Re: [RFH] revision limiting sometimes ignored

From: Linus Torvalds <torvalds@linux-foundation.org>
Date: Feb 6, 2008, 00:52 UTC
Message-ID: <alpine.LFD.1.00.0802051648410.2967@woody.linux-foundation.org>
In-Reply-To: <7vprvb6k9u.fsf@gitster.siamese.dyndns.org>

On Tue, 5 Feb 2008, Junio C Hamano wrote:

Show 6 quoted lines

> >
> >  - make commit warn if any parent commit date is in the future from the 
> >    current commit date (allow a *small* fudge factor here, say 5 minutes).
> 
> Hmmm.  In other words, you are punished for trying to build on
> top of somebody else who screwed up.  That sucks.

Well, I was actually thinking that the most reasonable thing to do is that if you pull from somebody, and you get this warning, you send an email saying "you suck, I will not pull your broken crap".

But the real problem is that you might be importing it from some external legacy SCM entity, and then you can't say "you suck, I won't pull", because the whole point is that external entity obviously *does* suck, and you want to simply stop using it. And then the "I won't pull" isn't an option ;)

So yeah, I don't think the warnings really work, if only because of that "import from crappy CVS repo" issue.

But the revision.c change might be worth it, if only as a slight band-aid for the current issue. It won't fix the original problem, though (because that broken repo had a five *year* clock skew, not an hour :)

I'll continue to think about whether I can come up with some sane heuristic that allows non-broken cases to not go all the way up to the root.

		Linus