With Float32 inputs a single bit of difference can cause all digits to be written to a string:
Showoff.showoff(Float32[0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.90000004], :plain)
# ["0.30000001", "0.40000001", "0.50000000", "0.60000002", "0.69999999", "0.80000001", "0.90000004"]
nextfloat(0.9f0)
# 0.90000004f0
With Float32 inputs a single bit of difference can cause all digits to be written to a string: