Re: [PATCH v2 9/9] read_loose_refs(): read refs using resolve_ref_recursively()

On 02/10/2017 08:22 PM, Junio C Hamano wrote:

Michael Haggerty [off-list ref] writes:

quoted

[...]

OK, but one thing puzzles me...

quoted

@@ -1390,27 +1390,6 @@ static struct ref_store *main_ref_store;
 static struct hashmap submodule_ref_stores;
 
 /*
- * Return the ref_store instance for the specified submodule (or the
- * main repository if submodule is NULL). If that ref_store hasn't
- * been initialized yet, return NULL.
- */
-static struct ref_store *lookup_ref_store(const char *submodule)
-{
-	struct submodule_hash_entry *entry;
-
-	if (!submodule)
-		return main_ref_store;
-
-	if (!submodule_ref_stores.tablesize)
-		/* It's initialized on demand in register_ref_store(). */
-		return NULL;
-
-	entry = hashmap_get_from_hash(&submodule_ref_stores,
-				      strhash(submodule), submodule);
-	return entry ? entry->refs : NULL;
-}
-
-/*
  * Register the specified ref_store to be the one that should be used
  * for submodule (or the main repository if submodule is NULL). It is
  * a fatal error to call this function twice for the same submodule.

@@ -1451,6 +1430,27 @@ static struct ref_store *ref_store_init(const char *submodule)
 	return refs;
 }
 
+/*
+ * Return the ref_store instance for the specified submodule (or the
+ * main repository if submodule is NULL). If that ref_store hasn't
+ * been initialized yet, return NULL.
+ */
+static struct ref_store *lookup_ref_store(const char *submodule)
+{
+	struct submodule_hash_entry *entry;
+
+	if (!submodule)
+		return main_ref_store;
+
+	if (!submodule_ref_stores.tablesize)
+		/* It's initialized on demand in register_ref_store(). */
+		return NULL;
+
+	entry = hashmap_get_from_hash(&submodule_ref_stores,
+				      strhash(submodule), submodule);
+	return entry ? entry->refs : NULL;
+}
+

I somehow thought that we had an early "reorder the code" step to
avoid hunks like these?  Am I missing some subtle changes made while
moving the function down?

You are quite right; thanks for noticing. I forgot to un-move this
function when re-rolling. These two hunks can be discarded (the function
text is unchanged).

I pushed the fixed commit to branch `submodule-hash` in my fork [1]. If
you'd like me to send it to the mailing list again, please let me know.

Michael

[1] https://github.com/mhagger/git