From: Junio C Hamano <gitster@pobox.com>
Date: Thu, 03 May 2012 20:32:39 GMT
Subject: Re: Index format v5
Message-ID: <7vr4v1dmzc.fsf@alter.siamese.dyndns.org>
In-Reply-To: <87k40tgix2.fsf@thomas.inf.ethz.ch>

Thomas Rast <trast@student.ethz.ch> writes:

> Thomas Rast <trast@student.ethz.ch> writes:
>
>> Comparisons for equality never have false negatives with any hash
>> function.  Collisions are false positives.
>
> Bah, shouldn't hit "send" so fast.  There goes my karma :-D

Heh, don't take "karma" thing too seriously. I was merely trying to be
funny and I seem to have failed the attempt.

It depends on what question you are asking, and against "Did this change?"
question, saying "No, it didn't" when the right answer is "Yes, it did" is
a false negative.  You are asking for a false negative if you substitute
byte-for-byte comparison with comparison of hashed results of the values
that needs to be compared.

But as I said, mtime is the primary thing that protects us, so it may not
be such a big deal.