On Mon, Nov 28, 2022 at 10:53:19AM -0800, Jonathan Tan wrote: > > In general, I think partial clones tend to know which filters were used > > to create them, because we save that filter in the config. Would it be > > reasonable before lazy-fetching to say "I am looking for an object of > > type X; would my configured filters have skipped such an object?". > > > > Then not only would you get the behavior you want for commits (which are > > never skipped), but a blob:none clone would not try to lazy-fetch trees > > (but a tree:depth one would lazy-fetch both trees and blobs). > > > > The gotcha I'd worry about is that the config doesn't necessarily match > > how the repository was originally created. There is nothing right now > > saying you cannot partial-clone with one filter, then change the config > > going forward. > > Thanks for weighing in. In the general case, we indeed do know what kind of > object we're fetching, so it does make sense to generalize my idea to trees and > blobs as well. On the other hand, though, besides the issue that the user may > subsequently change the config, the benefits of distinguishing between a blob > and tree are not that great, I think - we would fail fast when, say, a tree > is missing due to object store corruption when trying to check out something > in a blob:none clone, but the same object store corruption probably would > mean that commits are missing too, so even if we just did this for commits, we > would already fail equally as fast. Because of this (and the user being able > to change the config), it makes sense to me to handle only the commit case, at > least for now. > > Having said that, this is not forwards incompatible with restricting lazy fetch > for trees and blobs, so if we see fit later to do that, we can still do it. Yeah, I think your reasoning is sound. Certainly it seems reasonable to start with commit objects, since we know they can never be filtered, and see how that fares in practice. And then think about handling other object types later (or never; this is really just a "we might catch corruption sooner" nice-to-have, and not a correctness problem in normal use). -Peff
