Stephan Szabo <sszabo@megazone23.bigpanda.com> writes:
>> A compromise answer might be to offer a SET variable that selects the
>> Microsoft-compatible misimplementation. Would that fly?
> It would for me. I'd rather have the default be the spec correct behavior
> and let people configure their server to follow the misinterpretation.
Yes, I agree.
> Is the conversion just the hack in the grammar rules for
> a_expr '=' a_expr?
AFAIK that's it. It might be a little tricky to do cleanly,
though, since gram.y shouldn't look at SET vars. Really, the
conversion would have to be done in analyze.c instead.
regards, tom lane