Peter Eisentraut <peter_e@gmx.net> writes:
> On fre, 2012-02-17 at 10:19 -0500, Tom Lane wrote:
>> That's still a cache, you've just defaulted on your obligation to think
>> about what conditions require the cache to be flushed. (In the case at
>> hand, the trigger for a cache rebuild would probably need to be a glibc
>> package update, which we have no way of knowing about.)
> We basically hardwire locale behavior at initdb time, so computing this
> then and storing it somewhere for eternity would be consistent.
Well, only if we could cache every locale-related libc inquiry we ever
make. Locking down only part of the behavior does not sound like a
plan.
regards, tom lane