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

Re: Trying to push into empty repo, get fatal: bad revision 'HEAD'

From
Junio C Hamano <gitster@pobox.com>
Date
Apr 1, 2015, 06:15 UTC
Message-ID
<xmqq4mp0nzla.fsf@gitster.dls.corp.google.com>
In-Reply-To
<xmqq8uecocjg.fsf@gitster.dls.corp.google.com>
Junio C Hamano <gitster@pobox.com> writes:
> A fix (or is it an enhancement) would probably look like this.
> ...

And this time with a pair of tests. It probably should be extended to make sure it fails when the pushed HEAD records paths that are floating in the working tree (as that would mean overwrting them), but it is getting late and I am lazy, so ... ;-)

-- >8 --
Subject: [PATCH] push-to-deploy: allow pushing into an unborn branch and updating it

Setting receive.denycurrentbranch to updateinstead and pushing into the current branch, when the working tree and the index is truly clean, is supposed to reset the working tree and the index to match the tree of the pushed commit. This did not work when pushing into an unborn branch.

The code that drives push-to-checkout hook needs no change, as the interface is defined so that hook can decide what to do when the push is coming to an unborn branch and take an appropriate action since the beginning.

Signed-off-by: Junio C Hamano <gitster@pobox.com>
---
 builtin/receive-pack.c | 21 ++++++++++++++++++-
 t/t5516-fetch-push.sh  | 55 +++++++++++++++++++++++++++++++++++++++++++++++++-
 2 files changed, 74 insertions(+), 2 deletions(-)
diff --git a/builtin/receive-pack.c b/builtin/receive-pack.c
index fc8ec9c..0c0a261 100644
--- a/builtin/receive-pack.c
+++ b/builtin/receive-pack.c
@@ -733,6 +733,22 @@ static int update_shallow_ref(struct command *cmd, struct shallow_info *si)
 	return 0;
 }
 
+/*
+ * NEEDSWORK: we should consolidate various implementions of "are we
+ * on an unborn branch?" test into one, and make the unified one more
+ * robust. !get_sha1() based check used here and elsewhere would not
+ * allow us to tell an unborn branch from corrupt ref, for example.
+ * For the purpose of fixing "deploy-to-update does not work when
+ * pushing into an empty repository" issue, this should suffice for
+ * now.
+ */
+static int head_has_history(void)
+{
+	unsigned char sha1[20];
+
+	return !get_sha1("HEAD", sha1);
+}
+
 static const char *push_to_deploy(unsigned char *sha1,
 				  struct argv_array *env,
 				  const char *work_tree)
@@ -745,7 +761,7 @@ static const char *push_to_deploy(unsigned char *sha1,
 	};
 	const char *diff_index[] = {
 		"diff-index", "--quiet", "--cached", "--ignore-submodules",
-		"HEAD", "--", NULL
+		NULL, "--", NULL
 	};
 	const char *read_tree[] = {
 		"read-tree", "-u", "-m", NULL, NULL
@@ -772,6 +788,9 @@ static const char *push_to_deploy(unsigned char *sha1,
 	if (run_command(&child))
 		return "Working directory has unstaged changes";
 
+	/* diff-index with either HEAD or an empty tree */
+	diff_index[4] = head_has_history() ? "HEAD" : EMPTY_TREE_SHA1_HEX;
+
 	child_process_init(&child);
 	child.argv = diff_index;
 	child.env = env->argv;
diff --git a/t/t5516-fetch-push.sh b/t/t5516-fetch-push.sh
index e4436c1..329d7d4 100755
--- a/t/t5516-fetch-push.sh
+++ b/t/t5516-fetch-push.sh
@@ -1430,8 +1430,22 @@ test_expect_success 'receive.denyCurrentBranch = updateInstead' '
 		test $(git -C .. rev-parse HEAD^^) = $(git rev-parse HEAD) &&
 		git diff --quiet &&
 		test fifth = "$(cat path3)"
-	)
+	) &&
 
+	# (5) push into void
+	rm -fr void &&
+	git init void &&
+	(
+		cd void &&
+		git config receive.denyCurrentBranch updateInstead
+	) &&
+	git push void master &&
+	(
+		cd void &&
+		test $(git -C .. rev-parse master) = $(git rev-parse HEAD) &&
+		git diff --quiet &&
+		git diff --cached --quiet
+	)
 '
 
 test_expect_success 'updateInstead with push-to-checkout hook' '
@@ -1494,6 +1508,45 @@ test_expect_success 'updateInstead with push-to-checkout hook' '
 		test "$(cat path5)" = irrelevant &&
 		test "$(git diff --name-only --cached HEAD)" = path5 &&
 		test $(git -C .. rev-parse HEAD) = $(git rev-parse HEAD)
+	) &&
+
+	# push into void
+	rm -fr void &&
+	git init void &&
+	(
+		cd void &&
+		git config receive.denyCurrentBranch updateInstead &&
+		write_script .git/hooks/push-to-checkout <<-\EOF
+		if git rev-parse --quiet --verify HEAD
+		then
+			has_head=yes
+			echo >&2 updating from $(git rev-parse HEAD)
+		else
+			has_head=no
+			echo >&2 pushing into void
+		fi
+		echo >&2 updating to "$1"
+
+		git update-index -q --refresh &&
+		case "$has_head" in
+		yes)
+			git read-tree -u -m HEAD "$1" ;;
+		no)
+			git read-tree -u -m "$1" ;;
+		esac || {
+			status=$?
+			echo >&2 read-tree failed
+			exit $status
+		}
+		EOF
+	) &&
+
+	git push void master &&
+	(
+		cd void &&
+		git diff --quiet &&
+		git diff --cached --quiet &&
+		test $(git -C .. rev-parse HEAD) = $(git rev-parse HEAD)
 	)
 '
 
-- 
2.4.0-rc0-179-ge750d7b
Previous: Junio C HamanoNext: Johannes Schindelin
Message 4 of 7 in “Trying to push into empty repo, get fatal: bad revision 'HEAD'”
  1. Samuel WilliamsMar 31, 2015
  2. Junio C HamanoApr 1, 2015
  3. Junio C HamanoApr 1, 2015
  4. Junio C HamanoApr 1, 2015
  5. Johannes SchindelinApr 1, 2015
  6. Junio C HamanoApr 1, 2015
  7. Johannes SchindelinApr 2, 2015

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.