Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Add metadata retrieved from the context to the user agent when a new HTTP client is created #2789

Merged
merged 6 commits into from
Jan 23, 2025

Conversation

MatteoPologruto
Copy link
Contributor

@MatteoPologruto MatteoPologruto commented Dec 18, 2024

Please check if the PR fulfills these requirements

See how to contribute

  • The PR has no duplicates (please search among the Pull Requests
    before creating one)
  • The PR follows
    our contributing guidelines
  • Tests for the changes have been added (for bug fixes / features)
  • Docs have been added / updated (for bug fixes / features)
  • UPGRADING.md has been updated with a migration guide (for breaking changes)
  • configuration.schema.json updated if new parameters are added.

What kind of change does this PR introduce?

Bug fix

What is the current behavior?

The user agent does not specify if the CLI runs in daemon mode. It is always something like arduino-cli/git-snapshot (amd64; windows; go1.23.2) Commit:84fc413a.

What is the new behavior?

The user agent obtained from the context metadata is propagated to thenetwork.user_agent_ext if it's empty and the CLI runs in daemon mode. The complete user agent is arduino-cli/git-snapshot arduino-ide/2.3.4 grpc-node-js/1.9.5 daemon (amd64; windows; go1.23.4) Commit:f3dc127e.

Does this PR introduce a breaking change, and is titled accordingly?

Other information

@MatteoPologruto MatteoPologruto added topic: code Related to content of the project itself type: imperfection Perceived defect in any part of project labels Dec 18, 2024
@MatteoPologruto MatteoPologruto self-assigned this Dec 18, 2024
@MatteoPologruto MatteoPologruto marked this pull request as ready for review December 18, 2024 11:23
Copy link

codecov bot commented Dec 18, 2024

Codecov Report

Attention: Patch coverage is 88.88889% with 3 lines in your changes missing coverage. Please review.

Project coverage is 67.71%. Comparing base (e9092cc) to head (2e7d6c2).
Report is 2 commits behind head on master.

Files with missing lines Patch % Lines
commands/service_check_for_updates.go 0.00% 3 Missing ⚠️
Additional details and impacted files
@@            Coverage Diff             @@
##           master    #2789      +/-   ##
==========================================
+ Coverage   67.69%   67.71%   +0.01%     
==========================================
  Files         238      238              
  Lines       22388    22392       +4     
==========================================
+ Hits        15156    15163       +7     
+ Misses       6036     6034       -2     
+ Partials     1196     1195       -1     
Flag Coverage Δ
unit 67.71% <88.88%> (+0.01%) ⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

Copy link
Contributor

@alessio-perugini alessio-perugini left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Left some comments/
It would be cool if we're able to test it.

Maybe we could use a proxy HTTP, and inspect the call that arrives are with the expected User agent. Something similar to: https://github.com/arduino/arduino-cli/blob/master/internal/cli/configuration/network_test.go#L29-L51

internal/cli/daemon/daemon.go Outdated Show resolved Hide resolved
internal/cli/daemon/daemon.go Outdated Show resolved Hide resolved
internal/cli/daemon/daemon.go Outdated Show resolved Hide resolved
internal/cli/daemon/daemon.go Outdated Show resolved Hide resolved
This allows the extraction of the user-agent in a single place. Also it
forces the context passing on all operations that requires access to
network.
@MatteoPologruto
Copy link
Contributor Author

1hsiii

@cmaglie

Copy link
Contributor

@alessio-perugini alessio-perugini left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I don't want to block this over some minor refactoring in the integration test part.

If possible I would extract the hardcoded assertion logic in the HTTPServeFile.

The implementation is LGTM. I've also tested locally with the usage of network.proxy and without and it correctly sends the UserAgent

Comment on lines 126 to 140
return CreateEnvForDaemonWithUserAgent(t, "cli-test/0.0.0")
}

// CreateEnvForDaemonWithUserAgent performs the minimum required operations to start the arduino-cli daemon.
// It returns a testsuite.Environment and an ArduinoCLI client to perform the integration tests.
// The Environment must be disposed by calling the CleanUp method via defer.
func CreateEnvForDaemonWithUserAgent(t *testing.T, userAgent string) (*Environment, *ArduinoCLI) {
env := NewEnvironment(t)

cli := NewArduinoCliWithinEnvironment(env, &ArduinoCLIConfig{
ArduinoCLIPath: FindRepositoryRootPath(t).Join("arduino-cli"),
UseSharedStagingFolder: true,
})

_ = cli.StartDaemon(false)
_ = cli.StartDaemon(false, userAgent)
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Suggested change
return CreateEnvForDaemonWithUserAgent(t, "cli-test/0.0.0")
}
// CreateEnvForDaemonWithUserAgent performs the minimum required operations to start the arduino-cli daemon.
// It returns a testsuite.Environment and an ArduinoCLI client to perform the integration tests.
// The Environment must be disposed by calling the CleanUp method via defer.
func CreateEnvForDaemonWithUserAgent(t *testing.T, userAgent string) (*Environment, *ArduinoCLI) {
env := NewEnvironment(t)
cli := NewArduinoCliWithinEnvironment(env, &ArduinoCLIConfig{
ArduinoCLIPath: FindRepositoryRootPath(t).Join("arduino-cli"),
UseSharedStagingFolder: true,
})
_ = cli.StartDaemon(false)
_ = cli.StartDaemon(false, userAgent)
env := NewEnvironment(t)
cli := NewArduinoCliWithinEnvironment(env, &ArduinoCLIConfig{
ArduinoCLIPath: FindRepositoryRootPath(t).Join("arduino-cli"),
UseSharedStagingFolder: true,
})
_ = cli.StartDaemon(false)

CreateEnvForDaemonWithUserAgent this function is used only by the CreateEnvForDaemon

I would simplify and put everything in the same function as it was before.
The useragent will be hardcoded inside the StartDaemon

@@ -410,7 +417,7 @@ func (cli *ArduinoCLI) run(stdoutBuff, stderrBuff io.Writer, stdinBuff io.Reader
}

// StartDaemon starts the Arduino CLI daemon. It returns the address of the daemon.
func (cli *ArduinoCLI) StartDaemon(verbose bool) string {
func (cli *ArduinoCLI) StartDaemon(verbose bool, userAgent string) string {
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Suggested change
func (cli *ArduinoCLI) StartDaemon(verbose bool, userAgent string) string {
func (cli *ArduinoCLI) StartDaemon(verbose bool) string {

I cannot think of uses cases that we need to pass different userAgent. The StartDaemon currently is only called in the CreateEnvForDaemon. It's responsibility is to create the grpclient and initializing grpc options.

My suggestion is to just hardcode the userAgent in the grpc option. See later suggestion

conn, err := grpc.NewClient(
cli.daemonAddr,
grpc.WithTransportCredentials(insecure.NewCredentials()),
grpc.WithUserAgent(userAgent),
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Suggested change
grpc.WithUserAgent(userAgent),
grpc.WithUserAgent("cli-test/0.0.0"),

I'd simply hardcode the user-agent here.

Comment on lines 29 to 41
func (env *Environment) HTTPServeFile(port uint16, path *paths.Path, isDaemon bool) *url.URL {
t := env.T()
mux := http.NewServeMux()
mux.HandleFunc("/"+path.Base(), func(w http.ResponseWriter, r *http.Request) {
http.ServeFile(w, r, path.String())
if isDaemon {
// Test that the user-agent contains metadata from the context when the CLI is in daemon mode
userAgent := r.Header.Get("User-Agent")
require.Contains(t, userAgent, "arduino-cli/git-snapshot")
require.Contains(t, userAgent, "cli-test/0.0.0")
require.Contains(t, userAgent, "grpc-go")
}
})
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The HTTPServeFile should just serve files, and not perform any assertion based on a boolean flag.
I understand why this was done, as it is convenient to do so, but if in he future we have to make assertions based on some other http header, this function will quickly be polluted with stuff that goes out of scope of this function.

I suggest to rollback these changes, and creating a dedicated HTTP server that asserts that. I would put such server inside the dedicated test of TestDaemonUserAgent.

See suggestion over the TestDaemonUserAgent.
Another approach could be to simply pass some assertions function as a callback, or create another abstraction.

Comment on lines 558 to 585
func TestDaemonUserAgent(t *testing.T) {
env, cli := integrationtest.CreateEnvForDaemon(t)
defer env.CleanUp()

// Set up an http server to serve our custom index file
// The user-agent is tested inside the HTTPServeFile function
test_index := paths.New("..", "testdata", "test_index.json")
url := env.HTTPServeFile(8000, test_index, true)

grpcInst := cli.Create()
require.NoError(t, grpcInst.Init("", "", func(ir *commands.InitResponse) {
fmt.Printf("INIT> %v\n", ir.GetMessage())
}))

// Set extra indexes
err := cli.SetValue("board_manager.additional_urls", `["http://127.0.0.1:8000/test_index.json"]`)
require.NoError(t, err)

{
cl, err := grpcInst.UpdateIndex(context.Background(), false)
require.NoError(t, err)
res, err := analyzeUpdateIndexClient(t, cl)
require.NoError(t, err)
require.Len(t, res, 2)
require.True(t, res[url.String()].GetSuccess())
}
}

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Here, if possible would be nice to extract the assertion code that is done by passing the true flag in the HTTPServerFile. Or make an abstraction that passes some assertion functions.

Another possible idea is by "proxing" the request to the file server:

diff --git a/internal/integrationtest/daemon/daemon_test.go b/internal/integrationtest/daemon/daemon_test.go
index 7d800c84..982ba8bb 100644
--- a/internal/integrationtest/daemon/daemon_test.go
+++ b/internal/integrationtest/daemon/daemon_test.go
@@ -20,6 +20,10 @@ import (
 	"errors"
 	"fmt"
 	"io"
+	"maps"
+	"net/http"
+	"net/http/httptest"
+	"strings"
 	"testing"
 	"time"
 
@@ -562,15 +566,41 @@ func TestDaemonUserAgent(t *testing.T) {
 	// Set up an http server to serve our custom index file
 	// The user-agent is tested inside the HTTPServeFile function
 	test_index := paths.New("..", "testdata", "test_index.json")
-	url := env.HTTPServeFile(8000, test_index, true)
+	url := env.HTTPServeFile(8000, test_index)
+	ts := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		// Test that the user-agent contains metadata from the context when the CLI is in daemon mode
+		userAgent := r.Header.Get("User-Agent")
+
+		require.Contains(t, userAgent, "cli-test/0.0.0")
+		require.Contains(t, userAgent, "grpc-go")
+		// Depends on how we built the client we may have git-snapshot or 0.0.0-git in dev releases
+		require.Condition(t, func() (success bool) {
+			return strings.Contains(userAgent, "arduino-cli/git-snapshot") ||
+				strings.Contains(userAgent, "arduino-cli/0.0.0-git")
+		})
+
+		proxiedReq, err := http.NewRequest(r.Method, url.String(), r.Body)
+		require.NoError(t, err)
+		maps.Copy(proxiedReq.Header, r.Header)
+
+		proxiedResp, err := http.DefaultTransport.RoundTrip(proxiedReq)
+		require.NoError(t, err)
+		defer proxiedResp.Body.Close()
+
+		// Copy the headers from the proxy response to the original response
+		maps.Copy(r.Header, proxiedReq.Header)
+		w.WriteHeader(proxiedResp.StatusCode)
+		io.Copy(w, proxiedResp.Body)
+	}))
+	defer ts.Close()
 
 	grpcInst := cli.Create()
 	require.NoError(t, grpcInst.Init("", "", func(ir *commands.InitResponse) {
 		fmt.Printf("INIT> %v\n", ir.GetMessage())
 	}))
 
-	// Set extra indexes
-	err := cli.SetValue("board_manager.additional_urls", `["http://127.0.0.1:8000/test_index.json"]`)
+	additionalURL := ts.URL + "/test_index.json"
+	err := cli.SetValue("board_manager.additional_urls", fmt.Sprintf(`["%s"]`, additionalURL))
 	require.NoError(t, err)
 
 	{
@@ -579,7 +609,7 @@ func TestDaemonUserAgent(t *testing.T) {
 		res, err := analyzeUpdateIndexClient(t, cl)
 		require.NoError(t, err)
 		require.Len(t, res, 2)
-		require.True(t, res[url.String()].GetSuccess())
+		require.True(t, res[additionalURL].GetSuccess())
 	}
 }
 

@MatteoPologruto MatteoPologruto changed the title Add daemon to the user agent when the CLI is started in daemon mode Add metadata retrieved from the context to the user agent when a new HTTP client is created Jan 23, 2025
@MatteoPologruto MatteoPologruto merged commit c5cfe4a into arduino:master Jan 23, 2025
98 checks passed
@MatteoPologruto MatteoPologruto deleted the user-agent branch January 23, 2025 16:41
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
topic: code Related to content of the project itself type: imperfection Perceived defect in any part of project
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants