lol this worked first try wtf

9 files changed, 184 insertions, 139 deletions

diff --git a/src/cache.c b/src/cache.c
index e7f333577..6eb18dae5 100644
--- a/src/cache.c
+++ b/src/cache.c
@@ -11,100 +11,147 @@
 #include "thread-utils.h"
 #include "util.h"
 #include "cache.h"
+#include "odb.h"
+#include "object.h"
 #include "git2/oid.h"
 
-int git_cache_init(git_cache *cache, size_t size, git_cached_obj_freeptr free_ptr)
-{
-	if (size < 8)
-		size = 8;
-	size = git__size_t_powerof2(size);
-
-	cache->size_mask = size - 1;
-	cache->lru_count = 0;
-	cache->free_obj = free_ptr;
+GIT__USE_OIDMAP
 
+int git_cache_init(git_cache *cache)
+{
+	cache->map = git_oidmap_alloc();
 	git_mutex_init(&cache->lock);
-
-	cache->nodes = git__malloc(size * sizeof(git_cached_obj *));
-	GITERR_CHECK_ALLOC(cache->nodes);
-
-	memset(cache->nodes, 0x0, size * sizeof(git_cached_obj *));
 	return 0;
 }
 
 void git_cache_free(git_cache *cache)
 {
-	size_t i;
-
-	for (i = 0; i < (cache->size_mask + 1); ++i) {
-		if (cache->nodes[i] != NULL)
-			git_cached_obj_decref(cache->nodes[i], cache->free_obj);
-	}
-
+	git_oidmap_free(cache->map);
 	git_mutex_free(&cache->lock);
-	git__free(cache->nodes);
 }
 
-void *git_cache_get(git_cache *cache, const git_oid *oid)
+static bool cache_should_store(git_cached_obj *entry)
 {
-	uint32_t hash;
-	git_cached_obj *node = NULL, *result = NULL;
+	return true;
+}
 
-	memcpy(&hash, oid->id, sizeof(hash));
+static void *cache_get(git_cache *cache, const git_oid *oid, unsigned int flags)
+{
+	khiter_t pos;
+	git_cached_obj *entry = NULL;
 
-	if (git_mutex_lock(&cache->lock)) {
-		giterr_set(GITERR_THREAD, "unable to lock cache mutex");
+	if (git_mutex_lock(&cache->lock) < 0)
 		return NULL;
-	}
 
-	{
-		node = cache->nodes[hash & cache->size_mask];
+	pos = kh_get(oid, cache->map, oid);
+	if (pos != kh_end(cache->map)) {
+		entry = kh_val(cache->map, pos);
 
-		if (node != NULL && git_oid_cmp(&node->oid, oid) == 0) {
-			git_cached_obj_incref(node);
-			result = node;
+		if (flags && entry->flags != flags) {
+			entry = NULL;
+		} else {
+			git_cached_obj_incref(entry);
 		}
 	}
+
 	git_mutex_unlock(&cache->lock);
 
-	return result;
+	return entry;
 }
 
-void *git_cache_try_store(git_cache *cache, void *_entry)
+static void *cache_store(git_cache *cache, git_cached_obj *entry)
 {
-	git_cached_obj *entry = _entry;
-	uint32_t hash;
+	khiter_t pos;
 
-	memcpy(&hash, &entry->oid, sizeof(uint32_t));
+	git_cached_obj_incref(entry);
 
-	if (git_mutex_lock(&cache->lock)) {
-		giterr_set(GITERR_THREAD, "unable to lock cache mutex");
-		return NULL;
-	}
+	if (!cache_should_store(entry))
+		return entry;
+
+	if (git_mutex_lock(&cache->lock) < 0)
+		return entry;
 
-	{
-		git_cached_obj *node = cache->nodes[hash & cache->size_mask];
+	pos = kh_get(oid, cache->map, &entry->oid);
 
-		/* increase the refcount on this object, because
-		 * the cache now owns it */
-		git_cached_obj_incref(entry);
+	/* not found */
+	if (pos == kh_end(cache->map)) {
+		int rval;
 
-		if (node == NULL) {
-			cache->nodes[hash & cache->size_mask] = entry;
-		} else if (git_oid_cmp(&node->oid, &entry->oid) == 0) {
-			git_cached_obj_decref(entry, cache->free_obj);
-			entry = node;
+		pos = kh_put(oid, cache->map, &entry->oid, &rval);
+		if (rval >= 0) {
+			kh_key(cache->map, pos) = &entry->oid;
+			kh_val(cache->map, pos) = entry;
+			git_cached_obj_incref(entry);
+		}
+	}
+	/* found */
+	else {
+		git_cached_obj *stored_entry = kh_val(cache->map, pos);
+
+		if (stored_entry->flags == entry->flags) {
+			git_cached_obj_decref(entry);
+			git_cached_obj_incref(stored_entry);
+			entry = stored_entry;
+		} else if (stored_entry->flags == GIT_CACHE_STORE_RAW &&
+			entry->flags == GIT_CACHE_STORE_PARSED) {
+			git_cached_obj_decref(stored_entry);
+			git_cached_obj_incref(entry);
+
+			kh_key(cache->map, pos) = &entry->oid;
+			kh_val(cache->map, pos) = entry;
 		} else {
-			git_cached_obj_decref(node, cache->free_obj);
-			cache->nodes[hash & cache->size_mask] = entry;
+			/* NO OP */
 		}
-
-		/* increase the refcount again, because we are
-		 * returning it to the user */
-		git_cached_obj_incref(entry);
-
 	}
-	git_mutex_unlock(&cache->lock);
 
+	git_mutex_unlock(&cache->lock);
 	return entry;
 }
+
+void *git_cache_store_raw(git_cache *cache, git_odb_object *entry)
+{
+	entry->cached.flags = GIT_CACHE_STORE_RAW;
+	return cache_store(cache, (git_cached_obj *)entry);
+}
+
+void *git_cache_store_parsed(git_cache *cache, git_object *entry)
+{
+	entry->cached.flags = GIT_CACHE_STORE_PARSED;
+	return cache_store(cache, (git_cached_obj *)entry);
+}
+
+git_odb_object *git_cache_get_raw(git_cache *cache, const git_oid *oid)
+{
+	return cache_get(cache, oid, GIT_CACHE_STORE_RAW);
+}
+
+git_object *git_cache_get_parsed(git_cache *cache, const git_oid *oid)
+{
+	return cache_get(cache, oid, GIT_CACHE_STORE_PARSED);
+}
+
+void *git_cache_get_any(git_cache *cache, const git_oid *oid)
+{
+	return cache_get(cache, oid, GIT_CACHE_STORE_ANY);
+}
+
+void git_cached_obj_decref(void *_obj)
+{
+	git_cached_obj *obj = _obj;
+
+	if (git_atomic_dec(&obj->refcount) == 0) {
+		switch (obj->flags) {
+		case GIT_CACHE_STORE_RAW:
+			git_odb_object__free(_obj);
+			break;
+
+		case GIT_CACHE_STORE_PARSED:
+			git_object__free(_obj);
+			break;
+
+		default:
+			git__free(_obj);
+			break;
+		}
+	}
+}
diff --git a/src/cache.h b/src/cache.h
index 7034ec268..f30af9c3e 100644
--- a/src/cache.h
+++ b/src/cache.h
@@ -12,30 +12,35 @@
 #include "git2/odb.h"
 
 #include "thread-utils.h"
+#include "oidmap.h"
 
-#define GIT_DEFAULT_CACHE_SIZE 128
-
-typedef void (*git_cached_obj_freeptr)(void *);
+enum {
+	GIT_CACHE_STORE_ANY = 0,
+	GIT_CACHE_STORE_RAW = 1,
+	GIT_CACHE_STORE_PARSED = 2
+};
 
 typedef struct {
 	git_oid oid;
 	git_atomic refcount;
+	uint32_t flags;
 } git_cached_obj;
 
 typedef struct {
-	git_cached_obj **nodes;
+	git_oidmap *map;
 	git_mutex lock;
-
 	unsigned int lru_count;
-	size_t size_mask;
-	git_cached_obj_freeptr free_obj;
 } git_cache;
 
-int git_cache_init(git_cache *cache, size_t size, git_cached_obj_freeptr free_ptr);
+int git_cache_init(git_cache *cache);
 void git_cache_free(git_cache *cache);
 
-void *git_cache_try_store(git_cache *cache, void *entry);
-void *git_cache_get(git_cache *cache, const git_oid *oid);
+void *git_cache_store_raw(git_cache *cache, git_odb_object *entry);
+void *git_cache_store_parsed(git_cache *cache, git_object *entry);
+
+git_odb_object *git_cache_get_raw(git_cache *cache, const git_oid *oid);
+git_object *git_cache_get_parsed(git_cache *cache, const git_oid *oid);
+void *git_cache_get_any(git_cache *cache, const git_oid *oid);
 
 GIT_INLINE(void) git_cached_obj_incref(void *_obj)
 {
@@ -43,12 +48,6 @@ GIT_INLINE(void) git_cached_obj_incref(void *_obj)
 	git_atomic_inc(&obj->refcount);
 }
 
-GIT_INLINE(void) git_cached_obj_decref(void *_obj, git_cached_obj_freeptr free_obj)
-{
-	git_cached_obj *obj = _obj;
-
-	if (git_atomic_dec(&obj->refcount) == 0)
-		free_obj(obj);
-}
+void git_cached_obj_decref(void *_obj);
 
 #endif
diff --git a/src/object.c b/src/object.c
index 80fe51152..5542ebc8e 100644
--- a/src/object.c
+++ b/src/object.c
@@ -121,12 +121,13 @@ int git_object__from_odb_object(
 		break;
 	}
 
-	if (error < 0)
+	if (error < 0) {
 		git_object__free(object);
-	else
-		*object_out = git_cache_try_store(&repo->objects, object);
+		return error;
+	}
 
-	return error;
+	*object_out = git_cache_store_parsed(&repo->objects, object);
+	return 0;
 }
 
 int git_object_lookup_prefix(
@@ -154,27 +155,38 @@ int git_object_lookup_prefix(
 		len = GIT_OID_HEXSZ;
 
 	if (len == GIT_OID_HEXSZ) {
+		git_cached_obj *cached = NULL;
+
 		/* We want to match the full id : we can first look up in the cache,
 		 * since there is no need to check for non ambiguousity
 		 */
-		object = git_cache_get(&repo->objects, id);
-		if (object != NULL) {
-			if (type != GIT_OBJ_ANY && type != object->type) {
-				git_object_free(object);
-				giterr_set(GITERR_INVALID, "The requested type does not match the type in ODB");
-				return GIT_ENOTFOUND;
+		cached = git_cache_get_any(&repo->objects, id);
+		if (cached != NULL) {
+			if (cached->flags == GIT_CACHE_STORE_PARSED) {
+				object = (git_object *)cached;
+
+				if (type != GIT_OBJ_ANY && type != object->type) {
+					git_object_free(object);
+					giterr_set(GITERR_INVALID,
+						"The requested type does not match the type in ODB");
+					return GIT_ENOTFOUND;
+				}
+
+				*object_out = object;
+				return 0;
+			} else if (cached->flags == GIT_CACHE_STORE_RAW) {
+				odb_obj = (git_odb_object *)cached;
+			} else {
+				assert(!"Wrong caching type in the global object cache");
 			}
-
-			*object_out = object;
-			return 0;
+		} else {
+			/* Object was not found in the cache, let's explore the backends.
+			 * We could just use git_odb_read_unique_short_oid,
+			 * it is the same cost for packed and loose object backends,
+			 * but it may be much more costly for sqlite and hiredis.
+			 */
+			error = git_odb_read(&odb_obj, odb, id);
 		}
-
-		/* Object was not found in the cache, let's explore the backends.
-		 * We could just use git_odb_read_unique_short_oid,
-		 * it is the same cost for packed and loose object backends,
-		 * but it may be much more costly for sqlite and hiredis.
-		 */
-		error = git_odb_read(&odb_obj, odb, id);
 	} else {
 		git_oid short_oid;
 
@@ -245,7 +257,7 @@ void git_object_free(git_object *object)
 	if (object == NULL)
 		return;
 
-	git_cached_obj_decref((git_cached_obj *)object, git_object__free);
+	git_cached_obj_decref(object);
 }
 
 const git_oid *git_object_id(const git_object *obj)
diff --git a/src/odb.c b/src/odb.c
index d6ea8c29a..821fbd70c 100644
--- a/src/odb.c
+++ b/src/odb.c
@@ -14,6 +14,7 @@
 #include "odb.h"
 #include "delta-apply.h"
 #include "filter.h"
+#include "repository.h"
 
 #include "git2/odb_backend.h"
 #include "git2/oid.h"
@@ -34,7 +35,15 @@ typedef struct
 	ino_t disk_inode;
 } backend_internal;
 
-size_t git_odb__cache_size = GIT_DEFAULT_CACHE_SIZE;
+static git_cache *odb_cache(git_odb *odb)
+{
+	if (odb->rc.owner != NULL) {
+		git_repository *owner = odb->rc.owner;
+		return &owner->objects;
+	}
+
+	return &odb->own_cache;
+}
 
 static int load_alternates(git_odb *odb, const char *objects_dir, int alternate_depth);
 
@@ -83,10 +92,8 @@ static git_odb_object *new_odb_object(const git_oid *oid, git_rawobj *source)
 	return object;
 }
 
-static void free_odb_object(void *o)
+void git_odb_object__free(git_odb_object *object)
 {
-	git_odb_object *object = (git_odb_object *)o;
-
 	if (object != NULL) {
 		git__free(object->raw.data);
 		git__free(object);
@@ -118,7 +125,7 @@ void git_odb_object_free(git_odb_object *object)
 	if (object == NULL)
 		return;
 
-	git_cached_obj_decref((git_cached_obj *)object, &free_odb_object);
+	git_cached_obj_decref(object);
 }
 
 int git_odb__hashfd(git_oid *out, git_file fd, size_t size, git_otype type)
@@ -355,9 +362,8 @@ int git_odb_new(git_odb **out)
 	git_odb *db = git__calloc(1, sizeof(*db));
 	GITERR_CHECK_ALLOC(db);
 
-	if (git_cache_init(&db->cache, git_odb__cache_size, &free_odb_object) < 0 ||
-		git_vector_init(&db->backends, 4, backend_sort_cmp) < 0)
-	{
+	if (git_cache_init(&db->own_cache) < 0 ||
+		git_vector_init(&db->backends, 4, backend_sort_cmp) < 0) {
 		git__free(db);
 		return -1;
 	}
@@ -559,7 +565,7 @@ static void odb_free(git_odb *db)
 	}
 
 	git_vector_free(&db->backends);
-	git_cache_free(&db->cache);
+	git_cache_free(&db->own_cache);
 	git__free(db);
 }
 
@@ -580,7 +586,7 @@ int git_odb_exists(git_odb *db, const git_oid *id)
 
 	assert(db && id);
 
-	if ((object = git_cache_get(&db->cache, id)) != NULL) {
+	if ((object = git_cache_get_raw(odb_cache(db), id)) != NULL) {
 		git_odb_object_free(object);
 		return (int)true;
 	}
@@ -630,7 +636,7 @@ int git_odb__read_header_or_object(
 
 	assert(db && id && out && len_p && type_p);
 
-	if ((object = git_cache_get(&db->cache, id)) != NULL) {
+	if ((object = git_cache_get_raw(odb_cache(db), id)) != NULL) {
 		*len_p = object->raw.len;
 		*type_p = object->raw.type;
 		*out = object;
@@ -678,7 +684,7 @@ int git_odb_read(git_odb_object **out, git_odb *db, const git_oid *id)
 		return GIT_ENOTFOUND;
 	}
 
-	*out = git_cache_get(&db->cache, id);
+	*out = git_cache_get_raw(odb_cache(db), id);
 	if (*out != NULL)
 		return 0;
 
@@ -704,7 +710,7 @@ attempt_lookup:
 	if (error && error != GIT_PASSTHROUGH)
 		return error;
 
-	*out = git_cache_try_store(&db->cache, new_odb_object(id, &raw));
+	*out = git_cache_store_raw(odb_cache(db), new_odb_object(id, &raw));
 	return 0;
 }
 
@@ -727,7 +733,7 @@ int git_odb_read_prefix(
 		len = GIT_OID_HEXSZ;
 
 	if (len == GIT_OID_HEXSZ) {
-		*out = git_cache_get(&db->cache, short_id);
+		*out = git_cache_get_raw(odb_cache(db), short_id);
 		if (*out != NULL)
 			return 0;
 	}
@@ -768,7 +774,7 @@ attempt_lookup:
 	if (!found)
 		return git_odb__error_notfound("no match for prefix", short_id);
 
-	*out = git_cache_try_store(&db->cache, new_odb_object(&found_full_oid, &raw));
+	*out = git_cache_store_raw(odb_cache(db), new_odb_object(&found_full_oid, &raw));
 	return 0;
 }
 
diff --git a/src/odb.h b/src/odb.h
index 7c018cc50..9abf594e8 100644
--- a/src/odb.h
+++ b/src/odb.h
@@ -36,9 +36,11 @@ struct git_odb_object {
 struct git_odb {
 	git_refcount rc;
 	git_vector backends;
-	git_cache cache;
+	git_cache own_cache;
 };
 
+void git_odb_object__free(git_odb_object *object);
+
 /*
  * Hash a git_rawobj internally.
  * The `git_rawobj` is supposed to be previously initialized
diff --git a/src/oidmap.h b/src/oidmap.h
index 40274cd19..dfa951af3 100644
--- a/src/oidmap.h
+++ b/src/oidmap.h
@@ -21,10 +21,8 @@ typedef khash_t(oid) git_oidmap;
 
 GIT_INLINE(khint_t) hash_git_oid(const git_oid *oid)
 {
-	int i;
-	khint_t h = 0;
-	for (i = 0; i < 20; ++i)
-		h = (h << 5) - h + oid->id[i];
+	khint_t h;
+	memcpy(&h, oid, sizeof(khint_t));
 	return h;
 }
 
diff --git a/src/repository.c b/src/repository.c
index a16f69da4..cda75b85f 100644
--- a/src/repository.c
+++ b/src/repository.c
@@ -119,7 +119,7 @@ static git_repository *repository_alloc(void)
 
 	memset(repo, 0x0, sizeof(git_repository));
 
-	if (git_cache_init(&repo->objects, GIT_DEFAULT_CACHE_SIZE, &git_object__free) < 0) {
+	if (git_cache_init(&repo->objects) < 0) {
 		git__free(repo);
 		return NULL;
 	}
diff --git a/src/util.c b/src/util.c
index 8e83d298e..c4a8c786d 100644
--- a/src/util.c
+++ b/src/util.c
@@ -93,14 +93,6 @@ int git_libgit2_opts(int key, ...)
 		if ((error = config_level_to_futils_dir(va_arg(ap, int))) >= 0)
 			error = git_futils_dirs_set(error, va_arg(ap, const char *));
 		break;
-
-	case GIT_OPT_GET_ODB_CACHE_SIZE:
-		*(va_arg(ap, size_t *)) = git_odb__cache_size;
-		break;
-
-	case GIT_OPT_SET_ODB_CACHE_SIZE:
-		git_odb__cache_size = va_arg(ap, size_t);
-		break;
 	}
 
 	va_end(ap);
diff --git a/tests-clar/core/opts.c b/tests-clar/core/opts.c
index 907339d51..3173c648b 100644
--- a/tests-clar/core/opts.c
+++ b/tests-clar/core/opts.c
@@ -16,15 +16,4 @@ void test_core_opts__readwrite(void)
 	git_libgit2_opts(GIT_OPT_GET_MWINDOW_SIZE, &new_val);
 
 	cl_assert(new_val == old_val);
-
-	git_libgit2_opts(GIT_OPT_GET_ODB_CACHE_SIZE, &old_val);
-
-	cl_assert(old_val == GIT_DEFAULT_CACHE_SIZE);
-
-	git_libgit2_opts(GIT_OPT_SET_ODB_CACHE_SIZE, (size_t)GIT_DEFAULT_CACHE_SIZE*2);
-	git_libgit2_opts(GIT_OPT_GET_ODB_CACHE_SIZE, &new_val);
-
-	cl_assert(new_val == (GIT_DEFAULT_CACHE_SIZE*2));
-
-	git_libgit2_opts(GIT_OPT_GET_ODB_CACHE_SIZE, &old_val);
 }