> I noticed one bug, or at least instance of bit rot: the float8 code> attempts to parse "Infinity" and "-Infinity" specially. Unfortunately,> this doesn't actually work:
Infinity doesn't work for float4 either.