Re: Interesting failure mode for initdb - Mailing list pgsql-hackers

I said:
> That stuff comes out on stdout; all of the interesting stuff is on
> stderr.

Actually, given the -o option all of the interesting stuff will go to
wherever -o says.

At this stage of the release cycle I suppose we must resist the
temptation to define -o '|command' as doing a popen(), but maybe
for 7.2 something could be done with "-o '|grep -v ^DEBUG'" ...
        regards, tom lane