git/list[1] front-page[2] threads[3] people[4] search[5] about
 

[PATCH 1/2] merge: fix numerus bugs around "trivial merge" area

From
Junio C Hamano <gitster@pobox.com>
Date
Aug 23, 2008, 19:56 UTC
Message-ID
<7vfxovy06u.fsf_-_@gitster.siamese.dyndns.org>
In-Reply-To
<7vk5e7y0a0.fsf@gitster.siamese.dyndns.org>

The "trivial merge" codepath wanted to optimize itself by making an internal call to the read-tree machinery, but it did not read the index before doing so, and the codepath was never exercised. Incidentally, this failure to read the index upfront meant that the safety to refuse doing anything when the index is unmerged did not kick in, either.

These two problem are fixed by using read_cache_unmerged() that does read the index before checking if it is unmerged at the beginning of cmd_merge().

The primary logic of the merge, however, had an assumption that the process never read the index in-core, and write_cache_as_tree() call it makes from write_tree_trivial() will always read from the on-disk index the strategies created and write it out as a tree. This assumption is now broken by the above fix. It now calls discard_cache() before calling write_tree_trivial() when it wants to write the on-disk index as a tree to fix this issue.

When multiple strategies are tried, their results are evaluated by reading the resulting index and inspecting it. The codepath needs to make a call to read_cache() for each successful strategy, and for that to work, they need to discard_cache() the one from the previous round.

Also the "trivial merge" forgot that the current commit is one of the parents of the resulting commit.

This still has breakage in the way it writes out the resulting tree out of the trivial merge codepath, which is a topic of the next patch. One test in t7605-merge-resolve.sh exposes this breakage.

Signed-off-by: Junio C Hamano <gitster@pobox.com>
---
 builtin-merge.c            |   16 +++++++++-------
 t/t3030-merge-recursive.sh |   11 +++++++++++
 t/t7600-merge.sh           |    9 +++++++++
 t/t7605-merge-resolve.sh   |    6 ++++--
 4 files changed, 33 insertions(+), 9 deletions(-)
diff --git a/builtin-merge.c b/builtin-merge.c
index a201c66..b280444 100644
--- a/builtin-merge.c
+++ b/builtin-merge.c
@@ -564,8 +564,6 @@ static int checkout_fast_forward(unsigned char *head, unsigned char *remote)
 	struct dir_struct dir;
 	struct lock_file *lock_file = xcalloc(1, sizeof(struct lock_file));
 
-	if (read_cache_unmerged())
-		die("you need to resolve your current index first");
 	refresh_cache(REFRESH_QUIET);
 
 	fd = hold_locked_index(lock_file, 1);
@@ -651,13 +649,15 @@ static void add_strategies(const char *string, unsigned attr)
 static int merge_trivial(void)
 {
 	unsigned char result_tree[20], result_commit[20];
-	struct commit_list parent;
+	struct commit_list *parent = xmalloc(sizeof(struct commit_list *));
 
 	write_tree_trivial(result_tree);
 	printf("Wonderful.\n");
-	parent.item = remoteheads->item;
-	parent.next = NULL;
-	commit_tree(merge_msg.buf, result_tree, &parent, result_commit);
+	parent->item = lookup_commit(head);
+	parent->next = xmalloc(sizeof(struct commit_list *));
+	parent->next->item = remoteheads->item;
+	parent->next->next = NULL;
+	commit_tree(merge_msg.buf, result_tree, parent, result_commit);
 	finish(result_commit, "In-index merge");
 	drop_save();
 	return 0;
@@ -743,6 +743,7 @@ static int evaluate_result(void)
 	int cnt = 0;
 	struct rev_info rev;
 
+	discard_cache();
 	if (read_cache() < 0)
 		die("failed to read the cache");
 
@@ -776,7 +777,7 @@ int cmd_merge(int argc, const char **argv, const char *prefix)
 	struct commit_list **remotes = &remoteheads;
 
 	setup_work_tree();
-	if (unmerged_cache())
+	if (read_cache_unmerged())
 		die("You are in the middle of a conflicted merge.");
 
 	/*
@@ -1073,6 +1074,7 @@ int cmd_merge(int argc, const char **argv, const char *prefix)
 		}
 
 		/* Automerge succeeded. */
+		discard_cache();
 		write_tree_trivial(result_tree);
 		automerge_was_ok = 1;
 		break;
diff --git a/t/t3030-merge-recursive.sh b/t/t3030-merge-recursive.sh
index aff3603..f288015 100755
--- a/t/t3030-merge-recursive.sh
+++ b/t/t3030-merge-recursive.sh
@@ -269,6 +269,17 @@ test_expect_success 'merge-recursive result' '
 
 '
 
+test_expect_success 'fail if the index has unresolved entries' '
+
+	rm -fr [abcd] &&
+	git checkout -f "$c1" &&
+
+	test_must_fail git merge "$c5" &&
+	test_must_fail git merge "$c5" 2> out &&
+	grep "You are in the middle of a conflicted merge" out
+
+'
+
 test_expect_success 'merge-recursive remove conflict' '
 
 	rm -fr [abcd] &&
diff --git a/t/t7600-merge.sh b/t/t7600-merge.sh
index fee8fb7..dbc90bc 100755
--- a/t/t7600-merge.sh
+++ b/t/t7600-merge.sh
@@ -498,4 +498,13 @@ test_expect_success 'merge fast-forward in a dirty tree' '
 
 test_debug 'gitk --all'
 
+test_expect_success 'in-index merge' '
+	git reset --hard c0 &&
+	git merge --no-ff -s resolve c1 > out &&
+	grep "Wonderful." out &&
+	verify_parents $c0 $c1
+'
+
+test_debug 'gitk --all'
+
 test_done
diff --git a/t/t7605-merge-resolve.sh b/t/t7605-merge-resolve.sh
index ee21a10..5c53608 100755
--- a/t/t7605-merge-resolve.sh
+++ b/t/t7605-merge-resolve.sh
@@ -27,7 +27,7 @@ test_expect_success 'setup' '
 	git tag c3
 '
 
-test_expect_success 'merge c1 to c2' '
+test_expect_failure 'merge c1 to c2' '
 	git reset --hard c1 &&
 	git merge -s resolve c2 &&
 	test "$(git rev-parse c1)" != "$(git rev-parse HEAD)" &&
@@ -36,7 +36,9 @@ test_expect_success 'merge c1 to c2' '
 	git diff --exit-code &&
 	test -f c0.c &&
 	test -f c1.c &&
-	test -f c2.c
+	test -f c2.c &&
+	test 3 = $(git ls-tree -r HEAD | wc -l) &&
+	test 3 = $(git ls-files | wc -l)
 '
 
 test_expect_success 'merge c2 to c3 (fails)' '
-- 
1.6.0.51.g078ae
Previous: Junio C HamanoNext: Junio C Hamano
Message 9 of 22 in “[bug] git `next' does not do trivial merges”
  1. Paolo BonziniAug 22, 2008
  2. Jeff KingAug 22, 2008
  3. Miklos VajnaAug 23, 2008
  4. Fix in-index merge.Miklos Vajna, Aug 23, 2008
  5. builtin-merge: fail properly when we are in the middle of a conflicted mergeMiklos Vajna, Aug 23, 2008
  6. Junio C HamanoAug 23, 2008
  7. builtin-merge: fail properly when we are in the middle of a conflicted mergeMiklos Vajna, Aug 23, 2008
  8. Junio C HamanoAug 23, 2008
  9. 1/2 merge: fix numerus bugs around "trivial merge" areaJunio C Hamano, Aug 23, 2008
  10. Junio C HamanoAug 24, 2008
  11. builtin-merge: avoid run_command_v_opt() for recursive and subtreeMiklos Vajna, Aug 28, 2008
  12. 2/2 unpack_trees(): protect the handcrafted in-core index from read_cache()Junio C Hamano, Aug 23, 2008
  13. Junio C HamanoAug 23, 2008
  14. Junio C HamanoAug 23, 2008
  15. Fix in-index merge.Miklos Vajna, Aug 23, 2008
  16. Paolo BonziniAug 23, 2008
  17. Junio C HamanoAug 23, 2008
  18. Junio C HamanoAug 23, 2008
  19. [RFH] two and half potential fixlets to the in-core index handlingJunio C Hamano, Aug 23, 2008
  20. Linus TorvaldsAug 23, 2008
  21. Junio C HamanoAug 23, 2008
  22. Junio C HamanoAug 23, 2008

Read the whole thread, see it on lore, or plain text.

$ cat FOOTERMessages come from the public archive at lore.kernel.org/git, fetched every hour. The front page is chosen and written each morning by an AI editor and can be wrong; the threads themselves are the record. About and API. For agents: an MCP server at https://gitlist.dev/mcp, and any thread, story or person page as Markdown by adding .md to its URL (or sending Accept: text/markdown). Details in /llms.txt.