[PATCH 00/11] Large blob fixes

These patches make sure we avoid keeping whole blob in memory, at
least in common cases. Blob-only streaming code paths are opened to
accomplish that.

I don't quite like having three different implementations for
checking sha-1 signature (one on git_istream, one on packed_git and
the other one in index-pack) but I failed to see how to unify them.

Making archive-zip work with stream can be difficult. But at least tar
format works. Good enough for me.

Nguyễn Thái Ngọc Duy (11):
  Add more large blob test cases
  Factor out and export large blob writing code to arbitrary file
    handle
  cat-file: use streaming interface to print blobs
  parse_object: special code path for blobs to avoid putting whole
    object in memory
  show: use streaming interface for showing blobs
  index-pack --verify: skip sha-1 collision test
  index-pack: split second pass obj handling into own function
  index-pack: reduce memory usage when the pack has large blobs
  pack-check: do not unpack blobs
  archive: support streaming large files to a tar archive
  fsck: use streaming interface for writing lost-found blobs

 archive-tar.c        |   35 +++++++++++++---
 archive-zip.c        |    9 ++--
 archive.c            |   51 ++++++++++++++++--------
 archive.h            |   11 ++++-
 builtin/cat-file.c   |   22 ++++++++++
 builtin/fsck.c       |    8 +---
 builtin/index-pack.c |  108 +++++++++++++++++++++++++++++++++++++------------
 builtin/log.c        |    9 ++++-
 cache.h              |    5 ++-
 entry.c              |   39 ++++++++++++------
 fast-import.c        |    2 +-
 object.c             |   11 +++++
 pack-check.c         |   21 +++++++++-
 sha1_file.c          |   78 +++++++++++++++++++++++++++++++-----
 t/t1050-large.sh     |   59 +++++++++++++++++++++++++++-
 wrapper.c            |   27 +++++++++++-
 16 files changed, 400 insertions(+), 95 deletions(-)

-- 
1.7.3.1.256.g2539c.dirty