Moves the point where we filter blob reuse candidates based on algorithm (when the users asks for gzip/zstd, or when the user specifies a v2s2 manifest format and Zstd is not acceptable). Before, we would prioritize and return up to 5 candidates regardless of algorithm, and then the transport would filter unwanted algorithms — in the worst case, getting not a single acceptable algorithm. Now, we first filter out unacceptable candidates, and only afterwards prioritize and return up to 5 candidates to the transport. That should improve the chances of reusing a known blob.
Moves the responsibility for turning compression algorithm names into internal objects from the transports to the BlobInfoCache2 implementation. This will, again, allow using c/image/docker as a ~light-weight registry client without dragging in c/storage/pkg/chunked.
Structures the code to give us an opportunity to add even more logic to the algorithm filtering/matching. E.g. we will want to return a known-zstd:chunked layer where we don’t trust the TOC as a zstd candidate; and we will want to return trusted zstd:chunked candidates where the user asked for just zstd. That will come in the future.
rhatdan
commented
6 months ago
This:
gzip/zstd, or when the user specifies av2s2manifest format and Zstd is not acceptable). Before, we would prioritize and return up to 5 candidates regardless of algorithm, and then the transport would filter unwanted algorithms — in the worst case, getting not a single acceptable algorithm. Now, we first filter out unacceptable candidates, and only afterwards prioritize and return up to 5 candidates to the transport. That should improve the chances of reusing a known blob.BlobInfoCache2implementation. This will, again, allow usingc/image/dockeras a ~light-weight registry client without dragging inc/storage/pkg/chunked.See individual commit messages for details.
Cc: @giuseppe