sm-sitemap-plugin/routes.go

package sitemap

import (
	"errors"
	"log/slog"
	"net"
	"net/http"
	"net/netip"
	"strings"

	"git.golem15.com/golem15/sm-sitemap-plugin/classes"
	"git.golem15.com/golem15/sm-sitemap-plugin/models"
	"git.golem15.com/golem15/summercms/modules/backpack"
	"git.golem15.com/golem15/summercms/modules/pact"
	"git.golem15.com/golem15/summercms/modules/signpost"
	"git.golem15.com/golem15/summercms/modules/surf"
	"gorm.io/gorm"
)

// The response headers of the PHP routes (live responses).
const (
	contentTypeXML = "application/xml"
	cacheControl   = "no-cache, private"
)

// Routes mounts GET /sitemap.xml at the application root, outside any
// throttle group and without the backend middleware: search engines fetch it
// anonymously.
func (p *Plugin) Routes(r pact.Router) error {
	r.Get("/sitemap.xml", p.sitemapXML)
	return nil
}

// sitemapXML serves the definition with the fixed theme key (D-11): 404
// when there is none, an empty application/xml body when it has no items,
// else the generated XML.
func (p *Plugin) sitemapXML(w http.ResponseWriter, r *http.Request) {
	app := p.app
	db, err := database(app)
	if err != nil {
		http.Error(w, "Internal Server Error", http.StatusInternalServerError)
		return
	}
	var defs []models.Definition
	if err := db.WithContext(r.Context()).Where("theme = ?", models.DefinitionTheme).Order("id ASC").Limit(1).Find(&defs).Error; err != nil {
		slog.Default().Error("golem15.sitemap: load definition", "error", err)
		http.Error(w, "Internal Server Error", http.StatusInternalServerError)
		return
	}
	if len(defs) == 0 {
		slog.Default().Info("golem15.sitemap: No sitemap definition was found. Try creating one first.")
		http.NotFound(w, r)
		return
	}
	gen := &classes.Generator{Root: backendURL(app, r)}
	if reg, ok := app.Lookup[*signpost.Registry](); ok && reg != nil {
		gen.Resolver = reg
	}
	xml, err := gen.Generate(r.Context(), defs[0].Items())
	if err != nil {
		slog.Default().Error("golem15.sitemap: generate", "error", err)
		http.Error(w, "Internal Server Error", http.StatusInternalServerError)
		return
	}
	w.Header().Set("Content-Type", contentTypeXML)
	w.Header().Set("Cache-Control", cacheControl)
	w.WriteHeader(http.StatusOK)
	_, _ = w.Write([]byte(xml))
}

// database returns the app's published *gorm.DB.
func database(app *backpack.App) (*gorm.DB, error) {
	if app == nil {
		return nil, errors.New("golem15.sitemap: app is not booted")
	}
	db, ok := app.Lookup[*gorm.DB]()
	if !ok || db == nil {
		return nil, errors.New("golem15.sitemap: database is not configured")
	}
	return db, nil
}

// backendURL is the root URL the sitemap's absolute URLs start with, PHP
// url('/') right-trimmed of slashes: app.url when it is set, else the
// request's scheme and Host. The scheme is https for a TLS connection, or
// when the immediate peer is in http.trusted_proxies and its last
// X-Forwarded-Proto value says https; a client that sends the header
// directly is ignored.
func backendURL(app *backpack.App, r *http.Request) string {
	if app != nil && app.Config != nil {
		if u := strings.TrimRight(strings.TrimSpace(app.Config.String("app.url")), "/"); u != "" {
			return u
		}
	}
	scheme := "http"
	if isHTTPS(app, r) {
		scheme = "https"
	}
	return strings.TrimRight(scheme+"://"+r.Host, "/")
}

func isHTTPS(app *backpack.App, r *http.Request) bool {
	if r.TLS != nil {
		return true
	}
	if app == nil || app.Config == nil {
		return false
	}
	trusted := surf.TrustedProxies(app.Config)
	if len(trusted) == 0 || !peerTrusted(r.RemoteAddr, trusted) {
		return false
	}
	values := r.Header.Values("X-Forwarded-Proto")
	if len(values) == 0 {
		return false
	}
	hops := strings.Split(values[len(values)-1], ",")
	return strings.EqualFold(strings.TrimSpace(hops[len(hops)-1]), "https")
}

func peerTrusted(remoteAddr string, trusted []netip.Prefix) bool {
	host, _, err := net.SplitHostPort(remoteAddr)
	if err != nil {
		host = remoteAddr
	}
	addr, err := netip.ParseAddr(strings.Trim(host, "[]"))
	if err != nil {
		return false
	}
	addr = addr.Unmap()
	for _, p := range trusted {
		if p.Contains(addr) {
			return true
		}
	}
	return false
}