round() does not comply with PDG rounding
The rounding method does not comply with the procedure stated in Phys. Rev. D 98, 030001, Introduction 5.3.
A difference arises for
>>> Quantity("1.23456 +- 0.950").str(significant_digits=0.354)
'1.2 +- 0.9'
In that case, the error 950 should be rounded up to 1000, thus giving a two digit uncertaintiy 1.23 +- 1.0
.