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

Re: What's cooking in git.git (Aug 2008, #05; Tue, 19)

From
Junio C Hamano <gitster@pobox.com>
Date
Sep 2, 2008, 20:02 UTC
Message-ID
<7v1w02l3jc.fsf@gitster.siamese.dyndns.org>
In-Reply-To
<20080901010612.GC16514@genesis.frugalware.org>
Miklos Vajna <vmiklos@frugalware.org> writes:
> I tried to let cmd_revert() use merge_trees() only and not
> merge_recursive_generic(), but something is fishy with it.
> t3501-revert-cherry-pick passes fine, but t3404-rebase-interactive fail,
> becase once we have a conflict, git diff --name-status says 'M' for the
> given file and not 'U', which is obviously wrong.

I think this is because you are forgotting to write the index file out. You are calling "git commit -n" at the end if a cherry-pick/revert is clean, or giving the user back an unmerged state in the index otherwise. In either case, after merge_trees() gives you a potentially unmerged index back in-core, you need to write it to the $GIT_INDEX_FILE for later users.

Also notice that your output is awfully silent. You are forgetting to flush the buffered output that is kept in "obuf".

Please check what other extra things merge_recursive() function does that merge_trees() doesn't; the above two are what I've spotted by code inspection, but there may be others. It could be an indication of incorrect interface layering when the builtin was refactored.

I found it a bit disturbing that "index_only" and "call_depth" were not part of merge_options structure. The machinery is not yet meant to be reentrant, so we may rely the previous call to the function (either merge_recursive() or merge_trees() revert them to a sane initial value, but it is a bit unnerving, having to depend on that assumption.

I think your code happily dereferences NULL when picking the root commit is involved, although I do not think the failure in the test suite you saw is related to this omission.

Here is a partial fix to address the above issues I noticed on top of your version; untested.

 builtin-revert.c |   46 ++++++++++++++++++++++++++++++++++++----------
 1 files changed, 36 insertions(+), 10 deletions(-)
diff --git i/builtin-revert.c w/builtin-revert.c
index 3071518..41f3ca2 100644
--- i/builtin-revert.c
+++ w/builtin-revert.c
@@ -234,16 +234,27 @@ static int index_is_dirty(void)
 	return !!DIFF_OPT_TST(&rev.diffopt, HAS_CHANGES);
 }
 
+static struct tree *empty_tree(void)
+{
+	struct tree *tree = xcalloc(1, sizeof(struct tree));
+
+	tree->object.parsed = 1;
+	tree->object.type = OBJ_TREE;
+	pretend_sha1_file(NULL, 0, OBJ_TREE, tree->object.sha1);
+	return tree;
+}
+
 static int revert_or_cherry_pick(int argc, const char **argv)
 {
 	unsigned char head[20];
 	struct commit *base, *next, *parent;
-	int i;
+	int i, index_fd, clean;
 	char *oneline, *reencoded_message = NULL;
 	const char *message, *encoding;
 	const char *defmsg = xstrdup(git_path("MERGE_MSG"));
 	struct merge_options o;
-	struct tree *result;
+	struct tree *result, *next_tree, *base_tree, *head_tree;
+	static struct lock_file index_lock;
 
 	git_config(git_default_config, NULL);
 	me = action == REVERT ? "revert" : "cherry-pick";
@@ -254,6 +265,10 @@ static int revert_or_cherry_pick(int argc, const char **argv)
 	if (action == REVERT && !no_replay)
 		die("revert is incompatible with replay");
 
+	index_fd = hold_locked_index(&index_lock, 1);
+
+	if (read_cache() < 0)
+		die("git %s: failed to read the index", me);
 	if (no_commit) {
 		/*
 		 * We do not intend to commit immediately.  We just want to
@@ -266,12 +281,10 @@ static int revert_or_cherry_pick(int argc, const char **argv)
 	} else {
 		if (get_sha1("HEAD", head))
 			die ("You do not have a valid HEAD");
-		if (read_cache() < 0)
-			die("could not read the index");
 		if (index_is_dirty())
 			die ("Dirty index: cannot %s", me);
-		discard_cache();
 	}
+	discard_cache();
 
 	if (!commit->parents) {
 		if (action == REVERT)
@@ -305,6 +318,10 @@ static int revert_or_cherry_pick(int argc, const char **argv)
 		die ("Cannot get commit message for %s",
 				sha1_to_hex(commit->object.sha1));
 
+	if (parent && parse_commit(parent) < 0)
+		die("%s: cannot parse parent commit %s",
+		    me, sha1_to_hex(parent->object.sha1));
+
 	/*
 	 * "commit" is an existing commit.  We would want to apply
 	 * the difference it introduces since its first parent "prev"
@@ -351,11 +368,20 @@ static int revert_or_cherry_pick(int argc, const char **argv)
 	init_merge_options(&o);
 	o.branch1 = "HEAD";
 	o.branch2 = oneline;
-	parse_commit(next);
-	parse_commit(base);
-	if (!merge_trees(&o, lookup_commit_reference_gently(head, 0)->tree,
-				next->tree, base->tree, &result) ||
-		write_cache_as_tree(head, 0, NULL)) {
+
+	head_tree = parse_tree_indirect(head);
+	next_tree = next ? next->tree : empty_tree();
+	base_tree = base ? base->tree : empty_tree();
+
+	clean = merge_trees(&o,
+			    head_tree,
+			    next_tree, base_tree, &result);
+
+	if (write_cache(index_fd, active_cache, active_nr) ||
+	    commit_locked_index(&index_lock))
+		die("%s: Unable to write new index file", me);
+
+	if (!clean) {
 		add_to_msg("\nConflicts:\n\n");
 		for (i = 0; i < active_nr;) {
 			struct cache_entry *ce = active_cache[i++];
Previous: Miklos VajnaNext: Junio C Hamano
Message 19 of 20 in “What's cooking in git.git (Aug 2008, #05; Tue, 19)”
  1. Junio C HamanoAug 19, 2008
  2. Johannes SixtAug 19, 2008
  3. Andreas FärberAug 19, 2008
  4. Miklos VajnaAug 19, 2008
  5. Junio C HamanoAug 19, 2008
  6. Miklos VajnaAug 19, 2008
  7. Junio C HamanoAug 19, 2008
  8. Miklos VajnaAug 20, 2008
  9. merge-recursive: introduce merge_optionsMiklos Vajna, Aug 25, 2008
  10. Junio C HamanoAug 25, 2008
  11. merge-recursive: introduce merge_optionsMiklos Vajna, Aug 25, 2008
  12. Junio C HamanoAug 28, 2008
  13. merge-recursive: fix subtree mergeMiklos Vajna, Aug 30, 2008
  14. Junio C HamanoAug 30, 2008
  15. Junio C HamanoAug 30, 2008
  16. Miklos VajnaAug 31, 2008
  17. Miklos VajnaSep 1, 2008
  18. builtin-revert: use merge_recursive_generic()Miklos Vajna, Sep 1, 2008
  19. Junio C HamanoSep 2, 2008
  20. Junio C HamanoSep 2, 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.