On Thu, Dec 01 2022, Siddharth Asthana wrote: > 2. Make one call to `oid_object_info_extended()` to get the type of the > object. Then, if the object type is either of commit or tag, make a > call to `read_object_file()` to read the contents of the object. > [...] > + buf = read_object_file(&data->oid, &data->type, &data->size); Could you please change this to: buf = repo_read_object_file(the_repository, &data->oid, &data->type, &data->size); I.e. just use the repo_*() variant. We have been trying for a long time to migrate away from these, and new code should use the non-macro variants. I noticed this because I have a local topic to do that migration, which has a semantic conflict with this.
