Re: [RFC/PATCH] fsck: do not canonicalize modes in trees we are checking

[-cc Kirill, as his address seem out-of-date]

On Tue, Sep 23, 2014 at 04:23:43PM +0000, Edward Thomson wrote:

On Tue, Sep 23, 2014 at 11:47:51AM -0400, Jeff King wrote:

quoted

As far as I can tell, fsck's mode-checking has been totally broken
basically forever. Which makes me a little nervous to fix it. :)
linux.git does have some bogus modes, but they are 100664, which is
specifically ignored here unless "fsck --strict" is in effect.

I'm in favor of checking the mode in fsck, at least when --strict.  
But I would suggest we be lax (by default) about other likely-to-exist
but strictly invalid modes to prevent peoples previously workable
repositories from being now broken.

I have, for example, encountered 100775 in the wild, and would argue that
like 100644, it should probably not fail unless we are in --strict mode.

Yeah, I'd agree with that. The big question is: what breakage have we
seen in the wild? :)

I think treating 100775 the same as 100664 makes sense (want to do a
patch?). Do we know of any others? I guess we can collect them as time
goes on and reports come in. That's not the nicest thing for people with
such repos, but then again, their repos _are_ broken (and it's only
really a showstopper if they are trying to push to somebody with
receive.fsckObjects turned on).

-Peff