Greg Stark <gsstark@mit.edu> writes:
> If it matches an SQL keyword after being downcased the old fashioned way, then
> it's an SQL keyword. If not then the locale-aware tolower() would be
> appropriate for tables, columns, etc.
That's exactly what we do already. The complaint was that the
locale-aware downcasing is broken (not to put it too finely) in Turkish
locales, leading to unexpected/unwanted results for identifiers that are
not keywords. My own opinion is that the correct response is to fix the
Turkish locale tables, but I can see where that might be beyond the
skills of the average Postgres user. Thus I thought a reasonable
compromise would be to override the locale for the handling of A-Z,
allowing it to determine what happens to high-bit-set characters only.
regards, tom lane