gvfs: allow "virtualizing" objects

The idea is to allow blob objects to be missing from the local repository,
and to load them lazily on demand.

After discussing this idea on the mailing list, we will rename the feature
to "lazy clone" and work more on this.

Signed-off-by: Ben Peart <[email protected]>
Signed-off-by: Johannes Schindelin <[email protected]>

Author: Ben Peart

config.c

@@ -1678,6 +1678,11 @@ int git_default_core_config(const char *var, const char *value,
 		return 0;
 	}
 
+	if (!strcmp(var, "core.virtualizeobjects")) {
+		core_virtualize_objects = git_config_bool(var, value);
+		return 0;
+	}
+
 	/* Add other config variables here and to Documentation/config.txt. */
 	return platform_core_config(var, value, ctx, cb);
 }

connected.c

@@ -1,6 +1,7 @@
 #define USE_THE_REPOSITORY_VARIABLE
 
 #include "git-compat-util.h"
+#include "environment.h"
 #include "gettext.h"
 #include "hex.h"
 #include "gvfs.h"
@@ -52,6 +53,8 @@ int check_connected(oid_iterate_fn fn, void *cb_data,
 	 */
 	if (gvfs_config_is_set(GVFS_FETCH_SKIP_REACHABILITY_AND_UPLOADPACK))
 		return 0;
+	if (core_virtualize_objects)
+		return 0;
 
 	if (!opt)
 		opt = &defaults;

environment.c

@@ -75,6 +75,7 @@ int core_gvfs;
 int merge_log_config = -1;
 int precomposed_unicode = -1; /* see probe_utf8_pathname_composition() */
 unsigned long pack_size_limit_cfg;
+int core_virtualize_objects;
 int max_allowed_tree_depth =
 #ifdef _MSC_VER
 	/*

environment.h

@@ -226,5 +226,6 @@ extern const char *comment_line_str;
 extern char *comment_line_str_to_free;
 extern int auto_comment_line_char;
 
+extern int core_virtualize_objects;
 # endif /* USE_THE_REPOSITORY_VARIABLE */
 #endif /* ENVIRONMENT_H */

object-file.c

@@ -40,6 +40,8 @@
 #include "fsck.h"
 #include "loose.h"
 #include "object-file-convert.h"
+#include "trace.h"
+#include "hook.h"
 
 /* The maximum size for an object header. */
 #define MAX_HEADER_LEN 32
@@ -1620,6 +1622,20 @@ void disable_obj_read_lock(void)
 	pthread_mutex_destroy(&obj_read_mutex);
 }
 
+static int run_read_object_hook(struct repository *r, const struct object_id *oid)
+{
+	struct run_hooks_opt opt = RUN_HOOKS_OPT_INIT;
+	int ret;
+	uint64_t start;
+
+	start = getnanotime();
+	strvec_push(&opt.args, oid_to_hex(oid));
+	ret = run_hooks_opt(r, "read-object", &opt);
+	trace_performance_since(start, "run_read_object_hook");
+
+	return ret;
+}
+
 int fetch_if_missing = 1;
 
 static int do_oid_object_info_extended(struct repository *r,
@@ -1632,6 +1648,7 @@ static int do_oid_object_info_extended(struct repository *r,
 	int rtype;
 	const struct object_id *real = oid;
 	int already_retried = 0;
+	int tried_hook = 0;
 
 
 	if (flags & OBJECT_INFO_LOOKUP_REPLACE)
@@ -1643,6 +1660,7 @@ static int do_oid_object_info_extended(struct repository *r,
 	if (!oi)
 		oi = &blank_oi;
 
+retry:
 	co = find_cached_object(real);
 	if (co) {
 		if (oi->typep)
@@ -1674,6 +1692,11 @@ static int do_oid_object_info_extended(struct repository *r,
 			reprepare_packed_git(r);
 			if (find_pack_entry(r, real, &e))
 				break;
+			if (core_virtualize_objects && !tried_hook) {
+				tried_hook = 1;
+				if (!run_read_object_hook(r, oid))
+					goto retry;
+			}
 		}
 
 		/*