rugged 0.19.0 → 0.28.0

data/vendor/libgit2/src/pack.c

@@ -5,10 +5,10 @@
  * a Linking Exception. For full terms see the included COPYING file.
  */
 
-#include "common.h"
-#include "odb.h"
 #include "pack.h"
-#include "delta-apply.h"
+
+#include "odb.h"
+#include "delta.h"
 #include "sha1_lookup.h"
 #include "mwindow.h"
 #include "fileops.h"
@@ -18,13 +18,13 @@
 
 static int packfile_open(struct git_pack_file *p);
 static git_off_t nth_packed_object_offset(const struct git_pack_file *p, uint32_t n);
-int packfile_unpack_compressed(
+static int packfile_unpack_compressed(
 		git_rawobj *obj,
 		struct git_pack_file *p,
 		git_mwindow **w_curs,
 		git_off_t *curpos,
 		size_t size,
-		git_otype type);
+		git_object_t type);
 
 /* Can find the offset of an object given
  * a prefix of an identifier.
@@ -42,7 +42,7 @@ static int pack_entry_find_offset(
 
 static int packfile_error(const char *message)
 {
-	giterr_set(GITERR_ODB, "Invalid pack file - %s", message);
+	git_error_set(GIT_ERROR_ODB, "invalid pack file - %s", message);
 	return -1;
 }
 
@@ -56,6 +56,7 @@ static git_pack_cache_entry *new_cache_object(git_rawobj *source)
 	if (!e)
 		return NULL;
 
+	git_atomic_inc(&e->refcount);
 	memcpy(&e->raw, source, sizeof(git_rawobj));
 
 	return e;
@@ -74,32 +75,27 @@ static void free_cache_object(void *o)
 
 static void cache_free(git_pack_cache *cache)
 {
-	khiter_t k;
+	git_pack_cache_entry *entry;
 
 	if (cache->entries) {
-		for (k = kh_begin(cache->entries); k != kh_end(cache->entries); k++) {
-			if (kh_exist(cache->entries, k))
-				free_cache_object(kh_value(cache->entries, k));
-		}
+		git_offmap_foreach_value(cache->entries, entry, {
+			free_cache_object(entry);
+		});
 
 		git_offmap_free(cache->entries);
-		git_mutex_free(&cache->lock);
+		cache->entries = NULL;
 	}
-
-	memset(cache, 0, sizeof(*cache));
 }
 
 static int cache_init(git_pack_cache *cache)
 {
-	memset(cache, 0, sizeof(*cache));
-
 	cache->entries = git_offmap_alloc();
-	GITERR_CHECK_ALLOC(cache->entries);
+	GIT_ERROR_CHECK_ALLOC(cache->entries);
 
 	cache->memory_limit = GIT_PACK_CACHE_MEMORY_LIMIT;
 
 	if (git_mutex_init(&cache->lock)) {
-		giterr_set(GITERR_OS, "Failed to initialize pack cache mutex");
+		git_error_set(GIT_ERROR_OS, "failed to initialize pack cache mutex");
 
 		git__free(cache->entries);
 		cache->entries = NULL;
@@ -112,15 +108,15 @@ static int cache_init(git_pack_cache *cache)
 
 static git_pack_cache_entry *cache_get(git_pack_cache *cache, git_off_t offset)
 {
-	khiter_t k;
 	git_pack_cache_entry *entry = NULL;
+	size_t k;
 
 	if (git_mutex_lock(&cache->lock) < 0)
 		return NULL;
 
-	k = kh_get(off, cache->entries, offset);
-	if (k != kh_end(cache->entries)) { /* found it */
-		entry = kh_value(cache->entries, k);
+	k = git_offmap_lookup_index(cache->entries, offset);
+	if (git_offmap_valid_index(cache->entries, k)) { /* found it */
+		entry = git_offmap_value_at(cache->entries, k);
 		git_atomic_inc(&entry->refcount);
 		entry->last_usage = cache->use_ctr++;
 	}
@@ -132,28 +128,27 @@ static git_pack_cache_entry *cache_get(git_pack_cache *cache, git_off_t offset)
 /* Run with the cache lock held */
 static void free_lowest_entry(git_pack_cache *cache)
 {
+	git_off_t offset;
 	git_pack_cache_entry *entry;
-	khiter_t k;
-
-	for (k = kh_begin(cache->entries); k != kh_end(cache->entries); k++) {
-		if (!kh_exist(cache->entries, k))
-			continue;
-
-		entry = kh_value(cache->entries, k);
 
+	git_offmap_foreach(cache->entries, offset, entry, {
 		if (entry && entry->refcount.val == 0) {
 			cache->memory_used -= entry->raw.len;
-			kh_del(off, cache->entries, k);
+			git_offmap_delete(cache->entries, offset);
 			free_cache_object(entry);
 		}
-	}
+	});
 }
 
-static int cache_add(git_pack_cache *cache, git_rawobj *base, git_off_t offset)
+static int cache_add(
+		git_pack_cache_entry **cached_out,
+		git_pack_cache *cache,
+		git_rawobj *base,
+		git_off_t offset)
 {
 	git_pack_cache_entry *entry;
 	int error, exists = 0;
-	khiter_t k;
+	size_t k;
 
 	if (base->len > GIT_PACK_CACHE_SIZE_LIMIT)
 		return -1;
@@ -161,19 +156,22 @@ static int cache_add(git_pack_cache *cache, git_rawobj *base, git_off_t offset)
 	entry = new_cache_object(base);
 	if (entry) {
 		if (git_mutex_lock(&cache->lock) < 0) {
-			giterr_set(GITERR_OS, "failed to lock cache");
+			git_error_set(GIT_ERROR_OS, "failed to lock cache");
+			git__free(entry);
 			return -1;
 		}
 		/* Add it to the cache if nobody else has */
-		exists = kh_get(off, cache->entries, offset) != kh_end(cache->entries);
+		exists = git_offmap_exists(cache->entries, offset);
 		if (!exists) {
 			while (cache->memory_used + base->len > cache->memory_limit)
 				free_lowest_entry(cache);
 
-			k = kh_put(off, cache->entries, offset, &error);
+			k = git_offmap_put(cache->entries, offset, &error);
 			assert(error != 0);
-			kh_value(cache->entries, k) = entry;
+			git_offmap_set_value_at(cache->entries, k, entry);
 			cache->memory_used += entry->raw.len;
+
+			*cached_out = entry;
 		}
 		git_mutex_unlock(&cache->lock);
 		/* Somebody beat us to adding it into the cache */
@@ -219,7 +217,7 @@ static int pack_index_check(const char *path, struct git_pack_file *p)
 
 	if (p_fstat(fd, &st) < 0) {
 		p_close(fd);
-		giterr_set(GITERR_OS, "Unable to stat pack index '%s'", path);
+		git_error_set(GIT_ERROR_OS, "unable to stat pack index '%s'", path);
 		return -1;
 	}
 
@@ -228,7 +226,7 @@ static int pack_index_check(const char *path, struct git_pack_file *p)
 		(idx_size = (size_t)st.st_size) < 4 * 256 + 20 + 20)
 	{
 		p_close(fd);
-		giterr_set(GITERR_ODB, "Invalid pack index '%s'", path);
+		git_error_set(GIT_ERROR_ODB, "invalid pack index '%s'", path);
 		return -1;
 	}
 
@@ -312,9 +310,9 @@ static int pack_index_check(const char *path, struct git_pack_file *p)
 
 static int pack_index_open(struct git_pack_file *p)
 {
-	char *idx_name;
 	int error = 0;
-	size_t name_len, base_len;
+	size_t name_len;
+	git_buf idx_name;
 
 	if (p->index_version > -1)
 		return 0;
@@ -322,20 +320,25 @@ static int pack_index_open(struct git_pack_file *p)
 	name_len = strlen(p->pack_name);
 	assert(name_len > strlen(".pack")); /* checked by git_pack_file alloc */
 
-	if ((idx_name = git__malloc(name_len)) == NULL)
+	if (git_buf_init(&idx_name, name_len) < 0)
 		return -1;
 
-	base_len = name_len - strlen(".pack");
-	memcpy(idx_name, p->pack_name, base_len);
-	memcpy(idx_name + base_len, ".idx", sizeof(".idx"));
+	git_buf_put(&idx_name, p->pack_name, name_len - strlen(".pack"));
+	git_buf_puts(&idx_name, ".idx");
+	if (git_buf_oom(&idx_name)) {
+		git_buf_dispose(&idx_name);
+		return -1;
+	}
 
-	if ((error = git_mutex_lock(&p->lock)) < 0)
+	if ((error = git_mutex_lock(&p->lock)) < 0) {
+		git_buf_dispose(&idx_name);
 		return error;
+	}
 
 	if (p->index_version == -1)
-		error = pack_index_check(idx_name, p);
+		error = pack_index_check(idx_name.ptr, p);
 
-	git__free(idx_name);
+	git_buf_dispose(&idx_name);
 
 	git_mutex_unlock(&p->lock);
 
@@ -355,17 +358,54 @@ static unsigned char *pack_window_open(
 	 * pointless to ask for an offset into the middle of that
 	 * hash, and the pack_window_contains function above wouldn't match
 	 * don't allow an offset too close to the end of the file.
+	 *
+	 * Don't allow a negative offset, as that means we've wrapped
+	 * around.
 	 */
 	if (offset > (p->mwf.size - 20))
 		return NULL;
+	if (offset < 0)
+		return NULL;
 
 	return git_mwindow_open(&p->mwf, w_cursor, offset, 20, left);
  }
 
+/*
+ * The per-object header is a pretty dense thing, which is
+ *  - first byte: low four bits are "size",
+ *    then three bits of "type",
+ *    with the high bit being "size continues".
+ *  - each byte afterwards: low seven bits are size continuation,
+ *    with the high bit being "size continues"
+ */
+size_t git_packfile__object_header(unsigned char *hdr, size_t size, git_object_t type)
+{
+	unsigned char *hdr_base;
+	unsigned char c;
+
+	assert(type >= GIT_OBJECT_COMMIT && type <= GIT_OBJECT_REF_DELTA);
+
+	/* TODO: add support for chunked objects; see git.git 6c0d19b1 */
+
+	c = (unsigned char)((type << 4) | (size & 15));
+	size >>= 4;
+	hdr_base = hdr;
+
+	while (size) {
+		*hdr++ = c | 0x80;
+		c = size & 0x7f;
+		size >>= 7;
+	}
+	*hdr++ = c;
+
+	return (hdr - hdr_base);
+}
+
+
 static int packfile_unpack_header1(
 		unsigned long *usedp,
 		size_t *sizep,
-		git_otype *type,
+		git_object_t *type,
 		const unsigned char *buf,
 		unsigned long len)
 {
@@ -378,11 +418,14 @@ static int packfile_unpack_header1(
 	size = c & 15;
 	shift = 4;
 	while (c & 0x80) {
-		if (len <= used)
+		if (len <= used) {
+			git_error_set(GIT_ERROR_ODB, "buffer too small");
 			return GIT_EBUFS;
+		}
 
 		if (bitsizeof(long) <= shift) {
 			*usedp = 0;
+			git_error_set(GIT_ERROR_ODB, "packfile corrupted");
 			return -1;
 		}
 
@@ -398,7 +441,7 @@ static int packfile_unpack_header1(
 
 int git_packfile_unpack_header(
 		size_t *size_p,
-		git_otype *type_p,
+		git_object_t *type_p,
 		git_mwindow_file *mwf,
 		git_mwindow **w_curs,
 		git_off_t *curpos)
@@ -432,45 +475,44 @@ int git_packfile_unpack_header(
 
 int git_packfile_resolve_header(
 		size_t *size_p,
-		git_otype *type_p,
+		git_object_t *type_p,
 		struct git_pack_file *p,
 		git_off_t offset)
 {
 	git_mwindow *w_curs = NULL;
 	git_off_t curpos = offset;
 	size_t size;
-	git_otype type;
+	git_object_t type;
 	git_off_t base_offset;
 	int error;
 
 	error = git_packfile_unpack_header(&size, &type, &p->mwf, &w_curs, &curpos);
-	git_mwindow_close(&w_curs);
 	if (error < 0)
 		return error;
 
-	if (type == GIT_OBJ_OFS_DELTA || type == GIT_OBJ_REF_DELTA) {
+	if (type == GIT_OBJECT_OFS_DELTA || type == GIT_OBJECT_REF_DELTA) {
 		size_t base_size;
-		git_rawobj delta;
+		git_packfile_stream stream;
+
 		base_offset = get_delta_base(p, &w_curs, &curpos, type, offset);
 		git_mwindow_close(&w_curs);
-		error = packfile_unpack_compressed(&delta, p, &w_curs, &curpos, size, type);
-		git_mwindow_close(&w_curs);
-		if (error < 0)
+		if ((error = git_packfile_stream_open(&stream, p, curpos)) < 0)
 			return error;
-		error = git__delta_read_header(delta.data, delta.len, &base_size, size_p);
-		git__free(delta.data);
+		error = git_delta_read_header_fromstream(&base_size, size_p, &stream);
+		git_packfile_stream_dispose(&stream);
 		if (error < 0)
 			return error;
-	} else
+	} else {
 		*size_p = size;
+		base_offset = 0;
+	}
 
-	while (type == GIT_OBJ_OFS_DELTA || type == GIT_OBJ_REF_DELTA) {
+	while (type == GIT_OBJECT_OFS_DELTA || type == GIT_OBJECT_REF_DELTA) {
 		curpos = base_offset;
 		error = git_packfile_unpack_header(&size, &type, &p->mwf, &w_curs, &curpos);
-		git_mwindow_close(&w_curs);
 		if (error < 0)
 			return error;
-		if (type != GIT_OBJ_OFS_DELTA && type != GIT_OBJ_REF_DELTA)
+		if (type != GIT_OBJECT_OFS_DELTA && type != GIT_OBJECT_REF_DELTA)
 			break;
 		base_offset = get_delta_base(p, &w_curs, &curpos, type, base_offset);
 		git_mwindow_close(&w_curs);
@@ -480,72 +522,101 @@ int git_packfile_resolve_header(
 	return error;
 }
 
-static int packfile_unpack_delta(
-		git_rawobj *obj,
-		struct git_pack_file *p,
-		git_mwindow **w_curs,
-		git_off_t *curpos,
-		size_t delta_size,
-		git_otype delta_type,
-		git_off_t obj_offset)
+#define SMALL_STACK_SIZE 64
+
+/**
+ * Generate the chain of dependencies which we need to get to the
+ * object at `off`. `chain` is used a stack, popping gives the right
+ * order to apply deltas on. If an object is found in the pack's base
+ * cache, we stop calculating there.
+ */
+static int pack_dependency_chain(git_dependency_chain *chain_out,
+				 git_pack_cache_entry **cached_out, git_off_t *cached_off,
+				 struct pack_chain_elem *small_stack, size_t *stack_sz,
+				 struct git_pack_file *p, git_off_t obj_offset)
 {
-	git_off_t base_offset, base_key;
-	git_rawobj base, delta;
-	git_pack_cache_entry *cached = NULL;
-	int error, found_base = 0;
+	git_dependency_chain chain = GIT_ARRAY_INIT;
+	git_mwindow *w_curs = NULL;
+	git_off_t curpos = obj_offset, base_offset;
+	int error = 0, use_heap = 0;
+	size_t size, elem_pos;
+	git_object_t type;
+
+	elem_pos = 0;
+	while (true) {
+		struct pack_chain_elem *elem;
+		git_pack_cache_entry *cached = NULL;
+
+		/* if we have a base cached, we can stop here instead */
+		if ((cached = cache_get(&p->bases, obj_offset)) != NULL) {
+			*cached_out = cached;
+			*cached_off = obj_offset;
+			break;
+		}
 
-	base_offset = get_delta_base(p, w_curs, curpos, delta_type, obj_offset);
-	git_mwindow_close(w_curs);
-	if (base_offset == 0)
-		return packfile_error("delta offset is zero");
-	if (base_offset < 0) /* must actually be an error code */
-		return (int)base_offset;
+		/* if we run out of space on the small stack, use the array */
+		if (elem_pos == SMALL_STACK_SIZE) {
+			git_array_init_to_size(chain, elem_pos);
+			GIT_ERROR_CHECK_ARRAY(chain);
+			memcpy(chain.ptr, small_stack, elem_pos * sizeof(struct pack_chain_elem));
+			chain.size = elem_pos;
+			use_heap = 1;
+		}
 
-	if (!p->bases.entries && (cache_init(&p->bases) < 0))
-		return -1;
+		curpos = obj_offset;
+		if (!use_heap) {
+			elem = &small_stack[elem_pos];
+		} else {
+			elem = git_array_alloc(chain);
+			if (!elem) {
+				error = -1;
+				goto on_error;
+			}
+		}
 
-	base_key = base_offset; /* git_packfile_unpack modifies base_offset */
-	if ((cached = cache_get(&p->bases, base_offset)) != NULL) {
-		memcpy(&base, &cached->raw, sizeof(git_rawobj));
-		found_base = 1;
-	}
+		elem->base_key = obj_offset;
 
-	if (!cached) { /* have to inflate it */
-		error = git_packfile_unpack(&base, p, &base_offset);
+		error = git_packfile_unpack_header(&size, &type, &p->mwf, &w_curs, &curpos);
 
-		/*
-		 * TODO: git.git tries to load the base from other packfiles
-		 * or loose objects.
-		 *
-		 * We'll need to do this in order to support thin packs.
-		 */
 		if (error < 0)
-			return error;
-	}
+			goto on_error;
 
-	error = packfile_unpack_compressed(&delta, p, w_curs, curpos, delta_size, delta_type);
-	git_mwindow_close(w_curs);
+		elem->offset = curpos;
+		elem->size = size;
+		elem->type = type;
+		elem->base_key = obj_offset;
 
-	if (error < 0) {
-		if (!found_base)
-			git__free(base.data);
-		return error;
+		if (type != GIT_OBJECT_OFS_DELTA && type != GIT_OBJECT_REF_DELTA)
+			break;
+
+		base_offset = get_delta_base(p, &w_curs, &curpos, type, obj_offset);
+		git_mwindow_close(&w_curs);
+
+		if (base_offset == 0) {
+			error = packfile_error("delta offset is zero");
+			goto on_error;
+		}
+		if (base_offset < 0) { /* must actually be an error code */
+			error = (int)base_offset;
+			goto on_error;
+		}
+
+		/* we need to pass the pos *after* the delta-base bit */
+		elem->offset = curpos;
+
+		/* go through the loop again, but with the new object */
+		obj_offset = base_offset;
+		elem_pos++;
 	}
 
-	obj->type = base.type;
-	error = git__delta_apply(obj, base.data, base.len, delta.data, delta.len);
-	if (error < 0)
-		goto on_error;
 
-	if (found_base)
-		git_atomic_dec(&cached->refcount);
-	else if (cache_add(&p->bases, &base, base_key) < 0)
-		git__free(base.data);
+	*stack_sz = elem_pos + 1;
+	*chain_out = chain;
+	return error;
 
 on_error:
-	git__free(delta.data);
-
-	return error; /* error set by git__delta_apply */
+	git_array_clear(chain);
+	return error;
 }
 
 int git_packfile_unpack(
@@ -555,48 +626,145 @@ int git_packfile_unpack(
 {
 	git_mwindow *w_curs = NULL;
 	git_off_t curpos = *obj_offset;
-	int error;
-
-	size_t size = 0;
-	git_otype type;
+	int error, free_base = 0;
+	git_dependency_chain chain = GIT_ARRAY_INIT;
+	struct pack_chain_elem *elem = NULL, *stack;
+	git_pack_cache_entry *cached = NULL;
+	struct pack_chain_elem small_stack[SMALL_STACK_SIZE];
+	size_t stack_size = 0, elem_pos, alloclen;
+	git_object_t base_type;
 
 	/*
 	 * TODO: optionally check the CRC on the packfile
 	 */
 
+	error = pack_dependency_chain(&chain, &cached, obj_offset, small_stack, &stack_size, p, *obj_offset);
+	if (error < 0)
+		return error;
+
 	obj->data = NULL;
 	obj->len = 0;
-	obj->type = GIT_OBJ_BAD;
+	obj->type = GIT_OBJECT_INVALID;
 
-	error = git_packfile_unpack_header(&size, &type, &p->mwf, &w_curs, &curpos);
-	git_mwindow_close(&w_curs);
+	/* let's point to the right stack */
+	stack = chain.ptr ? chain.ptr : small_stack;
 
-	if (error < 0)
-		return error;
+	elem_pos = stack_size;
+	if (cached) {
+		memcpy(obj, &cached->raw, sizeof(git_rawobj));
+		base_type = obj->type;
+		elem_pos--;	/* stack_size includes the base, which isn't actually there */
+	} else {
+		elem = &stack[--elem_pos];
+		base_type = elem->type;
+	}
 
-	switch (type) {
-	case GIT_OBJ_OFS_DELTA:
-	case GIT_OBJ_REF_DELTA:
-		error = packfile_unpack_delta(
-				obj, p, &w_curs, &curpos,
-				size, type, *obj_offset);
+	switch (base_type) {
+	case GIT_OBJECT_COMMIT:
+	case GIT_OBJECT_TREE:
+	case GIT_OBJECT_BLOB:
+	case GIT_OBJECT_TAG:
+		if (!cached) {
+			curpos = elem->offset;
+			error = packfile_unpack_compressed(obj, p, &w_curs, &curpos, elem->size, elem->type);
+			git_mwindow_close(&w_curs);
+			base_type = elem->type;
+		}
+		if (error < 0)
+			goto cleanup;
 		break;
+	case GIT_OBJECT_OFS_DELTA:
+	case GIT_OBJECT_REF_DELTA:
+		error = packfile_error("dependency chain ends in a delta");
+		goto cleanup;
+	default:
+		error = packfile_error("invalid packfile type in header");
+		goto cleanup;
+	}
 
-	case GIT_OBJ_COMMIT:
-	case GIT_OBJ_TREE:
-	case GIT_OBJ_BLOB:
-	case GIT_OBJ_TAG:
-		error = packfile_unpack_compressed(
-				obj, p, &w_curs, &curpos,
-				size, type);
-		break;
+	/*
+	 * Finding the object we want a cached base element is
+	 * problematic, as we need to make sure we don't accidentally
+	 * give the caller the cached object, which it would then feel
+	 * free to free, so we need to copy the data.
+	 */
+	if (cached && stack_size == 1) {
+		void *data = obj->data;
 
-	default:
-		error = packfile_error("invalid packfile type in header");;
-		break;
+		GIT_ERROR_CHECK_ALLOC_ADD(&alloclen, obj->len, 1);
+		obj->data = git__malloc(alloclen);
+		GIT_ERROR_CHECK_ALLOC(obj->data);
+
+		memcpy(obj->data, data, obj->len + 1);
+		git_atomic_dec(&cached->refcount);
+		goto cleanup;
+	}
+
+	/* we now apply each consecutive delta until we run out */
+	while (elem_pos > 0 && !error) {
+		git_rawobj base, delta;
+
+		/*
+		 * We can now try to add the base to the cache, as
+		 * long as it's not already the cached one.
+		 */
+		if (!cached)
+			free_base = !!cache_add(&cached, &p->bases, obj, elem->base_key);
+
+		elem = &stack[elem_pos - 1];
+		curpos = elem->offset;
+		error = packfile_unpack_compressed(&delta, p, &w_curs, &curpos, elem->size, elem->type);
+		git_mwindow_close(&w_curs);
+
+		if (error < 0) {
+			/* We have transferred ownership of the data to the cache. */
+			obj->data = NULL;
+			break;
+		}
+
+		/* the current object becomes the new base, on which we apply the delta */
+		base = *obj;
+		obj->data = NULL;
+		obj->len = 0;
+		obj->type = GIT_OBJECT_INVALID;
+
+		error = git_delta_apply(&obj->data, &obj->len, base.data, base.len, delta.data, delta.len);
+		obj->type = base_type;
+
+		/*
+		 * We usually don't want to free the base at this
+		 * point, as we put it into the cache in the previous
+		 * iteration. free_base lets us know that we got the
+		 * base object directly from the packfile, so we can free it.
+		 */
+		git__free(delta.data);
+		if (free_base) {
+			free_base = 0;
+			git__free(base.data);
+		}
+
+		if (cached) {
+			git_atomic_dec(&cached->refcount);
+			cached = NULL;
+		}
+
+		if (error < 0)
+			break;
+
+		elem_pos--;
+	}
+
+cleanup:
+	if (error < 0) {
+		git__free(obj->data);
+		if (cached)
+			git_atomic_dec(&cached->refcount);
 	}
 
-	*obj_offset = curpos;
+	if (elem)
+		*obj_offset = curpos;
+
+	git_array_clear(chain);
 	return error;
 }
 
@@ -625,8 +793,7 @@ int git_packfile_stream_open(git_packfile_stream *obj, struct git_pack_file *p,
 	obj->zstream.next_out = Z_NULL;
 	st = inflateInit(&obj->zstream);
 	if (st != Z_OK) {
-		git__free(obj);
-		giterr_set(GITERR_ZLIB, "Failed to inflate packfile");
+		git_error_set(GIT_ERROR_ZLIB, "failed to init packfile stream");
 		return -1;
 	}
 
@@ -657,7 +824,7 @@ ssize_t git_packfile_stream_read(git_packfile_stream *obj, void *buffer, size_t
 	written = len - obj->zstream.avail_out;
 
 	if (st != Z_OK && st != Z_STREAM_END) {
-		giterr_set(GITERR_ZLIB, "Failed to inflate packfile");
+		git_error_set(GIT_ERROR_ZLIB, "error reading from the zlib stream");
 		return -1;
 	}
 
@@ -673,36 +840,38 @@ ssize_t git_packfile_stream_read(git_packfile_stream *obj, void *buffer, size_t
 
 }
 
-void git_packfile_stream_free(git_packfile_stream *obj)
+void git_packfile_stream_dispose(git_packfile_stream *obj)
 {
 	inflateEnd(&obj->zstream);
 }
 
-int packfile_unpack_compressed(
+static int packfile_unpack_compressed(
 	git_rawobj *obj,
 	struct git_pack_file *p,
 	git_mwindow **w_curs,
 	git_off_t *curpos,
 	size_t size,
-	git_otype type)
+	git_object_t type)
 {
+	size_t buf_size;
 	int st;
 	z_stream stream;
 	unsigned char *buffer, *in;
 
-	buffer = git__calloc(1, size + 1);
-	GITERR_CHECK_ALLOC(buffer);
+	GIT_ERROR_CHECK_ALLOC_ADD(&buf_size, size, 1);
+	buffer = git__calloc(1, buf_size);
+	GIT_ERROR_CHECK_ALLOC(buffer);
 
 	memset(&stream, 0, sizeof(stream));
 	stream.next_out = buffer;
-	stream.avail_out = (uInt)size + 1;
+	stream.avail_out = (uInt)buf_size;
 	stream.zalloc = use_git_alloc;
 	stream.zfree = use_git_free;
 
 	st = inflateInit(&stream);
 	if (st != Z_OK) {
 		git__free(buffer);
-		giterr_set(GITERR_ZLIB, "Failed to inflate packfile");
+		git_error_set(GIT_ERROR_ZLIB, "failed to init zlib stream on unpack");
 
 		return -1;
 	}
@@ -729,7 +898,7 @@ int packfile_unpack_compressed(
 
 	if ((st != Z_STREAM_END) || stream.total_out != size) {
 		git__free(buffer);
-		giterr_set(GITERR_ZLIB, "Failed to inflate packfile");
+		git_error_set(GIT_ERROR_ZLIB, "error inflating zlib stream");
 		return -1;
 	}
 
@@ -747,7 +916,7 @@ git_off_t get_delta_base(
 	struct git_pack_file *p,
 	git_mwindow **w_curs,
 	git_off_t *curpos,
-	git_otype type,
+	git_object_t type,
 	git_off_t delta_obj_offset)
 {
 	unsigned int left = 0;
@@ -765,36 +934,43 @@ git_off_t get_delta_base(
 	 * that is assured. An OFS_DELTA longer than the hash size
 	 * is stupid, as then a REF_DELTA would be smaller to store.
 	 */
-	if (type == GIT_OBJ_OFS_DELTA) {
+	if (type == GIT_OBJECT_OFS_DELTA) {
 		unsigned used = 0;
 		unsigned char c = base_info[used++];
-		base_offset = c & 127;
+		size_t unsigned_base_offset = c & 127;
 		while (c & 128) {
 			if (left <= used)
 				return GIT_EBUFS;
-			base_offset += 1;
-			if (!base_offset || MSB(base_offset, 7))
+			unsigned_base_offset += 1;
+			if (!unsigned_base_offset || MSB(unsigned_base_offset, 7))
 				return 0; /* overflow */
 			c = base_info[used++];
-			base_offset = (base_offset << 7) + (c & 127);
+			unsigned_base_offset = (unsigned_base_offset << 7) + (c & 127);
 		}
-		base_offset = delta_obj_offset - base_offset;
-		if (base_offset <= 0 || base_offset >= delta_obj_offset)
+		if (unsigned_base_offset == 0 || (size_t)delta_obj_offset <= unsigned_base_offset)
 			return 0; /* out of bound */
+		base_offset = delta_obj_offset - unsigned_base_offset;
 		*curpos += used;
-	} else if (type == GIT_OBJ_REF_DELTA) {
+	} else if (type == GIT_OBJECT_REF_DELTA) {
 		/* If we have the cooperative cache, search in it first */
 		if (p->has_cache) {
-			khiter_t k;
 			git_oid oid;
+			size_t k;
 
 			git_oid_fromraw(&oid, base_info);
-			k = kh_get(oid, p->idx_cache, &oid);
-			if (k != kh_end(p->idx_cache)) {
+			k = git_oidmap_lookup_index(p->idx_cache, &oid);
+			if (git_oidmap_valid_index(p->idx_cache, k)) {
 				*curpos += 20;
-				return ((struct git_pack_entry *)kh_value(p->idx_cache, k))->offset;
+				return ((struct git_pack_entry *)git_oidmap_value_at(p->idx_cache, k))->offset;
+			} else {
+				/* If we're building an index, don't try to find the pack
+				 * entry; we just haven't seen it yet.  We'll make
+				 * progress again in the next loop.
+				 */
+				return GIT_PASSTHROUGH;
 			}
 		}
+
 		/* The base entry _must_ be in the same pack */
 		if (pack_entry_find_offset(&base_offset, &unused, p, (git_oid *)base_info, GIT_OID_HEXSZ) < 0)
 			return packfile_error("base entry delta is not in the same pack");
@@ -811,6 +987,18 @@ git_off_t get_delta_base(
  *
  ***********************************************************/
 
+void git_packfile_close(struct git_pack_file *p, bool unlink_packfile)
+{
+	if (p->mwf.fd >= 0) {
+		git_mwindow_free_all_locked(&p->mwf);
+		p_close(p->mwf.fd);
+		p->mwf.fd = -1;
+	}
+
+	if (unlink_packfile)
+		p_unlink(p->pack_name);
+}
+
 void git_packfile_free(struct git_pack_file *p)
 {
 	if (!p)
@@ -818,16 +1006,14 @@ void git_packfile_free(struct git_pack_file *p)
 
 	cache_free(&p->bases);
 
-	git_mwindow_free_all(&p->mwf);
-
-	if (p->mwf.fd != -1)
-		p_close(p->mwf.fd);
+	git_packfile_close(p, false);
 
 	pack_index_free(p);
 
 	git__free(p->bad_object_sha1);
 
 	git_mutex_free(&p->lock);
+	git_mutex_free(&p->bases.lock);
 	git__free(p);
 }
 
@@ -839,7 +1025,7 @@ static int packfile_open(struct git_pack_file *p)
 	unsigned char *idx_sha1;
 
 	if (p->index_version == -1 && pack_index_open(p) < 0)
-		return git_odb__error_notfound("failed to open packfile", NULL);
+		return git_odb__error_notfound("failed to open packfile", NULL, 0);
 
 	/* if mwf opened by another thread, return now */
 	if (git_mutex_lock(&p->lock) < 0)
@@ -901,9 +1087,10 @@ static int packfile_open(struct git_pack_file *p)
 	return 0;
 
 cleanup:
-	giterr_set(GITERR_OS, "Invalid packfile '%s'", p->pack_name);
+	git_error_set(GIT_ERROR_OS, "invalid packfile '%s'", p->pack_name);
 
-	p_close(p->mwf.fd);
+	if (p->mwf.fd >= 0)
+		p_close(p->mwf.fd);
 	p->mwf.fd = -1;
 
 	git_mutex_unlock(&p->lock);
@@ -911,19 +1098,39 @@ cleanup:
 	return -1;
 }
 
+int git_packfile__name(char **out, const char *path)
+{
+	size_t path_len;
+	git_buf buf = GIT_BUF_INIT;
+
+	path_len = strlen(path);
+
+	if (path_len < strlen(".idx"))
+		return git_odb__error_notfound("invalid packfile path", NULL, 0);
+
+	if (git_buf_printf(&buf, "%.*s.pack", (int)(path_len - strlen(".idx")), path) < 0)
+		return -1;
+
+	*out = git_buf_detach(&buf);
+	return 0;
+}
+
 int git_packfile_alloc(struct git_pack_file **pack_out, const char *path)
 {
 	struct stat st;
 	struct git_pack_file *p;
-	size_t path_len = path ? strlen(path) : 0;
+	size_t path_len = path ? strlen(path) : 0, alloc_len;
 
 	*pack_out = NULL;
 
 	if (path_len < strlen(".idx"))
-		return git_odb__error_notfound("invalid packfile path", NULL);
+		return git_odb__error_notfound("invalid packfile path", NULL, 0);
 
-	p = git__calloc(1, sizeof(*p) + path_len + 2);
-	GITERR_CHECK_ALLOC(p);
+	GIT_ERROR_CHECK_ALLOC_ADD(&alloc_len, sizeof(*p), path_len);
+	GIT_ERROR_CHECK_ALLOC_ADD(&alloc_len, alloc_len, 2);
+
+	p = git__calloc(1, alloc_len);
+	GIT_ERROR_CHECK_ALLOC(p);
 
 	memcpy(p->pack_name, path, path_len + 1);
 
@@ -939,12 +1146,11 @@ int git_packfile_alloc(struct git_pack_file **pack_out, const char *path)
 			p->pack_keep = 1;
 
 		memcpy(p->pack_name + root_len, ".pack", sizeof(".pack"));
-		path_len = path_len - strlen(".idx") + strlen(".pack");
 	}
 
 	if (p_stat(p->pack_name, &st) < 0 || !S_ISREG(st.st_mode)) {
 		git__free(p);
-		return git_odb__error_notfound("packfile not found", NULL);
+		return git_odb__error_notfound("packfile not found", NULL, 0);
 	}
 
 	/* ok, it looks sane as far as we can check without
@@ -957,15 +1163,15 @@ int git_packfile_alloc(struct git_pack_file **pack_out, const char *path)
 	p->index_version = -1;
 
 	if (git_mutex_init(&p->lock)) {
-		giterr_set(GITERR_OS, "Failed to initialize packfile mutex");
+		git_error_set(GIT_ERROR_OS, "failed to initialize packfile mutex");
 		git__free(p);
 		return -1;
 	}
 
-	/* see if we can parse the sha1 oid in the packfile name */
-	if (path_len < 40 ||
-		git_oid_fromstr(&p->sha1, path + path_len - GIT_OID_HEXSZ) < 0)
-		memset(&p->sha1, 0x0, GIT_OID_RAWSZ);
+	if (cache_init(&p->bases) < 0) {
+		git__free(p);
+		return -1;
+	}
 
 	*pack_out = p;
 
@@ -981,6 +1187,7 @@ int git_packfile_alloc(struct git_pack_file **pack_out, const char *path)
 static git_off_t nth_packed_object_offset(const struct git_pack_file *p, uint32_t n)
 {
 	const unsigned char *index = p->index_map.data;
+	const unsigned char *end = index + p->index_map.len;
 	index += 4 * 256;
 	if (p->index_version == 1) {
 		return ntohl(*((uint32_t *)(index + 24 * n)));
@@ -991,6 +1198,11 @@ static git_off_t nth_packed_object_offset(const struct git_pack_file *p, uint32_
 		if (!(off & 0x80000000))
 			return off;
 		index += p->num_objects * 4 + (off & 0x7fffffff) * 8;
+
+		/* Make sure we're not being sent out of bounds */
+		if (index >= end - 8)
+			return -1;
+
 		return (((uint64_t)ntohl(*((uint32_t *)(index + 0)))) << 32) |
 					ntohl(*((uint32_t *)(index + 4)));
 	}
@@ -1007,10 +1219,9 @@ int git_pack_foreach_entry(
 {
 	const unsigned char *index = p->index_map.data, *current;
 	uint32_t i;
+	int error = 0;
 
 	if (index == NULL) {
-		int error;
-
 		if ((error = pack_index_open(p)) < 0)
 			return error;
 
@@ -1027,7 +1238,6 @@ int git_pack_foreach_entry(
 
 	if (p->oids == NULL) {
 		git_vector offsets, oids;
-		int error;
 
 		if ((error = git_vector_init(&oids, p->num_objects, NULL)))
 			return error;
@@ -1049,15 +1259,16 @@ int git_pack_foreach_entry(
 			git_vector_foreach(&offsets, i, current)
 				git_vector_insert(&oids, (void*)&current[4]);
 		}
+
 		git_vector_free(&offsets);
-		p->oids = (git_oid **)oids.contents;
+		p->oids = (git_oid **)git_vector_detach(NULL, NULL, &oids);
 	}
 
 	for (i = 0; i < p->num_objects; i++)
-		if (cb(p->oids[i], data))
-			return GIT_EUSER;
+		if ((error = cb(p->oids[i], data)) != 0)
+			return git_error_set_after_callback(error);
 
-	return 0;
+	return error;
 }
 
 static int pack_entry_find_offset(
@@ -1067,10 +1278,11 @@ static int pack_entry_find_offset(
 	const git_oid *short_oid,
 	size_t len)
 {
-	const uint32_t *level1_ofs = p->index_map.data;
-	const unsigned char *index = p->index_map.data;
+	const uint32_t *level1_ofs;
+	const unsigned char *index;
 	unsigned hi, lo, stride;
 	int pos, found = 0;
+	git_off_t offset;
 	const unsigned char *current = 0;
 
 	*offset_out = 0;
@@ -1081,11 +1293,11 @@ static int pack_entry_find_offset(
 		if ((error = pack_index_open(p)) < 0)
 			return error;
 		assert(p->index_map.data);
-
-		index = p->index_map.data;
-		level1_ofs = p->index_map.data;
 	}
 
+	index = p->index_map.data;
+	level1_ofs = p->index_map.data;
+
 	if (p->index_version > 1) {
 		level1_ofs += 2;
 		index += 8;
@@ -1107,8 +1319,7 @@ static int pack_entry_find_offset(
 		short_oid->id[0], short_oid->id[1], short_oid->id[2], lo, hi, p->num_objects);
 #endif
 
-	/* Use git.git lookup code */
-	pos = sha1_entry_pos(index, stride, 0, lo, hi, p->num_objects, short_oid->id);
+	pos = sha1_position(index, stride, lo, hi, short_oid->id);
 
 	if (pos >= 0) {
 		/* An object matching exactly the oid was found */
@@ -1136,11 +1347,16 @@ static int pack_entry_find_offset(
 	}
 
 	if (!found)
-		return git_odb__error_notfound("failed to find offset for pack entry", short_oid);
+		return git_odb__error_notfound("failed to find offset for pack entry", short_oid, len);
 	if (found > 1)
 		return git_odb__error_ambiguous("found multiple offsets for pack entry");
 
-	*offset_out = nth_packed_object_offset(p, pos);
+	if ((offset = nth_packed_object_offset(p, pos)) < 0) {
+		git_error_set(GIT_ERROR_ODB, "packfile index is corrupt");
+		return -1;
+	}
+
+	*offset_out = offset;
 	git_oid_fromraw(found_oid, current);
 
 #ifdef INDEX_DEBUG_LOOKUP