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

[PATCH v2 4/7] builtin/clone: fix bundle URIs with mismatching object formats

From
Patrick Steinhardt <ps@pks.im>
Date
Dec 12, 2023, 07:00 UTC
Message-ID
<3bef564b57ec67ecb54ae8f24396e48ad5967582.1702361370.git.ps@pks.im>
In-Reply-To
<cover.1702361370.git.ps@pks.im>

We create the reference database in git-clone(1) quite early before connecting to the remote repository. Given that we do not yet know about the object format that the remote repository uses at that point in time the consequence is that the refdb may be initialized with the wrong object format.

This is not a problem in the context of the files backend as we do not encode the object format anywhere, and furthermore the only reference that we write between initializing the refdb and learning about the object format is the "HEAD" symref. It will become a problem though once we land the reftable backend, which indeed does require to know about the proper object format at the time of creation. We thus need to rearrange the logic in git-clone(1) so that we only initialize the refdb once we have learned about the actual object format.

As a first step, move listing of remote references to happen earlier, which also allow us to set up the hash algorithm of the repository earlier now. While we aim to execute this logic as late as possible until after most of the setup has happened already, detection of the object format and thus later the setup of the reference database must happen before any other logic that may spawn Git commands or otherwise these Git commands may not recognize the repository as such.

The first Git step where we expect the repository to be fully initalized is when we fetch bundles via bundle URIs. Funny enough, the comments there also state that "the_repository must match the cloned repo", which is indeed not necessarily the case for the hash algorithm right now. So in practice it is the right thing to detect the remote's object format before downloading bundle URIs anyway, and not doing so causes clones with bundle URIs to fail when the local default object format does not match the remote repository's format.

Unfortunately though, this creates a new issue: downloading bundles may take a long time, so if we list refs beforehand they might've grown stale meanwhile. It is not clear how to solve this issue except for a second reference listing though after we have downloaded the bundles, which may be an expensive thing to do.

Arguably though, it's preferable to have a staleness issue compared to being unable to clone a repository altogether.

Signed-off-by: Patrick Steinhardt <ps@pks.im>
---
 builtin/clone.c             | 48 ++++++++++++++++++-------------------
 t/t5558-clone-bundle-uri.sh | 18 ++++++++++++++
 2 files changed, 41 insertions(+), 25 deletions(-)
diff --git a/builtin/clone.c b/builtin/clone.c
index c6357af949..d188650881 100644
--- a/builtin/clone.c
+++ b/builtin/clone.c
@@ -1266,6 +1266,26 @@ int cmd_clone(int argc, const char **argv, const char *prefix)
 	if (transport->smart_options && !deepen && !filter_options.choice)
 		transport->smart_options->check_self_contained_and_connected = 1;
 
+	strvec_push(&transport_ls_refs_options.ref_prefixes, "HEAD");
+	refspec_ref_prefixes(&remote->fetch,
+			     &transport_ls_refs_options.ref_prefixes);
+	if (option_branch)
+		expand_ref_prefix(&transport_ls_refs_options.ref_prefixes,
+				  option_branch);
+	if (!option_no_tags)
+		strvec_push(&transport_ls_refs_options.ref_prefixes,
+			    "refs/tags/");
+
+	refs = transport_get_remote_refs(transport, &transport_ls_refs_options);
+
+	/*
+	 * Now that we know what algorithm the remote side is using, let's set
+	 * ours to the same thing.
+	 */
+	hash_algo = hash_algo_by_ptr(transport_get_hash_algo(transport));
+	initialize_repository_version(hash_algo, 1);
+	repo_set_hash_algo(the_repository, hash_algo);
+
 	/*
 	 * Before fetching from the remote, download and install bundle
 	 * data from the --bundle-uri option.
@@ -1281,24 +1301,7 @@ int cmd_clone(int argc, const char **argv, const char *prefix)
 				bundle_uri);
 		else if (has_heuristic)
 			git_config_set_gently("fetch.bundleuri", bundle_uri);
-	}
-
-	strvec_push(&transport_ls_refs_options.ref_prefixes, "HEAD");
-	refspec_ref_prefixes(&remote->fetch,
-			     &transport_ls_refs_options.ref_prefixes);
-	if (option_branch)
-		expand_ref_prefix(&transport_ls_refs_options.ref_prefixes,
-				  option_branch);
-	if (!option_no_tags)
-		strvec_push(&transport_ls_refs_options.ref_prefixes,
-			    "refs/tags/");
-
-	refs = transport_get_remote_refs(transport, &transport_ls_refs_options);
-
-	if (refs)
-		mapped_refs = wanted_peer_refs(refs, &remote->fetch);
-
-	if (!bundle_uri) {
+	} else {
 		/*
 		* Populate transport->got_remote_bundle_uri and
 		* transport->bundle_uri. We might get nothing.
@@ -1319,13 +1322,8 @@ int cmd_clone(int argc, const char **argv, const char *prefix)
 		}
 	}
 
-		/*
-		 * Now that we know what algorithm the remote side is using,
-		 * let's set ours to the same thing.
-		 */
-	hash_algo = hash_algo_by_ptr(transport_get_hash_algo(transport));
-	initialize_repository_version(hash_algo, 1);
-	repo_set_hash_algo(the_repository, hash_algo);
+	if (refs)
+		mapped_refs = wanted_peer_refs(refs, &remote->fetch);
 
 	if (mapped_refs) {
 		/*
diff --git a/t/t5558-clone-bundle-uri.sh b/t/t5558-clone-bundle-uri.sh
index 996a08e90c..1ca5f745e7 100755
--- a/t/t5558-clone-bundle-uri.sh
+++ b/t/t5558-clone-bundle-uri.sh
@@ -33,6 +33,15 @@ test_expect_success 'clone with path bundle' '
 	test_cmp expect actual
 '
 
+test_expect_success 'clone with path bundle and non-default hash' '
+	test_when_finished "rm -rf clone-path-non-default-hash" &&
+	GIT_DEFAULT_HASH=sha256 git clone --bundle-uri="clone-from/B.bundle" \
+		clone-from clone-path-non-default-hash &&
+	git -C clone-path-non-default-hash rev-parse refs/bundles/topic >actual &&
+	git -C clone-from rev-parse topic >expect &&
+	test_cmp expect actual
+'
+
 test_expect_success 'clone with file:// bundle' '
 	git clone --bundle-uri="file://$(pwd)/clone-from/B.bundle" \
 		clone-from clone-file &&
@@ -284,6 +293,15 @@ test_expect_success 'clone HTTP bundle' '
 	test_config -C clone-http log.excludedecoration refs/bundle/
 '
 
+test_expect_success 'clone HTTP bundle with non-default hash' '
+	test_when_finished "rm -rf clone-http-non-default-hash" &&
+	GIT_DEFAULT_HASH=sha256 git clone --bundle-uri="$HTTPD_URL/B.bundle" \
+		"$HTTPD_URL/smart/fetch.git" clone-http-non-default-hash &&
+	git -C clone-http-non-default-hash rev-parse refs/bundles/topic >actual &&
+	git -C clone-from rev-parse topic >expect &&
+	test_cmp expect actual
+'
+
 test_expect_success 'clone bundle list (HTTP, no heuristic)' '
 	test_when_finished rm -f trace*.txt &&
 

-- 
2.43.GIT
Previous: Patrick SteinhardtNext: Patrick Steinhardt
Message 29 of 33 in “clone: fix init of refdb with wrong object format”
  1. 0/7 clone: fix init of refdb with wrong object formatPatrick Steinhardt, Dec 6, 2023
  2. 1/7 setup: extract function to create the refdbPatrick Steinhardt, Dec 6, 2023
  3. Karthik NayakDec 6, 2023
  4. Patrick SteinhardtDec 7, 2023
  5. Junio C HamanoDec 8, 2023
  6. Patrick SteinhardtDec 11, 2023
  7. Karthik NayakDec 11, 2023
  8. 2/7 setup: allow skipping creation of the refdbPatrick Steinhardt, Dec 6, 2023
  9. 3/7 remote-curl: rediscover repository when fetching refsPatrick Steinhardt, Dec 6, 2023
  10. Junio C HamanoDec 8, 2023
  11. Patrick SteinhardtDec 11, 2023
  12. 4/7 builtin/clone: fix bundle URIs with mismatching object formatsPatrick Steinhardt, Dec 6, 2023
  13. Karthik NayakDec 6, 2023
  14. Patrick SteinhardtDec 7, 2023
  15. Junio C HamanoDec 8, 2023
  16. 5/7 builtin/clone: set up sparse checkout laterPatrick Steinhardt, Dec 6, 2023
  17. 6/7 builtin/clone: skip reading HEAD when retrieving remotePatrick Steinhardt, Dec 6, 2023
  18. 7/7 builtin/clone: create the refdb with the correct object formatPatrick Steinhardt, Dec 6, 2023
  19. Patrick SteinhardtDec 6, 2023
  20. Junio C HamanoDec 10, 2023
  21. Patrick SteinhardtDec 11, 2023
  22. Junio C HamanoDec 11, 2023
  23. Patrick SteinhardtDec 11, 2023
  24. brian m. carlsonDec 11, 2023
  25. 0/7 clone: fix init of refdb with wrong object formatPatrick Steinhardt, Dec 12, 2023
  26. 1/7 setup: extract function to create the refdbPatrick Steinhardt, Dec 12, 2023
  27. 2/7 setup: allow skipping creation of the refdbPatrick Steinhardt, Dec 12, 2023
  28. 3/7 remote-curl: rediscover repository when fetching refsPatrick Steinhardt, Dec 12, 2023
  29. 4/7 builtin/clone: fix bundle URIs with mismatching object formatsPatrick Steinhardt, Dec 12, 2023
  30. 5/7 builtin/clone: set up sparse checkout laterPatrick Steinhardt, Dec 12, 2023
  31. 6/7 builtin/clone: skip reading HEAD when retrieving remotePatrick Steinhardt, Dec 12, 2023
  32. 7/7 builtin/clone: create the refdb with the correct object formatPatrick Steinhardt, Dec 12, 2023
  33. Junio C HamanoDec 12, 2023

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.