Re: Record comparison compiler warning

Alvaro Herrera <alvherre(at)2ndquadrant(dot)com> writes:
> Stefan Kaltenbrunner wrote:
>> http://www.postgresql.org/message-id/raw/1381949353(dot)78943(dot)YahooMailNeo(at)web162902(dot)mail(dot)bf1(dot)yahoo(dot)com

> I would blame Bruce's MUA, or surrounding configuration, for this
> problem. It looks fine in mine, and as far as I can see, Kevin's
> message correctly declares the email to be in Latin-1 quoted-printable
> encoding, which declares A0 to mean non-breaking space.

I agree with Bruce: this patch is broken. A0 may be a non-breaking space,
but the fact remains that it isn't a space, and since we're talking about
a diff, the whitespace needs to be the same as what it is in the original
file. (I believe that some of those whitespace runs involve tabs not just
spaces, making the diff even more wrong.)

Admittedly, if you're just eyeballing it, it might look fine. But try
feeding it to "patch" and you'll find out it ain't.

regards, tom lane