Skip to content

fix inconsistent uses of the_repo in parse_object()'s call chain

Matheus Tavares git 1 version In review

Get series with b4
b4 am [email protected]

Versions newest first

  1. v1 Latest7 patches 8 repliesIn review

Patches in v1 No applies recorded yet

#Subject and filesSizeRepliesApplied
1/7diff: make diff_populate_filespec() honor its repo argumentdiff.c+1 -10 replies
2/7cache-tree: use given repo's hash_algo at verify_one()cache-tree.c+1 -10 replies
3/7pack-check: use given repo's hash_algo at verify_packfile()pack-check.c+5 -50 replies
4/7streaming: allow open_istream() to handle any repoarchive-tar.c archive-zip.c builtin/index-pack.c +4 more+28 -211 reply
5/7sha1-file: pass git_hash_algo to write_object_file_prepare()sha1-file.c+12 -82 replies
6/7sha1-file: pass git_hash_algo to hash_object_file()apply.c builtin/index-pack.c builtin/replace.c +8 more+36 -250 replies
7/7sha1-file: allow check_object_signature() to handle any repobuiltin/fast-export.c builtin/index-pack.c builtin/mktag.c +4 more+27 -190 replies

Reviews trailers per patch

No Reviewed-by, Acked-by or Tested-by trailers for v1 yet.

Read the v1 thread

Cover letter [PATCH 0/7] fix inconsistent uses of the_repo in parse_object()'s call chain

The motivation for this patchset is another series I'm working on, to make git-grep stop adding submodules' odbs to the alternates list. With that, parse_object() will have to be called with the subdmodules' struct repository. But it seemed that this function doesn't pass on the received repo to some inner calls, which, instead, always use the_repository. This series seeks to fix these inconsistencies.

Note: I also tried to replace some uses of the_hash_algo with the struct git_hash_algo from the received repository, for consitency. (In practice, I'm not sure if this is very useful right now, but maybe it will be relevant for the change to SHA256?) Still, many functions in parse_object()'s call chain call oid_to_hex(), which uses the_hash_algo. Since changing this would require a much bigger operation, I decided to leave it as is, for now.

Note II: Despite receiving a repo through the apply_state struct, apply.c:apply_binary() call functions which uses the_repository internally. Because of that, I used the_hash_algo in this function in patch 6. Should I change it to use apply_state->repo->hash_algo anyway?

travis build: https://travis-ci.org/matheustavares/git/builds/644022000

Matheus Tavares (7):
diff: make diff_populate_filespec() honor its repo argument
cache-tree: use given repo's hash_algo at verify_one()
pack-check: use given repo's hash_algo at verify_packfile()
streaming: allow open_istream() to handle any repo
sha1-file: pass git_hash_algo to write_object_file_prepare()
sha1-file: pass git_hash_algo to hash_object_file()
sha1-file: allow check_object_signature() to handle any repo

 apply.c                  |  6 +++--
 archive-tar.c            |  6 ++---
 archive-zip.c            |  3 ++-
 builtin/fast-export.c    |  3 ++-
 builtin/index-pack.c     | 10 +++++---
 builtin/mktag.c          |  7 +++--
 builtin/pack-objects.c   |  3 ++-
 builtin/replace.c        |  3 ++-
 builtin/unpack-objects.c |  3 ++-
 cache-tree.c             | 11 +++++---
 cache.h                  |  3 ++-
 convert.c                |  2 +-
 diff.c                   |  2 +-
 diffcore-rename.c        |  4 +--
 dir.c                    |  4 +--
 log-tree.c               |  3 ++-
 object-store.h           |  5 ++--
 object.c                 |  5 ++--
 pack-check.c             | 12 ++++-----
 sha1-file.c              | 55 ++++++++++++++++++++++------------------
 streaming.c              | 28 ++++++++++----------
 streaming.h              |  4 ++-
 22 files changed, 106 insertions(+), 76 deletions(-)
Signature · 1 line
2.25.0