation from the bottom of the flame. There is no way that the cup of a tea light gets hot enough to ignite wood, polished or not.
The flame is a different matter. It will ignite flammable material which is in or close to the flame. That is, not separated from it by an insulating layer. When all the wax has gone the flame goes out. Quite often the flame goes out before the wax has all gone becasue of the way the wick burns away before the wax pool has all vapourised. I note the comments about reports of tea lights causing fires, but to say that the heat of a normal tea light resting on a flammable surface caused a fire is implausible. You can try resting one on the palm of your hand as it burns down to discover how unlikely that is.
I'm not saying that you shouldn't use glass cups in tea light holders, but criticism of those who don't should be based on science, not anecdote.