Merge pull request #2376 from libgit2/cmn/remote-symref

Add support for the symref extension

11 files changed, 284 insertions, 98 deletions

diff --git a/include/git2/net.h b/include/git2/net.h
index e70ba1f71..a727696a2 100644
--- a/include/git2/net.h
+++ b/include/git2/net.h
@@ -41,6 +41,11 @@ struct git_remote_head {
 	git_oid oid;
 	git_oid loid;
 	char *name;
+	/**
+	 * If the server send a symref mapping for this ref, this will
+	 * point to the target.
+	 */
+	char *symref_target;
 };
 
 /**
diff --git a/include/git2/remote.h b/include/git2/remote.h
index 07cd2e7c6..28771ac42 100644
--- a/include/git2/remote.h
+++ b/include/git2/remote.h
@@ -623,6 +623,24 @@ GIT_EXTERN(int) git_remote_is_valid_name(const char *remote_name);
 */
 GIT_EXTERN(int) git_remote_delete(git_remote *remote);
 
+/**
+ * Retrieve the name of the remote's default branch
+ *
+ * The default branch of a repository is the branch which HEAD points
+ * to. If the remote does not support reporting this information
+ * directly, it performs the guess as git does; that is, if there are
+ * multiple branches which point to the same commit, the first one is
+ * chosen. If the master branch is a candidate, it wins.
+ *
+ * This function must only be called after connecting.
+ *
+ * @param out the buffern in which to store the reference name
+ * @param remote the remote
+ * @return 0, GIT_ENOTFOUND if the remote does not have any references
+ * or none of them point to HEAD's commit, or an error message.
+ */
+GIT_EXTERN(int) git_remote_default_branch(git_buf *out, git_remote *remote);
+
 /** @} */
 GIT_END_DECL
 #endif
diff --git a/src/clone.c b/src/clone.c
index 9ac9eb2a1..f19771c1e 100644
--- a/src/clone.c
+++ b/src/clone.c
@@ -105,54 +105,6 @@ static int create_tracking_branch(
 		git_reference_name(*branch));
 }
 
-struct head_info {
-	git_repository *repo;
-	git_oid remote_head_oid;
-	git_buf branchname;
-	const git_refspec *refspec;
-	bool found;
-};
-
-static int reference_matches_remote_head(
-	const char *reference_name,
-	void *payload)
-{
-	struct head_info *head_info = (struct head_info *)payload;
-	git_oid oid;
-	int error;
-
-	/* TODO: Should we guard against references
-	 * which name doesn't start with refs/heads/ ?
-	 */
-
-	error = git_reference_name_to_id(&oid, head_info->repo, reference_name);
-	if (error == GIT_ENOTFOUND) {
-		/* If the reference doesn't exists, it obviously cannot match the
-		 * expected oid. */
-		giterr_clear();
-		return 0;
-	}
-
-	if (!error && !git_oid__cmp(&head_info->remote_head_oid, &oid)) {
-		/* Determine the local reference name from the remote tracking one */
-		error = git_refspec_rtransform(
-			&head_info->branchname, head_info->refspec, reference_name);
-
-		if (!error &&
-			git_buf_len(&head_info->branchname) > 0 &&
-			!(error = git_buf_sets(
-				&head_info->branchname,
-				git_buf_cstr(&head_info->branchname) +
-				strlen(GIT_REFS_HEADS_DIR))))
-		{
-			head_info->found = true;
-			error = GIT_ITEROVER;
-		}
-	}
-
-	return error;
-}
-
 static int update_head_to_new_branch(
 	git_repository *repo,
 	const git_oid *target,
@@ -161,6 +113,10 @@ static int update_head_to_new_branch(
 	const char *reflog_message)
 {
 	git_reference *tracking_branch = NULL;
+
+	if (!git__prefixcmp(name, GIT_REFS_HEADS_DIR))
+		name += strlen(GIT_REFS_HEADS_DIR);
+
 	int error = create_tracking_branch(&tracking_branch, repo, target, name,
 			signature, reflog_message);
 
@@ -184,12 +140,13 @@ static int update_head_to_remote(
 		const git_signature *signature,
 		const char *reflog_message)
 {
-	int error = 0;
+	int error = 0, found_branch = 0;
 	size_t refs_len;
-	git_refspec dummy_spec;
+	git_refspec dummy_spec, *refspec;
 	const git_remote_head *remote_head, **refs;
-	struct head_info head_info;
+	const git_oid *remote_head_id;
 	git_buf remote_master_name = GIT_BUF_INIT;
+	git_buf branch = GIT_BUF_INIT;
 
 	if ((error = git_remote_ls(&refs, &refs_len, remote)) < 0)
 		return error;
@@ -199,63 +156,45 @@ static int update_head_to_remote(
 		return setup_tracking_config(
 			repo, "master", GIT_REMOTE_ORIGIN, GIT_REFS_HEADS_MASTER_FILE);
 
+	error = git_remote_default_branch(&branch, remote);
+	if (error == GIT_ENOTFOUND) {
+		git_buf_puts(&branch, GIT_REFS_HEADS_MASTER_FILE);
+	} else {
+		found_branch = 1;
+	}
+
 	/* Get the remote's HEAD. This is always the first ref in the list. */
 	remote_head = refs[0];
 	assert(remote_head);
 
-	memset(&head_info, 0, sizeof(head_info));
-	git_oid_cpy(&head_info.remote_head_oid, &remote_head->oid);
-	head_info.repo = repo;
-	head_info.refspec =
-		git_remote__matching_refspec(remote, GIT_REFS_HEADS_MASTER_FILE);
+	remote_head_id = &remote_head->oid;
+	refspec = git_remote__matching_refspec(remote, git_buf_cstr(&branch));
 
-	if (head_info.refspec == NULL) {
+	if (refspec == NULL) {
 		memset(&dummy_spec, 0, sizeof(git_refspec));
-		head_info.refspec = &dummy_spec;
+		refspec = &dummy_spec;
 	}
 
 	/* Determine the remote tracking reference name from the local master */
 	if ((error = git_refspec_transform(
 		&remote_master_name,
-		head_info.refspec,
-		GIT_REFS_HEADS_MASTER_FILE)) < 0)
+		refspec,
+		git_buf_cstr(&branch))) < 0)
 		return error;
 
-	/* Check to see if the remote HEAD points to the remote master */
-	error = reference_matches_remote_head(
-		git_buf_cstr(&remote_master_name), &head_info);
-	if (error < 0 && error != GIT_ITEROVER)
-		goto cleanup;
-
-	if (head_info.found) {
+	if (found_branch) {
 		error = update_head_to_new_branch(
 			repo,
-			&head_info.remote_head_oid,
-			git_buf_cstr(&head_info.branchname),
-			signature, reflog_message);
-		goto cleanup;
-	}
-
-	/* Not master. Check all the other refs. */
-	error = git_reference_foreach_name(
-		repo, reference_matches_remote_head, &head_info);
-	if (error < 0 && error != GIT_ITEROVER)
-		goto cleanup;
-
-	if (head_info.found) {
-		error = update_head_to_new_branch(
-			repo,
-			&head_info.remote_head_oid,
-			git_buf_cstr(&head_info.branchname),
+			remote_head_id,
+			git_buf_cstr(&branch),
 			signature, reflog_message);
 	} else {
 		error = git_repository_set_head_detached(
-			repo, &head_info.remote_head_oid, signature, reflog_message);
+			repo, remote_head_id, signature, reflog_message);
 	}
 
-cleanup:
 	git_buf_free(&remote_master_name);
-	git_buf_free(&head_info.branchname);
+	git_buf_free(&branch);
 	return error;
 }
 
diff --git a/src/remote.c b/src/remote.c
index bdc4791a9..f2e2e2f7a 100644
--- a/src/remote.c
+++ b/src/remote.c
@@ -1885,3 +1885,50 @@ int git_remote_delete(git_remote *remote)
 
 	return 0;
 }
+
+int git_remote_default_branch(git_buf *out, git_remote *remote)
+{
+	const git_remote_head **heads;
+	const git_remote_head *guess = NULL;
+	const git_oid *head_id;
+	size_t heads_len, i;
+	int error;
+
+	if ((error = git_remote_ls(&heads, &heads_len, remote)) < 0)
+		return error;
+
+	if (heads_len == 0)
+		return GIT_ENOTFOUND;
+
+	git_buf_sanitize(out);
+	/* the first one must be HEAD so if that has the symref info, we're done */
+	if (heads[0]->symref_target)
+		return git_buf_puts(out, heads[0]->symref_target);
+
+	/*
+	 * If there's no symref information, we have to look over them
+	 * and guess. We return the first match unless the master
+	 * branch is a candidate. Then we return the master branch.
+	 */
+	head_id = &heads[0]->oid;
+
+	for (i = 1; i < heads_len; i++) {
+		if (git_oid_cmp(head_id, &heads[i]->oid))
+			continue;
+
+		if (!guess) {
+			guess = heads[i];
+			continue;
+		}
+
+		if (!git__strcmp(GIT_REFS_HEADS_MASTER_FILE, heads[i]->name)) {
+			guess = heads[i];
+			break;
+		}
+	}
+
+	if (!guess)
+		return GIT_ENOTFOUND;
+
+	return git_buf_puts(out, guess->name);
+}
diff --git a/src/transports/local.c b/src/transports/local.c
index 2c17e6271..8d3619388 100644
--- a/src/transports/local.c
+++ b/src/transports/local.c
@@ -43,14 +43,19 @@ typedef struct {
 static int add_ref(transport_local *t, const char *name)
 {
 	const char peeled[] = "^{}";
-	git_oid head_oid;
+	git_reference *ref, *resolved;
 	git_remote_head *head;
+	git_oid obj_id;
 	git_object *obj = NULL, *target = NULL;
 	git_buf buf = GIT_BUF_INIT;
 	int error;
 
-	error = git_reference_name_to_id(&head_oid, t->repo, name);
+	if ((error = git_reference_lookup(&ref, t->repo, name)) < 0)
+		return error;
+
+	error = git_reference_resolve(&resolved, ref);
 	if (error < 0) {
+		git_reference_free(ref);
 		if (!strcmp(name, GIT_HEAD_FILE) && error == GIT_ENOTFOUND) {
 			/* This is actually okay.  Empty repos often have a HEAD that
 			 * points to a nonexistent "refs/heads/master". */
@@ -60,13 +65,22 @@ static int add_ref(transport_local *t, const char *name)
 		return error;
 	}
 
+	git_oid_cpy(&obj_id, git_reference_target(resolved));
+	git_reference_free(resolved);
+
 	head = git__calloc(1, sizeof(git_remote_head));
 	GITERR_CHECK_ALLOC(head);
 
 	head->name = git__strdup(name);
 	GITERR_CHECK_ALLOC(head->name);
 
-	git_oid_cpy(&head->oid, &head_oid);
+	git_oid_cpy(&head->oid, &obj_id);
+
+	if (git_reference_type(ref) == GIT_REF_SYMBOLIC) {
+		head->symref_target = git__strdup(git_reference_symbolic_target(ref));
+		GITERR_CHECK_ALLOC(head->symref_target);
+	}
+	git_reference_free(ref);
 
 	if ((error = git_vector_insert(&t->refs, head)) < 0) {
 		git__free(head->name);
@@ -176,7 +190,7 @@ static int path_from_url_or_path(git_buf *local_path_out, const char *url_or_pat
 
 /*
  * Try to open the url as a git directory. The direction doesn't
- * matter in this case because we're calulating the heads ourselves.
+ * matter in this case because we're calculating the heads ourselves.
  */
 static int local_connect(
 	git_transport *transport,
diff --git a/src/transports/smart.c b/src/transports/smart.c
index 7fb41f788..c9845bff9 100644
--- a/src/transports/smart.c
+++ b/src/transports/smart.c
@@ -7,6 +7,7 @@
 #include "git2.h"
 #include "smart.h"
 #include "refs.h"
+#include "refspec.h"
 
 static int git_smart__recv_cb(gitno_buffer *buf)
 {
@@ -63,7 +64,7 @@ static int git_smart__set_callbacks(
 	return 0;
 }
 
-int git_smart__update_heads(transport_smart *t)
+int git_smart__update_heads(transport_smart *t, git_vector *symrefs)
 {
 	size_t i;
 	git_pkt *pkt;
@@ -74,6 +75,25 @@ int git_smart__update_heads(transport_smart *t)
 		if (pkt->type != GIT_PKT_REF)
 			continue;
 
+		if (symrefs) {
+			git_refspec *spec;
+			git_buf buf = GIT_BUF_INIT;
+			size_t j;
+			int error;
+
+			git_vector_foreach(symrefs, j, spec) {
+				git_buf_clear(&buf);
+				if (git_refspec_src_matches(spec, ref->head.name) &&
+				    !(error = git_refspec_transform(&buf, spec, ref->head.name)))
+					ref->head.symref_target = git_buf_detach(&buf);
+			}
+
+			git_buf_free(&buf);
+
+			if (error < 0)
+				return error;
+		}
+
 		if (git_vector_insert(&t->heads, &ref->head) < 0)
 			return -1;
 	}
@@ -81,6 +101,19 @@ int git_smart__update_heads(transport_smart *t)
 	return 0;
 }
 
+static void free_symrefs(git_vector *symrefs)
+{
+	git_refspec *spec;
+	size_t i;
+
+	git_vector_foreach(symrefs, i, spec) {
+		git_refspec__free(spec);
+		git__free(spec);
+	}
+
+	git_vector_free(symrefs);
+}
+
 static int git_smart__connect(
 	git_transport *transport,
 	const char *url,
@@ -94,6 +127,7 @@ static int git_smart__connect(
 	int error;
 	git_pkt *pkt;
 	git_pkt_ref *first;
+	git_vector symrefs;
 	git_smart_service_t service;
 
 	if (git_smart__reset_stream(t, true) < 0)
@@ -147,8 +181,11 @@ static int git_smart__connect(
 
 	first = (git_pkt_ref *)git_vector_get(&t->refs, 0);
 
+	if ((error = git_vector_init(&symrefs, 1, NULL)) < 0)
+		return error;
+
 	/* Detect capabilities */
-	if (git_smart__detect_caps(first, &t->caps) < 0)
+	if (git_smart__detect_caps(first, &t->caps, &symrefs) < 0)
 		return -1;
 
 	/* If the only ref in the list is capabilities^{} with OID_ZERO, remove it */
@@ -159,7 +196,9 @@ static int git_smart__connect(
 	}
 
 	/* Keep a list of heads for _ls */
-	git_smart__update_heads(t);
+	git_smart__update_heads(t, &symrefs);
+
+	free_symrefs(&symrefs);
 
 	if (t->rpc && git_smart__reset_stream(t, false) < 0)
 		return -1;
diff --git a/src/transports/smart.h b/src/transports/smart.h
index a2b6b2a71..f1fc29520 100644
--- a/src/transports/smart.h
+++ b/src/transports/smart.h
@@ -23,6 +23,7 @@
 #define GIT_CAP_DELETE_REFS "delete-refs"
 #define GIT_CAP_REPORT_STATUS "report-status"
 #define GIT_CAP_THIN_PACK "thin-pack"
+#define GIT_CAP_SYMREF "symref"
 
 enum git_pkt_type {
 	GIT_PKT_CMD,
@@ -154,7 +155,7 @@ typedef struct {
 
 /* smart_protocol.c */
 int git_smart__store_refs(transport_smart *t, int flushes);
-int git_smart__detect_caps(git_pkt_ref *pkt, transport_smart_caps *caps);
+int git_smart__detect_caps(git_pkt_ref *pkt, transport_smart_caps *caps, git_vector *symrefs);
 int git_smart__push(git_transport *transport, git_push *push);
 
 int git_smart__negotiate_fetch(
@@ -174,7 +175,7 @@ int git_smart__download_pack(
 int git_smart__negotiation_step(git_transport *transport, void *data, size_t len);
 int git_smart__get_push_stream(transport_smart *t, git_smart_subtransport_stream **out);
 
-int git_smart__update_heads(transport_smart *t);
+int git_smart__update_heads(transport_smart *t, git_vector *symrefs);
 
 /* smart_pkt.c */
 int git_pkt_parse_line(git_pkt **head, const char *line, const char **out, size_t len);
diff --git a/src/transports/smart_pkt.c b/src/transports/smart_pkt.c
index e9376ae6f..b5f9d6dbe 100644
--- a/src/transports/smart_pkt.c
+++ b/src/transports/smart_pkt.c
@@ -433,6 +433,7 @@ void git_pkt_free(git_pkt *pkt)
 	if (pkt->type == GIT_PKT_REF) {
 		git_pkt_ref *p = (git_pkt_ref *) pkt;
 		git__free(p->head.name);
+		git__free(p->head.symref_target);
 	}
 
 	if (pkt->type == GIT_PKT_OK) {
diff --git a/src/transports/smart_protocol.c b/src/transports/smart_protocol.c
index 5dd6bab24..bab0cf113 100644
--- a/src/transports/smart_protocol.c
+++ b/src/transports/smart_protocol.c
@@ -78,7 +78,52 @@ int git_smart__store_refs(transport_smart *t, int flushes)
 	return flush;
 }
 
-int git_smart__detect_caps(git_pkt_ref *pkt, transport_smart_caps *caps)
+static int append_symref(const char **out, git_vector *symrefs, const char *ptr)
+{
+	int error;
+	const char *end;
+	git_buf buf = GIT_BUF_INIT;
+	git_refspec *mapping;
+
+	ptr += strlen(GIT_CAP_SYMREF);
+	if (*ptr != '=')
+		goto on_invalid;
+
+	ptr++;
+	if (!(end = strchr(ptr, ' ')) &&
+	    !(end = strchr(ptr, '\0')))
+		goto on_invalid;
+
+	if ((error = git_buf_put(&buf, ptr, end - ptr)) < 0)
+		return error;
+
+	/* symref mapping has refspec format */
+	mapping = git__malloc(sizeof(git_refspec));
+	GITERR_CHECK_ALLOC(mapping);
+
+	error = git_refspec__parse(mapping, git_buf_cstr(&buf), true);
+	git_buf_free(&buf);
+
+	/* if the error isn't OOM, then it's a parse error; let's use a nicer message */
+	if (error < 0) {
+		if (giterr_last()->klass != GITERR_NOMEMORY)
+			goto on_invalid;
+
+		return error;
+	}
+
+	if ((error = git_vector_insert(symrefs, mapping)) < 0)
+		return error;
+
+	*out = end;
+	return 0;
+
+on_invalid:
+	giterr_set(GITERR_NET, "remote sent invalid symref");
+	return -1;
+}
+
+int git_smart__detect_caps(git_pkt_ref *pkt, transport_smart_caps *caps, git_vector *symrefs)
 {
 	const char *ptr;
 
@@ -141,6 +186,15 @@ int git_smart__detect_caps(git_pkt_ref *pkt, transport_smart_caps *caps)
 			continue;
 		}
 
+		if (!git__prefixcmp(ptr, GIT_CAP_SYMREF)) {
+			int error;
+
+			if ((error = append_symref(&ptr, symrefs, ptr)) < 0)
+				return error;
+
+			continue;
+		}
+
 		/* We don't know this capability, so skip it */
 		ptr = strchr(ptr, ' ');
 	}
@@ -969,7 +1023,7 @@ int git_smart__push(git_transport *transport, git_push *push)
 		if (error < 0)
 			goto done;
 
-		error = git_smart__update_heads(t);
+		error = git_smart__update_heads(t, NULL);
 	}
 
 done:
diff --git a/tests/network/remote/defaultbranch.c b/tests/network/remote/defaultbranch.c
new file mode 100644
index 000000000..fa3a329db
--- /dev/null
+++ b/tests/network/remote/defaultbranch.c
@@ -0,0 +1,50 @@
+#include "clar_libgit2.h"
+#include "buffer.h"
+#include "refspec.h"
+#include "remote.h"
+
+static git_remote *g_remote;
+static git_repository *g_repo_a, *g_repo_b;
+
+void test_network_remote_defaultbranch__initialize(void)
+{
+	g_repo_a = cl_git_sandbox_init("testrepo.git");
+	cl_git_pass(git_repository_init(&g_repo_b, "repo-b.git", true));
+	cl_git_pass(git_remote_create(&g_remote, g_repo_b, "origin", git_repository_path(g_repo_a)));
+}
+
+void test_network_remote_defaultbranch__cleanup(void)
+{
+	git_remote_free(g_remote);
+	git_repository_free(g_repo_b);
+
+	cl_git_sandbox_cleanup();
+	cl_fixture_cleanup("repo-b.git");
+}
+
+static void assert_default_branch(const char *should)
+{
+	git_buf name = GIT_BUF_INIT;
+
+	cl_git_pass(git_remote_connect(g_remote, GIT_DIRECTION_FETCH));
+	cl_git_pass(git_remote_default_branch(&name, g_remote));
+	cl_assert_equal_s(should, name.ptr);
+	git_buf_free(&name);
+}
+
+void test_network_remote_defaultbranch__master(void)
+{
+	assert_default_branch("refs/heads/master");
+}
+
+void test_network_remote_defaultbranch__master_does_not_win(void)
+{
+	cl_git_pass(git_repository_set_head(g_repo_a, "refs/heads/not-good", NULL, NULL));
+	assert_default_branch("refs/heads/not-good");
+}
+
+void test_network_remote_defaultbranch__master_on_detached(void)
+{
+	cl_git_pass(git_repository_detach_head(g_repo_a, NULL, NULL));
+	assert_default_branch("refs/heads/master");
+}
diff --git a/tests/online/fetch.c b/tests/online/fetch.c
index c54ec5673..f03a6faa6 100644
--- a/tests/online/fetch.c
+++ b/tests/online/fetch.c
@@ -184,3 +184,21 @@ void test_online_fetch__ls_disconnected(void)
 
 	git_remote_free(remote);
 }
+
+void test_online_fetch__remote_symrefs(void)
+{
+	const git_remote_head **refs;
+	size_t refs_len;
+	git_remote *remote;
+
+	cl_git_pass(git_remote_create(&remote, _repo, "test",
+				"http://github.com/libgit2/TestGitRepository.git"));
+	cl_git_pass(git_remote_connect(remote, GIT_DIRECTION_FETCH));
+	git_remote_disconnect(remote);
+	cl_git_pass(git_remote_ls(&refs, &refs_len, remote));
+
+	cl_assert_equal_s("HEAD", refs[0]->name);
+	cl_assert_equal_s("refs/heads/master", refs[0]->symref_target);
+
+	git_remote_free(remote);
+}