From 8cd898f04fd89208442502af4849648f6af23a52 Mon Sep 17 00:00:00 2001
From: Johannes Schindelin <johannes.schindelin@gmx.de>
Date: Mon, 26 Apr 2021 22:08:18 +0200
Subject: [PATCH 01/29] scalar: implement a minimal JSON parser

No grown-up C project comes without their own JSON parser.

Just kidding!

We need to parse a JSON result when determining which cache server to
use. It would appear that searching for needles `"CacheServers":[`,
`"Url":"` and `"GlobalDefault":true` _happens_ to work right now, it is
fragile as it depends on no whitespace padding and on the order of the
fields remaining as-is.

Let's implement a super simple JSON parser (at the cost of being
slightly inefficient) for that purpose. To avoid allocating a ton of
memory, we implement a callback-based one. And to save on complexity,
let's not even bother validating the input properly (we will just go
ahead and instead rely on Azure Repos to produce correct JSON).

Note: An alternative would have been to use existing solutions such as
JSON-C, CentiJSON or JSMN. However, they are all a lot larger than the
current solution; The smallest, JSMN, which does not even provide parsed
string values (something we actually need) weighs in with 471 lines,
while we get away with 182 + 29 lines for the C and the header file,
respectively.

Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
---
 Makefile                            |   3 +-
 contrib/buildsystems/CMakeLists.txt |   2 +-
 json-parser.c                       | 182 ++++++++++++++++++++++++++++
 json-parser.h                       |  29 +++++
 4 files changed, 214 insertions(+), 2 deletions(-)
 create mode 100644 json-parser.c
 create mode 100644 json-parser.h

diff --git a/Makefile b/Makefile
index b9694fc500..2bbe46542b 100644
--- a/Makefile
+++ b/Makefile
@@ -2711,6 +2711,7 @@ GIT_OBJS += git.o
 .PHONY: git-objs
 git-objs: $(GIT_OBJS)
 
+SCALAR_OBJS := json-parser.o
 SCALAR_OBJS += scalar.o
 .PHONY: scalar-objs
 scalar-objs: $(SCALAR_OBJS)
@@ -2859,7 +2860,7 @@ $(REMOTE_CURL_PRIMARY): remote-curl.o http.o http-walker.o GIT-LDFLAGS $(GITLIBS
 	$(QUIET_LINK)$(CC) $(ALL_CFLAGS) -o $@ $(ALL_LDFLAGS) $(filter %.o,$^) \
 		$(CURL_LIBCURL) $(EXPAT_LIBEXPAT) $(LIBS)
 
-scalar$X: scalar.o GIT-LDFLAGS $(GITLIBS)
+scalar$X: $(SCALAR_OBJS) GIT-LDFLAGS $(GITLIBS)
 	$(QUIET_LINK)$(CC) $(ALL_CFLAGS) -o $@ $(ALL_LDFLAGS) \
 		$(filter %.o,$^) $(LIBS)
 
diff --git a/contrib/buildsystems/CMakeLists.txt b/contrib/buildsystems/CMakeLists.txt
index 41193f844c..a60cc1c9b9 100644
--- a/contrib/buildsystems/CMakeLists.txt
+++ b/contrib/buildsystems/CMakeLists.txt
@@ -797,7 +797,7 @@ target_link_libraries(git-sh-i18n--envsubst common-main)
 add_executable(git-shell ${CMAKE_SOURCE_DIR}/shell.c)
 target_link_libraries(git-shell common-main)
 
-add_executable(scalar ${CMAKE_SOURCE_DIR}/scalar.c)
+add_executable(scalar ${CMAKE_SOURCE_DIR}/scalar.c ${CMAKE_SOURCE_DIR}/json-parser.c)
 target_link_libraries(scalar common-main)
 
 if(CURL_FOUND)
diff --git a/json-parser.c b/json-parser.c
new file mode 100644
index 0000000000..30799e17dc
--- /dev/null
+++ b/json-parser.c
@@ -0,0 +1,182 @@
+#include "cache.h"
+#include "json-parser.h"
+
+static int reset_iterator(struct json_iterator *it)
+{
+	it->p = it->begin = it->json;
+	strbuf_release(&it->key);
+	strbuf_release(&it->string_value);
+	it->type = JSON_NULL;
+	return -1;
+}
+
+static int parse_json_string(struct json_iterator *it, struct strbuf *out)
+{
+	const char *begin = it->p;
+
+	if (*(it->p)++ != '"')
+		return error("expected double quote: '%.*s'", 5, begin),
+			reset_iterator(it);
+
+	strbuf_reset(&it->string_value);
+#define APPEND(c) strbuf_addch(out, c)
+	while (*it->p != '"') {
+		switch (*it->p) {
+		case '\0':
+			return error("incomplete string: '%s'", begin),
+				reset_iterator(it);
+		case '\\':
+			it->p++;
+			if (*it->p == '\\' || *it->p == '"')
+				APPEND(*it->p);
+			else if (*it->p == 'b')
+				APPEND(8);
+			else if (*it->p == 't')
+				APPEND(9);
+			else if (*it->p == 'n')
+				APPEND(10);
+			else if (*it->p == 'f')
+				APPEND(12);
+			else if (*it->p == 'r')
+				APPEND(13);
+			else if (*it->p == 'u') {
+				unsigned char binary[2];
+				int i;
+
+				if (hex_to_bytes(binary, it->p + 1, 2) < 0)
+					return error("invalid: '%.*s'",
+						     6, it->p - 1),
+						reset_iterator(it);
+				it->p += 4;
+
+				i = (binary[0] << 8) | binary[1];
+				if (i < 0x80)
+					APPEND(i);
+				else if (i < 0x0800) {
+					APPEND(0xc0 | ((i >> 6) & 0x1f));
+					APPEND(0x80 | (i & 0x3f));
+				} else if (i < 0x10000) {
+					APPEND(0xe0 | ((i >> 12) & 0x0f));
+					APPEND(0x80 | ((i >> 6) & 0x3f));
+					APPEND(0x80 | (i & 0x3f));
+				} else {
+					APPEND(0xf0 | ((i >> 18) & 0x07));
+					APPEND(0x80 | ((i >> 12) & 0x3f));
+					APPEND(0x80 | ((i >> 6) & 0x3f));
+					APPEND(0x80 | (i & 0x3f));
+				}
+			}
+			break;
+		default:
+			APPEND(*it->p);
+		}
+		it->p++;
+	}
+
+	it->end = it->p++;
+	return 0;
+}
+
+static void skip_whitespace(struct json_iterator *it)
+{
+	while (isspace(*it->p))
+		it->p++;
+}
+
+int iterate_json(struct json_iterator *it)
+{
+	skip_whitespace(it);
+	it->begin = it->p;
+
+	switch (*it->p) {
+	case '\0':
+		return reset_iterator(it), 0;
+	case 'n':
+		if (!starts_with(it->p, "null"))
+			return error("unexpected value: %.*s", 4, it->p),
+				reset_iterator(it);
+		it->type = JSON_NULL;
+		it->end = it->p = it->begin + 4;
+		break;
+	case 't':
+		if (!starts_with(it->p, "true"))
+			return error("unexpected value: %.*s", 4, it->p),
+				reset_iterator(it);
+		it->type = JSON_TRUE;
+		it->end = it->p = it->begin + 4;
+		break;
+	case 'f':
+		if (!starts_with(it->p, "false"))
+			return error("unexpected value: %.*s", 5, it->p),
+				reset_iterator(it);
+		it->type = JSON_FALSE;
+		it->end = it->p = it->begin + 5;
+		break;
+	case '-': case '.':
+	case '0': case '1': case '2': case '3': case '4':
+	case '5': case '6': case '7': case '8': case '9':
+		it->type = JSON_NUMBER;
+		it->end = it->p = it->begin + strspn(it->p, "-.0123456789");
+		break;
+	case '"':
+		it->type = JSON_STRING;
+		if (parse_json_string(it, &it->string_value) < 0)
+			return -1;
+		break;
+	case '[': {
+		const char *save = it->begin;
+		size_t key_offset = it->key.len;
+		int i = 0, res;
+
+		for (it->p++, skip_whitespace(it); *it->p != ']'; i++) {
+			strbuf_addf(&it->key, "[%d]", i);
+
+			if ((res = iterate_json(it)))
+				return reset_iterator(it), res;
+			strbuf_setlen(&it->key, key_offset);
+
+			skip_whitespace(it);
+			if (*it->p == ',')
+				it->p++;
+		}
+
+		it->type = JSON_ARRAY;
+		it->begin = save;
+		it->end = it->p;
+		it->p++;
+		break;
+	}
+	case '{': {
+		const char *save = it->begin;
+		size_t key_offset = it->key.len;
+		int res;
+
+		strbuf_addch(&it->key, '.');
+		for (it->p++, skip_whitespace(it); *it->p != '}'; ) {
+			strbuf_setlen(&it->key, key_offset + 1);
+			if (parse_json_string(it, &it->key) < 0)
+				return -1;
+			skip_whitespace(it);
+			if (*(it->p)++ != ':')
+				return error("expected colon: %.*s", 5, it->p),
+					reset_iterator(it);
+
+			if ((res = iterate_json(it)))
+				return res;
+
+			skip_whitespace(it);
+			if (*it->p == ',')
+				it->p++;
+		}
+		strbuf_setlen(&it->key, key_offset);
+
+		it->type = JSON_OBJECT;
+		it->begin = save;
+		it->end = it->p;
+		it->p++;
+		break;
+	}
+	}
+
+	return it->fn(it);
+}
diff --git a/json-parser.h b/json-parser.h
new file mode 100644
index 0000000000..ce1fdc5ee2
--- /dev/null
+++ b/json-parser.h
@@ -0,0 +1,29 @@
+#ifndef JSON_PARSER_H
+#define JSON_PARSER_H
+
+#include "strbuf.h"
+
+struct json_iterator {
+	const char *json, *p, *begin, *end;
+	struct strbuf key, string_value;
+	enum {
+		JSON_NULL = 0,
+		JSON_FALSE,
+		JSON_TRUE,
+		JSON_NUMBER,
+		JSON_STRING,
+		JSON_ARRAY,
+		JSON_OBJECT
+	} type;
+	int (*fn)(struct json_iterator *it);
+	void *fn_data;
+};
+#define JSON_ITERATOR_INIT(json_, fn_, fn_data_) { \
+	.json = json_, .p = json_, \
+	.key = STRBUF_INIT, .string_value = STRBUF_INIT, \
+	.fn = fn_, .fn_data = fn_data_ \
+}
+
+int iterate_json(struct json_iterator *it);
+
+#endif

From baa7fd8ac5e497da02eb0e182deaf3b79abd1b14 Mon Sep 17 00:00:00 2001
From: Johannes Schindelin <johannes.schindelin@gmx.de>
Date: Tue, 24 Aug 2021 21:05:42 +0200
Subject: [PATCH 02/29] scalar clone: support GVFS-enabled remote repositories

With this change, we come a big step closer to feature parity with
Scalar: this allows cloning from Azure Repos (which do not support
partial clones at time of writing).

We use the just-implemented JSON parser to parse the response we got
from the `gvfs/config` endpoint; Please note that this response might,
or might not, contain information about a cache server. The presence or
absence of said cache server, however, has nothing to do with the
ability to speak the GVFS protocol (but the presence of the
`gvfs/config` endpoint does that).

An alternative considered during the development of this patch was to
perform simple string matching instead of parsing the JSON-formatted
data; However, this would have been fragile, as the response contains
free-form text (e.g. the repository's description) which might contain
parts that would confuse a simple string matcher (but not a proper JSON
parser).

Note: we need to limit the re-try logic in `git clone` to handle only
the non-GVFS case: the call to `set_config()` to un-set the partial
clone settings would otherwise fail because those settings would not
exist in the GVFS protocol case. This will at least give us a clearer
reason why such a fetch fails.

Co-authored-by: Derrick Stolee <dstolee@microsoft.com>
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
---
 diagnose.c |   8 ++++
 scalar.c   | 126 +++++++++++++++++++++++++++++++++++++++++++++++++++--
 2 files changed, 131 insertions(+), 3 deletions(-)

diff --git a/diagnose.c b/diagnose.c
index 8f26569896..b4c2dc3c29 100644
--- a/diagnose.c
+++ b/diagnose.c
@@ -7,6 +7,7 @@
 #include "strvec.h"
 #include "object-store.h"
 #include "packfile.h"
+#include "config.h"
 
 struct archive_dir {
 	const char *path;
@@ -214,6 +215,7 @@ int create_diagnostics_archive(struct strbuf *zip_path, enum diagnose_mode mode)
 	struct strvec archiver_args = STRVEC_INIT;
 	char **argv_copy = NULL;
 	int stdout_fd = -1, archiver_fd = -1;
+	char *cache_server_url = NULL;
 	struct strbuf buf = STRBUF_INIT;
 	int res, i;
 	struct archive_dir archive_dirs[] = {
@@ -249,6 +251,11 @@ int create_diagnostics_archive(struct strbuf *zip_path, enum diagnose_mode mode)
 	get_version_info(&buf, 1);
 
 	strbuf_addf(&buf, "Repository root: %s\n", the_repository->worktree);
+
+	git_config_get_string("gvfs.cache-server", &cache_server_url);
+	strbuf_addf(&buf, "Cache Server: %s\n\n",
+		    cache_server_url ? cache_server_url : "None");
+
 	get_disk_info(&buf);
 	write_or_die(stdout_fd, buf.buf, buf.len);
 	strvec_pushf(&archiver_args,
@@ -306,6 +313,7 @@ diagnose_cleanup:
 	free(argv_copy);
 	strvec_clear(&archiver_args);
 	strbuf_release(&buf);
+	free(cache_server_url);
 
 	return res;
 }
diff --git a/scalar.c b/scalar.c
index dc266cdb35..f53130cb48 100644
--- a/scalar.c
+++ b/scalar.c
@@ -14,6 +14,7 @@
 #include "dir.h"
 #include "packfile.h"
 #include "help.h"
+#include "json-parser.h"
 
 static void setup_enlistment_directory(int argc, const char **argv,
 				       const char * const *usagestr,
@@ -331,6 +332,80 @@ static int set_config(const char *fmt, ...)
 	return res;
 }
 
+/* Find N for which .CacheServers[N].GlobalDefault == true */
+static int get_cache_server_index(struct json_iterator *it)
+{
+	const char *p;
+	char *q;
+	long l;
+
+	if (it->type == JSON_TRUE &&
+	    skip_iprefix(it->key.buf, ".CacheServers[", &p) &&
+	    (l = strtol(p, &q, 10)) >= 0 && p != q &&
+	    !strcasecmp(q, "].GlobalDefault")) {
+		*(long *)it->fn_data = l;
+		return 1;
+	}
+
+	return 0;
+}
+
+struct cache_server_url_data {
+	char *key, *url;
+};
+
+/* Get .CacheServers[N].Url */
+static int get_cache_server_url(struct json_iterator *it)
+{
+	struct cache_server_url_data *data = it->fn_data;
+
+	if (it->type == JSON_STRING &&
+	    !strcasecmp(data->key, it->key.buf)) {
+		data->url = strbuf_detach(&it->string_value, NULL);
+		return 1;
+	}
+
+	return 0;
+}
+
+/*
+ * If `cache_server_url` is `NULL`, print the list to `stdout`.
+ *
+ * Since `gvfs-helper` requires a Git directory, this _must_ be run in
+ * a worktree.
+ */
+static int supports_gvfs_protocol(const char *url, char **cache_server_url)
+{
+	struct child_process cp = CHILD_PROCESS_INIT;
+	struct strbuf out = STRBUF_INIT;
+
+	cp.git_cmd = 1;
+	strvec_pushl(&cp.args, "gvfs-helper", "--remote", url, "config", NULL);
+	if (!pipe_command(&cp, NULL, 0, &out, 512, NULL, 0)) {
+		long l = 0;
+		struct json_iterator it =
+			JSON_ITERATOR_INIT(out.buf, get_cache_server_index, &l);
+		struct cache_server_url_data data = { .url = NULL };
+
+		if (iterate_json(&it) < 0) {
+			strbuf_release(&out);
+			return error("JSON parse error");
+		}
+		data.key = xstrfmt(".CacheServers[%ld].Url", l);
+		it.fn = get_cache_server_url;
+		it.fn_data = &data;
+		if (iterate_json(&it) < 0) {
+			strbuf_release(&out);
+			return error("JSON parse error");
+		}
+		*cache_server_url = data.url;
+		free(data.key);
+		return 1;
+	}
+	strbuf_release(&out);
+	return 0; /* error out quietly */
+}
+
 static char *remote_default_branch(const char *url)
 {
 	struct child_process cp = CHILD_PROCESS_INIT;
@@ -432,6 +507,8 @@ static int cmd_clone(int argc, const char **argv)
 {
 	const char *branch = NULL;
 	int full_clone = 0, single_branch = 0, show_progress = isatty(2);
+	const char *cache_server_url = NULL;
+	char *default_cache_server_url = NULL;
 	struct option clone_options[] = {
 		OPT_STRING('b', "branch", &branch, N_("<branch>"),
 			   N_("branch to checkout after clone")),
@@ -440,6 +517,9 @@ static int cmd_clone(int argc, const char **argv)
 		OPT_BOOL(0, "single-branch", &single_branch,
 			 N_("only download metadata for the branch that will "
 			    "be checked out")),
+		OPT_STRING(0, "cache-server-url", &cache_server_url,
+			   N_("<url>"),
+			   N_("the url or friendly name of the cache server")),
 		OPT_END(),
 	};
 	const char * const clone_usage[] = {
@@ -450,6 +530,7 @@ 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);
 
@@ -512,13 +593,46 @@ static int cmd_clone(int argc, const char **argv)
 	    set_config("remote.origin.fetch="
 		       "+refs/heads/%s:refs/remotes/origin/%s",
 		       single_branch ? branch : "*",
-		       single_branch ? branch : "*") ||
-	    set_config("remote.origin.promisor=true") ||
-	    set_config("remote.origin.partialCloneFilter=blob:none")) {
+		       single_branch ? branch : "*")) {
 		res = error(_("could not configure remote in '%s'"), dir);
 		goto cleanup;
 	}
 
+	if (set_config("credential.https://dev.azure.com.useHttpPath=true")) {
+		res = error(_("could not configure credential.useHttpPath"));
+		goto cleanup;
+	}
+
+	gvfs_protocol = cache_server_url ||
+			supports_gvfs_protocol(url, &default_cache_server_url);
+
+	if (gvfs_protocol) {
+		if (!cache_server_url)
+			cache_server_url = default_cache_server_url;
+		if (set_config("core.useGVFSHelper=true") ||
+		    set_config("core.gvfs=150") ||
+		    set_config("http.version=HTTP/1.1")) {
+			res = error(_("could not turn on GVFS helper"));
+			goto cleanup;
+		}
+		if (cache_server_url &&
+		    set_config("gvfs.cache-server=%s", cache_server_url)) {
+			res = error(_("could not configure cache server"));
+			goto cleanup;
+		}
+		if (cache_server_url)
+			fprintf(stderr, "Cache server URL: %s\n",
+				cache_server_url);
+	} else {
+		if (set_config("core.useGVFSHelper=false") ||
+		    set_config("remote.origin.promisor=true") ||
+		    set_config("remote.origin.partialCloneFilter=blob:none")) {
+			res = error(_("could not configure partial clone in "
+				      "'%s'"), dir);
+			goto cleanup;
+		}
+	}
+
 	if (!full_clone &&
 	    (res = run_git("sparse-checkout", "init", "--cone", NULL)))
 		goto cleanup;
@@ -529,6 +643,11 @@ static int cmd_clone(int argc, const char **argv)
 	if ((res = run_git("fetch", "--quiet",
 				show_progress ? "--progress" : "--no-progress",
 				"origin", NULL))) {
+		if (gvfs_protocol) {
+			res = error(_("failed to prefetch commits and trees"));
+			goto cleanup;
+		}
+
 		warning(_("partial clone failed; attempting full clone"));
 
 		if (set_config("remote.origin.promisor") ||
@@ -561,6 +680,7 @@ cleanup:
 	free(enlistment);
 	free(dir);
 	strbuf_release(&buf);
+	free(default_cache_server_url);
 	return res;
 }
 

From 89f76e0bdb2fc972ca0fa47161783fb84ee3cc68 Mon Sep 17 00:00:00 2001
From: Johannes Schindelin <johannes.schindelin@gmx.de>
Date: Fri, 16 Apr 2021 19:47:05 +0200
Subject: [PATCH 03/29] test-gvfs-protocol: also serve smart protocol

This comes in handy, as we want to verify that `scalar clone` also works
against a GVFS-enabled remote repository.

Note that we have to set `MSYS2_ENV_CONV_EXCL` to prevent MSYS2 from
mangling `PATH_TRANSLATED`: The value _does_ look like a Unix-style
path, but no, MSYS2 must not be allowed to convert that into a Windows
path: `http-backend` needs it in the unmodified form. (The MSYS2 runtime
comes in when `git` is run via `bin-wrappers/git`, which is a shell
script.)

Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
---
 t/helper/test-gvfs-protocol.c | 49 +++++++++++++++++++++++++++++++++++
 1 file changed, 49 insertions(+)

diff --git a/t/helper/test-gvfs-protocol.c b/t/helper/test-gvfs-protocol.c
index 97b3d363bf..a437248a10 100644
--- a/t/helper/test-gvfs-protocol.c
+++ b/t/helper/test-gvfs-protocol.c
@@ -1481,6 +1481,8 @@ done:
 
 static enum worker_result dispatch(struct req *req)
 {
+	static regex_t *smart_http_regex;
+	static int initialized;
 	const char *method;
 	enum worker_result wr;
 
@@ -1529,6 +1531,53 @@ static enum worker_result dispatch(struct req *req)
 			return do__gvfs_prefetch__get(req);
 	}
 
+	if (!initialized) {
+		smart_http_regex = xmalloc(sizeof(*smart_http_regex));
+		if (regcomp(smart_http_regex, "^/(HEAD|info/refs|"
+			    "objects/info/[^/]+|git-(upload|receive)-pack)$",
+			    REG_EXTENDED)) {
+			warning("could not compile smart HTTP regex");
+			smart_http_regex = NULL;
+		}
+		initialized = 1;
+	}
+
+	if (smart_http_regex &&
+	    !regexec(smart_http_regex, req->uri_base.buf, 0, NULL, 0)) {
+		const char *ok = "HTTP/1.1 200 OK\r\n";
+		struct child_process cp = CHILD_PROCESS_INIT;
+		int i, res;
+
+		if (write(1, ok, strlen(ok)) < 0)
+			return error(_("could not send '%s'"), ok);
+
+		strvec_pushf(&cp.env, "REQUEST_METHOD=%s", method);
+		strvec_pushf(&cp.env, "PATH_TRANSLATED=%s",
+			     req->uri_base.buf);
+		/* Prevent MSYS2 from "converting to a Windows path" */
+		strvec_pushf(&cp.env,
+			     "MSYS2_ENV_CONV_EXCL=PATH_TRANSLATED");
+		strvec_push(&cp.env, "SERVER_PROTOCOL=HTTP/1.1");
+		if (req->quest_args.len)
+			strvec_pushf(&cp.env, "QUERY_STRING=%s",
+				     req->quest_args.buf);
+		for (i = 0; i < req->header_list.nr; i++) {
+			const char *header = req->header_list.items[i].string;
+			if (!strncasecmp("Content-Type: ", header, 14))
+				strvec_pushf(&cp.env, "CONTENT_TYPE=%s",
+					     header + 14);
+			else if (!strncasecmp("Content-Length: ", header, 16))
+				strvec_pushf(&cp.env, "CONTENT_LENGTH=%s",
+					     header + 16);
+		}
+		cp.git_cmd = 1;
+		strvec_push(&cp.args, "http-backend");
+		res = run_command(&cp);
+		close(1);
+		close(0);
+		return !!res;
+	}
+
 	return send_http_error(1, 501, "Not Implemented", -1,
 			       WR_OK | WR_HANGUP);
 }

From 390c135a0786d5e0f799884183ea830c2bcd9d8d Mon Sep 17 00:00:00 2001
From: Johannes Schindelin <johannes.schindelin@gmx.de>
Date: Mon, 26 Apr 2021 17:31:40 +0200
Subject: [PATCH 04/29] gvfs-helper: add the `endpoint` command

We already have the `config` command that accesses the `gvfs/config`
endpoint.

To implement `scalar`, we also need to be able to access the `vsts/info`
endpoint. Let's add a command to do precisely that.

Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
---
 gvfs-helper.c | 61 +++++++++++++++++++++++++++++++++++++++++++++------
 1 file changed, 54 insertions(+), 7 deletions(-)

diff --git a/gvfs-helper.c b/gvfs-helper.c
index ce89edc33c..e8e6fbbdd9 100644
--- a/gvfs-helper.c
+++ b/gvfs-helper.c
@@ -202,6 +202,12 @@
 //            [2] Documentation/technical/long-running-process-protocol.txt
 //            [3] See GIT_TRACE_PACKET
 //
+//     endpoint
+//
+//            Fetch the given endpoint from the main Git server (specifying
+//            `gvfs/config` as endpoint is idempotent to the `config`
+//            command mentioned above).
+//
 //////////////////////////////////////////////////////////////////
 
 #include "cache.h"
@@ -3110,18 +3116,20 @@ static void do_req__with_fallback(const char *url_component,
  *
  * Return server's response buffer.  This is probably a raw JSON string.
  */
-static void do__http_get__gvfs_config(struct gh__response_status *status,
-				      struct strbuf *config_data)
+static void do__http_get__simple_endpoint(struct gh__response_status *status,
+					  struct strbuf *response,
+					  const char *endpoint,
+					  const char *tr2_label)
 {
 	struct gh__request_params params = GH__REQUEST_PARAMS_INIT;
 
-	strbuf_addstr(&params.tr2_label, "GET/config");
+	strbuf_addstr(&params.tr2_label, tr2_label);
 
 	params.b_is_post = 0;
 	params.b_write_to_file = 0;
 	/* cache-servers do not handle gvfs/config REST calls */
 	params.b_permit_cache_server_if_defined = 0;
-	params.buffer = config_data;
+	params.buffer = response;
 	params.objects_mode = GH__OBJECTS_MODE__NONE;
 
 	params.object_count = 1; /* a bit of a lie */
@@ -3143,15 +3151,22 @@ static void do__http_get__gvfs_config(struct gh__response_status *status,
 		 * see any need to report progress on the upload side of
 		 * the GET.  So just report progress on the download side.
 		 */
-		strbuf_addstr(&params.progress_base_phase3_msg,
-			      "Receiving gvfs/config");
+		strbuf_addf(&params.progress_base_phase3_msg,
+			    "Receiving %s", endpoint);
 	}
 
-	do_req__with_fallback("gvfs/config", &params, status);
+	do_req__with_fallback(endpoint, &params, status);
 
 	gh__request_params__release(&params);
 }
 
+static void do__http_get__gvfs_config(struct gh__response_status *status,
+				      struct strbuf *config_data)
+{
+	do__http_get__simple_endpoint(status, config_data, "gvfs/config",
+				      "GET/config");
+}
+
 static void setup_gvfs_objects_progress(struct gh__request_params *params,
 					unsigned long num, unsigned long den)
 {
@@ -3596,6 +3611,35 @@ static enum gh__error_code do_sub_cmd__config(int argc, const char **argv)
 	return ec;
 }
 
+static enum gh__error_code do_sub_cmd__endpoint(int argc, const char **argv)
+{
+	struct gh__response_status status = GH__RESPONSE_STATUS_INIT;
+	struct strbuf data = STRBUF_INIT;
+	enum gh__error_code ec = GH__ERROR_CODE__OK;
+	const char *endpoint;
+
+	if (argc != 2)
+		return GH__ERROR_CODE__ERROR;
+	endpoint = argv[1];
+
+	trace2_cmd_mode(endpoint);
+
+	finish_init(0);
+
+	do__http_get__simple_endpoint(&status, &data, endpoint, endpoint);
+	ec = status.ec;
+
+	if (ec == GH__ERROR_CODE__OK)
+		printf("%s\n", data.buf);
+	else
+		error("config: %s", status.error_message.buf);
+
+	gh__response_status__release(&status);
+	strbuf_release(&data);
+
+	return ec;
+}
+
 /*
  * Read a list of objects from stdin and fetch them as a series of
  * single object HTTP GET requests.
@@ -4087,6 +4131,9 @@ static enum gh__error_code do_sub_cmd(int argc, const char **argv)
 	if (!strcmp(argv[0], "config"))
 		return do_sub_cmd__config(argc, argv);
 
+	if (!strcmp(argv[0], "endpoint"))
+		return do_sub_cmd__endpoint(argc, argv);
+
 	if (!strcmp(argv[0], "prefetch"))
 		return do_sub_cmd__prefetch(argc, argv);
 

From d3cb241ee8f460896eccb21149438d468d4361d2 Mon Sep 17 00:00:00 2001
From: Johannes Schindelin <johannes.schindelin@gmx.de>
Date: Sat, 15 May 2021 00:04:20 +0200
Subject: [PATCH 05/29] dir_inside_of(): handle directory separators correctly

On Windows, both the forward slash and the backslash are directory
separators. Which means that `a\b\c` really is inside `a/b`. Therefore,
we need to special-case the directory separators in the helper function
`cmp_icase()` that is used in the loop in `dir_inside_of()`.

Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
---
 dir.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/dir.c b/dir.c
index 01329773c1..c89f17618c 100644
--- a/dir.c
+++ b/dir.c
@@ -3127,6 +3127,8 @@ static int cmp_icase(char a, char b)
 {
 	if (a == b)
 		return 0;
+	if (is_dir_sep(a))
+		return is_dir_sep(b) ? 0 : -1;
 	if (ignore_case)
 		return toupper(a) - toupper(b);
 	return a - b;

From 173785c0a36b350503e2812b8fabcd96b8c6c7b9 Mon Sep 17 00:00:00 2001
From: Johannes Schindelin <johannes.schindelin@gmx.de>
Date: Thu, 6 May 2021 14:35:12 +0200
Subject: [PATCH 06/29] scalar: disable authentication in unattended mode

Modified to remove call to is_unattended() that has not been implemented
yet.

Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
---
 t/t9210-scalar.sh | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/t/t9210-scalar.sh b/t/t9210-scalar.sh
index 4432a30d10..543f2e1d3a 100755
--- a/t/t9210-scalar.sh
+++ b/t/t9210-scalar.sh
@@ -7,6 +7,13 @@ test_description='test the `scalar` command'
 GIT_TEST_MAINT_SCHEDULER="crontab:test-tool crontab cron.txt,launchctl:true,schtasks:true"
 export GIT_TEST_MAINT_SCHEDULER
 
+# Do not write any files outside the trash directory
+Scalar_UNATTENDED=1
+export Scalar_UNATTENDED
+
+GIT_ASKPASS=true
+export GIT_ASKPASS
+
 test_expect_success 'scalar shows a usage' '
 	test_expect_code 129 scalar -h
 '

From ba664115e34fa1c9e9498ff08d7c08e38b8410b1 Mon Sep 17 00:00:00 2001
From: Johannes Schindelin <johannes.schindelin@gmx.de>
Date: Mon, 3 May 2021 23:41:28 +0200
Subject: [PATCH 07/29] scalar: do initialize `gvfs.sharedCache`

This finalizes the port of the `QueryVstsInfo()` function: we already
taught `gvfs-helper` to access the `vsts/info` endpoint on demand, we
implemented proper JSON parsing, and now it is time to hook it all up.

To that end, we also provide a default local cache root directory. It
works the same way as the .NET version of Scalar: it uses

    C:\scalarCache on Windows,

    ~/.scalarCache/ on macOS and

    ~/.cache/scalar on Linux

Modified to include call to is_unattended() that was removed from a
previous commit.

Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
---
 Documentation/scalar.txt |  15 +++-
 diagnose.c               |   9 +-
 scalar.c                 | 182 ++++++++++++++++++++++++++++++++++++++-
 3 files changed, 200 insertions(+), 6 deletions(-)

diff --git a/Documentation/scalar.txt b/Documentation/scalar.txt
index f33436c7f6..e7f6d39397 100644
--- a/Documentation/scalar.txt
+++ b/Documentation/scalar.txt
@@ -8,7 +8,9 @@ scalar - A tool for managing large Git repositories
 SYNOPSIS
 --------
 [verse]
-scalar clone [--single-branch] [--branch <main-branch>] [--full-clone] <url> [<enlistment>]
+scalar clone [--single-branch] [--branch <main-branch>] [--full-clone]
+	[--local-cache-path <path>] [--cache-server-url <url>]
+	<url> [<enlistment>]
 scalar list
 scalar register [<enlistment>]
 scalar unregister [<enlistment>]
@@ -84,6 +86,17 @@ cloning. If the HEAD at the remote did not point at any branch when
 	A sparse-checkout is initialized by default. This behavior can be
 	turned off via `--full-clone`.
 
+--local-cache-path <path>::
+    Override the path to the local cache root directory; Pre-fetched objects
+    are stored into a repository-dependent subdirectory of that path.
++
+The default is `<drive>:\.scalarCache` on Windows (on the same drive as the
+clone), and `~/.scalarCache` on macOS.
+
+--cache-server-url <url>::
+    Retrieve missing objects from the specified remote, which is expected to
+    understand the GVFS protocol.
+
 List
 ~~~~
 
diff --git a/diagnose.c b/diagnose.c
index b4c2dc3c29..c6839da136 100644
--- a/diagnose.c
+++ b/diagnose.c
@@ -215,7 +215,7 @@ int create_diagnostics_archive(struct strbuf *zip_path, enum diagnose_mode mode)
 	struct strvec archiver_args = STRVEC_INIT;
 	char **argv_copy = NULL;
 	int stdout_fd = -1, archiver_fd = -1;
-	char *cache_server_url = NULL;
+	char *cache_server_url = NULL, *shared_cache = NULL;
 	struct strbuf buf = STRBUF_INIT;
 	int res, i;
 	struct archive_dir archive_dirs[] = {
@@ -253,8 +253,10 @@ int create_diagnostics_archive(struct strbuf *zip_path, enum diagnose_mode mode)
 	strbuf_addf(&buf, "Repository root: %s\n", the_repository->worktree);
 
 	git_config_get_string("gvfs.cache-server", &cache_server_url);
-	strbuf_addf(&buf, "Cache Server: %s\n\n",
-		    cache_server_url ? cache_server_url : "None");
+	git_config_get_string("gvfs.sharedCache", &shared_cache);
+	strbuf_addf(&buf, "Cache Server: %s\nLocal Cache: %s\n\n",
+		    cache_server_url ? cache_server_url : "None",
+		    shared_cache ? shared_cache : "None");
 
 	get_disk_info(&buf);
 	write_or_die(stdout_fd, buf.buf, buf.len);
@@ -314,6 +316,7 @@ diagnose_cleanup:
 	strvec_clear(&archiver_args);
 	strbuf_release(&buf);
 	free(cache_server_url);
+	free(shared_cache);
 
 	return res;
 }
diff --git a/scalar.c b/scalar.c
index f53130cb48..f11db89cba 100644
--- a/scalar.c
+++ b/scalar.c
@@ -16,6 +16,10 @@
 #include "help.h"
 #include "json-parser.h"
 
+static int is_unattended(void) {
+	return git_env_bool("Scalar_UNATTENDED", 0);
+}
+
 static void setup_enlistment_directory(int argc, const char **argv,
 				       const char * const *usagestr,
 				       const struct option *options,
@@ -100,6 +104,19 @@ static int run_git(const char *arg, ...)
 	return res;
 }
 
+static const char *ensure_absolute_path(const char *path, char **absolute)
+{
+	struct strbuf buf = STRBUF_INIT;
+
+	if (is_absolute_path(path))
+		return path;
+
+	strbuf_realpath_forgiving(&buf, path, 1);
+	free(*absolute);
+	*absolute = strbuf_detach(&buf, NULL);
+	return *absolute;
+}
+
 struct scalar_config {
 	const char *key;
 	const char *value;
@@ -406,6 +423,87 @@ static int supports_gvfs_protocol(const char *url, char **cache_server_url)
 	return 0; /* error out quietly */
 }
 
+static char *default_cache_root(const char *root)
+{
+	const char *env;
+
+	if (is_unattended())
+		return xstrfmt("%s/.scalarCache", root);
+
+#ifdef WIN32
+	(void)env;
+	return xstrfmt("%.*s.scalarCache", offset_1st_component(root), root);
+#elif defined(__APPLE__)
+	if ((env = getenv("HOME")) && *env)
+		return xstrfmt("%s/.scalarCache", env);
+	return NULL;
+#else
+	if ((env = getenv("XDG_CACHE_HOME")) && *env)
+		return xstrfmt("%s/scalar", env);
+	if ((env = getenv("HOME")) && *env)
+		return xstrfmt("%s/.cache/scalar", env);
+	return NULL;
+#endif
+}
+
+static int get_repository_id(struct json_iterator *it)
+{
+	if (it->type == JSON_STRING &&
+	    !strcasecmp(".repository.id", it->key.buf)) {
+		*(char **)it->fn_data = strbuf_detach(&it->string_value, NULL);
+		return 1;
+	}
+
+	return 0;
+}
+
+/* Needs to run this in a worktree; gvfs-helper requires a Git repository */
+static char *get_cache_key(const char *url)
+{
+	struct child_process cp = CHILD_PROCESS_INIT;
+	struct strbuf out = STRBUF_INIT;
+	char *cache_key = NULL;
+
+	cp.git_cmd = 1;
+	strvec_pushl(&cp.args, "gvfs-helper", "--remote", url,
+		     "endpoint", "vsts/info", NULL);
+	if (!pipe_command(&cp, NULL, 0, &out, 512, NULL, 0)) {
+		char *id = NULL;
+		struct json_iterator it =
+			JSON_ITERATOR_INIT(out.buf, get_repository_id, &id);
+
+		if (iterate_json(&it) < 0)
+			warning("JSON parse error (%s)", out.buf);
+		else if (id)
+			cache_key = xstrfmt("id_%s", id);
+		free(id);
+	}
+
+	if (!cache_key) {
+		struct strbuf downcased = STRBUF_INIT;
+		int hash_algo_index = hash_algo_by_name("sha1");
+		const struct git_hash_algo *hash_algo = hash_algo_index < 0 ?
+			the_hash_algo : &hash_algos[hash_algo_index];
+		git_hash_ctx ctx;
+		unsigned char hash[GIT_MAX_RAWSZ];
+
+		strbuf_addstr(&downcased, url);
+		strbuf_tolower(&downcased);
+
+		hash_algo->init_fn(&ctx);
+		hash_algo->update_fn(&ctx, downcased.buf, downcased.len);
+		hash_algo->final_fn(hash, &ctx);
+
+		strbuf_release(&downcased);
+
+		cache_key = xstrfmt("url_%s",
+				    hash_to_hex_algop(hash, hash_algo));
+	}
+
+	strbuf_release(&out);
+	return cache_key;
+}
+
 static char *remote_default_branch(const char *url)
 {
 	struct child_process cp = CHILD_PROCESS_INIT;
@@ -503,12 +601,48 @@ void load_builtin_commands(const char *prefix, struct cmdnames *cmds)
 	die("not implemented");
 }
 
+static int init_shared_object_cache(const char *url,
+				    const char *local_cache_root)
+{
+	struct strbuf buf = STRBUF_INIT;
+	int res = 0;
+	char *cache_key = NULL, *shared_cache_path = NULL;
+
+	if (!(cache_key = get_cache_key(url))) {
+		res = error(_("could not determine cache key for '%s'"), url);
+		goto cleanup;
+	}
+
+	shared_cache_path = xstrfmt("%s/%s", local_cache_root, cache_key);
+	if (set_config("gvfs.sharedCache=%s", shared_cache_path)) {
+		res = error(_("could not configure shared cache"));
+		goto cleanup;
+	}
+
+	strbuf_addf(&buf, "%s/pack", shared_cache_path);
+	switch (safe_create_leading_directories(buf.buf)) {
+	case SCLD_OK: case SCLD_EXISTS:
+		break; /* okay */
+	default:
+		res = error_errno(_("could not initialize '%s'"), buf.buf);
+		goto cleanup;
+	}
+
+	write_file(git_path("objects/info/alternates"),"%s\n", shared_cache_path);
+
+	cleanup:
+	strbuf_release(&buf);
+	free(shared_cache_path);
+	free(cache_key);
+	return res;
+}
+
 static int cmd_clone(int argc, const char **argv)
 {
 	const char *branch = NULL;
 	int full_clone = 0, single_branch = 0, show_progress = isatty(2);
-	const char *cache_server_url = NULL;
-	char *default_cache_server_url = NULL;
+	const char *cache_server_url = NULL, *local_cache_root = NULL;
+	char *default_cache_server_url = NULL, *local_cache_root_abs = NULL;
 	struct option clone_options[] = {
 		OPT_STRING('b', "branch", &branch, N_("<branch>"),
 			   N_("branch to checkout after clone")),
@@ -520,6 +654,9 @@ static int cmd_clone(int argc, const char **argv)
 		OPT_STRING(0, "cache-server-url", &cache_server_url,
 			   N_("<url>"),
 			   N_("the url or friendly name of the cache server")),
+		OPT_STRING(0, "local-cache-path", &local_cache_root,
+			   N_("<path>"),
+			   N_("override the path for the local Scalar cache")),
 		OPT_END(),
 	};
 	const char * const clone_usage[] = {
@@ -560,8 +697,20 @@ static int cmd_clone(int argc, const char **argv)
 	if (is_directory(enlistment))
 		die(_("directory '%s' exists already"), enlistment);
 
+	ensure_absolute_path(enlistment, &enlistment);
+
 	dir = xstrfmt("%s/src", enlistment);
 
+	if (!local_cache_root)
+		local_cache_root = local_cache_root_abs =
+			default_cache_root(enlistment);
+	else
+		local_cache_root = ensure_absolute_path(local_cache_root,
+							&local_cache_root_abs);
+
+	if (!local_cache_root)
+		die(_("could not determine local cache root"));
+
 	strbuf_reset(&buf);
 	if (branch)
 		strbuf_addf(&buf, "init.defaultBranch=%s", branch);
@@ -581,8 +730,28 @@ static int cmd_clone(int argc, const char **argv)
 
 	setup_git_directory();
 
+	git_config(git_default_config, NULL);
+
+	/*
+	 * This `dir_inside_of()` call relies on git_config() having parsed the
+	 * newly-initialized repository config's `core.ignoreCase` value.
+	 */
+	if (dir_inside_of(local_cache_root, dir) >= 0) {
+		struct strbuf path = STRBUF_INIT;
+
+		strbuf_addstr(&path, enlistment);
+		if (chdir("../..") < 0 ||
+		    remove_dir_recursively(&path, 0) < 0)
+			die(_("'--local-cache-path' cannot be inside the src "
+			      "folder;\nCould not remove '%s'"), enlistment);
+
+		die(_("'--local-cache-path' cannot be inside the src folder"));
+	}
+
 	/* common-main already logs `argv` */
 	trace2_def_repo(the_repository);
+	trace2_data_intmax("scalar", the_repository, "unattended",
+			   is_unattended());
 
 	if (!branch && !(branch = remote_default_branch(url))) {
 		res = error(_("failed to get default branch for '%s'"), url);
@@ -607,6 +776,8 @@ static int cmd_clone(int argc, const char **argv)
 			supports_gvfs_protocol(url, &default_cache_server_url);
 
 	if (gvfs_protocol) {
+		if ((res = init_shared_object_cache(url, local_cache_root)))
+			goto cleanup;
 		if (!cache_server_url)
 			cache_server_url = default_cache_server_url;
 		if (set_config("core.useGVFSHelper=true") ||
@@ -681,6 +852,7 @@ cleanup:
 	free(dir);
 	strbuf_release(&buf);
 	free(default_cache_server_url);
+	free(local_cache_root_abs);
 	return res;
 }
 
@@ -1055,6 +1227,12 @@ int cmd_main(int argc, const char **argv)
 	struct strbuf scalar_usage = STRBUF_INIT;
 	int i;
 
+	if (is_unattended()) {
+		setenv("GIT_ASKPASS", "", 0);
+		setenv("GIT_TERMINAL_PROMPT", "false", 0);
+		git_config_push_parameter("credential.interactive=never");
+	}
+
 	while (argc > 1 && *argv[1] == '-') {
 		if (!strcmp(argv[1], "-C")) {
 			if (argc < 3)

From 0e8ac7a7f4ecfad92db00cb2fb793fcb09038119 Mon Sep 17 00:00:00 2001
From: Johannes Schindelin <johannes.schindelin@gmx.de>
Date: Tue, 1 Jun 2021 23:18:14 +0200
Subject: [PATCH 08/29] scalar diagnose: include shared cache info

Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
---
 diagnose.c | 81 +++++++++++++++++++++++++++++++++++++++++++++++++++++-
 1 file changed, 80 insertions(+), 1 deletion(-)

diff --git a/diagnose.c b/diagnose.c
index c6839da136..f47470f8cc 100644
--- a/diagnose.c
+++ b/diagnose.c
@@ -103,6 +103,39 @@ cleanup:
 	return dtype;
 }
 
+static void dir_stats(struct strbuf *buf, const char *path)
+{
+	DIR *dir = opendir(path);
+	struct dirent *e;
+	struct stat e_stat;
+	struct strbuf file_path = STRBUF_INIT;
+	size_t base_path_len;
+
+	if (!dir)
+		return;
+
+	strbuf_addstr(buf, "Contents of ");
+	strbuf_add_absolute_path(buf, path);
+	strbuf_addstr(buf, ":\n");
+
+	strbuf_add_absolute_path(&file_path, path);
+	strbuf_addch(&file_path, '/');
+	base_path_len = file_path.len;
+
+	while ((e = readdir(dir)) != NULL)
+		if (!is_dot_or_dotdot(e->d_name) && e->d_type == DT_REG) {
+			strbuf_setlen(&file_path, base_path_len);
+			strbuf_addstr(&file_path, e->d_name);
+			if (!stat(file_path.buf, &e_stat))
+				strbuf_addf(buf, "%-70s %16"PRIuMAX"\n",
+					    e->d_name,
+					    (uintmax_t)e_stat.st_size);
+		}
+
+	strbuf_release(&file_path);
+	closedir(dir);
+}
+
 static int count_files(struct strbuf *path)
 {
 	DIR *dir = opendir(path->buf);
@@ -216,7 +249,7 @@ int create_diagnostics_archive(struct strbuf *zip_path, enum diagnose_mode mode)
 	char **argv_copy = NULL;
 	int stdout_fd = -1, archiver_fd = -1;
 	char *cache_server_url = NULL, *shared_cache = NULL;
-	struct strbuf buf = STRBUF_INIT;
+	struct strbuf buf = STRBUF_INIT, path = STRBUF_INIT;
 	int res, i;
 	struct archive_dir archive_dirs[] = {
 		{ ".git", 0 },
@@ -288,6 +321,52 @@ int create_diagnostics_archive(struct strbuf *zip_path, enum diagnose_mode mode)
 		}
 	}
 
+	if (shared_cache) {
+		size_t path_len;
+
+		strbuf_reset(&buf);
+		strbuf_addf(&path, "%s/pack", shared_cache);
+		strbuf_reset(&buf);
+		strbuf_addstr(&buf, "--add-virtual-file=packs-cached.txt:");
+		dir_stats(&buf, path.buf);
+		strvec_push(&archiver_args, buf.buf);
+
+		strbuf_reset(&buf);
+		strbuf_addstr(&buf, "--add-virtual-file=objects-cached.txt:");
+		loose_objs_stats(&buf, shared_cache);
+		strvec_push(&archiver_args, buf.buf);
+
+		strbuf_reset(&path);
+		strbuf_addf(&path, "%s/info", shared_cache);
+		path_len = path.len;
+
+		if (is_directory(path.buf)) {
+			DIR *dir = opendir(path.buf);
+			struct dirent *e;
+
+			while ((e = readdir(dir))) {
+				if (!strcmp(".", e->d_name) || !strcmp("..", e->d_name))
+					continue;
+				if (e->d_type == DT_DIR)
+					continue;
+
+				strbuf_reset(&buf);
+				strbuf_addf(&buf, "--add-virtual-file=info/%s:", e->d_name);
+
+				strbuf_setlen(&path, path_len);
+				strbuf_addch(&path, '/');
+				strbuf_addstr(&path, e->d_name);
+
+				if (strbuf_read_file(&buf, path.buf, 0) < 0) {
+					res = error_errno(_("could not read '%s'"), path.buf);
+					goto diagnose_cleanup;
+				}
+				strvec_push(&archiver_args, buf.buf);
+			}
+			closedir(dir);
+		}
+	}
+
 	strvec_pushl(&archiver_args, "--prefix=",
 		     oid_to_hex(the_hash_algo->empty_tree), "--", NULL);
 

From ccc7fd9518639275558133d4b4c7a6db6214255e Mon Sep 17 00:00:00 2001
From: Johannes Schindelin <johasc@microsoft.com>
Date: Tue, 18 May 2021 22:48:24 +0200
Subject: [PATCH 09/29] git_config_set_multivar_in_file_gently(): add a lock
 timeout

In particular when multiple processes want to write to the config
simultaneously, it would come in handy to not fail immediately when
another process locked the config, but to gently try again.

This will help with Scalar's functional test suite which wants to
register multiple repositories for maintenance semi-simultaneously.

As not all code paths calling this function read the config (e.g. `git
config`), we have to read the config setting via
`git_config_get_ulong()`.

Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
---
 Documentation/config/core.txt | 9 +++++++++
 config.c                      | 8 +++++++-
 2 files changed, 16 insertions(+), 1 deletion(-)

diff --git a/Documentation/config/core.txt b/Documentation/config/core.txt
index ac2baa136e..cd4fcdba20 100644
--- a/Documentation/config/core.txt
+++ b/Documentation/config/core.txt
@@ -749,3 +749,12 @@ core.abbrev::
 	If set to "no", no abbreviation is made and the object names
 	are shown in their full length.
 	The minimum length is 4.
+
+core.configWriteLockTimeoutMS::
+	When processes try to write to the config concurrently, it is likely
+	that one process "wins" and the other process(es) fail to lock the
+	config file. By configuring a timeout larger than zero, Git can be
+	told to try to lock the config again a couple times within the
+	specified timeout. If the timeout is configure to zero (which is the
+	default), Git will fail immediately when the config is already
+	locked.
diff --git a/config.c b/config.c
index d82e00acc7..67800d2a0a 100644
--- a/config.c
+++ b/config.c
@@ -3202,6 +3202,7 @@ int git_config_set_multivar_in_file_gently(const char *config_filename,
 					   const char *value_pattern,
 					   unsigned flags)
 {
+	static unsigned long timeout_ms = ULONG_MAX;
 	int fd = -1, in_fd = -1;
 	int ret;
 	struct lock_file lock = LOCK_INIT;
@@ -3222,11 +3223,16 @@ int git_config_set_multivar_in_file_gently(const char *config_filename,
 	if (!config_filename)
 		config_filename = filename_buf = git_pathdup("config");
 
+	if ((long)timeout_ms < 0 &&
+	    git_config_get_ulong("core.configWriteLockTimeoutMS", &timeout_ms))
+		timeout_ms = 0;
+
 	/*
 	 * The lock serves a purpose in addition to locking: the new
 	 * contents of .git/config will be written into it.
 	 */
-	fd = hold_lock_file_for_update(&lock, config_filename, 0);
+	fd = hold_lock_file_for_update_timeout(&lock, config_filename, 0,
+					       timeout_ms);
 	if (fd < 0) {
 		error_errno(_("could not lock config file %s"), config_filename);
 		ret = CONFIG_NO_LOCK;

From 331635229ba02c3364aaa956875db1b5ced22a20 Mon Sep 17 00:00:00 2001
From: Johannes Schindelin <johannes.schindelin@gmx.de>
Date: Wed, 28 Apr 2021 13:56:16 +0200
Subject: [PATCH 10/29] scalar: only try GVFS protocol on https:// URLs

Well, technically also the http:// protocol is allowed _when testing_...

Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
---
 scalar.c | 45 +++++++++++++++++++++++++++++++++------------
 1 file changed, 33 insertions(+), 12 deletions(-)

diff --git a/scalar.c b/scalar.c
index f11db89cba..bf664b95bf 100644
--- a/scalar.c
+++ b/scalar.c
@@ -385,6 +385,13 @@ static int get_cache_server_url(struct json_iterator *it)
 	return 0;
 }
 
+static int can_url_support_gvfs(const char *url)
+{
+	return starts_with(url, "https://") ||
+		(git_env_bool("GIT_TEST_ALLOW_GVFS_VIA_HTTP", 0) &&
+		 starts_with(url, "http://"));
+}
+
 /*
  * If `cache_server_url` is `NULL`, print the list to `stdout`.
  *
@@ -396,6 +403,13 @@ static int supports_gvfs_protocol(const char *url, char **cache_server_url)
 	struct child_process cp = CHILD_PROCESS_INIT;
 	struct strbuf out = STRBUF_INIT;
 
+	/*
+	 * The GVFS protocol is only supported via https://; For testing, we
+	 * also allow http://.
+	 */
+	if (!can_url_support_gvfs(url))
+		return 0;
+
 	cp.git_cmd = 1;
 	strvec_pushl(&cp.args, "gvfs-helper", "--remote", url, "config", NULL);
 	if (!pipe_command(&cp, NULL, 0, &out, 512, NULL, 0)) {
@@ -464,19 +478,26 @@ static char *get_cache_key(const char *url)
 	struct strbuf out = STRBUF_INIT;
 	char *cache_key = NULL;
 
-	cp.git_cmd = 1;
-	strvec_pushl(&cp.args, "gvfs-helper", "--remote", url,
-		     "endpoint", "vsts/info", NULL);
-	if (!pipe_command(&cp, NULL, 0, &out, 512, NULL, 0)) {
-		char *id = NULL;
-		struct json_iterator it =
-			JSON_ITERATOR_INIT(out.buf, get_repository_id, &id);
+	/*
+	 * The GVFS protocol is only supported via https://; For testing, we
+	 * also allow http://.
+	 */
+	if (can_url_support_gvfs(url)) {
+		cp.git_cmd = 1;
+		strvec_pushl(&cp.args, "gvfs-helper", "--remote", url,
+			     "endpoint", "vsts/info", NULL);
+		if (!pipe_command(&cp, NULL, 0, &out, 512, NULL, 0)) {
+			char *id = NULL;
+			struct json_iterator it =
+				JSON_ITERATOR_INIT(out.buf, get_repository_id,
+						   &id);
 
-		if (iterate_json(&it) < 0)
-			warning("JSON parse error (%s)", out.buf);
-		else if (id)
-			cache_key = xstrfmt("id_%s", id);
-		free(id);
+			if (iterate_json(&it) < 0)
+				warning("JSON parse error (%s)", out.buf);
+			else if (id)
+				cache_key = xstrfmt("id_%s", id);
+			free(id);
+		}
 	}
 
 	if (!cache_key) {

From f153b31d8af280c52b32fb639f3d2850b4fb5ee3 Mon Sep 17 00:00:00 2001
From: Johannes Schindelin <johasc@microsoft.com>
Date: Tue, 18 May 2021 23:22:56 +0200
Subject: [PATCH 11/29] scalar: set the config write-lock timeout to 150ms

By default, Git fails immediately when locking a config file for writing
fails due to an existing lock. With this change, Scalar-registered
repositories will fall back to trying a couple times within a 150ms
timeout.

Signed-off-by: Johannes Schindelin <johasc@microsoft.com>
---
 scalar.c | 10 ++++++++++
 1 file changed, 10 insertions(+)

diff --git a/scalar.c b/scalar.c
index ca19b95ce4..a1cee98c64 100644
--- a/scalar.c
+++ b/scalar.c
@@ -164,6 +164,7 @@ static int set_recommended_config(int reconfigure)
 		{ "core.autoCRLF", "false" },
 		{ "core.safeCRLF", "false" },
 		{ "fetch.showForcedUpdates", "false" },
+		{ "core.configWriteLockTimeoutMS", "150" },
 		{ NULL, NULL },
 	};
 	int i;
@@ -205,6 +206,11 @@ static int set_recommended_config(int reconfigure)
 
 static int toggle_maintenance(int enable)
 {
+	unsigned long ul;
+
+	if (git_config_get_ulong("core.configWriteLockTimeoutMS", &ul))
+		git_config_push_parameter("core.configWriteLockTimeoutMS=150");
+
 	return run_git("maintenance",
 		       enable ? "start" : "unregister",
 		       enable ? NULL : "--force",
@@ -214,10 +220,14 @@ static int toggle_maintenance(int enable)
 static int add_or_remove_enlistment(int add)
 {
 	int res;
+	unsigned long ul;
 
 	if (!the_repository->worktree)
 		die(_("Scalar enlistments require a worktree"));
 
+	if (git_config_get_ulong("core.configWriteLockTimeoutMS", &ul))
+		git_config_push_parameter("core.configWriteLockTimeoutMS=150");
+
 	res = run_git("config", "--global", "--get", "--fixed-value",
 		      "scalar.repo", the_repository->worktree, NULL);
 

From 8ebf13a2f58f98584371cf113d72f8767957d718 Mon Sep 17 00:00:00 2001
From: Johannes Schindelin <johannes.schindelin@gmx.de>
Date: Fri, 16 Apr 2021 21:43:57 +0200
Subject: [PATCH 12/29] scalar: verify that we can use a GVFS-enabled
 repository

Azure Repos does not support partial clones at the moment, but it does
support the GVFS protocol. To that end, the Microsoft fork of Git has a
`gvfs-helper` command that is optionally used to perform essentially the
same functionality as partial clone.

Let's verify that `scalar clone` detects that situation and enables the
GVFS helper.

Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
---
 t/t9210-scalar.sh | 153 ++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 153 insertions(+)

diff --git a/t/t9210-scalar.sh b/t/t9210-scalar.sh
index 543f2e1d3a..6acf279afd 100755
--- a/t/t9210-scalar.sh
+++ b/t/t9210-scalar.sh
@@ -238,4 +238,157 @@ test_expect_success UNZIP 'scalar diagnose' '
 	grep "^Total: [1-9]" out
 '
 
+GIT_TEST_ALLOW_GVFS_VIA_HTTP=1
+export GIT_TEST_ALLOW_GVFS_VIA_HTTP
+
+test_set_port GIT_TEST_GVFS_PROTOCOL_PORT
+HOST_PORT=127.0.0.1:$GIT_TEST_GVFS_PROTOCOL_PORT
+PID_FILE="$(pwd)"/pid-file.pid
+SERVER_LOG="$(pwd)"/OUT.server.log
+
+test_atexit '
+	test -f "$PID_FILE" || return 0
+
+	# The server will shutdown automatically when we delete the pid-file.
+	rm -f "$PID_FILE"
+
+	test -z "$verbose$verbose_log" || {
+		echo "server log:"
+		cat "$SERVER_LOG"
+	}
+
+	# Give it a few seconds to shutdown (mainly to completely release the
+	# port before the next test start another instance and it attempts to
+	# bind to it).
+	for k in $(test_seq 5)
+	do
+		grep -q "Starting graceful shutdown" "$SERVER_LOG" &&
+		return 0 ||
+		sleep 1
+	done
+
+	echo "stop_gvfs_protocol_server: timeout waiting for server shutdown"
+	return 1
+'
+
+start_gvfs_enabled_http_server () {
+	GIT_HTTP_EXPORT_ALL=1 \
+	test-gvfs-protocol --verbose \
+		--listen=127.0.0.1 \
+		--port=$GIT_TEST_GVFS_PROTOCOL_PORT \
+		--reuseaddr \
+		--pid-file="$PID_FILE" \
+		2>"$SERVER_LOG" &
+
+	for k in 0 1 2 3 4
+	do
+		if test -f "$PID_FILE"
+		then
+			return 0
+		fi
+		sleep 1
+	done
+	return 1
+}
+
+test_expect_success 'start GVFS-enabled server' '
+	git config uploadPack.allowFilter false &&
+	git config uploadPack.allowAnySHA1InWant false &&
+	start_gvfs_enabled_http_server
+'
+
+test_expect_success '`scalar clone` with GVFS-enabled server' '
+	: the fake cache server requires fake authentication &&
+	git config --global core.askPass true &&
+	scalar clone --single-branch -- http://$HOST_PORT/ using-gvfs &&
+
+	: verify that the shared cache has been configured &&
+	cache_key="url_$(printf "%s" http://$HOST_PORT/ |
+		tr A-Z a-z |
+		test-tool sha1)" &&
+	echo "$(pwd)/using-gvfs/.scalarCache/$cache_key" >expect &&
+	git -C using-gvfs/src config gvfs.sharedCache >actual &&
+	test_cmp expect actual &&
+
+	second=$(git rev-parse --verify second:second.t) &&
+	(
+		cd using-gvfs/src &&
+		test_path_is_missing 1/2 &&
+		GIT_TRACE=$PWD/trace.txt git cat-file blob $second >actual &&
+		: verify that the gvfs-helper was invoked to fetch it &&
+		test_i18ngrep gvfs-helper trace.txt &&
+		echo "second" >expect &&
+		test_cmp expect actual
+	)
+'
+
+test_expect_success '`scalar register` parallel to worktree is unsupported' '
+	git init test-repo/src &&
+	mkdir -p test-repo/out &&
+
+	: parallel to worktree is unsupported &&
+	test_must_fail env GIT_CEILING_DIRECTORIES="$(pwd)" \
+		scalar register test-repo/out &&
+	test_must_fail git config --get --global --fixed-value \
+		maintenance.repo "$(pwd)/test-repo/src" &&
+	scalar list >scalar.repos &&
+	! grep -F "$(pwd)/test-repo/src" scalar.repos &&
+
+	: at enlistment root, i.e. parent of repository, is supported &&
+	GIT_CEILING_DIRECTORIES="$(pwd)" scalar register test-repo &&
+	git config --get --global --fixed-value \
+		maintenance.repo "$(pwd)/test-repo/src" &&
+	scalar list >scalar.repos &&
+	grep -F "$(pwd)/test-repo/src" scalar.repos &&
+
+	: scalar delete properly unregisters enlistment &&
+	scalar delete test-repo &&
+	test_must_fail git config --get --global --fixed-value \
+		maintenance.repo "$(pwd)/test-repo/src" &&
+	scalar list >scalar.repos &&
+	! grep -F "$(pwd)/test-repo/src" scalar.repos
+'
+
+test_expect_success '`scalar register` & `unregister` with existing repo' '
+	git init existing &&
+	scalar register existing &&
+	git config --get --global --fixed-value \
+		maintenance.repo "$(pwd)/existing" &&
+	scalar list >scalar.repos &&
+	grep -F "$(pwd)/existing" scalar.repos &&
+	scalar unregister existing &&
+	test_must_fail git config --get --global --fixed-value \
+		maintenance.repo "$(pwd)/existing" &&
+	scalar list >scalar.repos &&
+	! grep -F "$(pwd)/existing" scalar.repos
+'
+
+test_expect_success '`scalar unregister` with existing repo, deleted .git' '
+	scalar register existing &&
+	rm -rf existing/.git &&
+	scalar unregister existing &&
+	test_must_fail git config --get --global --fixed-value \
+		maintenance.repo "$(pwd)/existing" &&
+	scalar list >scalar.repos &&
+	! grep -F "$(pwd)/existing" scalar.repos
+'
+
+test_expect_success '`scalar register` existing repo with `src` folder' '
+	git init existing &&
+	mkdir -p existing/src &&
+	scalar register existing/src &&
+	scalar list >scalar.repos &&
+	grep -F "$(pwd)/existing" scalar.repos &&
+	scalar unregister existing &&
+	scalar list >scalar.repos &&
+	! grep -F "$(pwd)/existing" scalar.repos
+'
+
+test_expect_success '`scalar delete` with existing repo' '
+	git init existing &&
+	scalar register existing &&
+	scalar delete existing &&
+	test_path_is_missing existing
+'
+
 test_done

From 8fcf4abc75470e751282e40e6b95840e4d9fa63c Mon Sep 17 00:00:00 2001
From: Derrick Stolee <dstolee@microsoft.com>
Date: Wed, 16 Jun 2021 10:01:37 -0400
Subject: [PATCH 13/29] scalar: add docs from microsoft/scalar

These docs have been altered to fit the version implemented in C within
microsoft/git. This means in particular that the advanced.md file no
longer applied at all. Some other areas were removed or significantly
edited.

Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
---
 contrib/scalar/docs/faq.md             | 51 ++++++++++++++
 contrib/scalar/docs/getting-started.md | 93 ++++++++++++++++++++++++++
 contrib/scalar/docs/index.md           | 50 ++++++++++++++
 contrib/scalar/docs/philosophy.md      | 66 ++++++++++++++++++
 contrib/scalar/docs/troubleshooting.md | 20 ++++++
 5 files changed, 280 insertions(+)
 create mode 100644 contrib/scalar/docs/faq.md
 create mode 100644 contrib/scalar/docs/getting-started.md
 create mode 100644 contrib/scalar/docs/index.md
 create mode 100644 contrib/scalar/docs/philosophy.md
 create mode 100644 contrib/scalar/docs/troubleshooting.md

diff --git a/contrib/scalar/docs/faq.md b/contrib/scalar/docs/faq.md
new file mode 100644
index 0000000000..a14f78a996
--- /dev/null
+++ b/contrib/scalar/docs/faq.md
@@ -0,0 +1,51 @@
+Frequently Asked Questions
+==========================
+
+Using Scalar
+------------
+
+### I don't want a sparse clone, I want every file after I clone!
+
+Run `scalar clone --full-clone <url>` to initialize your repo to include
+every file. You can switch to a sparse-checkout later by running
+`git sparse-checkout init --cone`.
+
+### I already cloned without `--full-clone`. How do I get everything?
+
+Run `git sparse-checkout disable`.
+
+Scalar Design Decisions
+-----------------------
+
+There may be many design decisions within Scalar that are confusing at first
+glance. Some of them may cause friction when you use Scalar with your existing
+repos and existing habits.
+
+> Scalar has the most benefit when users design repositories
+> with efficient patterns.
+
+For example: Scalar uses the sparse-checkout feature to limit the size of the
+working directory within a large monorepo. It is designed to work efficiently
+with monorepos that are highly componentized, allowing most developers to
+need many fewer files in their daily work.
+
+### Why does `scalar clone` create a `<repo>/src` folder?
+
+Scalar uses a file system watcher to keep track of changes under this `src` folder.
+Any activity in this folder is assumed to be important to Git operations. By
+creating the `src` folder, we are making it easy for your build system to
+create output folders outside the `src` directory. We commonly see systems
+create folders for build outputs and package downloads. Scalar itself creates
+these folders during its builds.
+
+Your build system may create build artifacts such as `.obj` or `.lib` files
+next to your source code. These are commonly "hidden" from Git using
+`.gitignore` files. Having such artifacts in your source tree creates
+additional work for Git because it needs to look at these files and match them
+against the `.gitignore` patterns.
+
+By following the `src` pattern Scalar tries to establish and placing your build
+intermediates and outputs parallel with the `src` folder and not inside it,
+you can help optimize Git command performance for developers in the repository
+by limiting the number of files Git needs to consider for many common
+operations.
diff --git a/contrib/scalar/docs/getting-started.md b/contrib/scalar/docs/getting-started.md
new file mode 100644
index 0000000000..ef7ea07b0f
--- /dev/null
+++ b/contrib/scalar/docs/getting-started.md
@@ -0,0 +1,93 @@
+Getting Started
+===============
+
+Registering existing Git repos
+------------------------------
+
+To add a repository to the list of registered repos, run `scalar register [<path>]`.
+If `<path>` is not provided, then the "current repository" is discovered from
+the working directory by scanning the parent paths for a path containing a `.git`
+folder, possibly inside a `src` folder.
+
+To see which repositories are currently tracked by the service, run
+`scalar list`.
+
+Run `scalar unregister [<path>]` to remove the repo from this list.
+
+Creating a new Scalar clone
+---------------------------------------------------
+
+The `clone` verb creates a local enlistment of a remote repository using the
+partial clone feature available e.g. on GitHub.
+
+
+```
+scalar clone [options] <url> [<dir>]
+```
+
+Create a local copy of the repository at `<url>`. If specified, create the `<dir>`
+directory and place the repository there. Otherwise, the last section of the `<url>`
+will be used for `<dir>`.
+
+At the end, the repo is located at `<dir>/src`. By default, the sparse-checkout
+feature is enabled and the only files present are those in the root of your
+Git repository. Use `git sparse-checkout set` to expand the set of directories
+you want to see, or `git sparse-checkout disable` to expand to all files. You
+can explore the subdirectories outside your sparse-checkout specification using
+`git ls-tree HEAD`.
+
+### Sparse Repo Mode
+
+By default, Scalar reduces your working directory to only the files at the
+root of the repository. You need to add the folders you care about to build up
+to your working set.
+
+* `scalar clone <url>`
+  * Please choose the **Clone with HTTPS** option in the `Clone Repository` dialog in Azure Repos, not **Clone with SSH**.
+* `cd <root>\src`
+* At this point, your `src` directory only contains files that appear in your root
+  tree. No folders are populated.
+* Set the directory list for your sparse-checkout using:
+	1. `git sparse-checkout set <dir1> <dir2> ...`
+	2. `git sparse-checkout set --stdin < dir-list.txt`
+* Run git commands as you normally would.
+* To fully populate your working directory, run `git sparse-checkout disable`.
+
+If instead you want to start with all files on-disk, you can clone with the
+`--full-clone` option. To enable sparse-checkout after the fact, run
+`git sparse-checkout init --cone`. This will initialize your sparse-checkout
+patterns to only match the files at root.
+
+If you are unfamiliar with what directories are available in the repository,
+then you can run `git ls-tree -d --name-only HEAD` to discover the directories
+at root, or `git ls-tree -d --name-only HEAD <path>` to discover the directories
+in `<path>`.
+
+### Options
+
+These options allow a user to customize their initial enlistment.
+
+* `--full-clone`: If specified, do not initialize the sparse-checkout feature.
+  All files will be present in your `src` directory. This uses a Git partial
+  clone: blobs are downloaded on demand.
+
+* `--branch=<ref>`: Specify the branch to checkout after clone.
+
+### Advanced Options
+
+The options below are not intended for use by a typical user. These are
+usually used by build machines to create a temporary enlistment that
+operates on a single commit.
+
+* `--single-branch`: Use this option to only download metadata for the branch
+  that will be checked out. This is helpful for build machines that target
+  a remote with many branches. Any `git fetch` commands after the clone will
+  still ask for all branches.
+
+Removing a Scalar Clone
+-----------------------
+
+Since the `scalar clone` command sets up a file-system watcher (when available),
+that watcher could prevent deleting the enlistment. Run `scalar delete <path>`
+from outside of your enlistment to unregister the enlistment from the filesystem
+watcher and delete the enlistment at `<path>`.
diff --git a/contrib/scalar/docs/index.md b/contrib/scalar/docs/index.md
new file mode 100644
index 0000000000..f9f5ab06e0
--- /dev/null
+++ b/contrib/scalar/docs/index.md
@@ -0,0 +1,50 @@
+Scalar: Enabling Git at Scale
+=============================
+
+Scalar is a tool that helps Git scale to some of the largest Git repositories.
+It achieves this by enabling some advanced Git features, such as:
+
+* *Partial clone:* reduces time to get a working repository by not
+  downloading all Git objects right away.
+
+* *Background prefetch:* downloads Git object data from all remotes every
+  hour, reducing the amount of time for foreground `git fetch` calls.
+
+* *Sparse-checkout:* limits the size of your working directory.
+
+* *File system monitor:* tracks the recently modified files and eliminates
+  the need for Git to scan the entire worktree.
+
+* *Commit-graph:* accelerates commit walks and reachability calculations,
+   speeding up commands like `git log`.
+
+* *Multi-pack-index:* enables fast object lookups across many pack-files.
+
+* *Incremental repack:* Repacks the packed Git data into fewer pack-file
+  without disrupting concurrent commands by using the multi-pack-index.
+
+By running `scalar register` in any Git repo, Scalar will automatically enable
+these features for that repo (except partial clone) and start running suggested
+maintenance in the background using
+[the `git maintenance` feature](https://git-scm.com/docs/git-maintenance).
+
+Repos cloned with the `scalar clone` command use partial clone to significantly
+reduce the amount of data required to get started using a repository. By
+delaying all blob downloads until they are required, Scalar allows you to work
+with very large repositories quickly.
+
+Documentation
+-------------
+
+* [Getting Started](getting-started.md): Get started with Scalar.
+  Includes `scalar register`, `scalar unregister`, `scalar clone`, and
+  `scalar delete`.
+
+* [Troubleshooting](troubleshooting.md):
+  Collect diagnostic information or update custom settings. Includes
+  `scalar diagnose`.
+
+* [The Philosophy of Scalar](philosophy.md): Why does Scalar work the way
+  it does, and how do we make decisions about its future?
+
+* [Frequently Asked Questions](faq.md)
diff --git a/contrib/scalar/docs/philosophy.md b/contrib/scalar/docs/philosophy.md
new file mode 100644
index 0000000000..51486a75e4
--- /dev/null
+++ b/contrib/scalar/docs/philosophy.md
@@ -0,0 +1,66 @@
+The Philosophy of Scalar
+========================
+
+The team building Scalar has **opinions** about Git performance. Scalar
+takes out the guesswork by automatically configuring your Git repositories
+to take advantage of the latest and greatest features. It is difficult to
+say that these are the absolute best settings for every repository, but
+these settings do work for some of the largest repositories in the world.
+
+Scalar intends to do very little more than the standard Git client. We
+actively implement new features into Git instead of Scalar, then update
+Scalar only to configure those new settings. In particular, we ported
+features like background maintenance to Git to make Scalar simpler and
+make Git more powerful.
+
+Services such as GitHub support partial clone , a standard adopted by the Git
+project to download only part of the Git objects when cloning, and fetching
+further objects on demand. If your hosting service supports partial clone, then
+we absolutely recommend it as a way to greatly speed up your clone and fetch
+times and to reduce how much disk space your Git repository requires. Scalar
+will help with this!
+
+Most of the value of Scalar can be found in the core Git client. However, most
+of the advanced features that really optimize Git's performance are off by
+default for compatibility reasons. To really take advantage of Git's latest and
+greatest features, you either need to study the [`git config`
+documentation](https://git-scm.com/docs/git-config) and regularly read [the Git
+release notes](https://github.com/git/git/tree/master/Documentation/RelNotes).
+Even if you do all that work and customize your Git settings on your machines,
+you likely will want to share those settings with other team members. Or, you
+can just use Scalar!
+
+Using `scalar register` on an existing Git repository will give you these
+benefits:
+
+* Additional compression of your `.git/index` file.
+* Hourly background `git fetch` operations, keeping you in-sync with your
+  remotes.
+* Advanced data structures, such as the `commit-graph` and `multi-pack-index`
+  are updated automatically in the background.
+* If using macOS or Windows, then Scalar configures Git's builtin File System
+  Monitor, providing faster commands such as `git status` or `git add`.
+
+Additionally, if you use `scalar clone` to create a new repository, then
+you will automatically get these benefits:
+
+* Use Git's partial clone feature to only download the files you need for
+  your current checkout.
+* Use Git's [sparse-checkout feature][sparse-checkout] to minimize the
+  number of files required in your working directory.
+  [Read more about sparse-checkout here.][sparse-checkout-blog]
+* Create the Git repository inside `<repo-name>/src` to make it easy to
+  place build artifacts outside of the Git repository, such as in
+  `<repo-name>/bin` or `<repo-name>/packages`.
+
+We also admit that these **opinions** can always be improved! If you have
+an idea of how to improve our setup, consider
+[creating an issue](https://github.com/microsoft/scalar/issues/new) or
+contributing a pull request! Some [existing](https://github.com/microsoft/scalar/issues/382)
+[issues](https://github.com/microsoft/scalar/issues/388) have already
+improved our configuration settings and roadmap!
+
+[gvfs-protocol]: https://github.com/microsoft/VFSForGit/blob/HEAD/Protocol.md
+[microsoft-git]: https://github.com/microsoft/git
+[sparse-checkout]: https://git-scm.com/docs/git-sparse-checkout
+[sparse-checkout-blog]: https://github.blog/2020-01-17-bring-your-monorepo-down-to-size-with-sparse-checkout/
diff --git a/contrib/scalar/docs/troubleshooting.md b/contrib/scalar/docs/troubleshooting.md
new file mode 100644
index 0000000000..8ec56ad437
--- /dev/null
+++ b/contrib/scalar/docs/troubleshooting.md
@@ -0,0 +1,20 @@
+Troubleshooting
+===============
+
+Diagnosing Issues
+-----------------
+
+The `scalar diagnose` command collects logs and config details for the current
+repository. The resulting zip file helps root-cause issues.
+
+When run inside your repository, creates a zip file containing several important
+files for that repository. This includes:
+
+* Configuration files from your `.git` folder, such as the `config` file,
+  `index`, `hooks`, and `refs`.
+
+* A summary of your Git object database, including the number of loose objects
+  and the names and sizes of pack-files.
+
+As the `diagnose` command completes, it provides the path of the resulting
+zip file. This zip can be attached to bug reports to make the analysis easier.

From a53786cfc8297c210972332714f951ed4ec543c4 Mon Sep 17 00:00:00 2001
From: Johannes Schindelin <johannes.schindelin@gmx.de>
Date: Fri, 23 Apr 2021 16:12:33 +0200
Subject: [PATCH 14/29] scalar: add the `cache-server` command

This allows setting the GVFS-enabled cache server, or listing the one(s)
associated with the remote repository.

Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
---
 Documentation/scalar.txt |  22 +++++++++
 scalar.c                 | 102 ++++++++++++++++++++++++++++++++++++++-
 2 files changed, 123 insertions(+), 1 deletion(-)

diff --git a/Documentation/scalar.txt b/Documentation/scalar.txt
index e7f6d39397..0424d020d9 100644
--- a/Documentation/scalar.txt
+++ b/Documentation/scalar.txt
@@ -18,6 +18,7 @@ scalar run ( all | config | commit-graph | fetch | loose-objects | pack-files )
 scalar reconfigure [ --all | <enlistment> ]
 scalar diagnose [<enlistment>]
 scalar delete <enlistment>
+scalar cache-server ( --get | --set <url> | --list [<remote>] ) [<enlistment>]
 
 DESCRIPTION
 -----------
@@ -170,6 +171,27 @@ delete <enlistment>::
 	This subcommand lets you delete an existing Scalar enlistment from your
 	local file system, unregistering the repository.
 
+Cache-server
+~~~~~~~~~~~~
+
+cache-server ( --get | --set <url> | --list [<remote>] ) [<enlistment>]::
+    This command lets you query or set the GVFS-enabled cache server used
+    to fetch missing objects.
+
+--get::
+    This is the default command mode: query the currently-configured cache
+    server URL, if any.
+
+--list::
+    Access the `gvfs/info` endpoint of the specified remote (default:
+    `origin`) to figure out which cache servers are available, if any.
++
+In contrast to the `--get` command mode (which only accesses the local
+repository), this command mode triggers a request via the network that
+potentially requires authentication. If authentication is required, the
+configured credential helper is employed (see linkgit:git-credential[1]
+for details).
+
 SEE ALSO
 --------
 linkgit:git-clone[1], linkgit:git-maintenance[1].
diff --git a/scalar.c b/scalar.c
index bf664b95bf..37c135a799 100644
--- a/scalar.c
+++ b/scalar.c
@@ -15,6 +15,7 @@
 #include "packfile.h"
 #include "help.h"
 #include "json-parser.h"
+#include "remote.h"
 
 static int is_unattended(void) {
 	return git_env_bool("Scalar_UNATTENDED", 0);
@@ -349,6 +350,21 @@ static int set_config(const char *fmt, ...)
 	return res;
 }
 
+static int list_cache_server_urls(struct json_iterator *it)
+{
+	const char *p;
+	char *q;
+	long l;
+
+	if (it->type == JSON_STRING &&
+	    skip_iprefix(it->key.buf, ".CacheServers[", &p) &&
+	    (l = strtol(p, &q, 10)) >= 0 && p != q &&
+	    !strcasecmp(q, "].Url"))
+		printf("#%ld: %s\n", l, it->string_value.buf);
+
+	return 0;
+}
+
 /* Find N for which .CacheServers[N].GlobalDefault == true */
 static int get_cache_server_index(struct json_iterator *it)
 {
@@ -418,6 +434,16 @@ static int supports_gvfs_protocol(const char *url, char **cache_server_url)
 			JSON_ITERATOR_INIT(out.buf, get_cache_server_index, &l);
 		struct cache_server_url_data data = { .url = NULL };
 
+		if (!cache_server_url) {
+			it.fn = list_cache_server_urls;
+			if (iterate_json(&it) < 0) {
+				strbuf_release(&out);
+				return error("JSON parse error");
+			}
+			strbuf_release(&out);
+			return 0;
+		}
+
 		if (iterate_json(&it) < 0) {
 			strbuf_release(&out);
 			return error("JSON parse error");
@@ -434,7 +460,9 @@ static int supports_gvfs_protocol(const char *url, char **cache_server_url)
 		return 1;
 	}
 	strbuf_release(&out);
-	return 0; /* error out quietly */
+	/* error out quietly, unless we wanted to list URLs */
+	return cache_server_url ?
+		0 : error(_("Could not access gvfs/config endpoint"));
 }
 
 static char *default_cache_root(const char *root)
@@ -1226,6 +1254,77 @@ static int cmd_version(int argc, const char **argv)
 	return 0;
 }
 
+static int cmd_cache_server(int argc, const char **argv)
+{
+	int get = 0;
+	char *set = NULL, *list = NULL;
+	const char *default_remote = "(default)";
+	struct option options[] = {
+		OPT_BOOL(0, "get", &get,
+			 N_("get the configured cache-server URL")),
+		OPT_STRING(0, "set", &set, N_("URL"),
+			    N_("configure the cache-server to use")),
+		{ OPTION_STRING, 0, "list", &list, N_("remote"),
+		  N_("list the possible cache-server URLs"),
+		  PARSE_OPT_OPTARG, NULL, (intptr_t) default_remote },
+		OPT_END(),
+	};
+	const char * const usage[] = {
+		N_("scalar cache_server "
+		   "[--get | --set <url> | --list [<remote>]] [<enlistment>]"),
+		NULL
+	};
+	int res = 0;
+
+	argc = parse_options(argc, argv, NULL, options,
+			     usage, 0);
+
+	if (get + !!set + !!list > 1)
+		usage_msg_opt(_("--get/--set/--list are mutually exclusive"),
+			      usage, options);
+
+	setup_enlistment_directory(argc, argv, usage, options, NULL);
+
+	if (list) {
+		const char *name = list, *url = list;
+
+		if (list == default_remote)
+			list = NULL;
+
+		if (!list || !strchr(list, '/')) {
+			struct remote *remote;
+
+			/* Look up remote */
+			remote = remote_get(list);
+			if (!remote) {
+				error("no such remote: '%s'", name);
+				free(list);
+				return 1;
+			}
+			if (!remote->url) {
+				free(list);
+				return error(_("remote '%s' has no URLs"),
+					     name);
+			}
+			url = remote->url[0];
+		}
+		res = supports_gvfs_protocol(url, NULL);
+		free(list);
+	} else if (set) {
+		res = set_config("gvfs.cache-server=%s", set);
+		free(set);
+	} else {
+		char *url = NULL;
+
+		printf("Using cache server: %s\n",
+		       git_config_get_string("gvfs.cache-server", &url) ?
+		       "(undefined)" : url);
+		free(url);
+	}
+
+	return !!res;
+}
+
 static struct {
 	const char *name;
 	int (*fn)(int, const char **);
@@ -1240,6 +1339,7 @@ static struct {
 	{ "help", cmd_help },
 	{ "version", cmd_version },
 	{ "diagnose", cmd_diagnose },
+	{ "cache-server", cmd_cache_server },
 	{ NULL, NULL},
 };
 

From e46877cea37f84ec0e99b7896b90f55cb3c839b4 Mon Sep 17 00:00:00 2001
From: Johannes Schindelin <johannes.schindelin@gmx.de>
Date: Wed, 12 May 2021 17:59:58 +0200
Subject: [PATCH 15/29] scalar: add a test toggle to skip accessing the
 vsts/info endpoint

In Scalar's functional tests, we do not do anything with authentication.
Therefore, we do want to avoid accessing the `vsts/info` endpoint
because it requires authentication even on otherwise public
repositories.

Let's introduce the environment variable `SCALAR_TEST_SKIP_VSTS_INFO`
which can be set to `true` to simply skip that step (and force the
`url_*` style repository IDs instead of `id_*` whenever possible).

Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
---
 scalar.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/scalar.c b/scalar.c
index 37c135a799..3992b933c9 100644
--- a/scalar.c
+++ b/scalar.c
@@ -510,7 +510,8 @@ static char *get_cache_key(const char *url)
 	 * The GVFS protocol is only supported via https://; For testing, we
 	 * also allow http://.
 	 */
-	if (can_url_support_gvfs(url)) {
+	if (!git_env_bool("SCALAR_TEST_SKIP_VSTS_INFO", 0) &&
+	    can_url_support_gvfs(url)) {
 		cp.git_cmd = 1;
 		strvec_pushl(&cp.args, "gvfs-helper", "--remote", url,
 			     "endpoint", "vsts/info", NULL);

From a12ec8885faf1b261a4b8e54eef3d3feefa23df2 Mon Sep 17 00:00:00 2001
From: Johannes Schindelin <johannes.schindelin@gmx.de>
Date: Tue, 10 May 2022 13:43:05 +0200
Subject: [PATCH 16/29] scalar (Windows): use forward slashes as directory
 separators

Git traditionally uses those, not backslashes, ever.

Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
---
 scalar.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/scalar.c b/scalar.c
index a1cee98c64..a27167430a 100644
--- a/scalar.c
+++ b/scalar.c
@@ -41,6 +41,9 @@ static void setup_enlistment_directory(int argc, const char **argv,
 		die(_("need a working directory"));
 
 	strbuf_trim_trailing_dir_sep(&path);
+#ifdef GIT_WINDOWS_NATIVE
+	convert_slashes(path.buf);
+#endif
 
 	/* check if currently in enlistment root with src/ workdir */
 	len = path.len;

From 82e36c676e12cde202e2f5bfd510bd75c0caf329 Mon Sep 17 00:00:00 2001
From: Johannes Schindelin <johannes.schindelin@gmx.de>
Date: Tue, 25 Jan 2022 23:49:21 +0100
Subject: [PATCH 17/29] scalar: adjust documentation to the microsoft/git fork

Scalar in Microsoft's Git fork can do a little more than Scalar in
upstream Git: in Microsoft's Git, it supports the GVFS protocol so that
Scalar can clone from Azure DevOps.

Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
---
 contrib/scalar/docs/getting-started.md | 24 +++++++++++++++----
 contrib/scalar/docs/index.md           | 14 +++++++----
 contrib/scalar/docs/philosophy.md      | 33 +++++++++++++++-----------
 contrib/scalar/docs/troubleshooting.md | 20 ++++++++++++++++
 4 files changed, 68 insertions(+), 23 deletions(-)

diff --git a/contrib/scalar/docs/getting-started.md b/contrib/scalar/docs/getting-started.md
index ef7ea07b0f..d512533032 100644
--- a/contrib/scalar/docs/getting-started.md
+++ b/contrib/scalar/docs/getting-started.md
@@ -18,8 +18,9 @@ Creating a new Scalar clone
 ---------------------------------------------------
 
 The `clone` verb creates a local enlistment of a remote repository using the
-partial clone feature available e.g. on GitHub.
-
+partial clone feature available e.g. on GitHub, or using the
+[GVFS protocol](https://github.com/microsoft/VFSForGit/blob/HEAD/Protocol.md),
+such as Azure Repos.
 
 ```
 scalar clone [options] <url> [<dir>]
@@ -68,11 +69,26 @@ in `<path>`.
 These options allow a user to customize their initial enlistment.
 
 * `--full-clone`: If specified, do not initialize the sparse-checkout feature.
-  All files will be present in your `src` directory. This uses a Git partial
-  clone: blobs are downloaded on demand.
+  All files will be present in your `src` directory. This behaves very similar
+  to a Git partial clone in that blobs are downloaded on demand. However, it
+  will use the GVFS protocol to download all Git objects.
+
+* `--cache-server-url=<url>`: If specified, set the intended cache server to
+  the specified `<url>`. All object queries will use the GVFS protocol to this
+  `<url>` instead of the origin remote. If the remote supplies a list of
+  cache servers via the `<url>/gvfs/config` endpoint, then the `clone` command
+  will select a nearby cache server from that list.
 
 * `--branch=<ref>`: Specify the branch to checkout after clone.
 
+* `--local-cache-path=<path>`: Use this option to override the path for the
+  local Scalar cache. If not specified, then Scalar will select a default
+  path to share objects with your other enlistments. On Windows, this path
+  is a subdirectory of `<Volume>:\.scalarCache\`. On Mac, this path is a
+  subdirectory of `~/.scalarCache/`. The default cache path is recommended so
+  multiple enlistments of the same remote repository share objects on the
+  same device.
+
 ### Advanced Options
 
 The options below are not intended for use by a typical user. These are
diff --git a/contrib/scalar/docs/index.md b/contrib/scalar/docs/index.md
index f9f5ab06e0..4f56e2b0eb 100644
--- a/contrib/scalar/docs/index.md
+++ b/contrib/scalar/docs/index.md
@@ -28,10 +28,14 @@ these features for that repo (except partial clone) and start running suggested
 maintenance in the background using
 [the `git maintenance` feature](https://git-scm.com/docs/git-maintenance).
 
-Repos cloned with the `scalar clone` command use partial clone to significantly
-reduce the amount of data required to get started using a repository. By
-delaying all blob downloads until they are required, Scalar allows you to work
-with very large repositories quickly.
+Repos cloned with the `scalar clone` command use partial clone or the
+[GVFS protocol](https://github.com/microsoft/VFSForGit/blob/HEAD/Protocol.md)
+to significantly reduce the amount of data required to get started
+using a repository. By delaying all blob downloads until they are required,
+Scalar allows you to work with very large repositories quickly. The GVFS
+protocol allows a network of _cache servers_ to serve objects with lower
+latency and higher throughput. The cache servers also reduce load on the
+central server.
 
 Documentation
 -------------
@@ -42,7 +46,7 @@ Documentation
 
 * [Troubleshooting](troubleshooting.md):
   Collect diagnostic information or update custom settings. Includes
-  `scalar diagnose`.
+  `scalar diagnose` and `scalar cache-server`.
 
 * [The Philosophy of Scalar](philosophy.md): Why does Scalar work the way
   it does, and how do we make decisions about its future?
diff --git a/contrib/scalar/docs/philosophy.md b/contrib/scalar/docs/philosophy.md
index 51486a75e4..e3dfa025a2 100644
--- a/contrib/scalar/docs/philosophy.md
+++ b/contrib/scalar/docs/philosophy.md
@@ -13,22 +13,27 @@ Scalar only to configure those new settings. In particular, we ported
 features like background maintenance to Git to make Scalar simpler and
 make Git more powerful.
 
-Services such as GitHub support partial clone , a standard adopted by the Git
-project to download only part of the Git objects when cloning, and fetching
-further objects on demand. If your hosting service supports partial clone, then
-we absolutely recommend it as a way to greatly speed up your clone and fetch
-times and to reduce how much disk space your Git repository requires. Scalar
-will help with this!
+Scalar ships inside [a custom version of Git][microsoft-git], but we are
+working to make it available in other forks of Git. The only feature
+that is not intended to ever reach the standard Git client is Scalar's use
+of [the GVFS Protocol][gvfs-protocol], which is essentially an older
+version of [Git's partial clone feature](https://github.blog/2020-12-21-get-up-to-speed-with-partial-clone-and-shallow-clone/)
+that was available first in Azure Repos. Services such as GitHub support
+only partial clone instead of the GVFS protocol because that is the
+standard adopted by the Git project. If your hosting service supports
+partial clone, then we absolutely recommend it as a way to greatly speed
+up your clone and fetch times and to reduce how much disk space your Git
+repository requires. Scalar will help with this!
 
-Most of the value of Scalar can be found in the core Git client. However, most
-of the advanced features that really optimize Git's performance are off by
-default for compatibility reasons. To really take advantage of Git's latest and
-greatest features, you either need to study the [`git config`
-documentation](https://git-scm.com/docs/git-config) and regularly read [the Git
-release notes](https://github.com/git/git/tree/master/Documentation/RelNotes).
+If you don't use the GVFS Protocol, then most of the value of Scalar can
+be found in the core Git client. However, most of the advanced features
+that really optimize Git's performance are off by default for compatibility
+reasons. To really take advantage of Git's latest and greatest features,
+you either need to study the [`git config` documentation](https://git-scm.com/docs/git-config)
+and regularly read [the Git release notes](https://github.com/git/git/tree/master/Documentation/RelNotes).
 Even if you do all that work and customize your Git settings on your machines,
-you likely will want to share those settings with other team members. Or, you
-can just use Scalar!
+you likely will want to share those settings with other team members.
+Or, you can just use Scalar!
 
 Using `scalar register` on an existing Git repository will give you these
 benefits:
diff --git a/contrib/scalar/docs/troubleshooting.md b/contrib/scalar/docs/troubleshooting.md
index 8ec56ad437..c54d2438f2 100644
--- a/contrib/scalar/docs/troubleshooting.md
+++ b/contrib/scalar/docs/troubleshooting.md
@@ -18,3 +18,23 @@ files for that repository. This includes:
 
 As the `diagnose` command completes, it provides the path of the resulting
 zip file. This zip can be attached to bug reports to make the analysis easier.
+
+Modifying Configuration Values
+------------------------------
+
+The Scalar-specific configuration is only available for repos using the
+GVFS protocol.
+
+### Cache Server URL
+
+When using an enlistment cloned with `scalar clone` and the GVFS protocol,
+you will have a value called the cache server URL. Cache servers are a feature
+of the GVFS protocol to provide low-latency access to the on-demand object
+requests. This modifies the `gvfs.cache-server` setting in your local Git config
+file.
+
+Run `scalar cache-server --get` to see the current cache server.
+
+Run `scalar cache-server --list` to see the available cache server URLs.
+
+Run `scalar cache-server --set=<url>` to set your cache server to `<url>`.

From 20e9822be55481ed13c398399f85593987e4b4ae Mon Sep 17 00:00:00 2001
From: Derrick Stolee <dstolee@microsoft.com>
Date: Thu, 17 Jun 2021 11:40:09 -0400
Subject: [PATCH 18/29] scalar: add retry logic to run_git()

Use a fixed 3 tries total to see how that increases our chances of
success for subcommands such as 'git fetch'.

We special-case the `diagnose` command here: When 672196a3073
(scalar-diagnose: use 'git diagnose --mode=all', 2022-08-12) updated
'scalar diagnose' to run 'git diagnose' as a subprocess, it was passed
through the run_git() caller. We need to avoid repeating the call when
the underlying 'git diagnose' command fails.

Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
---
 scalar.c | 25 ++++++++++++++++++++-----
 1 file changed, 20 insertions(+), 5 deletions(-)

diff --git a/scalar.c b/scalar.c
index a27167430a..b029446107 100644
--- a/scalar.c
+++ b/scalar.c
@@ -70,20 +70,33 @@ static void setup_enlistment_directory(int argc, const char **argv,
 	strbuf_release(&path);
 }
 
+static int git_retries = 3;
+
 static int run_git(const char *arg, ...)
 {
-	struct child_process cmd = CHILD_PROCESS_INIT;
 	va_list args;
 	const char *p;
+	struct strvec argv = STRVEC_INIT;
+	int res = 0, attempts;
 
 	va_start(args, arg);
-	strvec_push(&cmd.args, arg);
+	strvec_push(&argv, arg);
 	while ((p = va_arg(args, const char *)))
-		strvec_push(&cmd.args, p);
+		strvec_push(&argv, p);
 	va_end(args);
 
-	cmd.git_cmd = 1;
-	return run_command(&cmd);
+	for (attempts = 0, res = 1;
+	     res && attempts < git_retries;
+	     attempts++) {
+		struct child_process cmd = CHILD_PROCESS_INIT;
+
+		cmd.git_cmd = 1;
+		strvec_pushv(&cmd.args, argv.v);
+		res = run_command(&cmd);
+	}
+
+	strvec_clear(&argv);
+	return res;
 }
 
 struct scalar_config {
@@ -569,6 +582,8 @@ static int cmd_diagnose(int argc, const char **argv)
 	setup_enlistment_directory(argc, argv, usage, options, &diagnostics_root);
 	strbuf_addstr(&diagnostics_root, "/.scalarDiagnostics");
 
+	/* Here, a failure should not repeat itself. */
+	git_retries = 1;
 	res = run_git("diagnose", "--mode=all", "-s", "%Y%m%d_%H%M%S",
 		      "-o", diagnostics_root.buf, NULL);
 

From af6812e753872ff1792a1b28162e4e667b28d018 Mon Sep 17 00:00:00 2001
From: Derrick Stolee <dstolee@microsoft.com>
Date: Mon, 21 Jun 2021 08:21:48 -0400
Subject: [PATCH 19/29] scalar: enable untracked cache unconditionally

Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
---
 scalar.c | 16 ----------------
 1 file changed, 16 deletions(-)

diff --git a/scalar.c b/scalar.c
index 3992b933c9..2ec60a998b 100644
--- a/scalar.c
+++ b/scalar.c
@@ -156,23 +156,7 @@ static int set_recommended_config(int reconfigure)
 		{ "core.FSCache", "true", 1 },
 		{ "core.multiPackIndex", "true", 1 },
 		{ "core.preloadIndex", "true", 1 },
-#ifndef WIN32
 		{ "core.untrackedCache", "true", 1 },
-#else
-		/*
-		 * Unfortunately, Scalar's Functional Tests demonstrated
-		 * that the untracked cache feature is unreliable on Windows
-		 * (which is a bummer because that platform would benefit the
-		 * most from it). For some reason, freshly created files seem
-		 * not to update the directory's `lastModified` time
-		 * immediately, but the untracked cache would need to rely on
-		 * that.
-		 *
-		 * Therefore, with a sad heart, we disable this very useful
-		 * feature on Windows.
-		 */
-		{ "core.untrackedCache", "false", 1 },
-#endif
 		{ "core.logAllRefUpdates", "true", 1 },
 		{ "credential.https://dev.azure.com.useHttpPath", "true", 1 },
 		{ "credential.validate", "false", 1 }, /* GCM4W-only */

From 88074c49040ec24d353e46b25f0acb932fff1aff Mon Sep 17 00:00:00 2001
From: Johannes Schindelin <johannes.schindelin@gmx.de>
Date: Thu, 27 May 2021 07:26:11 +0200
Subject: [PATCH 20/29] scalar: support the `config` command for backwards
 compatibility

The .NET version supported running `scalar config` to reconfigure the
current enlistment, and now the C port does, too.

Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
---
 scalar.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/scalar.c b/scalar.c
index b029446107..dc266cdb35 100644
--- a/scalar.c
+++ b/scalar.c
@@ -958,6 +958,9 @@ int cmd_main(int argc, const char **argv)
 		argv++;
 		argc--;
 
+		if (!strcmp(argv[0], "config"))
+			argv[0] = "reconfigure";
+
 		for (i = 0; builtins[i].name; i++)
 			if (!strcmp(builtins[i].name, argv[0]))
 				return !!builtins[i].fn(argc, argv);

From d45a014071241e8b4a4793df18942f57c1a8506f Mon Sep 17 00:00:00 2001
From: Johannes Schindelin <johannes.schindelin@gmx.de>
Date: Tue, 10 May 2022 00:29:55 +0200
Subject: [PATCH 21/29] scalar diagnose: accommodate Scalar's Functional Tests

Those tests specifically verify that the `.zip` file path is shown on
`stdout`. Let's do that again, under the assumption that there are
scripts out there that rely on this behavior.

Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
---
 diagnose.c | 11 +++++++----
 1 file changed, 7 insertions(+), 4 deletions(-)

diff --git a/diagnose.c b/diagnose.c
index f47470f8cc..09515e39bb 100644
--- a/diagnose.c
+++ b/diagnose.c
@@ -380,10 +380,13 @@ int create_diagnostics_archive(struct strbuf *zip_path, enum diagnose_mode mode)
 		goto diagnose_cleanup;
 	}
 
-	fprintf(stderr, "\n"
-		"Diagnostics complete.\n"
-		"All of the gathered info is captured in '%s'\n",
-		zip_path->buf);
+	strbuf_reset(&buf);
+	strbuf_addf(&buf, "\n"
+		    "Diagnostics complete.\n"
+		    "All of the gathered info is captured in '%s'\n",
+		    zip_path->buf);
+	write_or_die(stdout_fd, buf.buf, buf.len);
+	write_or_die(2, buf.buf, buf.len);
 
 diagnose_cleanup:
 	if (archiver_fd >= 0) {

From 34bc5685edb7c4fcd3f66b623424887f07362f1a Mon Sep 17 00:00:00 2001
From: Johannes Schindelin <johannes.schindelin@gmx.de>
Date: Tue, 24 Aug 2021 18:01:47 +0200
Subject: [PATCH 22/29] scalar: parse `clone --no-fetch-commits-and-trees` for
 backwards compatibility

This option does not do anything anymore, though.

Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
---
 scalar.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/scalar.c b/scalar.c
index 2ec60a998b..42a2e8c015 100644
--- a/scalar.c
+++ b/scalar.c
@@ -673,6 +673,7 @@ static int init_shared_object_cache(const char *url,
 
 static int cmd_clone(int argc, const char **argv)
 {
+	int dummy = 0;
 	const char *branch = NULL;
 	int full_clone = 0, single_branch = 0, show_progress = isatty(2);
 	const char *cache_server_url = NULL, *local_cache_root = NULL;
@@ -691,6 +692,8 @@ static int cmd_clone(int argc, const char **argv)
 		OPT_STRING(0, "local-cache-path", &local_cache_root,
 			   N_("<path>"),
 			   N_("override the path for the local Scalar cache")),
+		OPT_HIDDEN_BOOL(0, "no-fetch-commits-and-trees",
+				&dummy, N_("no longer used")),
 		OPT_END(),
 	};
 	const char * const clone_usage[] = {

From 8d5d5ca53a92281e83120c9555aec3505893bb3e Mon Sep 17 00:00:00 2001
From: Johannes Schindelin <johannes.schindelin@gmx.de>
Date: Tue, 8 Jun 2021 09:43:36 +0200
Subject: [PATCH 23/29] ci: run Scalar's Functional Tests

Scalar's Functional Test suite is pretty comprehensive, and caught more
than just one bug in the built-in FSMonitor that was missed by Git's own
test suite.

To benefit from this test suite, automatically run it on the `vfs-*`
and `features/*` branches.

Note: for simplicity, we're building Git from scratch in all matrix
jobs.

Also note: for speed, we are using `git-sdk-64-minimal`, even if it
lacks the `/bin/install` that we need to install Git's files; We're
providing a minimal shell script shim instead. Also, we do not need to
bother with the Tcl/Tk parts, therefore we're skipping them, too.

Finally, we use GIT_FORCE_UNTRACKED_CACHE in the functional tests, to
give the untracked cache a thorough work-out.

Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
---
 .github/workflows/scalar-functional-tests.yml | 220 ++++++++++++++++++
 1 file changed, 220 insertions(+)
 create mode 100644 .github/workflows/scalar-functional-tests.yml

diff --git a/.github/workflows/scalar-functional-tests.yml b/.github/workflows/scalar-functional-tests.yml
new file mode 100644
index 0000000000..7226054aeb
--- /dev/null
+++ b/.github/workflows/scalar-functional-tests.yml
@@ -0,0 +1,220 @@
+name: Scalar Functional Tests
+
+env:
+  SCALAR_REPOSITORY: microsoft/scalar
+  SCALAR_REF: main
+  DEBUG_WITH_TMATE: false
+  SCALAR_TEST_SKIP_VSTS_INFO: true
+
+on:
+  push:
+    branches: [ vfs-*, tentative/vfs-* ]
+  pull_request:
+    branches: [ vfs-*, features/* ]
+
+jobs:
+  scalar:
+    name: "Scalar Functional Tests"
+
+    strategy:
+      fail-fast: false
+      matrix:
+        # Order by runtime (in descending order)
+        os: [windows-2019, macos-11, ubuntu-20.04, ubuntu-22.04]
+        # Scalar.NET used to be tested using `features: [false, experimental]`
+        # But currently, Scalar/C ignores `feature.scalar` altogether, so let's
+        # save some electrons and run only one of them...
+        features: [ignored]
+        exclude:
+          # The built-in FSMonitor is not (yet) supported on Linux
+          - os: ubuntu-20.04
+            features: experimental
+          - os: ubuntu-22.04
+            features: experimental
+    runs-on: ${{ matrix.os }}
+
+    env:
+      BUILD_FRAGMENT: bin/Release/netcoreapp3.1
+      GIT_FORCE_UNTRACKED_CACHE: 1
+
+    steps:
+      - name: Check out Git's source code
+        uses: actions/checkout@v3
+
+      - name: Setup build tools on Windows
+        if: runner.os == 'Windows'
+        uses: git-for-windows/setup-git-for-windows-sdk@v1
+
+      - name: Provide a minimal `install` on Windows
+        if: runner.os == 'Windows'
+        shell: bash
+        run: |
+          test -x /usr/bin/install ||
+          tr % '\t' >/usr/bin/install <<-\EOF
+          #!/bin/sh
+
+          cmd=cp
+          while test $# != 0
+          do
+          %case "$1" in
+          %-d) cmd="mkdir -p";;
+          %-m) shift;; # ignore mode
+          %*) break;;
+          %esac
+          %shift
+          done
+
+          exec $cmd "$@"
+          EOF
+
+      - name: Install build dependencies for Git (Linux)
+        if: runner.os == 'Linux'
+        run: |
+          sudo apt-get update
+          sudo apt-get -q -y install libssl-dev libcurl4-openssl-dev gettext
+
+      - name: Build and install Git
+        shell: bash
+        env:
+          NO_TCLTK: Yup
+        run: |
+          # We do require a VFS version
+          def_ver="$(sed -n 's/DEF_VER=\(.*vfs.*\)/\1/p' GIT-VERSION-GEN)"
+          test -n "$def_ver"
+
+          # Ensure that `git version` reflects DEF_VER
+          case "$(git describe --match "v[0-9]*vfs*" HEAD)" in
+          ${def_ver%%.vfs.*}.vfs.*) ;; # okay, we can use this
+          *) git -c user.name=ci -c user.email=ci@github tag -m for-testing ${def_ver}.NNN.g$(git rev-parse --short HEAD);;
+          esac
+
+          SUDO=
+          extra=
+          case "${{ runner.os }}" in
+          Windows)
+            extra=DESTDIR=/c/Progra~1/Git
+            cygpath -aw "/c/Program Files/Git/cmd" >>$GITHUB_PATH
+            ;;
+          Linux)
+            SUDO=sudo
+            extra=prefix=/usr
+            ;;
+          macOS)
+            SUDO=sudo
+            extra=prefix=/usr/local
+            ;;
+          esac
+
+          $SUDO make -j5 $extra install
+
+      - name: Ensure that we use the built Git and Scalar
+        shell: bash
+        run: |
+          type -p git
+          git version
+          case "$(git version)" in *.vfs.*) echo Good;; *) exit 1;; esac
+          type -p scalar
+          scalar version
+          case "$(scalar version 2>&1)" in *.vfs.*) echo Good;; *) exit 1;; esac
+
+      - name: Check out Scalar's source code
+        uses: actions/checkout@v3
+        with:
+          fetch-depth: 0 # Indicate full history so Nerdbank.GitVersioning works.
+          path: scalar
+          repository: ${{ env.SCALAR_REPOSITORY }}
+          ref: ${{ env.SCALAR_REF }}
+
+      - name: Setup .NET Core
+        uses: actions/setup-dotnet@v3
+        with:
+          dotnet-version: '3.1.x'
+
+      - name: Install dependencies
+        run: dotnet restore
+        working-directory: scalar
+        env:
+          DOTNET_NOLOGO: 1
+
+      - name: Build
+        working-directory: scalar
+        run: dotnet build --configuration Release --no-restore -p:UseAppHost=true # Force generation of executable on macOS.
+
+      - name: Setup platform (Linux)
+        if: runner.os == 'Linux'
+        run: |
+          echo "BUILD_PLATFORM=${{ runner.os }}" >>$GITHUB_ENV
+          echo "TRACE2_BASENAME=Trace2.${{ github.run_id }}__${{ github.run_number }}__${{ matrix.os }}__${{ matrix.features }}" >>$GITHUB_ENV
+
+      - name: Setup platform (Mac)
+        if: runner.os == 'macOS'
+        run: |
+          echo 'BUILD_PLATFORM=Mac' >>$GITHUB_ENV
+          echo "TRACE2_BASENAME=Trace2.${{ github.run_id }}__${{ github.run_number }}__${{ matrix.os }}__${{ matrix.features }}" >>$GITHUB_ENV
+
+      - name: Setup platform (Windows)
+        if: runner.os == 'Windows'
+        run: |
+          echo "BUILD_PLATFORM=${{ runner.os }}" >>$env:GITHUB_ENV
+          echo 'BUILD_FILE_EXT=.exe' >>$env:GITHUB_ENV
+          echo "TRACE2_BASENAME=Trace2.${{ github.run_id }}__${{ github.run_number }}__${{ matrix.os }}__${{ matrix.features }}" >>$env:GITHUB_ENV
+
+      - name: Configure feature.scalar
+        run: git config --global feature.scalar ${{ matrix.features }}
+
+      - id: functional_test
+        name: Functional test
+        timeout-minutes: 60
+        working-directory: scalar
+        shell: bash
+        run: |
+          export GIT_TRACE2_EVENT="$PWD/$TRACE2_BASENAME/Event"
+          export GIT_TRACE2_PERF="$PWD/$TRACE2_BASENAME/Perf"
+          export GIT_TRACE2_EVENT_BRIEF=true
+          export GIT_TRACE2_PERF_BRIEF=true
+          mkdir -p "$TRACE2_BASENAME"
+          mkdir -p "$TRACE2_BASENAME/Event"
+          mkdir -p "$TRACE2_BASENAME/Perf"
+          git version --build-options
+          cd ../out
+          Scalar.FunctionalTests/$BUILD_FRAGMENT/Scalar.FunctionalTests$BUILD_FILE_EXT --test-scalar-on-path --test-git-on-path --timeout=300000 --full-suite
+
+      - name: Force-stop FSMonitor daemons and Git processes (Windows)
+        if: runner.os == 'Windows' && (success() || failure())
+        shell: bash
+        run: |
+          set -x
+          wmic process get CommandLine,ExecutablePath,HandleCount,Name,ParentProcessID,ProcessID
+          wmic process where "CommandLine Like '%fsmonitor--daemon %run'" delete
+          wmic process where "ExecutablePath Like '%git.exe'" delete
+
+      - id: trace2_zip_unix
+        if: runner.os != 'Windows' && ( success() || failure() ) && ( steps.functional_test.conclusion == 'success' || steps.functional_test.conclusion == 'failure' )
+        name: Zip Trace2 Logs (Unix)
+        shell: bash
+        working-directory: scalar
+        run: zip -q -r $TRACE2_BASENAME.zip $TRACE2_BASENAME/
+
+      - id: trace2_zip_windows
+        if: runner.os == 'Windows' && ( success() || failure() ) && ( steps.functional_test.conclusion == 'success' || steps.functional_test.conclusion == 'failure' )
+        name: Zip Trace2 Logs (Windows)
+        working-directory: scalar
+        run: Compress-Archive -DestinationPath ${{ env.TRACE2_BASENAME }}.zip -Path ${{ env.TRACE2_BASENAME }}
+
+      - name: Archive Trace2 Logs
+        if: ( success() || failure() ) && ( steps.trace2_zip_unix.conclusion == 'success' || steps.trace2_zip_windows.conclusion == 'success' )
+        uses: actions/upload-artifact@v3
+        with:
+          name: ${{ env.TRACE2_BASENAME }}.zip
+          path: scalar/${{ env.TRACE2_BASENAME }}.zip
+          retention-days: 3
+
+      # The GitHub Action `action-tmate` allows developers to connect to the running agent
+      # using SSH (it will be a `tmux` session; on Windows agents it will be inside the MSYS2
+      # environment in `C:\msys64`, therefore it can be slightly tricky to interact with
+      # Git for Windows, which runs a slightly incompatible MSYS2 runtime).
+      - name: action-tmate
+        if: env.DEBUG_WITH_TMATE == 'true' && failure()
+        uses: mxschmitt/action-tmate@v3
+        with:
+          limit-access-to-actor: true

From 822ec7a946b82c3acbfe5bd070338164b9867b8a Mon Sep 17 00:00:00 2001
From: Derrick Stolee <derrickstolee@github.com>
Date: Tue, 4 Oct 2022 08:01:17 -0400
Subject: [PATCH 24/29] abspath: make strip_last_path_component() global

The strip_last_component() method is helpful for finding the parent
directory of a path stored in a strbuf. Extract it to a global method
advertised in path.h. With that additional visibility, it is helpful to
rename it to be more specific to paths.

Signed-off-by: Derrick Stolee <derrickstolee@github.com>
---
 abspath.c | 6 +++---
 path.h    | 4 ++++
 2 files changed, 7 insertions(+), 3 deletions(-)

diff --git a/abspath.c b/abspath.c
index 7a5f114382..058e638e98 100644
--- a/abspath.c
+++ b/abspath.c
@@ -12,7 +12,7 @@ int is_directory(const char *path)
 }
 
 /* removes the last path component from 'path' except if 'path' is root */
-static void strip_last_component(struct strbuf *path)
+void strip_last_path_component(struct strbuf *path)
 {
 	size_t offset = offset_1st_component(path->buf);
 	size_t len = path->len;
@@ -117,7 +117,7 @@ static char *strbuf_realpath_1(struct strbuf *resolved, const char *path,
 			continue; /* '.' component */
 		} else if (next.len == 2 && !strcmp(next.buf, "..")) {
 			/* '..' component; strip the last path component */
-			strip_last_component(resolved);
+			strip_last_path_component(resolved);
 			continue;
 		}
 
@@ -169,7 +169,7 @@ static char *strbuf_realpath_1(struct strbuf *resolved, const char *path,
 				 * strip off the last component since it will
 				 * be replaced with the contents of the symlink
 				 */
-				strip_last_component(resolved);
+				strip_last_path_component(resolved);
 			}
 
 			/*
diff --git a/path.h b/path.h
index 0a59c85a62..db417a43bd 100644
--- a/path.h
+++ b/path.h
@@ -179,6 +179,10 @@ const char *git_path_auto_merge(struct repository *r);
 const char *git_path_fetch_head(struct repository *r);
 const char *git_path_shallow(struct repository *r);
 
+/**
+ * Remove the last path component from 'path' except if 'path' is root.
+ */
+void strip_last_path_component(struct strbuf *path);
 
 int ends_with_path_components(const char *path, const char *components);
 

From 61f997b2ad2112144121c0a09d3241ab7226c81f Mon Sep 17 00:00:00 2001
From: Derrick Stolee <derrickstolee@github.com>
Date: Tue, 24 May 2022 14:39:29 -0400
Subject: [PATCH 25/29] setup: add discover_git_directory_reason()

There are many reasons why discovering a Git directory may fail. In
particular, 8959555cee7 (setup_git_directory(): add an owner check for
the top-level directory, 2022-03-02) added ownership checks as a
security precaution.

Callers attempting to set up a Git directory may want to inform the user
about the reason for the failure. For that, expose the enum
discovery_result from within setup.c and into cache.h where
discover_git_directory() is defined.

I initially wanted to change the return type of discover_git_directory()
to be this enum, but several callers rely upon the "zero means success".
The two problems with this are:

1. The zero value of the enum is actually GIT_DIR_NONE, so nonpositive
   results are errors.

2. There are multiple successful states, so some positive results are
   successful.

Instead of updating all callers immediately, add a new method,
discover_git_directory_reason(), and convert discover_git_directory() to
be a thin shim on top of it.

Because there are extra checks that discover_git_directory_reason() does
after setup_git_directory_gently_1(), there are other modes that can be
returned for failure states. Add these modes to the enum, but be sure to
explicitly add them as BUG() states in the switch of
setup_git_directory_gently().

Signed-off-by: Derrick Stolee <derrickstolee@github.com>
---
 cache.h | 32 ++++++++++++++++++++++++++++++--
 setup.c | 32 +++++++++++---------------------
 2 files changed, 41 insertions(+), 23 deletions(-)

diff --git a/cache.h b/cache.h
index 1c689455b3..91e0b9947e 100644
--- a/cache.h
+++ b/cache.h
@@ -611,6 +611,30 @@ void set_git_work_tree(const char *tree);
 #define ALTERNATE_DB_ENVIRONMENT "GIT_ALTERNATE_OBJECT_DIRECTORIES"
 
 void setup_work_tree(void);
+
+/*
+ * discover_git_directory_reason() is similar to discover_git_directory(),
+ * except it returns an enum value instead. It is important to note that
+ * a zero-valued return here is actually GIT_DIR_NONE, which is different
+ * from discover_git_directory.
+ */
+enum discovery_result {
+	GIT_DIR_NONE = 0,
+	GIT_DIR_EXPLICIT,
+	GIT_DIR_DISCOVERED,
+	GIT_DIR_BARE,
+	/* these are errors */
+	GIT_DIR_HIT_CEILING = -1,
+	GIT_DIR_HIT_MOUNT_POINT = -2,
+	GIT_DIR_INVALID_GITFILE = -3,
+	GIT_DIR_INVALID_OWNERSHIP = -4,
+	GIT_DIR_DISALLOWED_BARE = -5,
+	GIT_DIR_INVALID_FORMAT = -6,
+	GIT_DIR_CWD_FAILURE = -7,
+};
+enum discovery_result discover_git_directory_reason(struct strbuf *commondir,
+						    struct strbuf *gitdir);
+
 /*
  * Find the commondir and gitdir of the repository that contains the current
  * working directory, without changing the working directory or other global
@@ -619,8 +643,12 @@ void setup_work_tree(void);
  * both have the same result appended to the buffer.  The return value is
  * either 0 upon success and non-zero if no repository was found.
  */
-int discover_git_directory(struct strbuf *commondir,
-			   struct strbuf *gitdir);
+static inline int discover_git_directory(struct strbuf *commondir,
+					 struct strbuf *gitdir)
+{
+	return discover_git_directory_reason(commondir, gitdir) <= 0;
+}
+
 const char *setup_git_directory_gently(int *);
 const char *setup_git_directory(void);
 char *prefix_path(const char *prefix, int len, const char *path);
diff --git a/setup.c b/setup.c
index 852b3d9141..ca3f83cc6d 100644
--- a/setup.c
+++ b/setup.c
@@ -1205,19 +1205,6 @@ static const char *allowed_bare_repo_to_string(
 	return NULL;
 }
 
-enum discovery_result {
-	GIT_DIR_NONE = 0,
-	GIT_DIR_EXPLICIT,
-	GIT_DIR_DISCOVERED,
-	GIT_DIR_BARE,
-	/* these are errors */
-	GIT_DIR_HIT_CEILING = -1,
-	GIT_DIR_HIT_MOUNT_POINT = -2,
-	GIT_DIR_INVALID_GITFILE = -3,
-	GIT_DIR_INVALID_OWNERSHIP = -4,
-	GIT_DIR_DISALLOWED_BARE = -5,
-};
-
 /*
  * We cannot decide in this function whether we are in the work tree or
  * not, since the config can only be read _after_ this function was called.
@@ -1368,21 +1355,22 @@ static enum discovery_result setup_git_directory_gently_1(struct strbuf *dir,
 	}
 }
 
-int discover_git_directory(struct strbuf *commondir,
-			   struct strbuf *gitdir)
+enum discovery_result discover_git_directory_reason(struct strbuf *commondir,
+						    struct strbuf *gitdir)
 {
 	struct strbuf dir = STRBUF_INIT, err = STRBUF_INIT;
 	size_t gitdir_offset = gitdir->len, cwd_len;
 	size_t commondir_offset = commondir->len;
 	struct repository_format candidate = REPOSITORY_FORMAT_INIT;
+	enum discovery_result result;
 
 	if (strbuf_getcwd(&dir))
-		return -1;
+		return GIT_DIR_CWD_FAILURE;
 
 	cwd_len = dir.len;
-	if (setup_git_directory_gently_1(&dir, gitdir, NULL, 0) <= 0) {
+	if ((result = setup_git_directory_gently_1(&dir, gitdir, NULL, 0)) <= 0) {
 		strbuf_release(&dir);
-		return -1;
+		return result;
 	}
 
 	/*
@@ -1412,7 +1400,7 @@ int discover_git_directory(struct strbuf *commondir,
 		strbuf_setlen(commondir, commondir_offset);
 		strbuf_setlen(gitdir, gitdir_offset);
 		clear_repository_format(&candidate);
-		return -1;
+		return GIT_DIR_INVALID_FORMAT;
 	}
 
 	/* take ownership of candidate.partial_clone */
@@ -1421,7 +1409,7 @@ int discover_git_directory(struct strbuf *commondir,
 	candidate.partial_clone = NULL;
 
 	clear_repository_format(&candidate);
-	return 0;
+	return result;
 }
 
 const char *setup_git_directory_gently(int *nongit_ok)
@@ -1513,9 +1501,11 @@ const char *setup_git_directory_gently(int *nongit_ok)
 		*nongit_ok = 1;
 		break;
 	case GIT_DIR_NONE:
+	case GIT_DIR_CWD_FAILURE:
+	case GIT_DIR_INVALID_FORMAT:
 		/*
 		 * As a safeguard against setup_git_directory_gently_1 returning
-		 * this value, fallthrough to BUG. Otherwise it is possible to
+		 * these values, fallthrough to BUG. Otherwise it is possible to
 		 * set startup_info->have_repository to 1 when we did nothing to
 		 * find a repository.
 		 */

From 16ca43d5c026f7a9bf5321a37cb2eab6845a3a18 Mon Sep 17 00:00:00 2001
From: Victoria Dye <vdye@github.com>
Date: Tue, 5 Apr 2022 13:42:23 -0700
Subject: [PATCH 26/29] scalar: upgrade to newest FSMonitor config setting

When FSMonitor was upstreamed, the 'core.useBuiltinFSMonitor' config was
deprecated and replaced with an overload of the 'core.fsmonitor' config
(i.e., if a boolean value was specified in 'core.fsmonitor', it is treated
the way 'core.useBuiltinFSMonitor' originally was). Because 'scalar
register' actively sets that config, use it to upgrade the deprecated config
setting.

Co-authored-by: Johannes Schindelin <Johannes.Schindelin@gmx.de>
Signed-off-by: Victoria Dye <vdye@github.com>
---
 scalar.c | 17 +++++++++++++++++
 1 file changed, 17 insertions(+)

diff --git a/scalar.c b/scalar.c
index 42a2e8c015..a295fb7033 100644
--- a/scalar.c
+++ b/scalar.c
@@ -189,6 +189,23 @@ static int set_recommended_config(int reconfigure)
 	int i;
 	char *value;
 
+	/*
+	 * If a user has "core.usebuiltinfsmonitor" enabled, try to switch to
+	 * the new (non-deprecated) setting (core.fsmonitor).
+	 */
+	if (!git_config_get_string("core.usebuiltinfsmonitor", &value)) {
+		char *dummy = NULL;
+		if (git_config_get_string("core.fsmonitor", &dummy) &&
+		    git_config_set_gently("core.fsmonitor", value) < 0)
+			return error(_("could not configure %s=%s"),
+				     "core.fsmonitor", value);
+		if (git_config_set_gently("core.usebuiltinfsmonitor", NULL) < 0)
+			return error(_("could not configure %s=%s"),
+				     "core.useBuiltinFSMonitor", "NULL");
+		free(value);
+		free(dummy);
+	}
+
 	for (i = 0; config[i].key; i++) {
 		if (set_scalar_config(config + i, reconfigure))
 			return error(_("could not configure %s=%s"),

From 35434e13331b56fda48df85a6dc8b045a2c84095 Mon Sep 17 00:00:00 2001
From: Derrick Stolee <derrickstolee@github.com>
Date: Fri, 30 Sep 2022 12:41:27 -0400
Subject: [PATCH 27/29] scalar: .scalarCache should live above enlistment

We should not be putting the .scalarCache inside the enlistment as a
sibling to the 'src' directory. This only happens in "unattended" mode,
but it also negates any benefit of a shared object cache because each
enlistment absolutely does not share any objects with others.

Move the shared object cache in this case to a level above the
enlistment, so at least there is some hope that it can be reused. This
is also critical to the upcoming --no-src option, since the shared
object cache cannot be located within the Git repository.

Signed-off-by: Derrick Stolee <derrickstolee@github.com>
---
 scalar.c          | 10 ++++++++--
 t/t9210-scalar.sh |  2 +-
 2 files changed, 9 insertions(+), 3 deletions(-)

diff --git a/scalar.c b/scalar.c
index 08bac10525..003aaeab9b 100644
--- a/scalar.c
+++ b/scalar.c
@@ -16,6 +16,7 @@
 #include "help.h"
 #include "json-parser.h"
 #include "remote.h"
+#include "path.h"
 
 static int is_unattended(void) {
 	return git_env_bool("Scalar_UNATTENDED", 0);
@@ -470,8 +471,13 @@ static char *default_cache_root(const char *root)
 {
 	const char *env;
 
-	if (is_unattended())
-		return xstrfmt("%s/.scalarCache", root);
+	if (is_unattended()) {
+		struct strbuf path = STRBUF_INIT;
+		strbuf_addstr(&path, root);
+		strip_last_path_component(&path);
+		strbuf_addstr(&path, "/.scalarCache");
+		return strbuf_detach(&path, NULL);
+	}
 
 #ifdef WIN32
 	(void)env;
diff --git a/t/t9210-scalar.sh b/t/t9210-scalar.sh
index 6acf279afd..548bff71ae 100755
--- a/t/t9210-scalar.sh
+++ b/t/t9210-scalar.sh
@@ -306,7 +306,7 @@ test_expect_success '`scalar clone` with GVFS-enabled server' '
 	cache_key="url_$(printf "%s" http://$HOST_PORT/ |
 		tr A-Z a-z |
 		test-tool sha1)" &&
-	echo "$(pwd)/using-gvfs/.scalarCache/$cache_key" >expect &&
+	echo "$(pwd)/.scalarCache/$cache_key" >expect &&
 	git -C using-gvfs/src config gvfs.sharedCache >actual &&
 	test_cmp expect actual &&
 

From 855b464348209b392f6d3024204e2bafbda03500 Mon Sep 17 00:00:00 2001
From: Derrick Stolee <derrickstolee@github.com>
Date: Mon, 23 May 2022 10:56:35 -0400
Subject: [PATCH 28/29] scalar reconfigure: help users remove buggy repos

When running 'scalar reconfigure -a', such as at install time, Scalar
has warning messages about the repository missing (or not containing a
.git directory). Failures can also happen while trying to modify the
repository-local config for that repository.

These warnings may seem confusing to users who don't understand what
they mean or how to stop them.

Add a warning that instructs the user how to remove the warning in
future installations.

Signed-off-by: Derrick Stolee <derrickstolee@github.com>
---
 scalar.c | 52 +++++++++++++++++++++++++++++++++++++---------------
 1 file changed, 37 insertions(+), 15 deletions(-)

diff --git a/scalar.c b/scalar.c
index a295fb7033..08bac10525 100644
--- a/scalar.c
+++ b/scalar.c
@@ -1026,6 +1026,7 @@ static int cmd_reconfigure(int argc, const char **argv)
 	git_config(get_scalar_repos, &scalar_repos);
 
 	for (i = 0; i < scalar_repos.nr; i++) {
+		int failed = 0;
 		const char *dir = scalar_repos.items[i].string;
 
 		strbuf_reset(&commondir);
@@ -1036,30 +1037,51 @@ static int cmd_reconfigure(int argc, const char **argv)
 
 			if (errno != ENOENT) {
 				warning_errno(_("could not switch to '%s'"), dir);
-				res = -1;
-				continue;
+				failed = -1;
+				goto loop_end;
 			}
 
 			strbuf_addstr(&buf, dir);
 			if (remove_deleted_enlistment(&buf))
-				res = error(_("could not remove stale "
-					      "scalar.repo '%s'"), dir);
+				failed = error(_("could not remove stale "
+						 "scalar.repo '%s'"), dir);
 			else
-				warning(_("removing stale scalar.repo '%s'"),
+				warning(_("removed stale scalar.repo '%s'"),
 					dir);
 			strbuf_release(&buf);
-		} else if (discover_git_directory(&commondir, &gitdir) < 0) {
-			warning_errno(_("git repository gone in '%s'"), dir);
-			res = -1;
-		} else {
-			git_config_clear();
+			goto loop_end;
+		}
 
-			the_repository = &r;
-			r.commondir = commondir.buf;
-			r.gitdir = gitdir.buf;
+		switch (discover_git_directory_reason(&commondir, &gitdir)) {
+		case GIT_DIR_INVALID_OWNERSHIP:
+			warning(_("repository at '%s' has different owner"), dir);
+			failed = -1;
+			goto loop_end;
 
-			if (set_recommended_config(1) < 0)
-				res = -1;
+		case GIT_DIR_DISCOVERED:
+			break;
+
+		default:
+			warning(_("repository not found in '%s'"), dir);
+			failed = -1;
+			break;
+		}
+
+		git_config_clear();
+
+		the_repository = &r;
+		r.commondir = commondir.buf;
+		r.gitdir = gitdir.buf;
+
+		if (set_recommended_config(1) < 0)
+			failed = -1;
+
+loop_end:
+		if (failed) {
+			res = failed;
+			warning(_("to unregister this repository from Scalar, run\n"
+				  "\tgit config --global --unset --fixed-value scalar.repo \"%s\""),
+				dir);
 		}
 	}
 

From 78fa0e8d6d76537eb05809aa109fcd37e59fe894 Mon Sep 17 00:00:00 2001
From: Derrick Stolee <derrickstolee@github.com>
Date: Fri, 30 Sep 2022 12:44:27 -0400
Subject: [PATCH 29/29] scalar: add --no-src option

Some users have strong aversions to Scalar's opinion that the repository
should be in a 'src' directory, even though it creates a clean slate for
placing build outputs in adjacent directories.

The --no-src option allows users to opt-out of the default behavior. The
parse-opt logic automatically figures out that '--src' is a possible
flag that negates '--no-src'.

Signed-off-by: Derrick Stolee <derrickstolee@github.com>
---
 Documentation/scalar.txt |  5 ++++-
 scalar.c                 | 14 ++++++++++++--
 t/t9210-scalar.sh        |  8 ++++++++
 3 files changed, 24 insertions(+), 3 deletions(-)

diff --git a/Documentation/scalar.txt b/Documentation/scalar.txt
index 0424d020d9..35f67801e9 100644
--- a/Documentation/scalar.txt
+++ b/Documentation/scalar.txt
@@ -9,7 +9,7 @@ SYNOPSIS
 --------
 [verse]
 scalar clone [--single-branch] [--branch <main-branch>] [--full-clone]
-	[--local-cache-path <path>] [--cache-server-url <url>]
+	[--local-cache-path <path>] [--cache-server-url <url>] [--[no-]src]
 	<url> [<enlistment>]
 scalar list
 scalar register [<enlistment>]
@@ -83,6 +83,9 @@ remote-tracking branch for the branch this option was used for the initial
 cloning. If the HEAD at the remote did not point at any branch when
 `--single-branch` clone was made, no remote-tracking branch is created.
 
+--no-src::
+	Skip adding a `src` directory within the target enlistment.
+
 --[no-]full-clone::
 	A sparse-checkout is initialized by default. This behavior can be
 	turned off via `--full-clone`.
diff --git a/scalar.c b/scalar.c
index 003aaeab9b..0b0d6297da 100644
--- a/scalar.c
+++ b/scalar.c
@@ -701,6 +701,8 @@ static int cmd_clone(int argc, const char **argv)
 	int full_clone = 0, single_branch = 0, show_progress = isatty(2);
 	const char *cache_server_url = NULL, *local_cache_root = NULL;
 	char *default_cache_server_url = NULL, *local_cache_root_abs = NULL;
+	const char *enlistment_parent;
+	int no_src = 0;
 	struct option clone_options[] = {
 		OPT_STRING('b', "branch", &branch, N_("<branch>"),
 			   N_("branch to checkout after clone")),
@@ -709,6 +711,8 @@ static int cmd_clone(int argc, const char **argv)
 		OPT_BOOL(0, "single-branch", &single_branch,
 			 N_("only download metadata for the branch that will "
 			    "be checked out")),
+		OPT_BOOL(0, "no-src", &no_src,
+			 N_("skip creating a 'src' directory")),
 		OPT_STRING(0, "cache-server-url", &cache_server_url,
 			   N_("<url>"),
 			   N_("the url or friendly name of the cache server")),
@@ -759,7 +763,13 @@ static int cmd_clone(int argc, const char **argv)
 
 	ensure_absolute_path(enlistment, &enlistment);
 
-	dir = xstrfmt("%s/src", enlistment);
+	if (!no_src) {
+		dir = xstrfmt("%s/src", enlistment);
+		enlistment_parent = "../..";
+	} else {
+		dir = xstrdup(enlistment);
+		enlistment_parent = "..";
+	}
 
 	if (!local_cache_root)
 		local_cache_root = local_cache_root_abs =
@@ -800,7 +810,7 @@ static int cmd_clone(int argc, const char **argv)
 		struct strbuf path = STRBUF_INIT;
 
 		strbuf_addstr(&path, enlistment);
-		if (chdir("../..") < 0 ||
+		if (chdir(enlistment_parent) < 0 ||
 		    remove_dir_recursively(&path, 0) < 0)
 			die(_("'--local-cache-path' cannot be inside the src "
 			      "folder;\nCould not remove '%s'"), enlistment);
diff --git a/t/t9210-scalar.sh b/t/t9210-scalar.sh
index 548bff71ae..43c4208cdc 100755
--- a/t/t9210-scalar.sh
+++ b/t/t9210-scalar.sh
@@ -391,4 +391,12 @@ test_expect_success '`scalar delete` with existing repo' '
 	test_path_is_missing existing
 '
 
+test_expect_success '`scalar clone --no-src`' '
+	scalar clone --src "file://$(pwd)" with-src &&
+	scalar clone --no-src "file://$(pwd)" without-src &&
+
+	test_path_is_dir with-src/src &&
+	test_path_is_missing without-src/src
+'
+
 test_done