Peter Eisentraut <peter_e@gmx.net> writes:
> Tom Lane writes:
>> ISTM that at one time we had considered actively discriminating against
>> -ffast-math in CFLAGS in configure ... but I don't see any sign in
>> current sources that we make any attempt to remove -ffast-math from
>> environment-supplied CFLAGS. Peter, do you think that would be a
>> reasonable thing to do?
> Would it be possible to write a short test case that exhibits this
> behavior?
Not sure; apparently it depends on optimization level, so I'd be
hesitant to assume that any short test case would reliably expose
the problem. Also, aren't you trying to avoid run-time tests in
configure?
But if you prefer a run-time test, I'll see if I can cons one up.
regards, tom lane