Provide configuration to allow camo-media proxying (#12802)

* Provide configuration to allow camo-media proxying

Fix #916

Signed-off-by: Andrew Thornton <[email protected]>
Co-authored-by: 6543 <[email protected]>
Co-authored-by: wxiaoguang <[email protected]>

Author: 3 people

custom/conf/app.example.ini

@@ -424,6 +424,23 @@ INTERNAL_TOKEN=
 ;; This cache will store the successfully hashed tokens in a LRU cache as a balance between performance and security.
 ;SUCCESSFUL_TOKENS_CACHE_SIZE = 20
 
+;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
+;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
+[camo]
+;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
+;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
+;;
+;; At the moment we only support images
+;;
+;; if the camo is enabled
+;ENABLED = false
+;; url to a camo image proxy, it **is required** if camo is enabled.
+;SERVER_URL =
+;; HMAC to encode urls with, it **is required** if camo is enabled.
+;HMAC_KEY =
+;; Set to true to use camo for https too lese only non https urls are proxyed
+;ALLWAYS = false
+
 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
 [oauth2]

docs/content/doc/advanced/config-cheat-sheet.en-us.md

@@ -513,7 +513,14 @@ Certain queues have defaults that override the defaults set in `[queue]` (this o
     - spec - use one or more special characters as ``!"#$%&'()*+,-./:;<=>?@[\\]^_`{|}~``
     - off - do not check password complexity
 - `PASSWORD_CHECK_PWN`: **false**: Check [HaveIBeenPwned](https://haveibeenpwned.com/Passwords) to see if a password has been exposed.
-- `SUCCESSFUL_TOKENS_CACHE_SIZE`: **20**: Cache successful token hashes. API tokens are stored in the DB as pbkdf2 hashes however, this means that there is a potentially significant hashing load when there are multiple API operations. This cache will store the successfully hashed tokens in a LRU cache as a balance between performance and security. 
+- `SUCCESSFUL_TOKENS_CACHE_SIZE`: **20**: Cache successful token hashes. API tokens are stored in the DB as pbkdf2 hashes however, this means that there is a potentially significant hashing load when there are multiple API operations. This cache will store the successfully hashed tokens in a LRU cache as a balance between performance and security.
+
+## Camo (`camo`)
+
+- `ENABLED`: **false**: Enable media proxy, we support images only at the moment.
+- `SERVER_URL`: **<empty>**: url of camo server, it **is required** if camo is enabled.
+- `HMAC_KEY`: **<empty>**: Provide the HMAC key for encoding urls, it **is required** if camo is enabled.
+- `ALLWAYS`: **false**: Set to true to use camo for https too lese only non https urls are proxyed
 
 ## OpenID (`openid`)
 

modules/markup/camo.go

@@ -0,0 +1,47 @@
+// Copyright 2022 The Gitea Authors. All rights reserved.
+// Use of this source code is governed by a MIT-style
+// license that can be found in the LICENSE file.
+
+package markup
+
+import (
+	"crypto/hmac"
+	"crypto/sha1"
+	"encoding/base64"
+	"net/url"
+	"strings"
+
+	"code.gitea.io/gitea/modules/setting"
+	"code.gitea.io/gitea/modules/util"
+)
+
+// CamoEncode encodes a lnk to fit with the go-camo and camo proxy links. The purposes of camo-proxy are:
+// 1. Allow accessing "http://" images on a HTTPS site by using the "https://" URLs provided by camo-proxy.
+// 2. Hide the visitor's real IP (protect privacy) when accessing external images.
+func CamoEncode(link string) string {
+	if strings.HasPrefix(link, setting.Camo.ServerURL) {
+		return link
+	}
+
+	mac := hmac.New(sha1.New, []byte(setting.Camo.HMACKey))
+	_, _ = mac.Write([]byte(link)) // hmac does not return errors
+	macSum := b64encode(mac.Sum(nil))
+	encodedURL := b64encode([]byte(link))
+
+	return util.URLJoin(setting.Camo.ServerURL, macSum, encodedURL)
+}
+
+func b64encode(data []byte) string {
+	return strings.TrimRight(base64.URLEncoding.EncodeToString(data), "=")
+}
+
+func camoHandleLink(link string) string {
+	if setting.Camo.Enabled {
+		lnkURL, err := url.Parse(link)
+		if err == nil && lnkURL.IsAbs() && !strings.HasPrefix(link, setting.AppURL) &&
+			(setting.Camo.Allways || lnkURL.Scheme != "https") {
+			return CamoEncode(link)
+		}
+	}
+	return link
+}

modules/markup/camo_test.go

@@ -0,0 +1,45 @@
+// Copyright 2022 The Gitea Authors. All rights reserved.
+// Use of this source code is governed by a MIT-style
+// license that can be found in the LICENSE file.
+
+package markup
+
+import (
+	"testing"
+
+	"code.gitea.io/gitea/modules/setting"
+
+	"github.com/stretchr/testify/assert"
+)
+
+func TestCamoHandleLink(t *testing.T) {
+	setting.AppURL = "https://gitea.com"
+	// Test media proxy
+	setting.Camo.Enabled = true
+	setting.Camo.ServerURL = "https://image.proxy"
+	setting.Camo.HMACKey = "geheim"
+
+	assert.Equal(t,
+		"https://gitea.com/img.jpg",
+		camoHandleLink("https://gitea.com/img.jpg"))
+	assert.Equal(t,
+		"https://testimages.org/img.jpg",
+		camoHandleLink("https://testimages.org/img.jpg"))
+	assert.Equal(t,
+		"https://image.proxy/eivin43gJwGVIjR9MiYYtFIk0mw/aHR0cDovL3Rlc3RpbWFnZXMub3JnL2ltZy5qcGc",
+		camoHandleLink("http://testimages.org/img.jpg"))
+
+	setting.Camo.Allways = true
+	assert.Equal(t,
+		"https://gitea.com/img.jpg",
+		camoHandleLink("https://gitea.com/img.jpg"))
+	assert.Equal(t,
+		"https://image.proxy/tkdlvmqpbIr7SjONfHNgEU622y0/aHR0cHM6Ly90ZXN0aW1hZ2VzLm9yZy9pbWcuanBn",
+		camoHandleLink("https://testimages.org/img.jpg"))
+	assert.Equal(t,
+		"https://image.proxy/eivin43gJwGVIjR9MiYYtFIk0mw/aHR0cDovL3Rlc3RpbWFnZXMub3JnL2ltZy5qcGc",
+		camoHandleLink("http://testimages.org/img.jpg"))
+
+	// Restore previous settings
+	setting.Camo.Enabled = false
+}

modules/markup/html.go

@@ -386,6 +386,7 @@ func visitNode(ctx *RenderContext, procs, textProcs []processor, node *html.Node
 
 					attr.Val = util.URLJoin(prefix, attr.Val)
 				}
+				attr.Val = camoHandleLink(attr.Val)
 				node.Attr[i] = attr
 			}
 		} else if node.Data == "a" {

modules/setting/setting.go

@@ -197,6 +197,13 @@ var (
 	PasswordCheckPwn                   bool
 	SuccessfulTokensCacheSize          int
 
+	Camo = struct {
+		Enabled   bool
+		ServerURL string `ini:"SERVER_URL"`
+		HMACKey   string `ini:"HMAC_KEY"`
+		Allways   bool
+	}{}
+
 	// UI settings
 	UI = struct {
 		ExplorePagingNum      int
@@ -1019,6 +1026,14 @@ func loadFromConf(allowEmpty bool, extraConfig string) {
 		log.Fatal("Failed to map API settings: %v", err)
 	} else if err = Cfg.Section("metrics").MapTo(&Metrics); err != nil {
 		log.Fatal("Failed to map Metrics settings: %v", err)
+	} else if err = Cfg.Section("camo").MapTo(&Camo); err != nil {
+		log.Fatal("Failed to map Camo settings: %v", err)
+	}
+
+	if Camo.Enabled {
+		if Camo.ServerURL == "" || Camo.HMACKey == "" {
+			log.Fatal(`Camo settings require "SERVER_URL" and HMAC_KEY`)
+		}
 	}
 
 	u := *appURL