Re: [PATCH 3/5] setup_revisions: remember whether a ref was positive or not

Sverre Rabbelier [off-list ref] writes:

Ok, knowing that, I am somewhat more inclined to have a look at this.
I'm still not sure that I know how to implement the design you
described though. I mean, I understand the general concept, but I have
no idea in what places it would need to hook in to make it work (ditto
on how to fix the diff bug).

It is entirely possible, as Dscho hinted in his response, that your "flag"
word may be sufficient to encode "where did this object (whether it is
marked as uninteresting or interesting) come from? what did the user want
to do with it?" information for all practical purposes, even though I
doubt one bit would be sufficient. So how about ignoring "diff A...B" vs
"diff ^C A B" in the meantime and see if your change to revision.c is
applicable to the same issue in "git bundle" as a starter?

I would imagine that "diff A...B" vs "diff ^C A B" case would be solved by
the former (^C is the common ancestor between A and B) add a flag
COMMON_ANCESTOR to pending object array entry that holds ^C,
SYMMETRIC_LEFT to the one with A and SYMMETRIC_RIGHT to the one with B,
while the latter would stuff one SINGLE_NEGATIVE (^C), and two
SINGLE_POSITIVE (A and B), and the caller would tell between the two, or
something like that.  And "bundle master^0..master" (and similarly your
export stream) would note the RHS of the dot operator with SYMMETRIC_RIGHT
so that "master", even though the commit object gets marked as
uninteresting at the end, can be recovered to be of interest to the end
user.

I am not suggesting the exact flag word names above, but just trying to
enumerate how many kinds would be needed). Your single-bit in the patch
would be SYMMETRIC_RIGHT in the above ugly notation, but you can call it
REF_INTERESTING or whatever.