Add sanitizer rules per renderer (#16110)

* Added sanitizer rules per renderer.

* Updated documentation.

Co-authored-by: techknowlogick <techknowlogick@gitea.io>

7 files changed, 114 insertions, 66 deletions

diff --git a/modules/markup/csv/csv.go b/modules/markup/csv/csv.go
index 6572b0ee1e..8a4df89511 100644
--- a/modules/markup/csv/csv.go
+++ b/modules/markup/csv/csv.go
@@ -10,6 +10,7 @@ import (
 	"html"
 	"io"
 	"io/ioutil"
+	"regexp"
 	"strconv"
 
 	"code.gitea.io/gitea/modules/csv"
@@ -38,6 +39,15 @@ func (Renderer) Extensions() []string {
 	return []string{".csv", ".tsv"}
 }
 
+// SanitizerRules implements markup.Renderer
+func (Renderer) SanitizerRules() []setting.MarkupSanitizerRule {
+	return []setting.MarkupSanitizerRule{
+		{Element: "table", AllowAttr: "class", Regexp: regexp.MustCompile(`data-table`)},
+		{Element: "th", AllowAttr: "class", Regexp: regexp.MustCompile(`line-num`)},
+		{Element: "td", AllowAttr: "class", Regexp: regexp.MustCompile(`line-num`)},
+	}
+}
+
 func writeField(w io.Writer, element, class, field string) error {
 	if _, err := io.WriteString(w, "<"); err != nil {
 		return err
diff --git a/modules/markup/external/external.go b/modules/markup/external/external.go
index 62814c9914..c849f505e7 100644
--- a/modules/markup/external/external.go
+++ b/modules/markup/external/external.go
@@ -30,7 +30,7 @@ func RegisterRenderers() {
 
 // Renderer implements markup.Renderer for external tools
 type Renderer struct {
-	setting.MarkupRenderer
+	*setting.MarkupRenderer
 }
 
 // Name returns the external tool name
@@ -48,6 +48,11 @@ func (p *Renderer) Extensions() []string {
 	return p.FileExtensions
 }
 
+// SanitizerRules implements markup.Renderer
+func (p *Renderer) SanitizerRules() []setting.MarkupSanitizerRule {
+	return p.MarkupSanitizerRules
+}
+
 func envMark(envName string) string {
 	if runtime.GOOS == "windows" {
 		return "%" + envName + "%"
diff --git a/modules/markup/html_test.go b/modules/markup/html_test.go
index 4cdd5798c8..8c3d2b5395 100644
--- a/modules/markup/html_test.go
+++ b/modules/markup/html_test.go
@@ -112,7 +112,7 @@ func TestRender_links(t *testing.T) {
 
 	defaultCustom := setting.Markdown.CustomURLSchemes
 	setting.Markdown.CustomURLSchemes = []string{"ftp", "magnet"}
-	ReplaceSanitizer()
+	InitializeSanitizer()
 	CustomLinkURLSchemes(setting.Markdown.CustomURLSchemes)
 
 	test(
@@ -192,7 +192,7 @@ func TestRender_links(t *testing.T) {
 
 	// Restore previous settings
 	setting.Markdown.CustomURLSchemes = defaultCustom
-	ReplaceSanitizer()
+	InitializeSanitizer()
 	CustomLinkURLSchemes(setting.Markdown.CustomURLSchemes)
 }
 
diff --git a/modules/markup/markdown/markdown.go b/modules/markup/markdown/markdown.go
index 87fae2a23b..cac2a180fa 100644
--- a/modules/markup/markdown/markdown.go
+++ b/modules/markup/markdown/markdown.go
@@ -199,7 +199,7 @@ func actualRender(ctx *markup.RenderContext, input io.Reader, output io.Writer)
 		}
 		_ = lw.Close()
 	}()
-	buf := markup.SanitizeReader(rd)
+	buf := markup.SanitizeReader(rd, "")
 	_, err := io.Copy(output, buf)
 	return err
 }
@@ -215,7 +215,7 @@ func render(ctx *markup.RenderContext, input io.Reader, output io.Writer) error
 		if log.IsDebug() {
 			log.Debug("Panic in markdown: %v\n%s", err, string(log.Stack(2)))
 		}
-		ret := markup.SanitizeReader(input)
+		ret := markup.SanitizeReader(input, "")
 		_, err = io.Copy(output, ret)
 		if err != nil {
 			log.Error("SanitizeReader failed: %v", err)
@@ -249,6 +249,11 @@ func (Renderer) Extensions() []string {
 	return setting.Markdown.FileExtensions
 }
 
+// SanitizerRules implements markup.Renderer
+func (Renderer) SanitizerRules() []setting.MarkupSanitizerRule {
+	return []setting.MarkupSanitizerRule{}
+}
+
 // Render implements markup.Renderer
 func (Renderer) Render(ctx *markup.RenderContext, input io.Reader, output io.Writer) error {
 	return render(ctx, input, output)
diff --git a/modules/markup/orgmode/orgmode.go b/modules/markup/orgmode/orgmode.go
index 851fc97f9a..7e9f1f45c5 100644
--- a/modules/markup/orgmode/orgmode.go
+++ b/modules/markup/orgmode/orgmode.go
@@ -13,6 +13,7 @@ import (
 
 	"code.gitea.io/gitea/modules/highlight"
 	"code.gitea.io/gitea/modules/markup"
+	"code.gitea.io/gitea/modules/setting"
 	"code.gitea.io/gitea/modules/util"
 
 	"github.com/alecthomas/chroma"
@@ -41,6 +42,11 @@ func (Renderer) Extensions() []string {
 	return []string{".org"}
 }
 
+// SanitizerRules implements markup.Renderer
+func (Renderer) SanitizerRules() []setting.MarkupSanitizerRule {
+	return []setting.MarkupSanitizerRule{}
+}
+
 // Render renders orgmode rawbytes to HTML
 func Render(ctx *markup.RenderContext, input io.Reader, output io.Writer) error {
 	htmlWriter := org.NewHTMLWriter()
diff --git a/modules/markup/renderer.go b/modules/markup/renderer.go
index d60c8ad710..04619caee3 100644
--- a/modules/markup/renderer.go
+++ b/modules/markup/renderer.go
@@ -81,6 +81,7 @@ type Renderer interface {
 	Name() string // markup format name
 	Extensions() []string
 	NeedPostProcess() bool
+	SanitizerRules() []setting.MarkupSanitizerRule
 	Render(ctx *RenderContext, input io.Reader, output io.Writer) error
 }
 
@@ -136,37 +137,32 @@ func render(ctx *RenderContext, renderer Renderer, input io.Reader, output io.Wr
 		_ = pw.Close()
 	}()
 
-	if renderer.NeedPostProcess() {
-		pr2, pw2 := io.Pipe()
-		defer func() {
-			_ = pr2.Close()
-			_ = pw2.Close()
-		}()
-
-		wg.Add(1)
-		go func() {
-			buf := SanitizeReader(pr2)
-			_, err = io.Copy(output, buf)
-			_ = pr2.Close()
-			wg.Done()
-		}()
-
-		wg.Add(1)
-		go func() {
+	pr2, pw2 := io.Pipe()
+	defer func() {
+		_ = pr2.Close()
+		_ = pw2.Close()
+	}()
+
+	wg.Add(1)
+	go func() {
+		buf := SanitizeReader(pr2, renderer.Name())
+		_, err = io.Copy(output, buf)
+		_ = pr2.Close()
+		wg.Done()
+	}()
+
+	wg.Add(1)
+	go func() {
+		if renderer.NeedPostProcess() {
 			err = PostProcess(ctx, pr, pw2)
-			_ = pr.Close()
-			_ = pw2.Close()
-			wg.Done()
-		}()
-	} else {
-		wg.Add(1)
-		go func() {
-			buf := SanitizeReader(pr)
-			_, err = io.Copy(output, buf)
-			_ = pr.Close()
-			wg.Done()
-		}()
-	}
+		} else {
+			_, err = io.Copy(pw2, pr)
+		}
+		_ = pr.Close()
+		_ = pw2.Close()
+		wg.Done()
+	}()
+
 	if err1 := renderer.Render(ctx, input, pw); err1 != nil {
 		return err1
 	}
diff --git a/modules/markup/sanitizer.go b/modules/markup/sanitizer.go
index 5611bd06ad..9342d65de5 100644
--- a/modules/markup/sanitizer.go
+++ b/modules/markup/sanitizer.go
@@ -19,8 +19,9 @@ import (
 // Sanitizer is a protection wrapper of *bluemonday.Policy which does not allow
 // any modification to the underlying policies once it's been created.
 type Sanitizer struct {
-	policy *bluemonday.Policy
-	init   sync.Once
+	defaultPolicy    *bluemonday.Policy
+	rendererPolicies map[string]*bluemonday.Policy
+	init             sync.Once
 }
 
 var sanitizer = &Sanitizer{}
@@ -30,47 +31,57 @@ var sanitizer = &Sanitizer{}
 // entire application lifecycle.
 func NewSanitizer() {
 	sanitizer.init.Do(func() {
-		ReplaceSanitizer()
+		InitializeSanitizer()
 	})
 }
 
-// ReplaceSanitizer replaces the current sanitizer to account for changes in settings
-func ReplaceSanitizer() {
-	sanitizer.policy = bluemonday.UGCPolicy()
+// InitializeSanitizer (re)initializes the current sanitizer to account for changes in settings
+func InitializeSanitizer() {
+	sanitizer.rendererPolicies = map[string]*bluemonday.Policy{}
+	sanitizer.defaultPolicy = createDefaultPolicy()
+
+	for name, renderer := range renderers {
+		sanitizerRules := renderer.SanitizerRules()
+		if len(sanitizerRules) > 0 {
+			policy := createDefaultPolicy()
+			addSanitizerRules(policy, sanitizerRules)
+			sanitizer.rendererPolicies[name] = policy
+		}
+	}
+}
+
+func createDefaultPolicy() *bluemonday.Policy {
+	policy := bluemonday.UGCPolicy()
 	// For Chroma markdown plugin
-	sanitizer.policy.AllowAttrs("class").Matching(regexp.MustCompile(`^is-loading$`)).OnElements("pre")
-	sanitizer.policy.AllowAttrs("class").Matching(regexp.MustCompile(`^(chroma )?language-[\w-]+$`)).OnElements("code")
+	policy.AllowAttrs("class").Matching(regexp.MustCompile(`^is-loading$`)).OnElements("pre")
+	policy.AllowAttrs("class").Matching(regexp.MustCompile(`^(chroma )?language-[\w-]+$`)).OnElements("code")
 
 	// Checkboxes
-	sanitizer.policy.AllowAttrs("type").Matching(regexp.MustCompile(`^checkbox$`)).OnElements("input")
-	sanitizer.policy.AllowAttrs("checked", "disabled", "data-source-position").OnElements("input")
+	policy.AllowAttrs("type").Matching(regexp.MustCompile(`^checkbox$`)).OnElements("input")
+	policy.AllowAttrs("checked", "disabled", "data-source-position").OnElements("input")
 
 	// Custom URL-Schemes
 	if len(setting.Markdown.CustomURLSchemes) > 0 {
-		sanitizer.policy.AllowURLSchemes(setting.Markdown.CustomURLSchemes...)
+		policy.AllowURLSchemes(setting.Markdown.CustomURLSchemes...)
 	}
 
 	// Allow classes for anchors
-	sanitizer.policy.AllowAttrs("class").Matching(regexp.MustCompile(`ref-issue`)).OnElements("a")
+	policy.AllowAttrs("class").Matching(regexp.MustCompile(`ref-issue`)).OnElements("a")
 
 	// Allow classes for task lists
-	sanitizer.policy.AllowAttrs("class").Matching(regexp.MustCompile(`task-list-item`)).OnElements("li")
+	policy.AllowAttrs("class").Matching(regexp.MustCompile(`task-list-item`)).OnElements("li")
 
 	// Allow icons
-	sanitizer.policy.AllowAttrs("class").Matching(regexp.MustCompile(`^icon(\s+[\p{L}\p{N}_-]+)+$`)).OnElements("i")
+	policy.AllowAttrs("class").Matching(regexp.MustCompile(`^icon(\s+[\p{L}\p{N}_-]+)+$`)).OnElements("i")
 
 	// Allow unlabelled labels
-	sanitizer.policy.AllowNoAttrs().OnElements("label")
+	policy.AllowNoAttrs().OnElements("label")
 
 	// Allow classes for emojis
-	sanitizer.policy.AllowAttrs("class").Matching(regexp.MustCompile(`emoji`)).OnElements("img")
+	policy.AllowAttrs("class").Matching(regexp.MustCompile(`emoji`)).OnElements("img")
 
 	// Allow icons, emojis, chroma syntax and keyword markup on span
-	sanitizer.policy.AllowAttrs("class").Matching(regexp.MustCompile(`^((icon(\s+[\p{L}\p{N}_-]+)+)|(emoji))$|^([a-z][a-z0-9]{0,2})$|^` + keywordClass + `$`)).OnElements("span")
-
-	// Allow data tables
-	sanitizer.policy.AllowAttrs("class").Matching(regexp.MustCompile(`data-table`)).OnElements("table")
-	sanitizer.policy.AllowAttrs("class").Matching(regexp.MustCompile(`line-num`)).OnElements("th", "td")
+	policy.AllowAttrs("class").Matching(regexp.MustCompile(`^((icon(\s+[\p{L}\p{N}_-]+)+)|(emoji))$|^([a-z][a-z0-9]{0,2})$|^` + keywordClass + `$`)).OnElements("span")
 
 	// Allow generally safe attributes
 	generalSafeAttrs := []string{"abbr", "accept", "accept-charset",
@@ -101,18 +112,29 @@ func ReplaceSanitizer() {
 		"abbr", "bdo", "cite", "dfn", "mark", "small", "span", "time", "wbr",
 	}
 
-	sanitizer.policy.AllowAttrs(generalSafeAttrs...).OnElements(generalSafeElements...)
+	policy.AllowAttrs(generalSafeAttrs...).OnElements(generalSafeElements...)
 
-	sanitizer.policy.AllowAttrs("itemscope", "itemtype").OnElements("div")
+	policy.AllowAttrs("itemscope", "itemtype").OnElements("div")
 
 	// FIXME: Need to handle longdesc in img but there is no easy way to do it
 
 	// Custom keyword markup
-	for _, rule := range setting.ExternalSanitizerRules {
-		if rule.Regexp != nil {
-			sanitizer.policy.AllowAttrs(rule.AllowAttr).Matching(rule.Regexp).OnElements(rule.Element)
-		} else {
-			sanitizer.policy.AllowAttrs(rule.AllowAttr).OnElements(rule.Element)
+	addSanitizerRules(policy, setting.ExternalSanitizerRules)
+
+	return policy
+}
+
+func addSanitizerRules(policy *bluemonday.Policy, rules []setting.MarkupSanitizerRule) {
+	for _, rule := range rules {
+		if rule.AllowDataURIImages {
+			policy.AllowDataURIImages()
+		}
+		if rule.Element != "" {
+			if rule.Regexp != nil {
+				policy.AllowAttrs(rule.AllowAttr).Matching(rule.Regexp).OnElements(rule.Element)
+			} else {
+				policy.AllowAttrs(rule.AllowAttr).OnElements(rule.Element)
+			}
 		}
 	}
 }
@@ -120,11 +142,15 @@ func ReplaceSanitizer() {
 // Sanitize takes a string that contains a HTML fragment or document and applies policy whitelist.
 func Sanitize(s string) string {
 	NewSanitizer()
-	return sanitizer.policy.Sanitize(s)
+	return sanitizer.defaultPolicy.Sanitize(s)
 }
 
 // SanitizeReader sanitizes a Reader
-func SanitizeReader(r io.Reader) *bytes.Buffer {
+func SanitizeReader(r io.Reader, renderer string) *bytes.Buffer {
 	NewSanitizer()
-	return sanitizer.policy.SanitizeReader(r)
+	policy, exist := sanitizer.rendererPolicies[renderer]
+	if !exist {
+		policy = sanitizer.defaultPolicy
+	}
+	return policy.SanitizeReader(r)
 }