commit-graph: allow cross-alternate chains
In an environment like a fork network, it is helpful to have a commit-graph chain that spans both the base repo and the fork repo. The fork is usually a small set of data on top of the large repo, but sometimes the fork is much larger. For example, git-for-windows/git has almost double the number of commits as git/git because it rebases its commits on every major version update. To allow cross-alternate commit-graph chains, we need a few pieces: 1. When looking for a graph-{hash}.graph file, check all alternates. 2. When merging commit-graph chains, do not merge across alternates. 3. When writing a new commit-graph chain based on a commit-graph file in another object directory, do not allow success if the base file has of the name "commit-graph" instead of "commit-graphs/graph-{hash}.graph". Signed-off-by: Derrick Stolee <dstolee@microsoft.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Derrick Stolee committed
Jun 18, 2019 at 11:14 UTC
c523035cbd8515d095dc15e9661fd896733bedbc
4 files changed
+123
-11
Documentation/technical/commit-graph.txt
+40
@@ -266,6 +266,42 @@ The merge strategy values (2 for the size multiple, 64,000 for the maximum
266
number of commits) could be extracted into config settings for full
267
flexibility.
268
269
+## Chains across multiple object directories
270
+
271
+In a repo with alternates, we look for the `commit-graph-chain` file starting
272
+in the local object directory and then in each alternate. The first file that
273
+exists defines our chain. As we look for the `graph-{hash}` files for
274
+each `{hash}` in the chain file, we follow the same pattern for the host
275
+directories.
276
+
277
+This allows commit-graphs to be split across multiple forks in a fork network.
278
+The typical case is a large "base" repo with many smaller forks.
279
+
280
+As the base repo advances, it will likely update and merge its commit-graph
281
+chain more frequently than the forks. If a fork updates their commit-graph after
282
+the base repo, then it should "reparent" the commit-graph chain onto the new
283
+chain in the base repo. When reading each `graph-{hash}` file, we track
284
+the object directory containing it. During a write of a new commit-graph file,
285
+we check for any changes in the source object directory and read the
286
+`commit-graph-chain` file for that source and create a new file based on those
287
+files. During this "reparent" operation, we necessarily need to collapse all
288
+levels in the fork, as all of the files are invalid against the new base file.
289
+
290
+It is crucial to be careful when cleaning up "unreferenced" `graph-{hash}.graph`
291
+files in this scenario. It falls to the user to define the proper settings for
292
+their custom environment:
293
+
294
+ 1. When merging levels in the base repo, the unreferenced files may still be
295
+ referenced by chains from fork repos.
296
+
297
+ 2. The expiry time should be set to a length of time such that every fork has
298
+ time to recompute their commit-graph chain to "reparent" onto the new base
299
+ file(s).
300
+
301
+ 3. If the commit-graph chain is updated in the base, the fork will not have
302
+ access to the new chain until its chain is updated to reference those files.
303
+ (This may change in the future [5].)
304
+
305
Related Links
306
-------------
307
[0] https://bugs.chromium.org/p/git/issues/detail?id=8
@@ -292,3 +328,7 @@ Related Links
328
329
[4] https://public-inbox.org/git/20180108154822.54829-1-git@jeffhostetler.com/T/#u
330
A patch to remove the ahead-behind calculation from 'status'.
331
+
332
+[5] https://public-inbox.org/git/f27db281-abad-5043-6d71-cbb083b1c877@gmail.com/
333
+ A discussion of a "two-dimensional graph position" that can allow reading
334
+ multiple commit-graph chains at the same time.
commit-graph.c
+45
-11
@@ -320,6 +320,9 @@ static struct commit_graph *load_commit_graph_v1(struct repository *r, const cha
320
struct commit_graph *g = load_commit_graph_one(graph_name);
321
free(graph_name);
322
323
+ if (g)
324
+ g->obj_dir = obj_dir;
325
+
326
return g;
327
}
328
@@ -379,9 +382,10 @@ static struct commit_graph *load_commit_graph_chain(struct repository *r, const
382
count = st.st_size / (the_hash_algo->hexsz + 1);
383
oids = xcalloc(count, sizeof(struct object_id));
384
382
- for (i = 0; i < count && valid; i++) {
383
- char *graph_name;
384
- struct commit_graph *g;
385
+ prepare_alt_odb(r);
386
+
387
+ for (i = 0; i < count; i++) {
388
+ struct object_directory *odb;
389
390
if (strbuf_getline_lf(&line, fp) == EOF)
391
break;
@@ -393,14 +397,29 @@ static struct commit_graph *load_commit_graph_chain(struct repository *r, const
397
break;
398
}
399
396
- graph_name = get_split_graph_filename(obj_dir, line.buf);
397
- g = load_commit_graph_one(graph_name);
398
- free(graph_name);
400
+ valid = 0;
401
+ for (odb = r->objects->odb; odb; odb = odb->next) {
402
+ char *graph_name = get_split_graph_filename(odb->path, line.buf);
403
+ struct commit_graph *g = load_commit_graph_one(graph_name);
404
400
- if (g && add_graph_to_chain(g, graph_chain, oids, i))
401
- graph_chain = g;
402
- else
403
- valid = 0;
405
+ free(graph_name);
406
+
407
+ if (g) {
408
+ g->obj_dir = odb->path;
409
+
410
+ if (add_graph_to_chain(g, graph_chain, oids, i)) {
411
+ graph_chain = g;
412
+ valid = 1;
413
+ }
414
+
415
+ break;
416
+ }
417
+ }
418
+
419
+ if (!valid) {
420
+ warning(_("unable to find all commit-graph files"));
421
+ break;
422
+ }
423
}
424
425
free(oids);
@@ -1418,7 +1437,7 @@ static int write_commit_graph_file(struct write_commit_graph_context *ctx)
1437
1438
if (ctx->split && ctx->base_graph_name && ctx->num_commit_graphs_after > 1) {
1439
char *new_base_hash = xstrdup(oid_to_hex(&ctx->new_base_graph->oid));
1421
- char *new_base_name = get_split_graph_filename(ctx->obj_dir, new_base_hash);
1440
+ char *new_base_name = get_split_graph_filename(ctx->new_base_graph->obj_dir, new_base_hash);
1441
1442
free(ctx->commit_graph_filenames_after[ctx->num_commit_graphs_after - 2]);
1443
free(ctx->commit_graph_hash_after[ctx->num_commit_graphs_after - 2]);
@@ -1493,6 +1512,9 @@ static void split_graph_merge_strategy(struct write_commit_graph_context *ctx)
1512
1513
while (g && (g->num_commits <= split_strategy_size_mult * num_commits ||
1514
num_commits > split_strategy_max_commits)) {
1515
+ if (strcmp(g->obj_dir, ctx->obj_dir))
1516
+ break;
1517
+
1518
num_commits += g->num_commits;
1519
g = g->base_graph;
1520
@@ -1501,6 +1523,18 @@ static void split_graph_merge_strategy(struct write_commit_graph_context *ctx)
1523
1524
ctx->new_base_graph = g;
1525
1526
+ if (ctx->num_commit_graphs_after == 2) {
1527
+ char *old_graph_name = get_commit_graph_filename(g->obj_dir);
1528
+
1529
+ if (!strcmp(g->filename, old_graph_name) &&
1530
+ strcmp(g->obj_dir, ctx->obj_dir)) {
1531
+ ctx->num_commit_graphs_after = 1;
1532
+ ctx->new_base_graph = NULL;
1533
+ }
1534
+
1535
+ free(old_graph_name);
1536
+ }
1537
+
1538
ALLOC_ARRAY(ctx->commit_graph_filenames_after, ctx->num_commit_graphs_after);
1539
ALLOC_ARRAY(ctx->commit_graph_hash_after, ctx->num_commit_graphs_after);
1540
commit-graph.h
+1
@@ -48,6 +48,7 @@ struct commit_graph {
48
uint32_t num_commits;
49
struct object_id oid;
50
char *filename;
51
+ const char *obj_dir;
52
53
uint32_t num_commits_in_base;
54
struct commit_graph *base_graph;
t/t5324-split-commit-graph.sh
+37
@@ -90,6 +90,21 @@ test_expect_success 'add more commits, and write a new base graph' '
90
graph_read_expect 12
91
'
92
93
+test_expect_success 'fork and fail to base a chain on a commit-graph file' '
94
+ test_when_finished rm -rf fork &&
95
+ git clone . fork &&
96
+ (
97
+ cd fork &&
98
+ rm .git/objects/info/commit-graph &&
99
+ echo "$(pwd)/../.git/objects" >.git/objects/info/alternates &&
100
+ test_commit new-commit &&
101
+ git commit-graph write --reachable --split &&
102
+ test_path_is_file $graphdir/commit-graph-chain &&
103
+ test_line_count = 1 $graphdir/commit-graph-chain &&
104
+ verify_chain_files_exist $graphdir
105
+ )
106
+'
107
+
108
test_expect_success 'add three more commits, write a tip graph' '
109
git reset --hard commits/3 &&
110
git merge merge/1 &&
@@ -132,4 +147,26 @@ test_expect_success 'add one commit, write a merged graph' '
147
148
graph_git_behavior 'merged commit-graph: commit 12 vs 6' commits/12 commits/6
149
150
+test_expect_success 'create fork and chain across alternate' '
151
+ git clone . fork &&
152
+ (
153
+ cd fork &&
154
+ git config core.commitGraph true &&
155
+ rm -rf $graphdir &&
156
+ echo "$(pwd)/../.git/objects" >.git/objects/info/alternates &&
157
+ test_commit 13 &&
158
+ git branch commits/13 &&
159
+ git commit-graph write --reachable --split &&
160
+ test_path_is_file $graphdir/commit-graph-chain &&
161
+ test_line_count = 3 $graphdir/commit-graph-chain &&
162
+ ls $graphdir/graph-*.graph >graph-files &&
163
+ test_line_count = 1 graph-files &&
164
+ git -c core.commitGraph=true rev-list HEAD >expect &&
165
+ git -c core.commitGraph=false rev-list HEAD >actual &&
166
+ test_cmp expect actual
167
+ )
168
+'
169
+
170
+graph_git_behavior 'alternate: commit 13 vs 6' commits/13 commits/6
171
+
172
test_done