scalar: make GVFS Protocol a forced choice

In the Office monorepo, we've recently had an uptick in issues with
`scalar clone`. These issues didn't make sense at first and seemed like
the users weren't using `microsoft/git` but instead the upstream
version's `scalar clone`. Instead of using GVFS cache servers, they were
attempting to use the Git protocol's partial clone (which times out).

It turns out that what's actually happening is that some network issue
is causing the connection with Azure DevOps to error out during the
`/gvfs/config` request. In the Git traces, we see the following error
during this request:

  (curl:56) Failure when receiving data from the peer [transient]

This isn't 100% of the time, but has increased enough to cause problems
for a variety of users.

The solution being proposed in this pull request is to remove the
fall-back mechanism and instead have an explicit choice to use the GVFS
protocol. To avoid significant disruption to Azure DevOps customers (the
vast majority of `microsoft/git` users who use `scalar clone` based on
my understanding), I added some inferring of a default value from the
clone URL.

This fallback mechanism was first implemented in the C# version of
Scalar in microsoft/scalar#339. This was an attempt to make the Scalar
client interesting to non-Azure DevOps customers, especially as GitHub
was about to launch the availability of partial clones. Now that the
`scalar` client is available upstream, users don't need the GVFS-enabled
version to get these benefits.

In addition, this will resolve #384 since those requests won't happen
against non-ADO URLs unless requested.

Signed-off-by: Derrick Stolee <[email protected]>

Author: derrickstolee

Documentation/scalar.adoc

@@ -116,6 +116,26 @@ clone), and `~/.scalarCache` on macOS.
 	Retrieve missing objects from the specified remote, which is expected to
 	understand the GVFS protocol.
 
+--[no-]gvfs-protocol::
+	When cloning from a `<url>` with either `dev.azure.com` or
+	`visualstudio.com` in the name, `scalar clone` will attempt to use the GVFS
+	Protocol to access Git objects, specifically from a cache server when
+	available, and will fail to clone if there is an error over that protocol.
+
+	To enable the GVFS Protocol regardless of the origin `<url>`, use
+	`--gvfs-protocol`. This will cause `scalar clone` to fail when the origin
+	server fails to provide a valid response to the `gvfs/config` endpoint.
+
+	To disable the GVFS Protocol, use `--no-gvfs-protocol` and `scalar clone`
+	will only use the Git protocol, starting with a partial clone. This can be
+	helpful if your `<url>` points to Azure Repos but the repository does not
+	have GVFS cache servers enabled. It is likely more efficient to use its
+	partial clone functionality through the Git protocol.
+
+	Previous versions of `scalar clone` could fall back to a partial clone over
+	the Git protocol if there is any issue gathering GVFS configuration
+	information from the origin server.
+
 List
 ~~~~
 

scalar.c

@@ -714,6 +714,8 @@ static int cmd_clone(int argc, const char **argv)
 	int src = 1, tags = 1, maintenance = 1;
 	const char *cache_server_url = NULL, *local_cache_root = NULL;
 	char *default_cache_server_url = NULL, *local_cache_root_abs = NULL;
+	int gvfs_protocol = -1;
+
 	struct option clone_options[] = {
 		OPT_STRING('b', "branch", &branch, N_("<branch>"),
 			   N_("branch to checkout after clone")),
@@ -728,6 +730,8 @@ static int cmd_clone(int argc, const char **argv)
 			 N_("specify if tags should be fetched during clone")),
 		OPT_BOOL(0, "maintenance", &maintenance,
 			 N_("specify if background maintenance should be enabled")),
+		OPT_BOOL(0, "gvfs-protocol", &gvfs_protocol,
+			 N_("force enable (or disable) the GVFS Protocol")),
 		OPT_STRING(0, "cache-server-url", &cache_server_url,
 			   N_("<url>"),
 			   N_("the url or friendly name of the cache server")),
@@ -747,7 +751,6 @@ static int cmd_clone(int argc, const char **argv)
 	char *enlistment = NULL, *dir = NULL;
 	struct strbuf buf = STRBUF_INIT;
 	int res;
-	int gvfs_protocol;
 
 	argc = parse_options(argc, argv, NULL, clone_options, clone_usage, 0);
 
@@ -860,8 +863,18 @@ static int cmd_clone(int argc, const char **argv)
 		goto cleanup;
 	}
 
-	gvfs_protocol = cache_server_url ||
-			supports_gvfs_protocol(url, &default_cache_server_url);
+	/* Is --[no-]gvfs-protocol unspecified? Infer from url. */
+	if (gvfs_protocol < 0) {
+		if (cache_server_url ||
+		    strstr(url, "dev.azure.com") ||
+		    strstr(url, "visualstudio.com"))
+			gvfs_protocol = 1;
+		else
+			gvfs_protocol = 0;
+	}
+
+	if (gvfs_protocol && !supports_gvfs_protocol(url, &default_cache_server_url))
+		die(_("failed to contact server via GVFS Protocol"));
 
 	if (gvfs_protocol) {
 		if ((res = init_shared_object_cache(url, local_cache_root)))

t/t9210-scalar.sh

@@ -393,7 +393,12 @@ test_expect_success '`scalar clone` with GVFS-enabled server' '
 	# We must set credential.interactive=true to bypass a setting
 	# in "scalar clone" that disables interactive credentials during
 	# an unattended command.
-	scalar -c credential.interactive=true clone --single-branch -- http://$HOST_PORT/ using-gvfs &&
+	GIT_TRACE2_EVENT="$(pwd)/clone-trace-with-gvfs" scalar \
+		-c credential.interactive=true \
+		clone --gvfs-protocol \
+		--single-branch -- http://$HOST_PORT/ using-gvfs &&
+
+	grep "GET/config(main)" <clone-trace-with-gvfs &&
 
 	: verify that the shared cache has been configured &&
 	cache_key="url_$(printf "%s" http://$HOST_PORT/ |
@@ -415,6 +420,24 @@ test_expect_success '`scalar clone` with GVFS-enabled server' '
 	)
 '
 
+test_expect_success '`scalar clone --no-gvfs-protocol` skips gvfs/config' '
+	# the fake cache server requires fake authentication &&
+	git config --global core.askPass true &&
+
+	# We must set credential.interactive=true to bypass a setting
+	# in "scalar clone" that disables interactive credentials during
+	# an unattended command.
+	GIT_TRACE2_EVENT="$(pwd)/clone-trace-no-gvfs" scalar \
+		-c credential.interactive=true \
+		clone --no-gvfs-protocol \
+		--single-branch -- http://$HOST_PORT/ skipping-gvfs &&
+
+	! grep "GET/config(main)" <clone-trace-no-gvfs &&
+	! git -C skipping-gvfs/src config core.gvfs &&
+
+	test_config -C skipping-gvfs/src remote.origin.partialclonefilter blob:none
+'
+
 test_expect_success '`scalar register` parallel to worktree is unsupported' '
 	git init test-repo/src &&
 	mkdir -p test-repo/out &&