Re: RFC: Design and code of partial clones (now, missing commits and trees OK) (part 3)

(part 3)

Additional overall comments on: https://github.com/jonathantanmy/git/commits/partialclone2

{} WRT the code in is_promised() [1]

[1] https://github.com/jonathantanmy/git/commit/7a9c2d9b6e2fce293817b595dee29a7eede0dddd#diff-5d5d5dc185ef37dc30bb7d9a7ae0c4e8R1960

    {} it looked like it was adding ALL promisor- and promised-objects to
       the "promised" OIDSET, rather than just promised-objects, but I
       could be mistaken.

    {} Is this iterating over ALL promisor-packfiles?

    {} It looked like this was being used by fsck and rev-list.  I have
       concerns about how big this OIDSET will get and how it will scale,
       since if we start with a partial-clone all packfiles will be
       promisor-packfiles.

    {} When iterating thru a tree object, you add everything that it
       references (everything in that folder).  This adds all of the
       child OIDs -- without regard to whether they are new to this
       version of the tree object. (Granted, this is hard to compute.)

       My concern is that this will add too many objects to the OIDSET.
       That is, a new tree object (because of a recent change to something
       in that folder) will also have the OIDs of the other *unchanged*
       files which may be present in an earlier non-provisor packfile
       from an earlier commit.

       I worry that this will grow the OIDSET to essentially include
       everything.  And possibly defeating its own purpose.  I could be
       wrong here, but that's my concern.

{} I'm not opposed to the .promisor file concept, but I have concerns
    that in practice all packfiles after a partial clone will be
    promisor-packfiles and therefore short-cut during fsck, so fsck
    still won't gain anything.

    It would help if there are also non-promisor packfiles present,
    but only for objects referenced by non-promisor packfiles.

    But then I also have to wonder whether we can even support non-promisor
    packfiles after starting with a partial clone -- because of the
    properties of received thin-packs on a non-partial fetch.
    

Thanks, Jeff