odb: add write_packfile, for_each_unique_abbrev, convert_object_id

Add three vtable methods to odb_source that were not part of the
recent ps/odb-sources and ps/object-counting series:

 - write_packfile: ingest a pack from a file descriptor. The files
   backend chooses between index-pack (large packs) and
   unpack-objects (small packs below fetch.unpackLimit). Options
   cover thin-pack fixing, promisor marking, fsck, lockfile
   capture, and shallow file passing.

 - for_each_unique_abbrev: iterate objects matching a hex prefix
   for disambiguation. Searches loose objects via oidtree, then
   multi-pack indices, then non-MIDX packs.

 - convert_object_id: translate between hash algorithms using the
   loose object map. Used during SHA-1 to SHA-256 migration.

Also add ODB_SOURCE_HELPER to the source type enum, preparing for
the helper backend in the next commit.

The write_packfile vtable method replaces the pattern where callers
spawn index-pack/unpack-objects directly. fast-import already uses
odb_write_packfile() and this allows non-files backends to handle
pack ingestion through their own mechanism.

Signed-off-by: Aaron Paterson <apaterson@pm.me>

Author: MayCXC

builtin/fast-import.c

@@ -518,31 +518,6 @@ static struct object_entry *insert_object(struct object_id *oid)
 	return e;
 }
 
-static void invalidate_pack_id(unsigned int id)
-{
-	unsigned long lu;
-	struct tag *t;
-	struct hashmap_iter iter;
-	struct object_entry *e;
-
-	hashmap_for_each_entry(&object_table, &iter, e, ent) {
-		if (e->pack_id == id)
-			e->pack_id = MAX_PACK_ID;
-	}
-
-	for (lu = 0; lu < branch_table_sz; lu++) {
-		struct branch *b;
-
-		for (b = branch_table[lu]; b; b = b->table_next_branch)
-			if (b->pack_id == id)
-				b->pack_id = MAX_PACK_ID;
-	}
-
-	for (t = first_tag; t; t = t->next_tag)
-		if (t->pack_id == id)
-			t->pack_id = MAX_PACK_ID;
-}
-
 static unsigned int hc_str(const char *s, size_t len)
 {
 	unsigned int r = 0;
@@ -786,86 +761,21 @@ static void start_packfile(void)
 	all_packs[pack_id] = p;
 }
 
-static const char *create_index(void)
-{
-	const char *tmpfile;
-	struct pack_idx_entry **idx, **c, **last;
-	struct object_entry *e;
-	struct object_entry_pool *o;
-
-	/* Build the table of object IDs. */
-	ALLOC_ARRAY(idx, object_count);
-	c = idx;
-	for (o = blocks; o; o = o->next_pool)
-		for (e = o->next_free; e-- != o->entries;)
-			if (pack_id == e->pack_id)
-				*c++ = &e->idx;
-	last = idx + object_count;
-	if (c != last)
-		die(_("internal consistency error creating the index"));
-
-	tmpfile = write_idx_file(the_repository, NULL, idx, object_count,
-				 &pack_idx_opts, pack_data->hash);
-	free(idx);
-	return tmpfile;
-}
-
-static char *keep_pack(const char *curr_index_name)
-{
-	static const char *keep_msg = "fast-import";
-	struct strbuf name = STRBUF_INIT;
-	int keep_fd;
-
-	odb_pack_name(pack_data->repo, &name, pack_data->hash, "keep");
-	keep_fd = safe_create_file_with_leading_directories(pack_data->repo,
-							    name.buf);
-	if (keep_fd < 0)
-		die_errno(_("cannot create keep file"));
-	write_or_die(keep_fd, keep_msg, strlen(keep_msg));
-	if (close(keep_fd))
-		die_errno(_("failed to write keep file"));
-
-	odb_pack_name(pack_data->repo, &name, pack_data->hash, "pack");
-	if (finalize_object_file(pack_data->repo, pack_data->pack_name, name.buf))
-		die(_("cannot store pack file"));
-
-	odb_pack_name(pack_data->repo, &name, pack_data->hash, "idx");
-	if (finalize_object_file(pack_data->repo, curr_index_name, name.buf))
-		die(_("cannot store index file"));
-	free((void *)curr_index_name);
-	return strbuf_detach(&name, NULL);
-}
-
 static void unkeep_all_packs(void)
 {
 	struct strbuf name = STRBUF_INIT;
 	int k;
 
 	for (k = 0; k < pack_id; k++) {
 		struct packed_git *p = all_packs[k];
+		if (!p)
+			continue;
 		odb_pack_name(p->repo, &name, p->hash, "keep");
 		unlink_or_warn(name.buf);
 	}
 	strbuf_release(&name);
 }
 
-static int loosen_small_pack(const struct packed_git *p)
-{
-	struct child_process unpack = CHILD_PROCESS_INIT;
-
-	if (lseek(p->pack_fd, 0, SEEK_SET) < 0)
-		die_errno(_("failed seeking to start of '%s'"), p->pack_name);
-
-	unpack.in = p->pack_fd;
-	unpack.git_cmd = 1;
-	unpack.stdout_to_stderr = 1;
-	strvec_push(&unpack.args, "unpack-objects");
-	if (!show_stats)
-		strvec_push(&unpack.args, "-q");
-
-	return run_command(&unpack);
-}
-
 static void end_packfile(void)
 {
 	static int running;
@@ -876,10 +786,7 @@ static void end_packfile(void)
 	running = 1;
 	clear_delta_base_cache();
 	if (object_count) {
-		struct odb_source_files *files = odb_source_files_downcast(pack_data->repo->objects->sources);
-		struct packed_git *new_p;
 		struct object_id cur_pack_oid;
-		char *idx_name;
 		int i;
 		struct branch *b;
 		struct tag *t;
@@ -891,26 +798,25 @@ static void end_packfile(void)
 					 object_count, cur_pack_oid.hash,
 					 pack_size);
 
-		if (object_count <= unpack_limit) {
-			if (!loosen_small_pack(pack_data)) {
-				invalidate_pack_id(pack_id);
-				goto discard_pack;
-			}
-		}
+		if (lseek(pack_data->pack_fd, 0, SEEK_SET) < 0)
+			die_errno(_("failed seeking to start of '%s'"),
+				  pack_data->pack_name);
 
-		close(pack_data->pack_fd);
-		idx_name = keep_pack(create_index());
+		if (odb_write_packfile(the_repository->objects,
+				       pack_data->pack_fd, NULL))
+			die(_("failed to ingest pack"));
 
-		/* Register the packfile with core git's machinery. */
-		new_p = packfile_store_load_pack(files->packed, idx_name, 1);
-		if (!new_p)
-			die(_("core Git rejected index %s"), idx_name);
-		all_packs[pack_id] = new_p;
-		free(idx_name);
+		/*
+		 * Non-files backends do not register a pack on disk,
+		 * so NULL out the slot to prevent use-after-free in
+		 * gfi_unpack_entry.
+		 */
+		all_packs[pack_id] = NULL;
 
 		/* Print the boundary */
 		if (pack_edges) {
-			fprintf(pack_edges, "%s:", new_p->pack_name);
+			fprintf(pack_edges, "pack-%s:",
+				hash_to_hex(pack_data->hash));
 			for (i = 0; i < branch_table_sz; i++) {
 				for (b = branch_table[i]; b; b = b->table_next_branch) {
 					if (b->pack_id == pack_id)
@@ -930,7 +836,6 @@ static void end_packfile(void)
 		pack_id++;
 	}
 	else {
-discard_pack:
 		close(pack_data->pack_fd);
 		unlink_or_warn(pack_data->pack_name);
 	}
@@ -1239,6 +1144,16 @@ static void *gfi_unpack_entry(
 {
 	enum object_type type;
 	struct packed_git *p = all_packs[oe->pack_id];
+	if (!p) {
+		/*
+		 * Pack was ingested by a non-files backend via
+		 * odb_write_packfile() and is no longer on disk.
+		 * Read the object back through the ODB instead.
+		 */
+		enum object_type odb_type;
+		return odb_read_object(the_repository->objects,
+				       &oe->idx.oid, &odb_type, sizep);
+	}
 	if (p == pack_data && p->pack_size < (pack_size + the_hash_algo->rawsz)) {
 		/* The object is stored in the packfile we are writing to
 		 * and we have modified it since the last time we scanned

object-name.c

@@ -20,6 +20,7 @@
 #include "packfile.h"
 #include "pretty.h"
 #include "object-file.h"
+#include "odb/source.h"
 #include "read-cache-ll.h"
 #include "repo-settings.h"
 #include "repository.h"
@@ -111,13 +112,28 @@ static enum cb_next match_prefix(const struct object_id *oid, void *arg)
 	return ds->ambiguous ? CB_BREAK : CB_CONTINUE;
 }
 
+static int disambiguate_cb(const struct object_id *oid,
+			   struct object_info *oi UNUSED, void *data)
+{
+	struct disambiguate_state *ds = data;
+	update_candidates(ds, oid);
+	return ds->ambiguous ? 1 : 0;
+}
+
 static void find_short_object_filename(struct disambiguate_state *ds)
 {
 	struct odb_source *source;
 
-	for (source = ds->repo->objects->sources; source && !ds->ambiguous; source = source->next)
-		oidtree_each(odb_source_loose_cache(source, &ds->bin_pfx),
-				&ds->bin_pfx, ds->len, match_prefix, ds);
+	for (source = ds->repo->objects->sources; source && !ds->ambiguous; source = source->next) {
+		if (source->for_each_unique_abbrev) {
+			odb_source_for_each_unique_abbrev(
+				source, &ds->bin_pfx, ds->len,
+				disambiguate_cb, ds);
+		} else {
+			oidtree_each(odb_source_loose_cache(source, &ds->bin_pfx),
+					&ds->bin_pfx, ds->len, match_prefix, ds);
+		}
+	}
 }
 
 static int match_hash(unsigned len, const unsigned char *a, const unsigned char *b)
@@ -208,15 +224,23 @@ static void find_short_packed_object(struct disambiguate_state *ds)
 
 	odb_prepare_alternates(ds->repo->objects);
 	for (source = ds->repo->objects->sources; source && !ds->ambiguous; source = source->next) {
-		struct multi_pack_index *m = get_multi_pack_index(source);
-		if (m)
-			unique_in_midx(m, ds);
+		if (source->for_each_unique_abbrev) {
+			odb_source_for_each_unique_abbrev(
+				source, &ds->bin_pfx, ds->len,
+				disambiguate_cb, ds);
+		} else {
+			struct multi_pack_index *m = get_multi_pack_index(source);
+			if (m)
+				unique_in_midx(m, ds);
+		}
 	}
 
-	repo_for_each_pack(ds->repo, p) {
-		if (ds->ambiguous)
-			break;
-		unique_in_pack(p, ds);
+	if (!ds->repo->objects->sources->for_each_unique_abbrev) {
+		repo_for_each_pack(ds->repo, p) {
+			if (ds->ambiguous)
+				break;
+			unique_in_pack(p, ds);
+		}
 	}
 }
 
@@ -796,19 +820,38 @@ static void find_abbrev_len_for_pack(struct packed_git *p,
 	mad->init_len = mad->cur_len;
 }
 
-static void find_abbrev_len_packed(struct min_abbrev_data *mad)
+static int abbrev_len_cb(const struct object_id *oid,
+			 struct object_info *oi UNUSED, void *data)
 {
-	struct packed_git *p;
+	struct min_abbrev_data *mad = data;
+	extend_abbrev_len(oid, mad);
+	return 0;
+}
 
+static void find_abbrev_len_packed(struct min_abbrev_data *mad)
+{
 	odb_prepare_alternates(mad->repo->objects);
-	for (struct odb_source *source = mad->repo->objects->sources; source; source = source->next) {
-		struct multi_pack_index *m = get_multi_pack_index(source);
-		if (m)
-			find_abbrev_len_for_midx(m, mad);
+
+	for (struct odb_source *source = mad->repo->objects->sources;
+	     source; source = source->next) {
+		if (source->for_each_unique_abbrev) {
+			mad->init_len = 0;
+			odb_source_for_each_unique_abbrev(
+				source, mad->oid, mad->cur_len,
+				abbrev_len_cb, mad);
+			mad->init_len = mad->cur_len;
+		} else {
+			struct multi_pack_index *m = get_multi_pack_index(source);
+			if (m)
+				find_abbrev_len_for_midx(m, mad);
+		}
 	}
 
-	repo_for_each_pack(mad->repo, p)
-		find_abbrev_len_for_pack(p, mad);
+	if (!mad->repo->objects->sources->for_each_unique_abbrev) {
+		struct packed_git *p;
+		repo_for_each_pack(mad->repo, p)
+			find_abbrev_len_for_pack(p, mad);
+	}
 }
 
 void strbuf_repo_add_unique_abbrev(struct strbuf *sb, struct repository *repo,

odb.c

@@ -981,6 +981,32 @@ int odb_write_object_stream(struct object_database *odb,
 	return odb_source_write_object_stream(odb->sources, stream, len, oid);
 }
 
+int odb_write_packfile(struct object_database *odb,
+		       int pack_fd,
+		       struct odb_write_packfile_options *opts)
+{
+	return odb_source_write_packfile(odb->sources, pack_fd, opts);
+}
+
+int odb_for_each_unique_abbrev(struct object_database *odb,
+			       const struct object_id *oid_prefix,
+			       unsigned int prefix_len,
+			       odb_for_each_object_cb cb,
+			       void *cb_data)
+{
+	int ret;
+
+	odb_prepare_alternates(odb);
+	for (struct odb_source *source = odb->sources; source; source = source->next) {
+		ret = odb_source_for_each_unique_abbrev(source, oid_prefix,
+							prefix_len, cb, cb_data);
+		if (ret)
+			return ret;
+	}
+
+	return 0;
+}
+
 struct object_database *odb_new(struct repository *repo,
 				const char *primary_source,
 				const char *secondary_sources)

odb.h

@@ -570,6 +570,25 @@ int odb_write_object_stream(struct object_database *odb,
 			    struct odb_write_stream *stream, size_t len,
 			    struct object_id *oid);
 
+/*
+ * Ingest a pack from a file descriptor into the primary source.
+ * Returns 0 on success, a negative error code otherwise.
+ */
+struct odb_write_packfile_options;
+int odb_write_packfile(struct object_database *odb,
+		       int pack_fd,
+		       struct odb_write_packfile_options *opts);
+
+/*
+ * Iterate over all objects across all sources whose ID starts with
+ * the given prefix. Used for object name disambiguation.
+ */
+int odb_for_each_unique_abbrev(struct object_database *odb,
+			       const struct object_id *oid_prefix,
+			       unsigned int prefix_len,
+			       odb_for_each_object_cb cb,
+			       void *cb_data);
+
 void parse_alternates(const char *string,
 		      int sep,
 		      const char *relative_base,