Re: [PATCH] apply: refuse touching a file beyond symlink

From: Jeff King <peff@peff.net>
Date: Jan 30, 2015, 20:48 UTC
Message-ID: <20150130204805.GA10616@peff.net>
In-Reply-To: <xmqqwq44auml.fsf@gitster.dls.corp.google.com>

On Fri, Jan 30, 2015 at 12:20:02PM -0800, Junio C Hamano wrote:

Show 13 quoted lines

> Jeff King <peff@peff.net> writes:
> 
> > I had the impression that we did not apply in any arbitrary order that
> > could work, but rather that we did deletions first followed by
> > additions. But I am fairly ignorant of the apply code.
> 
> No, you are thinking about the write-out of the finished result,
> which may have to turn existing directory to a file or vice versa on
> the filesystem, but that happens _after_ we decide what to turn into
> what else, completely in-core.
> 
> And the decision to determine what the input _means_ should not
> depend on the order of patches in the input.

Ah, OK. Yeah, doing it progressively can only be accurate if our name-checks follow the same order as applying, because we are checking against a particular state.

But could we instead pull this check to just before the write-out time? That is, to let any horrible thing happen in-core, as long as what we write out to the index and the filesystem is sane?

-Peff