convert git-remote-annex to not include old pushed refs in new bundle

Rather than requiring the last listed bundle in the manifest include all
refs that are in the remote, build up refs from each bundle listed in
the manifest.

This fixes a bug where pushing first a new branch foo from one clone,
and then pushing a new branch bar from another clone, caused the second
push to lose branch foo. Now the second push will add a new bundle, but
the foo ref in the bundle from the first push will still be used.

Pushing a deletion of a ref now has to delete all bundles and push a new
bundle with only the remaining refs in it.

In a "list for-push", it now has to unbundle all bundles, in order for a
deletion repush to have available all objects. (And a non-deletion push
can also rely on refs/namespaces/mine/ being up-to-date.)

It would have been possible to fix the bug by only making it do that
unbundling in "list for-push", without changing what's stored in the
bundles. But I think I prefer to populate the bundles this way. For one
thing, deleting a pushed ref now really deletes all data relating to it,
rather than leaving it present in old bundles. For another, it's easier
to explain since there is no special case for the last bundle. And, it
will often result in smaller bundles.

Note that further efficiency gains are possible with respect to what
objects are included in an incremental bundle. Two XXX comments
document how to reduce excess objects. It didn't seem worth implementing
those optimisations in this proof of concept code.

Sponsored-by: Brock Spratlen on Patreon
This commit is contained in:
Joey Hess 2024-04-30 13:51:43 -04:00
parent e5cfaf003c
commit fc37243ffe
No known key found for this signature in database
GPG key ID: DB12DB0FF05F8F38
2 changed files with 104 additions and 96 deletions

View file

@ -9,27 +9,30 @@ GITBUNDLE--sha256 is a git bundle.
An ordered list of bundle keys, one per line.
The last bundle in the list provides all refs that are currently stored in
the repository. The bundles before it in the list can incrementally provide
objects, but not refs.
# fetching
1. download GITMANIFEST for the uuid of the special remote
2. download each listed GITBUNDLE object that we don't have
3. `git bundle unpack` each bundle in order
4. `git fetch` from the last bundle listed in the manifest
3. `git fetch` from each new bundle in order
(note that later bundles can update refs from the versions in previous
bundles)
# pushing (incrementally)
1. create git bundle all refs that will be stored in the repository,
This is how pushes are usually done.
1. create git bundle of all refs that are being pushed and have changed,
and objects since the previously pushed refs
2. hash to calculate GITBUNDLE key
3. upload GITBUNDLE object
4. download current manifest
5. append GITBUNDLE key to manifest
# pushing (replacing incrementals with single bundle)
# pushing (full)
Note that this can be used to replace incrementals with a single bundle for
performance. It is also the only way to handle a push that deletes a
previously pushed ref.
1. create git bundle containing all refs stored in the repository, and all
objects