setup: add an escape hatch for "no more default hash algorithm" change

Partially revert c8aed5e8 (repository: stop setting SHA1 as the
default object hash, 2024-05-07), to keep end-user systems still
broken when we have gap in our test coverage but yet give them an
escape hatch to set the GIT_TEST_DEFAULT_HASH_ALGO environment
variable to "sha1" in order to revert to the previous behaviour, in
case we haven't done a thorough job in fixing the fallout from
c8aed5e8.  After we build confidence, we should remove the escape
hatch support, but we are not there yet after only fixing three
commands (hash-object, apply, and patch-id) in this series.

Due to the way the end-user facing GIT_DEFAULT_HASH environment
variable is used in our test suite, we unfortunately cannot reuse it
for this purpose.

Signed-off-by: Junio C Hamano <gitster@pobox.com>

diff --git a/repository.c b/repository.c
index 15c10015b056051c18ced7b5bc6cd910cc874a52..c62e3298789c6c29dd0f5e9aee0ee877bc71855b 100644 (file)
--- a/repository.c
+++ b/repository.c
@@ -19,6 +19,28 @@
 static struct repository the_repo;
 struct repository *the_repository = &the_repo;
 
+/*
+ * An escape hatch: if we hit a bug in the production code that fails
+ * to set an appropriate hash algorithm (most likely to happen when
+ * running outside a repository), we can tell the user who reported
+ * the crash to set the environment variable to "sha1" (all lowercase)
+ * to revert to the historical behaviour of defaulting to SHA-1.
+ */
+static void set_default_hash_algo(struct repository *repo)
+{
+       const char *hash_name;
+       int algo;
+
+       hash_name = getenv("GIT_TEST_DEFAULT_HASH_ALGO");
+       if (!hash_name)
+               return;
+       algo = hash_algo_by_name(hash_name);
+       if (algo == GIT_HASH_UNKNOWN)
+               return;
+
+       repo_set_hash_algo(repo, algo);
+}
+
 void initialize_repository(struct repository *repo)
 {
        repo->objects = raw_object_store_new();
@@ -26,6 +48,28 @@ void initialize_repository(struct repository *repo)
        repo->parsed_objects = parsed_object_pool_new();
        ALLOC_ARRAY(repo->index, 1);
        index_state_init(repo->index, repo);
+
+       /*
+        * When a command runs inside a repository, it learns what
+        * hash algorithm is in use from the repository, but some
+        * commands are designed to work outside a repository, yet
+        * they want to access the_hash_algo, if only for the length
+        * of the hashed value to see if their input looks like a
+        * plausible hash value.
+        *
+        * We are in the process of identifying such code paths and
+        * giving them an appropriate default individually; any
+        * unconverted code paths that try to access the_hash_algo
+        * will thus fail.  The end-users however have an escape hatch
+        * to set GIT_TEST_DEFAULT_HASH_ALGO environment variable to
+        * "sha1" to get back the old behaviour of defaulting to SHA-1.
+        *
+        * This escape hatch is deliberately kept unadvertised, so
+        * that they see crashes and we can get a report before
+        * telling them about it.
+        */
+       if (repo == the_repository)
+               set_default_hash_algo(repo);
 }
 
 static void expand_base_dir(char **out, const char *in,