Move to the layout REPO_POLICIES.md gives (closes #30)

The Go module moves from backend/ to the root (go.mod, go.sum, .golangci.yml, cmd/, internal/), the page's source to web/, and nginx's two files to configs/. backend/ is gone: its LICENSE and .dockerignore are deleted, and its .gitignore entries, README and scripts are folded into the root ones. script/build builds the server binary after the frontend; script/run and script/clean move up from backend/script/; make test, lint, fmt and fmt-check do what the backend's did. Vite's root is web/ and it still builds into dist/. go.mod ignores node_modules/, whose npm packages can ship .go files. script/bootstrap no longer installs gcc, which only the host Go tests needed. Model: opus-5-5
2026-10-14 12:01:32 +02:00
parent b1eefe0c29
commit 406e073aee
68 changed files with 333 additions and 530 deletions
@@ -0,0 +1,238 @@
+// Package config loads application configuration from
+// environment variables, .env files, and config files.
+package config
+
+import (
+	"errors"
+	"fmt"
+	"log/slog"
+	"math"
+	"net/netip"
+	"net/url"
+	"strconv"
+	"strings"
+
+	"sneak.berlin/go/netwatch/internal/globals"
+	"sneak.berlin/go/netwatch/internal/logger"
+
+	_ "github.com/joho/godotenv/autoload" // loads .env file
+	"github.com/spf13/viper"
+	"go.uber.org/fx"
+)
+
+// defaultTrustedProxies lists the networks whose forwarded
+// headers are honoured by default: IPv4 and IPv6 loopback,
+// for a reverse proxy on the same host, and the RFC1918
+// ranges. The container image does not use it:
+// bin/entrypoint.sh gives the server 127.0.0.1/32, since
+// nginx is its only client there.
+const defaultTrustedProxies = "127.0.0.1/32,::1/128," +
+	"10.0.0.0/8,172.16.0.0/12,192.168.0.0/16"
+
+// Default limits on stored reports; README.md gives the reasons for
+// these values.
+const (
+	defaultReportsPerMinute = 60
+	defaultDataDirMaxBytes  = 1 << 30 // 1 GiB
+)
+
+var (
+	errNotPositive = errors.New("must be a positive whole number")
+	errNotOrigin   = errors.New(
+		"must be an origin, scheme://host with an optional port",
+	)
+	errNotPort = errors.New("must be a port number, 1 to 65535")
+	errNotBool = errors.New("must be true or false")
+	errNotIP   = errors.New("must be an IP address, or empty")
+
+	errMetricsCredentials = errors.New(
+		"METRICS_USERNAME and METRICS_PASSWORD must be set together, " +
+			"or neither",
+	)
+	errMetricsUsernameColon = errors.New(
+		"METRICS_USERNAME must not contain \":\", " +
+			"which basic auth cannot carry in a user name",
+	)
+)
+
+// Params defines the dependencies for Config.
+type Params struct {
+	fx.In
+
+	Globals *globals.Globals
+	Logger  *logger.Logger
+}
+
+// Config holds the resolved application configuration.
+type Config struct {
+	BindAddress        string
+	CORSAllowedOrigins []string
+	DataDir            string
+	DataDirMaxBytes    int64
+	Debug              bool
+	MetricsPassword    string
+	MetricsUsername    string
+	Port               int
+	ReportsPerMinute   int
+	SentryDSN          string
+	TrustedProxies     []string
+	log                *slog.Logger
+	params             *Params
+}
+
+// New loads configuration from env, .env files, and config
+// files, returning a fully resolved Config. It fails, with an error
+// naming the setting, on a value the server cannot use, and on a
+// config file it finds but cannot read.
+func New(
+	_ fx.Lifecycle,
+	params Params,
+) (*Config, error) {
+	log := params.Logger.Get()
+	name := params.Globals.Appname
+
+	viper.SetConfigName(name)
+	viper.SetConfigType("yaml")
+	viper.AddConfigPath("/etc/" + name)
+	viper.AddConfigPath("$HOME/.config/" + name)
+
+	viper.AutomaticEnv()
+
+	// An empty CORS_ALLOWED_ORIGINS allows no other origin.
+	viper.SetDefault("CORS_ALLOWED_ORIGINS", "")
+	viper.SetDefault("DATA_DIR", "./data/reports")
+	viper.SetDefault("DATA_DIR_MAX_BYTES", defaultDataDirMaxBytes)
+	viper.SetDefault("DEBUG", "false")
+	// An empty BIND_ADDRESS listens on every interface.
+	viper.SetDefault("BIND_ADDRESS", "")
+	viper.SetDefault("PORT", "8080")
+	viper.SetDefault("REPORTS_PER_MINUTE", defaultReportsPerMinute)
+	viper.SetDefault("SENTRY_DSN", "")
+	viper.SetDefault("METRICS_USERNAME", "")
+	viper.SetDefault("METRICS_PASSWORD", "")
+	viper.SetDefault("TRUSTED_PROXIES", defaultTrustedProxies)
+
+	err := viper.ReadInConfig()
+	if err != nil {
+		var notFound viper.ConfigFileNotFoundError
+		if !errors.As(err, &notFound) {
+			return nil, fmt.Errorf("config file %s: %w",
+				viper.ConfigFileUsed(), err)
+		}
+	}
+
+	// Read with strconv: viper's GetInt and GetBool would read a value
+	// they cannot parse as 0 or false instead of failing.
+	port, err := strconv.Atoi(viper.GetString("PORT"))
+	if err != nil || port < 1 || port > math.MaxUint16 {
+		return nil, fmt.Errorf("PORT %q: %w",
+			viper.GetString("PORT"), errNotPort)
+	}
+
+	debug, err := strconv.ParseBool(viper.GetString("DEBUG"))
+	if err != nil {
+		return nil, fmt.Errorf("DEBUG %q: %w",
+			viper.GetString("DEBUG"), errNotBool)
+	}
+
+	s := &Config{
+		BindAddress:        viper.GetString("BIND_ADDRESS"),
+		CORSAllowedOrigins: splitList(viper.GetString("CORS_ALLOWED_ORIGINS")),
+		DataDir:            viper.GetString("DATA_DIR"),
+		DataDirMaxBytes:    viper.GetInt64("DATA_DIR_MAX_BYTES"),
+		Debug:              debug,
+		MetricsPassword:    viper.GetString("METRICS_PASSWORD"),
+		MetricsUsername:    viper.GetString("METRICS_USERNAME"),
+		Port:               port,
+		ReportsPerMinute:   viper.GetInt("REPORTS_PER_MINUTE"),
+		SentryDSN:          viper.GetString("SENTRY_DSN"),
+		TrustedProxies:     splitList(viper.GetString("TRUSTED_PROXIES")),
+		log:                log,
+		params:             &params,
+	}
+
+	err = s.check()
+	if err != nil {
+		return nil, err
+	}
+
+	if s.Debug {
+		params.Logger.EnableDebugLogging()
+		s.log = params.Logger.Get()
+	}
+
+	return s, nil
+}
+
+// check fails with an error naming the first setting here whose value
+// the server cannot use. New checks PORT and DEBUG as it reads them,
+// and the middleware checks TRUSTED_PROXIES as it parses it.
+func (s *Config) check() error {
+	// viper reads a value that is not a number as 0, so this also
+	// catches a mistyped setting.
+	if s.ReportsPerMinute <= 0 {
+		return fmt.Errorf("REPORTS_PER_MINUTE %q: %w",
+			viper.GetString("REPORTS_PER_MINUTE"), errNotPositive)
+	}
+
+	if s.DataDirMaxBytes <= 0 {
+		return fmt.Errorf("DATA_DIR_MAX_BYTES %q: %w",
+			viper.GetString("DATA_DIR_MAX_BYTES"), errNotPositive)
+	}
+
+	if s.BindAddress != "" {
+		_, err := netip.ParseAddr(s.BindAddress)
+		if err != nil {
+			return fmt.Errorf("BIND_ADDRESS %q: %w", s.BindAddress, errNotIP)
+		}
+	}
+
+	// The server records and serves metrics only with both set, so
+	// one alone is a mistake that would otherwise go unnoticed.
+	if (s.MetricsUsername == "") != (s.MetricsPassword == "") {
+		return errMetricsCredentials
+	}
+
+	// Basic auth splits the credentials at the first ":", so with one
+	// in the user name every request to /metrics would get 401.
+	if strings.Contains(s.MetricsUsername, ":") {
+		return errMetricsUsernameColon
+	}
+
+	return checkOrigins(s.CORSAllowedOrigins)
+}
+
+// checkOrigins fails on the first CORS_ALLOWED_ORIGINS entry that is
+// not a plain origin, scheme://host with an optional port, as browsers
+// send it; anything more, such as a trailing "/", would match no page.
+// go-chi/cors reads a "*" anywhere in an entry as a wildcard, so no
+// entry may contain one.
+func checkOrigins(origins []string) error {
+	for _, origin := range origins {
+		u, err := url.Parse(origin)
+		if err != nil || u.Scheme == "" || u.Host == "" ||
+			strings.Contains(origin, "*") ||
+			origin != u.Scheme+"://"+u.Host {
+			return fmt.Errorf("CORS_ALLOWED_ORIGINS %q: %w",
+				origin, errNotOrigin)
+		}
+	}
+
+	return nil
+}
+
+// splitList turns a comma-separated setting into a trimmed
+// slice, dropping empty entries.
+func splitList(raw string) []string {
+	parts := strings.Split(raw, ",")
+
+	out := make([]string, 0, len(parts))
+	for _, p := range parts {
+		p = strings.TrimSpace(p)
+		if p != "" {
+			out = append(out, p)
+		}
+	}
+
+	return out
+}
@@ -0,0 +1,147 @@
+package config_test
+
+import (
+	"strings"
+	"testing"
+
+	"sneak.berlin/go/netwatch/internal/config"
+	"sneak.berlin/go/netwatch/internal/globals"
+	"sneak.berlin/go/netwatch/internal/logger"
+
+	"go.uber.org/fx"
+)
+
+// requireConfigError builds the config as main does and fails the
+// test unless that fails with an error naming setting. It uses
+// fx.New, because fxtest.New fails the test itself on an error.
+func requireConfigError(t *testing.T, setting string) {
+	t.Helper()
+
+	app := fx.New(
+		fx.NopLogger,
+		fx.Provide(globals.New, logger.New, config.New),
+		fx.Invoke(func(*config.Config) {}),
+	)
+
+	err := app.Err()
+	if err == nil || !strings.Contains(err.Error(), setting) {
+		t.Fatalf("config error = %v, want one naming %s", err, setting)
+	}
+}
+
+// TestSettingsLoadAsGiven: valid values pass the checks and are used
+// as given. bin/entrypoint.sh starts the server with these
+// BIND_ADDRESS and PORT values.
+func TestSettingsLoadAsGiven(t *testing.T) {
+	t.Setenv("BIND_ADDRESS", "127.0.0.1")
+	t.Setenv("PORT", "8081")
+	t.Setenv("DEBUG", "true")
+
+	var cfg *config.Config
+
+	app := fx.New(
+		fx.NopLogger,
+		fx.Provide(globals.New, logger.New, config.New),
+		fx.Populate(&cfg),
+	)
+
+	err := app.Err()
+	if err != nil {
+		t.Fatalf("config error = %v", err)
+	}
+
+	if cfg.BindAddress != "127.0.0.1" || cfg.Port != 8081 || !cfg.Debug {
+		t.Fatalf("BindAddress, Port, Debug = %q, %d, %t; "+
+			"want \"127.0.0.1\", 8081, true",
+			cfg.BindAddress, cfg.Port, cfg.Debug)
+	}
+}
+
+// TestPortMustBeAPortNumber: viper reads a value that is not a number
+// as 0, on which the server would listen on a random port.
+func TestPortMustBeAPortNumber(t *testing.T) {
+	for _, value := range []string{"abc", "0", "65536", "8080.5"} {
+		t.Run(value, func(t *testing.T) {
+			t.Setenv("PORT", value)
+
+			requireConfigError(t, "PORT")
+		})
+	}
+}
+
+// TestDebugMustBeTrueOrFalse: viper reads any other value, such as
+// "yes", as false.
+func TestDebugMustBeTrueOrFalse(t *testing.T) {
+	t.Setenv("DEBUG", "yes")
+
+	requireConfigError(t, "DEBUG")
+}
+
+// TestBindAddressMustBeAnIPAddress: a host name would be looked up
+// only once the server starts listening, and a mistyped one would stop
+// it then with an error that does not name the setting.
+func TestBindAddressMustBeAnIPAddress(t *testing.T) {
+	t.Setenv("BIND_ADDRESS", "localhost")
+
+	requireConfigError(t, "BIND_ADDRESS")
+}
+
+// TestReportsPerMinuteMustBePositive: unchecked, zero would panic
+// when the routes are built, and a negative rate would lift the
+// limit.
+func TestReportsPerMinuteMustBePositive(t *testing.T) {
+	t.Setenv("REPORTS_PER_MINUTE", "0")
+
+	requireConfigError(t, "REPORTS_PER_MINUTE")
+}
+
+// TestDataDirMaxBytesMustBeANumber: viper reads a value that is not
+// a number, such as "1GB", as 0, which would refuse every report.
+func TestDataDirMaxBytesMustBeANumber(t *testing.T) {
+	t.Setenv("DATA_DIR_MAX_BYTES", "1GB")
+
+	requireConfigError(t, "DATA_DIR_MAX_BYTES")
+}
+
+// TestMetricsCredentialsGoTogether: with only one of the two set, the
+// server would quietly serve no metrics, so the start fails, naming
+// both.
+func TestMetricsCredentialsGoTogether(t *testing.T) {
+	for _, set := range []string{"METRICS_USERNAME", "METRICS_PASSWORD"} {
+		t.Run(set, func(t *testing.T) {
+			t.Setenv("METRICS_USERNAME", "")
+			t.Setenv("METRICS_PASSWORD", "")
+			t.Setenv(set, "prometheus")
+
+			requireConfigError(t, "METRICS_USERNAME")
+			requireConfigError(t, "METRICS_PASSWORD")
+		})
+	}
+}
+
+// TestMetricsUsernameMustNotContainColon: basic auth splits the
+// credentials at the first ":", so such a user name would get 401 on
+// every request to /metrics.
+func TestMetricsUsernameMustNotContainColon(t *testing.T) {
+	t.Setenv("METRICS_USERNAME", "prom:etheus")
+	t.Setenv("METRICS_PASSWORD", "secret")
+
+	requireConfigError(t, "METRICS_USERNAME")
+}
+
+// TestCORSAllowedOriginsMustBeOrigins: "*" would let every origin in,
+// and an entry that is not a plain origin would match no page.
+func TestCORSAllowedOriginsMustBeOrigins(t *testing.T) {
+	for _, entry := range []string{
+		"*",
+		"https://*.netwatch.example",
+		"netwatch.example",
+		"https://netwatch.example/",
+	} {
+		t.Run(entry, func(t *testing.T) {
+			t.Setenv("CORS_ALLOWED_ORIGINS", entry)
+
+			requireConfigError(t, "CORS_ALLOWED_ORIGINS")
+		})
+	}
+}