When running git-blame in a partial clone (--filter=blob:none), it fetches missing blob objects one at a time. This can result in thousands of serial fetch operations, making blame extremely slow, regardless of network latency. For example, in one large repository, blaming a single large file required fetching about 6500 objects. Each fetch requiring a round-trip means this operation would have taken something on the order of an hour to complete. The core issue appears to be in fill_origin_blob(), which is called individually for each blob needed during the blame process. While the blame algorithm does need blob contents to make detailed line-matching decisions, it seems like we don't necessarily need the contents just to determine which blobs we'llexamine. It seems like this could be optimized by batch-fetching the needed objects upfront, rather than fetching them one at a time. This would convert O(n) round-trips into a small number of batch fetches. Reproduction: 1. Create a partial clone with --filter=blob:none 2. Run git blame on a file with significant history 3. Observe serial fetching of objects in the trace output Let me know if you need any additional information to investigate this issue. —burke
