path-walk: mark trees and blobs as UNINTERESTING
When the input rev_info has UNINTERESTING starting points, we want to be sure that the UNINTERESTING flag is passed appropriately through the objects. To match how this is done in places such as 'git pack-objects', we use the mark_edges_uninteresting() method. This method has an option for using the "sparse" walk, which is similar in spirit to the path-walk API's walk. To be sure to keep it independent, add a new 'prune_all_uninteresting' option to the path_walk_info struct. To check how the UNINTERSTING flag is spread through our objects, extend the 'test-tool path-walk' command to output whether or not an object has that flag. This changes our tests significantly, including the removal of some objects that were previously visited due to the incomplete implementation. Signed-off-by: Derrick Stolee <stolee@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Derrick Stolee committed
Dec 20, 2024 at 16:21 UTC
6333e7ae0bb1027b3299d482be5a4a0937116ab0
5 files changed
+159
-22
Documentation/technical/api-path-walk.txt
+8
@@ -48,6 +48,14 @@ commits.
48
While it is possible to walk only commits in this way, consumers would be
49
better off using the revision walk API instead.
50
51
+`prune_all_uninteresting`::
52
+ By default, all reachable paths are emitted by the path-walk API.
53
+ This option allows consumers to declare that they are not
54
+ interested in paths where all included objects are marked with the
55
+ `UNINTERESTING` flag. This requires using the `boundary` option in
56
+ the revision walk so that the walk emits commits marked with the
57
+ `UNINTERESTING` flag.
58
+
59
Examples
60
--------
61
path-walk.c
+74
@@ -8,6 +8,7 @@
8
#include "dir.h"
9
#include "hashmap.h"
10
#include "hex.h"
11
+#include "list-objects.h"
12
#include "object.h"
13
#include "oid-array.h"
14
#include "revision.h"
@@ -23,6 +24,7 @@ static const char *root_path = "";
24
struct type_and_oid_list {
25
enum object_type type;
26
struct oid_array oids;
27
+ int maybe_interesting;
28
};
29
30
#define TYPE_AND_OID_LIST_INIT { \
@@ -142,6 +144,10 @@ static int add_tree_entries(struct path_walk_context *ctx,
144
strmap_put(&ctx->paths_to_lists, path.buf, list);
145
}
146
push_to_stack(ctx, path.buf);
147
+
148
+ if (!(o->flags & UNINTERESTING))
149
+ list->maybe_interesting = 1;
150
+
151
oid_array_append(&list->oids, &entry.oid);
152
}
153
@@ -169,6 +175,43 @@ static int walk_path(struct path_walk_context *ctx,
175
if (!list->oids.nr)
176
return 0;
177
178
+ if (ctx->info->prune_all_uninteresting) {
179
+ /*
180
+ * This is true if all objects were UNINTERESTING
181
+ * when added to the list.
182
+ */
183
+ if (!list->maybe_interesting)
184
+ return 0;
185
+
186
+ /*
187
+ * But it's still possible that the objects were set
188
+ * as UNINTERESTING after being added. Do a quick check.
189
+ */
190
+ list->maybe_interesting = 0;
191
+ for (size_t i = 0;
192
+ !list->maybe_interesting && i < list->oids.nr;
193
+ i++) {
194
+ if (list->type == OBJ_TREE) {
195
+ struct tree *t = lookup_tree(ctx->repo,
196
+ &list->oids.oid[i]);
197
+ if (t && !(t->object.flags & UNINTERESTING))
198
+ list->maybe_interesting = 1;
199
+ } else if (list->type == OBJ_BLOB) {
200
+ struct blob *b = lookup_blob(ctx->repo,
201
+ &list->oids.oid[i]);
202
+ if (b && !(b->object.flags & UNINTERESTING))
203
+ list->maybe_interesting = 1;
204
+ } else {
205
+ /* Tags are always interesting if visited. */
206
+ list->maybe_interesting = 1;
207
+ }
208
+ }
209
+
210
+ /* We have confirmed that all objects are UNINTERESTING. */
211
+ if (!list->maybe_interesting)
212
+ return 0;
213
+ }
214
+
215
/* Evaluate function pointer on this data, if requested. */
216
if ((list->type == OBJ_TREE && ctx->info->trees) ||
217
(list->type == OBJ_BLOB && ctx->info->blobs) ||
@@ -203,6 +246,26 @@ static void clear_paths_to_lists(struct strmap *map)
246
strmap_init(map);
247
}
248
249
+static struct repository *edge_repo;
250
+static struct type_and_oid_list *edge_tree_list;
251
+
252
+static void show_edge(struct commit *commit)
253
+{
254
+ struct tree *t = repo_get_commit_tree(edge_repo, commit);
255
+
256
+ if (!t)
257
+ return;
258
+
259
+ if (commit->object.flags & UNINTERESTING)
260
+ t->object.flags |= UNINTERESTING;
261
+
262
+ if (t->object.flags & SEEN)
263
+ return;
264
+ t->object.flags |= SEEN;
265
+
266
+ oid_array_append(&edge_tree_list->oids, &t->object.oid);
267
+}
268
+
269
static int setup_pending_objects(struct path_walk_info *info,
270
struct path_walk_context *ctx)
271
{
@@ -314,6 +377,7 @@ static int setup_pending_objects(struct path_walk_info *info,
377
if (tagged_blobs->oids.nr) {
378
const char *tagged_blob_path = "/tagged-blobs";
379
tagged_blobs->type = OBJ_BLOB;
380
+ tagged_blobs->maybe_interesting = 1;
381
push_to_stack(ctx, tagged_blob_path);
382
strmap_put(&ctx->paths_to_lists, tagged_blob_path, tagged_blobs);
383
} else {
@@ -325,6 +389,7 @@ static int setup_pending_objects(struct path_walk_info *info,
389
if (tags->oids.nr) {
390
const char *tag_path = "/tags";
391
tags->type = OBJ_TAG;
392
+ tags->maybe_interesting = 1;
393
push_to_stack(ctx, tag_path);
394
strmap_put(&ctx->paths_to_lists, tag_path, tags);
395
} else {
@@ -369,6 +434,7 @@ int walk_objects_by_path(struct path_walk_info *info)
434
/* Insert a single list for the root tree into the paths. */
435
CALLOC_ARRAY(root_tree_list, 1);
436
root_tree_list->type = OBJ_TREE;
437
+ root_tree_list->maybe_interesting = 1;
438
strmap_put(&ctx.paths_to_lists, root_path, root_tree_list);
439
push_to_stack(&ctx, root_path);
440
@@ -382,6 +448,14 @@ int walk_objects_by_path(struct path_walk_info *info)
448
if (prepare_revision_walk(info->revs))
449
die(_("failed to setup revision walk"));
450
451
+ /* Walk trees to mark them as UNINTERESTING. */
452
+ edge_repo = info->revs->repo;
453
+ edge_tree_list = root_tree_list;
454
+ mark_edges_uninteresting(info->revs, show_edge,
455
+ info->prune_all_uninteresting);
456
+ edge_repo = NULL;
457
+ edge_tree_list = NULL;
458
+
459
info->revs->blob_objects = info->revs->tree_objects = 0;
460
461
trace2_region_enter("path-walk", "pending-walk", info->revs->repo);
path-walk.h
+8
@@ -40,6 +40,14 @@ struct path_walk_info {
40
int trees;
41
int blobs;
42
int tags;
43
+
44
+ /**
45
+ * When 'prune_all_uninteresting' is set and a path has all objects
46
+ * marked as UNINTERESTING, then the path-walk will not visit those
47
+ * objects. It will not call path_fn on those objects and will not
48
+ * walk the children of such trees.
49
+ */
50
+ int prune_all_uninteresting;
51
};
52
53
#define PATH_WALK_INFO_INIT { \
t/helper/test-path-walk.c
+9
-3
@@ -50,10 +50,14 @@ static int emit_block(const char *path, struct oid_array *oids,
50
printf("%"PRIuMAX":%s:%s:EMPTY\n",
51
tdata->batch_nr, typestr, path);
52
53
- for (size_t i = 0; i < oids->nr; i++)
54
- printf("%"PRIuMAX":%s:%s:%s\n",
53
+ for (size_t i = 0; i < oids->nr; i++) {
54
+ struct object *o = lookup_unknown_object(the_repository,
55
+ &oids->oid[i]);
56
+ printf("%"PRIuMAX":%s:%s:%s%s\n",
57
tdata->batch_nr, typestr, path,
56
- oid_to_hex(&oids->oid[i]));
58
+ oid_to_hex(&oids->oid[i]),
59
+ o->flags & UNINTERESTING ? ":UNINTERESTING" : "");
60
+ }
61
62
tdata->batch_nr++;
63
return 0;
@@ -74,6 +78,8 @@ int cmd__path_walk(int argc, const char **argv)
78
N_("toggle inclusion of tag objects")),
79
OPT_BOOL(0, "trees", &info.trees,
80
N_("toggle inclusion of tree objects")),
81
+ OPT_BOOL(0, "prune", &info.prune_all_uninteresting,
82
+ N_("toggle pruning of uninteresting paths")),
83
OPT_END(),
84
};
85
t/t6601-path-walk.sh
+60
-19
@@ -211,11 +211,11 @@ test_expect_success 'topic, not base' '
211
0:commit::$(git rev-parse topic)
212
1:tree::$(git rev-parse topic^{tree})
213
2:tree:right/:$(git rev-parse topic:right)
214
- 3:blob:right/d:$(git rev-parse topic:right/d)
214
+ 3:blob:right/d:$(git rev-parse topic:right/d):UNINTERESTING
215
4:blob:right/c:$(git rev-parse topic:right/c)
216
- 5:tree:left/:$(git rev-parse topic:left)
217
- 6:blob:left/b:$(git rev-parse topic:left/b)
218
- 7:blob:a:$(git rev-parse topic:a)
216
+ 5:tree:left/:$(git rev-parse topic:left):UNINTERESTING
217
+ 6:blob:left/b:$(git rev-parse topic:left/b):UNINTERESTING
218
+ 7:blob:a:$(git rev-parse topic:a):UNINTERESTING
219
blobs:4
220
commits:1
221
tags:0
@@ -225,15 +225,38 @@ test_expect_success 'topic, not base' '
225
test_cmp_sorted expect out
226
'
227
228
+test_expect_success 'fourth, blob-tag2, not base' '
229
+ test-tool path-walk -- fourth blob-tag2 --not base >out &&
230
+
231
+ cat >expect <<-EOF &&
232
+ 0:commit::$(git rev-parse topic)
233
+ 1:tag:/tags:$(git rev-parse fourth)
234
+ 2:blob:/tagged-blobs:$(git rev-parse refs/tags/blob-tag2^{})
235
+ 3:tree::$(git rev-parse topic^{tree})
236
+ 4:tree:right/:$(git rev-parse topic:right)
237
+ 5:blob:right/d:$(git rev-parse base~1:right/d):UNINTERESTING
238
+ 6:blob:right/c:$(git rev-parse topic:right/c)
239
+ 7:tree:left/:$(git rev-parse base~1:left):UNINTERESTING
240
+ 8:blob:left/b:$(git rev-parse base~1:left/b):UNINTERESTING
241
+ 9:blob:a:$(git rev-parse base~1:a):UNINTERESTING
242
+ blobs:5
243
+ commits:1
244
+ tags:1
245
+ trees:3
246
+ EOF
247
+
248
+ test_cmp_sorted expect out
249
+'
250
+
251
test_expect_success 'topic, not base, only blobs' '
252
test-tool path-walk --no-trees --no-commits \
253
-- topic --not base >out &&
254
255
cat >expect <<-EOF &&
233
- 0:blob:right/d:$(git rev-parse topic:right/d)
256
+ 0:blob:right/d:$(git rev-parse topic:right/d):UNINTERESTING
257
1:blob:right/c:$(git rev-parse topic:right/c)
235
- 2:blob:left/b:$(git rev-parse topic:left/b)
236
- 3:blob:a:$(git rev-parse topic:a)
258
+ 2:blob:left/b:$(git rev-parse topic:left/b):UNINTERESTING
259
+ 3:blob:a:$(git rev-parse topic:a):UNINTERESTING
260
blobs:4
261
commits:0
262
tags:0
@@ -267,7 +290,7 @@ test_expect_success 'topic, not base, only trees' '
290
cat >expect <<-EOF &&
291
0:tree::$(git rev-parse topic^{tree})
292
1:tree:right/:$(git rev-parse topic:right)
270
- 2:tree:left/:$(git rev-parse topic:left)
293
+ 2:tree:left/:$(git rev-parse topic:left):UNINTERESTING
294
commits:0
295
blobs:0
296
tags:0
@@ -282,17 +305,17 @@ test_expect_success 'topic, not base, boundary' '
305
306
cat >expect <<-EOF &&
307
0:commit::$(git rev-parse topic)
285
- 0:commit::$(git rev-parse base~1)
308
+ 0:commit::$(git rev-parse base~1):UNINTERESTING
309
1:tree::$(git rev-parse topic^{tree})
287
- 1:tree::$(git rev-parse base~1^{tree})
310
+ 1:tree::$(git rev-parse base~1^{tree}):UNINTERESTING
311
2:tree:right/:$(git rev-parse topic:right)
289
- 2:tree:right/:$(git rev-parse base~1:right)
290
- 3:blob:right/d:$(git rev-parse base~1:right/d)
291
- 4:blob:right/c:$(git rev-parse base~1:right/c)
312
+ 2:tree:right/:$(git rev-parse base~1:right):UNINTERESTING
313
+ 3:blob:right/d:$(git rev-parse base~1:right/d):UNINTERESTING
314
+ 4:blob:right/c:$(git rev-parse base~1:right/c):UNINTERESTING
315
4:blob:right/c:$(git rev-parse topic:right/c)
293
- 5:tree:left/:$(git rev-parse base~1:left)
294
- 6:blob:left/b:$(git rev-parse base~1:left/b)
295
- 7:blob:a:$(git rev-parse base~1:a)
316
+ 5:tree:left/:$(git rev-parse base~1:left):UNINTERESTING
317
+ 6:blob:left/b:$(git rev-parse base~1:left/b):UNINTERESTING
318
+ 7:blob:a:$(git rev-parse base~1:a):UNINTERESTING
319
blobs:5
320
commits:2
321
tags:0
@@ -302,6 +325,27 @@ test_expect_success 'topic, not base, boundary' '
325
test_cmp_sorted expect out
326
'
327
328
+test_expect_success 'topic, not base, boundary with pruning' '
329
+ test-tool path-walk --prune -- --boundary topic --not base >out &&
330
+
331
+ cat >expect <<-EOF &&
332
+ 0:commit::$(git rev-parse topic)
333
+ 0:commit::$(git rev-parse base~1):UNINTERESTING
334
+ 1:tree::$(git rev-parse topic^{tree})
335
+ 1:tree::$(git rev-parse base~1^{tree}):UNINTERESTING
336
+ 2:tree:right/:$(git rev-parse topic:right)
337
+ 2:tree:right/:$(git rev-parse base~1:right):UNINTERESTING
338
+ 3:blob:right/c:$(git rev-parse base~1:right/c):UNINTERESTING
339
+ 3:blob:right/c:$(git rev-parse topic:right/c)
340
+ blobs:2
341
+ commits:2
342
+ tags:0
343
+ trees:4
344
+ EOF
345
+
346
+ test_cmp_sorted expect out
347
+'
348
+
349
test_expect_success 'trees are reported exactly once' '
350
test_when_finished "rm -rf unique-trees" &&
351
test_create_repo unique-trees &&
@@ -309,15 +353,12 @@ test_expect_success 'trees are reported exactly once' '
353
cd unique-trees &&
354
mkdir initial &&
355
test_commit initial/file &&
312
-
356
git switch -c move-to-top &&
357
git mv initial/file.t ./ &&
358
test_tick &&
359
git commit -m moved &&
317
-
360
git update-ref refs/heads/other HEAD
361
) &&
320
-
362
test-tool -C unique-trees path-walk -- --all >out &&
363
tree=$(git -C unique-trees rev-parse HEAD:) &&
364
grep "$tree" out >out-filtered &&