[v2,02/25] init-db: set the_repository->hash_algo early on

Commit Message

Han-Wen Nienhuys Aug. 16, 2021, 8:16 p.m. UTC

From: Han-Wen Nienhuys <[e-mail]>

The reftable backend needs to know the hash algorithm for writing the
initialization hash table.

The initial reftable contains a symref HEAD => "main" (or "master"), which is
agnostic to the size of hash value, but this is an exceptional circumstance, and
the reftable library does not cater to this exception. It insists that all
tables in the stack have a consistent format ID for the hash algorithm.

Call set_repo_hash_algo directly after calling validate_hash_algorithm() (which
reads $GIT_DEFAULT_HASH).

Helped-by: Junio C Hamano <[e-mail]>
Signed-off-by: Han-Wen Nienhuys <[e-mail]>
---
 builtin/init-db.c | 21 +++++++++++++++++++++
 1 file changed, 21 insertions(+)

Patch

diff --git a/builtin/init-db.c b/builtin/init-db.c
index 2167796ff2a..c2f03f6018e 100644
--- a/builtin/init-db.c
+++ b/builtin/init-db.c
@@ -425,6 +425,27 @@  int init_db(const char *git_dir, const char *real_git_dir,
 
 	validate_hash_algorithm(&repo_fmt, hash);
 
+	/*
+	 * At this point, the_repository we have in-core does not look
+	 * anything like one that we would see initialized in an already
+	 * working repository after calling setup_git_directory().
+	 *
+	 * Calling repository.c::initialize_the_repository() may have
+	 * prepared the .index .objects and .parsed_objects members, but
+	 * other members like .gitdir, .commondir, etc. have not been
+	 * initialized.
+	 *
+	 * Many API functions assume they are working with the_repository
+	 * that has sensibly been initialized, but because we haven't
+	 * really read from an existing repository, we need to hand-craft
+	 * the necessary members of the structure to get out of this
+	 * chicken-and-egg situation.
+	 *
+	 * For now, we update the hash algorithm member to what the
+	 * validate_hash_algorithm() call decided for us.
+	 */
+	repo_set_hash_algo(the_repository, repo_fmt.hash_algo);
+
 	reinit = create_default_files(template_dir, original_git_dir,
 				      initial_branch, &repo_fmt,
 				      flags & INIT_DB_QUIET);