> I think that refs in files-backend is a bit hard to update to > annotate with extra information like "is this merely promising or a > tip of the connectivity graph?" and agree that we should not make > refs promisors. OK, we agree on this :-) > Such an otherwise unreferenced blob could be made into a promised > object by the promisor repository, though, I think. It can include > a synthetic tree that points at such a blob in the pack stream data > if it knows that the resulting pack will be marked with the > .promisor bit, and by doing so, there is no need to update the > client side, no? Yes, that is possible, but I think the client side still needs to be updated, for the same reason as in my implementation (the client should check that the server sent all necessary objects - the blob itself in my case, and the synthetic tree in yours). I think that if we were planning to introduce a synthetic tree, though, it would be better just to include the object itself (as in my implementation). In both solutions, both the protocol and the server need to be updated, but existing clients will still work (minus the checking that would still be desirable with both solutions).
