[PATCH 0/7] speeding up cat-file by reordering object access

[Jeff King <peff@peff.net>]

This series is meant to replace the RFC discussion in:

  https://public-inbox.org/git/20180808231210.242120-1-jonathantanmy@google.com/

and

  https://public-inbox.org/git/20180808155045.GB1607@sigill.intra.peff.net/

The general idea is that accessing objects in packfile order is way
kinder to the delta base cache, and thus way more efficient. See patches
4 and 7 in particular for discussion and numbers.

I'm primarily interested in cat-file, so this series is focused there.
But there may be other callers of for_each_packed_object() who could
benefit. Most of the existing ones just care about getting the oid, so
they're better off as-is. It's possible the call in is_promisor_object()
could benefit, since it calls parse_object() on each entry it visits. I
didn't experiment with it.

  [1/7]: for_each_*_object: store flag definitions in a single location
  [2/7]: for_each_*_object: take flag arguments as enum
  [3/7]: for_each_*_object: give more comprehensive docstrings
  [4/7]: for_each_packed_object: support iterating in pack-order
  [5/7]: t1006: test cat-file --batch-all-objects with duplicates
  [6/7]: cat-file: rename batch_{loose,packed}_object callbacks
  [7/7]: cat-file: support "unordered" output for --batch-all-objects

 Documentation/git-cat-file.txt |  8 ++++
 builtin/cat-file.c             | 70 ++++++++++++++++++++++++++++------
 cache.h                        | 29 +++++++++++---
 commit-graph.c                 |  2 +-
 packfile.c                     | 24 +++++++++---
 packfile.h                     | 23 ++++++-----
 sha1-file.c                    |  3 +-
 t/t1006-cat-file.sh            | 17 ++++++++-
 8 files changed, 139 insertions(+), 37 deletions(-)

-Peff