Fujii Masao <masao.fujii@oss.nttdata.com> writes:
> Attached is the patch that Nathan proposed at [1] and I think that
> it's worth applying. I'd like to add this to next CommitFest.
> Thought?
I can't get excited about this in the least. For any "normal" use of
passwords, 100 bytes is surely far more than sufficient. Furthermore,
if there is someone out there for whom it isn't sufficient, they're not
going to want to build custom versions of Postgres to change it.
If we think that longer passwords are actually a thing to be concerned
about, then what we need to do is change all these places to support
expansible buffers. I'm not taking a position on whether that's worth
the trouble ... but I do take the position that just inserting a
#define is a waste of time.
regards, tom lane