On Wed, May 25, 2016 at 10:49:07PM +0000, Eric Wong wrote: > > Thanks for the analysis. I think this is definitely the problem. After > > fast-import finalizes a packfile (either at the end of the program or > > due to a checkpoint), it never discards its internal mapping of objects > > to pack locations. It _has_ to keep such a mapping before the pack is > > finalized, because git's regular object database doesn't know about it. > > But afterwards, it should be able to rely on the object database. > > Almost; but relying on marks is a problem since that set can contain > mark => object_entry mappings which the normal object database won't > know about. Ah, thanks for finding that. I had a feeling there might be other users of the object_entry structs, but I didn't dig. Given that and your other responses here, I agree that just invalidating the pack_id is probably the most sensible route. -Peff -- To unsubscribe from this list: send the line "unsubscribe git" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html