tree 8a1b6e6a668a609c28937c9dcddb71269412fbfc
parent f66d4e025059b734ba8da40ec059bb0fb8991306
author Jeff King <peff@peff.net> 1582518747 -0500
committer Junio C Hamano <gitster@pobox.com> 1582577753 -0800

pack-bitmap: use object_id when loading on-disk bitmaps

A pack bitmap file contains the index position of the commit for each
bitmap, which we then translate into an object id via
nth_packed_object_sha1(). In preparation for that function going away,
we can switch to the more type-safe nth_packed_object_id().

Note that even though the result ends up in an object_id this does incur
an extra copy of the hash (into our temporary object_id, and then into
the final malloc'd stored_bitmap struct). This shouldn't make any
measurable difference. If it did, we could avoid this copy _and_ the
copy of the rest of the items by allocating the stored_bitmap struct
beforehand and reading directly into it from the bitmap file. Or better
still, if this is a bottleneck, we could introduce an on-disk index to
the bitmap file so we don't have to read every single entry to use just
one of them. So it's not worth worrying about micro-optimizing out this
one hash copy.

Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>