feat: scan host LAN web devices via proxy

internal/webdevice/proxy.go

@@ -3,11 +3,13 @@ package webdevice
 import (
 	"errors"
 	"io"
+	"log"
 	"net/http"
 	"net/http/httputil"
 	"net/url"
 	"strconv"
 	"strings"
+	"time"
 )
 
 var (
@@ -36,6 +38,14 @@ func (s *Service) ProxyHTTP(w http.ResponseWriter, r *http.Request, targetIP, pr
 	rawQuery := r.URL.RawQuery
 
 	proxy := &httputil.ReverseProxy{
+		Transport: retryTransport{
+			base: &http.Transport{
+				DisableKeepAlives:     true,
+				ResponseHeaderTimeout: 5 * time.Second,
+			},
+			attempts: 3,
+			delay:    80 * time.Millisecond,
+		},
 		Director: func(req *http.Request) {
 			req.URL.Scheme = targetURL.Scheme
 			req.URL.Host = targetURL.Host
@@ -43,7 +53,8 @@ func (s *Service) ProxyHTTP(w http.ResponseWriter, r *http.Request, targetIP, pr
 			req.URL.RawPath = ""
 			req.URL.RawQuery = rawQuery
 			req.Host = targetURL.Host
-			req.Header.Del("Accept-Encoding")
+			req.Header = sanitizeProxyRequestHeader(req.Header, req.URL.Path)
+			req.Close = true
 		},
 		ModifyResponse: func(resp *http.Response) error {
 			proxyPrefix := "/proxy/web/" + targetIP
@@ -77,6 +88,7 @@ func (s *Service) ProxyHTTP(w http.ResponseWriter, r *http.Request, targetIP, pr
 			return nil
 		},
 		ErrorHandler: func(w http.ResponseWriter, r *http.Request, err error) {
+			log.Printf("web device proxy failed target=%s path=%s error=%v", targetURL.String(), r.URL.RequestURI(), err)
 			http.Error(w, "代理访问失败: "+err.Error(), http.StatusBadGateway)
 		},
 	}
@@ -85,6 +97,92 @@ func (s *Service) ProxyHTTP(w http.ResponseWriter, r *http.Request, targetIP, pr
 	return nil
 }
 
+type retryTransport struct {
+	base     http.RoundTripper
+	attempts int
+	delay    time.Duration
+}
+
+func (t retryTransport) RoundTrip(req *http.Request) (*http.Response, error) {
+	base := t.base
+	if base == nil {
+		base = http.DefaultTransport
+	}
+	attempts := t.attempts
+	if attempts < 1 {
+		attempts = 1
+	}
+
+	var lastErr error
+	for attempt := 0; attempt < attempts; attempt++ {
+		nextReq := req
+		if attempt > 0 {
+			nextReq = req.Clone(req.Context())
+			nextReq.Header = req.Header.Clone()
+			nextReq.Close = true
+		}
+
+		resp, err := base.RoundTrip(nextReq)
+		if err == nil {
+			return resp, nil
+		}
+		lastErr = err
+		if !shouldRetryProxyRequest(req, err) || attempt == attempts-1 {
+			return nil, err
+		}
+		time.Sleep(t.delay)
+	}
+	return nil, lastErr
+}
+
+func shouldRetryProxyRequest(req *http.Request, err error) bool {
+	if req == nil || err == nil {
+		return false
+	}
+	switch req.Method {
+	case http.MethodGet, http.MethodHead, http.MethodOptions:
+	default:
+		return false
+	}
+	message := strings.ToLower(err.Error())
+	return strings.Contains(message, "eof") ||
+		strings.Contains(message, "connection reset") ||
+		strings.Contains(message, "broken pipe")
+}
+
+func sanitizeProxyRequestHeader(source http.Header, upstreamPath string) http.Header {
+	header := make(http.Header)
+	copyHeaderValue(header, source, "Accept")
+	copyHeaderValue(header, source, "Content-Type")
+	copyHeaderValue(header, source, "Authorization")
+
+	userAgent := strings.TrimSpace(source.Get("User-Agent"))
+	if userAgent == "" {
+		userAgent = "Mozilla/5.0"
+	}
+	header.Set("User-Agent", userAgent)
+	header.Set("Connection", "close")
+
+	if !isLoginPagePath(upstreamPath) {
+		copyHeaderValue(header, source, "Cookie")
+	}
+	return header
+}
+
+func copyHeaderValue(target, source http.Header, key string) {
+	if value := source.Values(key); len(value) > 0 {
+		target.Del(key)
+		for _, item := range value {
+			target.Add(key, item)
+		}
+	}
+}
+
+func isLoginPagePath(path string) bool {
+	path = strings.ToLower(path)
+	return strings.HasSuffix(path, "/doc/page/login.asp") || path == "/doc/page/login.asp"
+}
+
 type closeNotifyWriter struct {
 	http.ResponseWriter
 }