On 26.08.21 16:00, Tom Lane wrote:
Generally speaking, we've been resistant to marking anything leakproof
unless it has a very small code footprint that can be easily audited.
In particular, anything that shares a lot of infrastructure with
not-leakproof functions seems quite hazardous. Even if you go through
the code and convince yourself that it's OK today, innocent changes
to the shared infrastructure could break the leakproofness tomorrow.
I think the complexity of the implementation of upper() and lower() is
on the same order as bttextcmp() and similar, so it wouldn't be totally
out of scope.
