Re: BUG #15080: ecpg on windows doesn't define HAVE_LONG_LONG_INT

Andrew Dunstan <andrew@dunslane.net> writes:
> Well that seems to have crashed and burned badly. I'm just going to 
> disable ecpg checks on this animal as suggested upthread.

Hmm ... this might be too much of a coincidence, but I can't help noticing
that the places that are going south with -D__USE_MINGW_ANSI_STDIO are
pretty nearly the same ones I just pointed to in
https://www.postgresql.org/message-id/21670.1526769114@sss.pgh.pa.us
as using "%lf".  I'd supposed that that was mostly compulsive neatnik-ism,
but is it possible that mingw's "ansi stdio" library is actually
sensitive to that?

            regards, tom lane