http: do not ignore proxy path

The documentation for `http.proxy` describes that option, and the
environment variables it overrides, as supporting "the syntax understood
by curl". curl allows SOCKS proxies to use a path to a Unix domain
socket, like `socks5h://localhost/path/to/socket.sock`. Git should
therefore include, if present, the path part of the proxy URL in what it
passes to libcurl.

Co-authored-by: Jeff King <peff@peff.net>
Signed-off-by: Ryan Hendrickson <ryan.hendrickson@alum.mit.edu>

Author: rhendric

Documentation/config/http.txt

@@ -5,8 +5,8 @@ http.proxy::
 	proxy string with a user name but no password, in which case git will
 	attempt to acquire one in the same way it does for other credentials. See
 	linkgit:gitcredentials[7] for more information. The syntax thus is
-	'[protocol://][user[:password]@]proxyhost[:port]'. This can be overridden
-	on a per-remote basis; see remote.<name>.proxy
+	'[protocol://][user[:password]@]proxyhost[:port][/path]'. This can be
+	overridden on a per-remote basis; see remote.<name>.proxy
 +
 Any proxy, however configured, must be completely transparent and must not
 modify, transform, or buffer the request or response in any way.  Proxies which

http.c

@@ -1227,6 +1227,8 @@ static CURL *get_curl_handle(void)
 		 */
 		curl_easy_setopt(result, CURLOPT_PROXY, "");
 	} else if (curl_http_proxy) {
+		struct strbuf proxy = STRBUF_INIT;
+
 		if (starts_with(curl_http_proxy, "socks5h"))
 			curl_easy_setopt(result,
 				CURLOPT_PROXYTYPE, CURLPROXY_SOCKS5_HOSTNAME);
@@ -1265,7 +1267,26 @@ static CURL *get_curl_handle(void)
 		if (!proxy_auth.host)
 			die("Invalid proxy URL '%s'", curl_http_proxy);
 
-		curl_easy_setopt(result, CURLOPT_PROXY, proxy_auth.host);
+		strbuf_addstr(&proxy, proxy_auth.host);
+		if (proxy_auth.path) {
+			curl_version_info_data *ver = curl_version_info(CURLVERSION_NOW);
+			if (ver->version_num < 0x075400)
+				die("libcurl 7.84 or later is required to support paths in proxy URLs");
+
+			if (!starts_with(proxy_auth.protocol, "socks"))
+				die("Invalid proxy URL '%s': only SOCKS proxies support paths",
+				    curl_http_proxy);
+
+			if (strcasecmp(proxy_auth.host, "localhost"))
+				die("Invalid proxy URL '%s': host must be localhost if a path is present",
+				    curl_http_proxy);
+
+			strbuf_addch(&proxy, '/');
+			strbuf_add_percentencode(&proxy, proxy_auth.path, 0);
+		}
+		curl_easy_setopt(result, CURLOPT_PROXY, proxy.buf);
+		strbuf_release(&proxy);
+
 		var_override(&curl_no_proxy, getenv("NO_PROXY"));
 		var_override(&curl_no_proxy, getenv("no_proxy"));
 		curl_easy_setopt(result, CURLOPT_NOPROXY, curl_no_proxy);

t/socks4-proxy.pl

@@ -0,0 +1,48 @@
+use strict;
+use IO::Select;
+use IO::Socket::UNIX;
+use IO::Socket::INET;
+
+my $path = shift;
+
+unlink($path);
+my $server = IO::Socket::UNIX->new(Listen => 1, Local => $path)
+	or die "unable to listen on $path: $!";
+
+$| = 1;
+print "ready\n";
+
+while (my $client = $server->accept()) {
+	sysread $client, my $buf, 8;
+	my ($version, $cmd, $port, $ip) = unpack 'CCnN', $buf;
+	next unless $version == 4; # socks4
+	next unless $cmd == 1; # TCP stream connection
+
+	# skip NUL-terminated id
+	while (sysread $client, my $char, 1) {
+		last unless ord($char);
+	}
+
+	# version(0), reply(5a == granted), port (ignored), ip (ignored)
+	syswrite $client, "\x00\x5a\x00\x00\x00\x00\x00\x00";
+
+	my $remote = IO::Socket::INET->new(PeerHost => $ip, PeerPort => $port)
+		or die "unable to connect to $ip/$port: $!";
+
+	my $io = IO::Select->new($client, $remote);
+	while ($io->count) {
+		for my $fh ($io->can_read(0)) {
+			for my $pair ([$client, $remote], [$remote, $client]) {
+				my ($from, $to) = @$pair;
+				next unless $fh == $from;
+
+				my $r = sysread $from, my $buf, 1024;
+				if (!defined $r || $r <= 0) {
+					$io->remove($from);
+					next;
+				}
+				syswrite $to, $buf;
+			}
+		}
+	}
+}

t/t5564-http-proxy.sh

@@ -39,4 +39,45 @@ test_expect_success 'clone can prompt for proxy password' '
 	expect_askpass pass proxuser
 '
 
+start_socks() {
+	mkfifo socks_output &&
+	{
+		"$PERL_PATH" "$TEST_DIRECTORY/socks4-proxy.pl" "$1" >socks_output &
+		echo $! > "$TRASH_DIRECTORY/socks.pid"
+	} &&
+	read line <socks_output &&
+	test "$line" = ready
+}
+
+# The %30 tests that the correct amount of percent-encoding is applied to the
+# proxy string passed to curl.
+test_lazy_prereq SOCKS_PROXY 'test_have_prereq PERL && start_socks "$TRASH_DIRECTORY/%30.sock"'
+
+test_atexit 'test ! -e "$TRASH_DIRECTORY/socks.pid" || kill "$(cat "$TRASH_DIRECTORY/socks.pid")"'
+
+test_expect_success SOCKS_PROXY 'clone via Unix socket' '
+	test_when_finished "rm -rf clone" &&
+	test_config_global http.proxy "socks4://localhost$PWD/%2530.sock" && {
+		{
+			GIT_TRACE_CURL=$PWD/trace git clone "$HTTPD_URL/smart/repo.git" clone 2>err &&
+			grep -i "SOCKS4 request granted." trace
+		} ||
+		grep "^fatal: libcurl 7\.84 or later" err
+	}
+'
+
+test_expect_success 'Unix socket requires socks*:' '
+	! git clone -c http.proxy=localhost/path https://example.com/repo.git 2>err && {
+		grep "^fatal: Invalid proxy URL '\''localhost/path'\'': only SOCKS proxies support paths" err ||
+		grep "^fatal: libcurl 7\.84 or later" err
+	}
+'
+
+test_expect_success 'Unix socket requires localhost' '
+	! git clone -c http.proxy=socks4://127.0.0.1/path https://example.com/repo.git 2>err && {
+		grep "^fatal: Invalid proxy URL '\''socks4://127\.0\.0\.1/path'\'': host must be localhost if a path is present" err ||
+		grep "^fatal: libcurl 7\.84 or later" err
+	}
+'
+
 test_done