cat-file: make remote-object-info allow-list adapt to the server
The static allow-list in expand_atom() is hardcoded to allow only "objectname" and "objectsize" for remote queries. This works because, up to this point, servers will either support object-info with name and size or they do not support them at all. As object-info gains new capabilities, we cannot expect different servers with different Git versions to have the same object-info capabilities. Therefore, the client needs to adapt its allow-list to what the server advertises. The client now: 1. Requests the protocol option that the placeholder refers to (i.e. "size" for "%(objectsize)"). 2. Drops any requested option that the server does not advertise in fetch_object_info(). 3. Maps the remaining advertised options back to their placeholders and populates remote_allowed_atoms. 4. Uses remote_allowed_atoms in expand_atom(), preserving the previous behavior for supported placeholders. For example, if the client requests "%(objectsize) %(objecttype)" and the server only supports 'size', then the client only requests 'size'. The server returns the size (i.e "42") "%(objectsize)" is expanded normally while "%(objecttype)" expands to an empty string: "42 " Note that the empty string expansion is only for known but unsupported placeholders. "%(objectcolor)" which doesn't exist would die(). This honors what for-each-ref does for known but inapplicable atoms (placeholders). Move object_info_options out of get_remote_info() so the caller which has data can select what options will be requested instead of requesting always size. Move batch_object_write() out so output is always produced. If there are no supported attributes, the output is a blank line. Include "type" in the object_info_options even though the client does not yet know how to parse the server's "type" capability. As a result, "type" is always filtered out, allowing the tests to verify that known but unsupported placeholders expand to an empty string. Since the filter removes options by swapping with the last element, the list is no longer kept sorted. Drop the pre-sort in fetch_object_info_via_pack() and use the unsorted string_list lookup for the response header. This has no effect in performance as the list can only be two entries long ('size' and 'type'). Mentored-by: Karthik Nayak <karthik.188@gmail.com> Mentored-by: Chandra Pratap <chandrapratap3519@gmail.com> Signed-off-by: Pablo Sabater <pabloosabaterr@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Pablo Sabater committed
Jul 24, 2026 at 12:54 UTC
dd1968bf84def5453886ab7320c44c18f4d03005
5 files changed
+113
-35
builtin/cat-file.c
+63
-33
@@ -338,15 +338,20 @@ struct expand_data {
338
* Flags about when an object info is being fetched from remote.
339
*/
340
unsigned is_remote:1;
341
-};
342
-
343
-#define EXPAND_DATA_INIT { .mode = S_IFINVALID, .type = OBJ_BAD }
341
345
-static const char *remote_object_info_atoms[] = {
346
- "objectname",
347
- "objectsize",
342
+ /*
343
+ * List of atoms (i.e. "objectsize") that the server supports. Built
344
+ * from the server's object-info advertised capabilities.
345
+ */
346
+ struct string_list remote_allowed_atoms;
347
};
348
349
+#define EXPAND_DATA_INIT { \
350
+ .mode = S_IFINVALID, \
351
+ .type = OBJ_BAD, \
352
+ .remote_allowed_atoms = STRING_LIST_INIT_NODUP, \
353
+}
354
+
355
static int is_atom(const char *atom, const char *s, int slen)
356
{
357
int alen = strlen(atom);
@@ -357,17 +362,12 @@ static int expand_atom(struct strbuf *sb, const char *atom, int len,
362
struct expand_data *data)
363
{
364
if (data->is_remote) {
360
- size_t i, allowed_nr = ARRAY_SIZE(remote_object_info_atoms);
361
- for (i = 0; i < allowed_nr; i++)
362
- if (is_atom(remote_object_info_atoms[i], atom, len))
365
+ size_t i;
366
+ for (i = 0; i < data->remote_allowed_atoms.nr; i++)
367
+ if (is_atom(data->remote_allowed_atoms.items[i].string,
368
+ atom, len))
369
break;
364
-
365
- /*
366
- * On remote, skip unsupported atoms returning an empty sb,
367
- * honoring how for-each-ref handles known but inapplicable
368
- * atoms (e.g. %(tagger)).
369
- */
370
- if (i == allowed_nr)
370
+ if (i == data->remote_allowed_atoms.nr)
371
return 1;
372
}
373
@@ -683,12 +683,12 @@ out:
683
static int get_remote_info(int argc,
684
const char **argv,
685
struct object_info **remote_object_info,
686
- struct oid_array *object_info_oids)
686
+ struct oid_array *object_info_oids,
687
+ struct string_list *object_info_options)
688
{
689
int retval = 0;
690
struct remote *remote = NULL;
691
struct object_id oid;
691
- struct string_list object_info_options = STRING_LIST_INIT_NODUP;
692
struct transport *gtransport;
693
694
remote = remote_get(argv[0]);
@@ -728,13 +728,10 @@ static int get_remote_info(int argc,
728
CALLOC_ARRAY(*remote_object_info, object_info_oids->nr);
729
gtransport->smart_options->object_info_oids = object_info_oids;
730
731
- string_list_append(&object_info_options, "size");
732
-
733
- gtransport->smart_options->object_info_options = &object_info_options;
731
+ gtransport->smart_options->object_info_options = object_info_options;
732
gtransport->smart_options->object_info_data = *remote_object_info;
733
retval = transport_fetch_object_info(gtransport);
734
cleanup:
737
- string_list_clear(&object_info_options, 0);
735
transport_disconnect(gtransport);
736
return retval;
737
}
@@ -820,6 +817,21 @@ static void parse_cmd_mailmap(struct batch_options *opt UNUSED,
817
load_mailmap();
818
}
819
820
+struct protocol_placeholder_entry {
821
+ const char *option;
822
+ const char *atom;
823
+};
824
+
825
+static const struct protocol_placeholder_entry remote_atom_map[] = {
826
+ {"size", "objectsize"},
827
+ {"type", "objecttype"},
828
+ /*
829
+ * Add new protocol options here. Even if the server doesn't support
830
+ * them the allow_list will drop them if the server doesn't advertise
831
+ * them.
832
+ */
833
+};
834
+
835
static void parse_cmd_remote_object_info(struct batch_options *opt,
836
const char *line, struct strbuf *output,
837
struct expand_data *data)
@@ -829,6 +841,7 @@ static void parse_cmd_remote_object_info(struct batch_options *opt,
841
char *line_to_split;
842
struct object_info *remote_object_info = NULL;
843
struct oid_array object_info_oids = OID_ARRAY_INIT;
844
+ struct string_list object_info_options = STRING_LIST_INIT_NODUP;
845
const char *saved_format = opt->format;
846
847
if (strlen(line) >= MAX_REMOTE_OBJ_INFO_LINE)
@@ -848,10 +861,22 @@ static void parse_cmd_remote_object_info(struct batch_options *opt,
861
die(_("remote-object-info supports at most %d objects"),
862
MAX_ALLOWED_OBJ_LIMIT);
863
864
+ if (data->info.sizep)
865
+ string_list_append(&object_info_options, "size");
866
+ if (data->info.typep)
867
+ string_list_append(&object_info_options, "type");
868
+
869
if (get_remote_info(count, argv, &remote_object_info,
852
- &object_info_oids))
870
+ &object_info_oids, &object_info_options))
871
die(_("failed to get object info from the remote: %s"), argv[0]);
872
873
+ string_list_clear(&data->remote_allowed_atoms, 0);
874
+ string_list_append(&data->remote_allowed_atoms, "objectname");
875
+ for (size_t i = 0; i < ARRAY_SIZE(remote_atom_map); i++)
876
+ if (unsorted_string_list_has_string(&object_info_options, remote_atom_map[i].option))
877
+ string_list_append(&data->remote_allowed_atoms,
878
+ remote_atom_map[i].atom);
879
+
880
data->skip_object_info = 1;
881
for (size_t i = 0; i < object_info_oids.nr; i++) {
882
data->oid = object_info_oids.oid[i];
@@ -862,25 +887,29 @@ static void parse_cmd_remote_object_info(struct batch_options *opt,
887
continue;
888
}
889
890
+ /*
891
+ * When reaching here, it means remote-object-info can retrieve
892
+ * information from server without downloading them.
893
+ */
894
if (remote_object_info[i].sizep) {
866
- /*
867
- * When reaching here, it means remote-object-info can retrieve
868
- * information from server without downloading them.
869
- */
895
data->size = *remote_object_info[i].sizep;
871
- opt->batch_mode = BATCH_MODE_INFO;
872
- data->is_remote = 1;
873
- batch_object_write(argv[i + 1], output, opt, data, NULL, 0);
874
- data->is_remote = 0;
875
- } else {
876
- report_object_status(opt, oid_to_hex(&data->oid), &data->oid, "missing");
896
}
897
+
898
+ if (remote_object_info[i].typep) {
899
+ data->type = *remote_object_info[i].typep;
900
+ }
901
+
902
+ opt->batch_mode = BATCH_MODE_INFO;
903
+ data->is_remote = 1;
904
+ batch_object_write(argv[i + 1], output, opt, data, NULL, 0);
905
+ data->is_remote = 0;
906
}
907
data->skip_object_info = 0;
908
opt->format = saved_format;
909
910
for (size_t i = 0; i < object_info_oids.nr; i++)
911
free_object_info_contents(&remote_object_info[i]);
912
+ string_list_clear(&object_info_options, 0);
913
free(line_to_split);
914
free(argv);
915
free(remote_object_info);
@@ -1200,6 +1229,7 @@ static int batch_objects(struct batch_options *opt)
1229
cleanup:
1230
strbuf_release(&input);
1231
strbuf_release(&output);
1232
+ string_list_clear(&data.remote_allowed_atoms, 0);
1233
cfg->warn_on_object_refname_ambiguity = save_warning;
1234
return retval;
1235
}
fetch-object-info.c
+19
-1
@@ -55,6 +55,24 @@ int fetch_object_info(const enum protocol_version version, struct object_info_ar
55
case protocol_v2:
56
if (!server_supports_v2("object-info"))
57
die(_("object-info capability is not enabled on the server"));
58
+ /*
59
+ * When removing an element from the list it gets swapped by the
60
+ * last element, iterate backwards to prevent elements skipping
61
+ * evaluation.
62
+ *
63
+ * object_info_options->nr can be safely casted without overflow
64
+ * because the number of options is a small known number (the
65
+ * supported placeholders which currently are size and type).
66
+ */
67
+ for (int i = (int)args->object_info_options->nr - 1; i >= 0; i--)
68
+ if (!server_supports_feature("object-info",
69
+ args->object_info_options->items[i].string, 0))
70
+ unsorted_string_list_delete_item(args->object_info_options, i, 0);
71
+
72
+ /*
73
+ * Even if no options are left, we still send the oid so we get
74
+ * at least an existence check.
75
+ */
76
send_object_info_request(fd_out, args);
77
break;
78
case protocol_v1:
@@ -71,7 +89,7 @@ int fetch_object_info(const enum protocol_version version, struct object_info_ar
89
return -1;
90
}
91
74
- if (!string_list_has_string(args->object_info_options, reader->line))
92
+ if (!unsorted_string_list_has_string(args->object_info_options, reader->line))
93
return -1;
94
95
if (!strcmp(reader->line, "size")) {
fetch-object-info.h
+3
@@ -14,6 +14,9 @@ struct object_info;
14
/*
15
* Sends git-cat-file object-info command into the request buf and read the
16
* results from packets.
17
+ *
18
+ * Modifies args->object_info_options, on return it contains only the supported
19
+ * options by the server.
20
*/
21
int fetch_object_info(enum protocol_version version, struct object_info_args *args,
22
struct packet_reader *reader, struct object_info *object_info_data,
t/t1017-cat-file-remote-object-info.sh
+28
@@ -271,6 +271,34 @@ test_expect_success 'unsupported placeholder on remote returns empty string' '
271
)
272
'
273
274
+test_expect_success 'requesting only objectname echoes back' '
275
+ (
276
+ set_transport_variables "$daemon_parent" &&
277
+ cd "$daemon_parent/daemon_client_empty" &&
278
+
279
+ echo $hello_oid >expect &&
280
+ git cat-file --batch-command="%(objectname)" >actual <<-EOF &&
281
+ remote-object-info "$GIT_DAEMON_URL/parent" $hello_oid
282
+ EOF
283
+ test_cmp expect actual
284
+ )
285
+'
286
+
287
+test_expect_success 'objectname goes through existence check' '
288
+ (
289
+ set_transport_variables "$daemon_parent" &&
290
+ cd "$daemon_parent/daemon_client_empty" &&
291
+
292
+ echo "$unstored_oid missing" >expect &&
293
+
294
+ git cat-file --batch-command="%(objectname)" >actual <<-EOF &&
295
+ remote-object-info "$GIT_DAEMON_URL/parent" $unstored_oid
296
+ EOF
297
+
298
+ test_cmp expect actual
299
+ )
300
+'
301
+
302
# Test --batch-command remote-object-info with 'git://' and
303
# transfer.advertiseobjectinfo set to false, i.e. server does not have object-info capability
304
test_expect_success 'batch-command remote-object-info git:// fails when transfer.advertiseobjectinfo=false' '
transport.c
-1
@@ -443,7 +443,6 @@ static int fetch_object_info_via_pack(struct transport *transport)
443
args.server_options = transport->server_options;
444
args.oids = transport->smart_options->object_info_oids;
445
args.object_info_options = transport->smart_options->object_info_options;
446
- string_list_sort(args.object_info_options);
446
447
connect_setup(transport, 0);
448
packet_reader_init(&reader, data->fd[0], NULL, 0,