Skip to contentSkip to conversation

push options: fail properly in the stateless case

PATCH In review Stefan Beller 2 participants 6 messages Feb 8, 2017 – Feb 10, 2017 1 message not in the archive Documentation

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 push options: fail properly in the stateless case

When using non-builtin protocols relying on a transport helper (such as http), push options are not propagated to the helper.

Fix this by propagating the push options to the transport helper and adding a test that push options using http fail properly.

---
 Documentation/gitremote-helpers.txt |  3 +++
 t/t5545-push-options.sh             | 15 +++++++++++++++
 transport-helper.c                  |  7 +++++++
 3 files changed, 25 insertions(+)
Documentation/gitremote-helpers.txt +3 -0
index 9e8681f9e1..6145d4d8df 100644
@@ -462,6 +462,9 @@ set by Git if the remote helper has the 'option' capability.
 'option pushcert {'true'|'false'}::
 	GPG sign pushes.
 
+'option push-option <c-string>::
+	Transmit this push option.
+
 SEE ALSO
 --------
 linkgit:git-remote[1]
t/t5545-push-options.sh +15 -0
index ea813b9383..9a57a7d8f2 100755
@@ -3,6 +3,8 @@
 test_description='pushing to a repository using push options'
 
 . ./test-lib.sh
+. "$TEST_DIRECTORY"/lib-httpd.sh
+start_httpd
 
 mk_repo_pair () {
 	rm -rf workbench upstream &&
@@ -100,4 +102,17 @@ test_expect_success 'two push options work' '
 	test_cmp expect upstream/.git/hooks/post-receive.push_options
 '
 
+test_expect_success 'push option denied properly by http remote helper' '\
+	mk_repo_pair &&
+	git -C upstream config receive.advertisePushOptions false &&
+	git -C upstream config http.receivepack true &&
+	cp -R upstream/.git "$HTTPD_DOCUMENT_ROOT_PATH"/upstream.git &&
+	git clone "$HTTPD_URL"/smart/upstream test_http_clone &&
+	test_commit -C test_http_clone one &&
+	test_must_fail git -C test_http_clone push --push-option=asdf origin master &&
+	git -C test_http_clone push origin master
+'
+
+stop_httpd
+
 test_done
transport-helper.c +7 -0
index 91aed35ebb..1258d6aedd 100644
@@ -826,6 +826,13 @@ static void set_common_push_options(struct transport *transport,
 		if (set_helper_option(transport, TRANS_OPT_PUSH_CERT, "if-asked") != 0)
 			die("helper %s does not support --signed=if-asked", name);
 	}
+
+	if (flags & TRANSPORT_PUSH_OPTIONS) {
+		struct string_list_item *item;
+		for_each_string_list_item(item, transport->push_options)
+			if (set_helper_option(transport, "push-option", item->string) != 0)
+				die("helper %s does not support 'push-option'", name);
+	}
 }
 
 static int push_refs_with_push(struct transport *transport,
Signature · 1 line
2.12.0.rc0.1.g018cb5e6f4
#1 lore raw
Quoted 4 lines · Stefan Beller

When using non-builtin protocols relying on a transport helper (such as http), push options are not propagated to the helper.

Fix this by propagating the push options to the transport helper and

The description up to this point is VERY readable and sensible. But that makes the title sound a bit strange. I read it as if it were saying "stateless case can never support push-options so fail if the caller attempts to use one", but that does not seem to be what is going on.

Quoted 1 line

adding a test that push options using http fail properly.

Sounds sensible. What end-user visible effect does this fix have? IOW, what feature do we use "push-option" for?

Ahh, OK, so you need to describe that there are two issues in order to be understood by the readers:

(1) the helper protocol does not propagate push-option
(2) the http helper is not prepared to handle push-option

You fix (1), and you take advantage of the fact (2) to ensure that
(1) is fixed in the new test.

With such an understanding, the title makes (sort of) sense and you wouldn't have to be asked "what end-user visible effect/benefit does this have?"

Documentation/gitremote-helpers.txt set by Git if the remote helper has the 'option' capability. · line 467

Quoted 3 lines

+'option push-option <c-string>::

Documentation/gitremote-helpers.txt +2 -0
+	Transmit this push option.
+

There is no "c-string" in the current documentation used or defined. The closest thing I found is

... that field will be quoted in the manner of a C string ...

in git-status page, but I do not think you send the value for an push-option after running quote_c_style(), so I am puzzled.

I'd rather see 'option push-option <string>' as the bullet item, and in its description say how arbitrary values (if you allow them, that is) can be used, e.g. "Transmit <string> encoded in such and such way a the value of the push-option".

#2 lore raw
Quoted 9 lines · Junio C Hamano
Quoted 4 lines · Stefan Beller

When using non-builtin protocols relying on a transport helper (such as http), push options are not propagated to the helper.

Fix this by propagating the push options to the transport helper and

The description up to this point is VERY readable and sensible. But that makes the title sound a bit strange.

Yes, the title was there first and then I massaged the commit message until it was readable. Originally I thought the issue is with stateless protocols, but that is wrong. The underlying issue is just the transport helper communication lacking.

Quoted 4 lines

I read it as if it were
saying "stateless case can never support push-options so fail if the caller attempts to use one", but that does not seem to be what is going on.

Indeed the subject is wrong.

Quoted 5 lines
Quoted 2 lines

adding a test that push options using http fail properly.

Sounds sensible. What end-user visible effect does this fix have? IOW, what feature do we use "push-option" for?

The Gerrit world started using push options for having a better git experience, not needing to navigate to the web UI, e.g.:

    # push for review and set me as a reviewer:
    git push --push-option [email protected]  ssh://gerrit...
    # same with http, it looks like it worked, but the push option
    # never made it to the server
    git push --push-option [email protected]  http://gerrit...

# This patch will make the second command fail, reporting # the http helper not supporting push options.

Quoted 7 lines

Ahh, OK, so you need to describe that there are two issues in order to be understood by the readers:

(1) the helper protocol does not propagate push-option
(2) the http helper is not prepared to handle push-option

Quoted 6 lines

You fix (1), and you take advantage of the fact (2) to ensure that
(1) is fixed in the new test.

With such an understanding, the title makes (sort of) sense and you wouldn't have to be asked "what end-user visible effect/benefit does this have?"

Your analysis is correct.

Quoted 5 lines

Documentation/gitremote-helpers.txt set by Git if the remote helper has the 'option' capability. · line 467

Quoted 4 lines

+'option push-option <c-string>::

Documentation/gitremote-helpers.txt +2 -0
+     Transmit this push option.
+
Quoted 7 lines

There is no "c-string" in the current documentation used or defined. The closest thing I found is

... that field will be quoted in the manner of a C string ...

in git-status page, but I do not think you send the value for an push-option after running quote_c_style(), so I am puzzled.

When implementing push options, we discussed that and according to Documentation/git-push:

The given string must not contain a NUL or LF character.

Quoted 5 lines

I'd rather see 'option push-option <string>' as the bullet item, and in its description say how arbitrary values (if you allow them, that is) can be used, e.g. "Transmit <string> encoded in such and such way a the value of the push-option".

okay.

#3 lore raw
Quoted 9 lines · Stefan Beller
Quoted 9 lines

Documentation/gitremote-helpers.txt set by Git if the remote helper has the 'option' capability. · line 467

Quoted 3 lines

+'option push-option <c-string>::

Documentation/gitremote-helpers.txt +2 -0
+     Transmit this push option.
+

There is no "c-string" in the current documentation used or defined. The closest thing I found is

... that field will be quoted in the manner of a C string ...

Quoted 7 lines
Quoted 2 lines

in git-status page, but I do not think you send the value for an push-option after running quote_c_style(), so I am puzzled.

When implementing push options, we discussed that and according to Documentation/git-push:

The given string must not contain a NUL or LF character.

OK, so "Transmit <string> as a push option" is sufficient, as the string is sent as-is. OK.

Thanks.

#4 lore raw

PATCH v2 1/1 push options: pass push options to the transport helper

In reply to Junio C Hamano

When using non-builtin protocols relying on a transport helper (such as http), push options are not propagated to the helper.

The user could ask for push options and a push would seemingly succeed, but the push options would never be transported to the server, misleading the users expectation.

Fix this by propagating the push options to the transport helper.

This is only addressing the first issue of
(1) the helper protocol does not propagate push-option
(2) the http helper is not prepared to handle push-option

Once we fix (2), the http transport helper can make use of push options as well, but that happens as a follow up. (1) is a bug fix, whereas (2) is a feature, which is why we only do (1) here.

---
 Documentation/gitremote-helpers.txt |  4 ++++
 t/t5545-push-options.sh             | 15 +++++++++++++++
 transport-helper.c                  |  7 +++++++
 3 files changed, 26 insertions(+)
Documentation/gitremote-helpers.txt +4 -0
index 9e8681f9e1..23474b1eab 100644
@@ -462,6 +462,10 @@ set by Git if the remote helper has the 'option' capability.
 'option pushcert {'true'|'false'}::
 	GPG sign pushes.
 
+'option push-option <string>::
+	Transmit <string> as a push option. As the a push option
+	must not contain LF or NUL characters, the string is not encoded.
+
 SEE ALSO
 --------
 linkgit:git-remote[1]
t/t5545-push-options.sh +15 -0
index ea813b9383..9a57a7d8f2 100755
@@ -3,6 +3,8 @@
 test_description='pushing to a repository using push options'
 
 . ./test-lib.sh
+. "$TEST_DIRECTORY"/lib-httpd.sh
+start_httpd
 
 mk_repo_pair () {
 	rm -rf workbench upstream &&
@@ -100,4 +102,17 @@ test_expect_success 'two push options work' '
 	test_cmp expect upstream/.git/hooks/post-receive.push_options
 '
 
+test_expect_success 'push option denied properly by http remote helper' '\
+	mk_repo_pair &&
+	git -C upstream config receive.advertisePushOptions false &&
+	git -C upstream config http.receivepack true &&
+	cp -R upstream/.git "$HTTPD_DOCUMENT_ROOT_PATH"/upstream.git &&
+	git clone "$HTTPD_URL"/smart/upstream test_http_clone &&
+	test_commit -C test_http_clone one &&
+	test_must_fail git -C test_http_clone push --push-option=asdf origin master &&
+	git -C test_http_clone push origin master
+'
+
+stop_httpd
+
 test_done
transport-helper.c +7 -0
index 91aed35ebb..1258d6aedd 100644
@@ -826,6 +826,13 @@ static void set_common_push_options(struct transport *transport,
 		if (set_helper_option(transport, TRANS_OPT_PUSH_CERT, "if-asked") != 0)
 			die("helper %s does not support --signed=if-asked", name);
 	}
+
+	if (flags & TRANSPORT_PUSH_OPTIONS) {
+		struct string_list_item *item;
+		for_each_string_list_item(item, transport->push_options)
+			if (set_helper_option(transport, "push-option", item->string) != 0)
+				die("helper %s does not support 'push-option'", name);
+	}
 }
 
 static int push_refs_with_push(struct transport *transport,
Signature · 1 line
2.12.0.rc0.1.g018cb5e6f4

A message that is not in the archive (sent off-list or not yet received): CAFzf2XzLNOKbNbkCfzpVJtN-ROW8m8WVBsWM-z7Bxdnv55b45Q@mail.gmail.com

#6 lore raw

Re: [PATCHv2] push options: pass push options to the transport helper

Quoted 1 line · Jonathan Nieder

Private reply because using HTML mail on phone:

So I presume putting it back on the list is fine.

Quoted 1 line

Does this need to be a remote helper capability?

No, as all other capabilities of the git-protocol are mapped 1:1 to the transport helper protocol for support, e.g. each transport helper has to handle this on its own, c.f. remote-curl.c:set_option (line 39 ff and 1065),

Quoted 2 lines

What happens with remote
helpers that don't understand the option?

The helper ought to print "unsupported"
(remote-curl.c:1065 for the http helper) and then the transport helper will take care of it (transport-helper.c: 265 and e.g. 820)

Quoted 3 lines

How do remote helpers communicate whether the server has said it accepts push options?

I guess the remote helper would communicate it the same way as communicating if the push succeeded? (i.e. reject non fast forward.)

Quoted 104 lines · On Wed, Feb 8, 2017, 14:04 Stefan Beller <[email protected]> wrote:
Quoted 101 lines

When using non-builtin protocols relying on a transport helper (such as http), push options are not propagated to the helper.

The user could ask for push options and a push would seemingly succeed, but the push options would never be transported to the server, misleading the users expectation.

Fix this by propagating the push options to the transport helper.

This is only addressing the first issue of
(1) the helper protocol does not propagate push-option
(2) the http helper is not prepared to handle push-option

Once we fix (2), the http transport helper can make use of push options as well, but that happens as a follow up. (1) is a bug fix, whereas (2) is a feature, which is why we only do (1) here.

---
 Documentation/gitremote-helpers.txt |  4 ++++
 t/t5545-push-options.sh             | 15 +++++++++++++++
 transport-helper.c                  |  7 +++++++
 3 files changed, 26 insertions(+)
Documentation/gitremote-helpers.txt +0 -0
index 9e8681f9e1..23474b1eab 100644
@@ -462,6 +462,10 @@ set by Git if the remote helper has the 'option'

capability.
'option pushcert {'true'|'false'}::
GPG sign pushes.

+'option push-option <string>::

Documentation/gitremote-helpers.txt +3 -0
+       Transmit <string> as a push option. As the a push option
+       must not contain LF or NUL characters, the string is not encoded.
+
 SEE ALSO
 --------
 linkgit:git-remote[1]
t/t5545-push-options.sh +10 -0
index ea813b9383..9a57a7d8f2 100755
@@ -3,6 +3,8 @@
 test_description='pushing to a repository using push options'
 
 . ./test-lib.sh
+. "$TEST_DIRECTORY"/lib-httpd.sh
+start_httpd
 
 mk_repo_pair () {
        rm -rf workbench upstream &&
@@ -100,4 +102,17 @@ test_expect_success 'two push options work' '
        test_cmp expect upstream/.git/hooks/post-receive.push_options
 '
 
+test_expect_success 'push option denied properly by http remote helper'
 '\
+       mk_repo_pair &&
+       git -C upstream config receive.advertisePushOptions false &&
+       git -C upstream config http.receivepack true &&
+       cp -R upstream/.git "$HTTPD_DOCUMENT_ROOT_PATH"/upstream.git &&
+       git clone "$HTTPD_URL"/smart/upstream test_http_clone &&
+       test_commit -C test_http_clone one &&
+       test_must_fail git -C test_http_clone push --push-option=asdf

origin master &&

t/t5545-push-options.sh +5 -0
+       git -C test_http_clone push origin master
+'
+
+stop_httpd
+
 test_done
transport-helper.c +5 -1
index 91aed35ebb..1258d6aedd 100644
@@ -826,6 +826,13 @@ static void set_common_push_options(struct transport
 *transport,
                if (set_helper_option(transport, TRANS_OPT_PUSH_CERT,
 "if-asked") != 0)
                        die("helper %s does not support
--signed=if-asked", name);
        }
+
+       if (flags & TRANSPORT_PUSH_OPTIONS) {
+               struct string_list_item *item;
+               for_each_string_list_item(item, transport->push_options)
+                       if (set_helper_option(transport, "push-option",

item->string) != 0)
+ die("helper %s does not support
'push-option'", name);
+ }
}

static int push_refs_with_push(struct transport *transport,
--
2.12.0.rc0.1.g018cb5e6f4