Do not render truncated links in markdown (#32980) (#32983)

Backport #32980 by wxiaoguang

Co-authored-by: wxiaoguang <wxiaoguang@gmail.com>

modules/markup/html_link.go

@@ -9,6 +9,7 @@ import (
 	"strings"
 
 	"code.gitea.io/gitea/modules/markup/common"
+	"code.gitea.io/gitea/modules/util"
 
 	"golang.org/x/net/html"
 	"golang.org/x/net/html/atom"
@@ -171,6 +172,10 @@ func linkProcessor(ctx *RenderContext, node *html.Node) {
 		}
 
 		uri := node.Data[m[0]:m[1]]
+		remaining := node.Data[m[1]:]
+		if util.IsLikelySplitLeftPart(remaining) {
+			return
+		}
 		replaceContent(node, m[0], m[1], createLink(ctx, uri, uri, "" /*link*/))
 		node = node.NextSibling.NextSibling
 	}

modules/markup/html_test.go

@@ -206,6 +206,16 @@ func TestRender_links(t *testing.T) {
 	test(
 		"ftps://gitea.com",
 		`<p>ftps://gitea.com</p>`)
+
+	t.Run("LinkSplit", func(t *testing.T) {
+		input, _ := util.SplitStringAtByteN("http://10.1.2.3", 12)
+		assert.Equal(t, "http://10…", input)
+		test(input, "<p>http://10…</p>")
+
+		input, _ = util.SplitStringAtByteN("http://10.1.2.3", 13)
+		assert.Equal(t, "http://10.…", input)
+		test(input, "<p>http://10.…</p>")
+	})
 }
 
 func TestRender_email(t *testing.T) {

modules/util/string.go

@@ -3,7 +3,10 @@
 
 package util
 
-import "unsafe"
+import (
+	"strings"
+	"unsafe"
+)
 
 func isSnakeCaseUpper(c byte) bool {
 	return 'A' <= c && c <= 'Z'
@@ -95,3 +98,15 @@ func UnsafeBytesToString(b []byte) string {
 func UnsafeStringToBytes(s string) []byte {
 	return unsafe.Slice(unsafe.StringData(s), len(s))
 }
+
+// SplitTrimSpace splits the string at given separator and trims leading and trailing space
+func SplitTrimSpace(input, sep string) []string {
+	input = strings.TrimSpace(input)
+	var stringList []string
+	for _, s := range strings.Split(input, sep) {
+		if s = strings.TrimSpace(s); s != "" {
+			stringList = append(stringList, s)
+		}
+	}
+	return stringList
+}

modules/util/string_test.go

@@ -45,3 +45,8 @@ func TestToSnakeCase(t *testing.T) {
 		assert.Equal(t, expected, ToSnakeCase(input))
 	}
 }
+
+func TestSplitTrimSpace(t *testing.T) {
+	assert.Equal(t, []string{"a", "b", "c"}, SplitTrimSpace("a\nb\nc", "\n"))
+	assert.Equal(t, []string{"a", "b"}, SplitTrimSpace("\r\na\n\r\nb\n\n", "\n"))
+}

modules/util/truncate.go

@@ -14,6 +14,10 @@ const (
 	asciiEllipsis = "..."
 )
 
+func IsLikelySplitLeftPart(s string) bool {
+	return strings.HasSuffix(s, utf8Ellipsis) || strings.HasSuffix(s, asciiEllipsis)
+}
+
 // SplitStringAtByteN splits a string at byte n accounting for rune boundaries. (Combining characters are not accounted for.)
 func SplitStringAtByteN(input string, n int) (left, right string) {
 	if len(input) <= n {
@@ -38,19 +42,3 @@ func SplitStringAtByteN(input string, n int) (left, right string) {
 
 	return input[:end] + utf8Ellipsis, utf8Ellipsis + input[end:]
 }
-
-// SplitTrimSpace splits the string at given separator and trims leading and trailing space
-func SplitTrimSpace(input, sep string) []string {
-	// Trim initial leading & trailing space
-	input = strings.TrimSpace(input)
-	// replace CRLF with LF
-	input = strings.ReplaceAll(input, "\r\n", "\n")
-
-	var stringList []string
-	for _, s := range strings.Split(input, sep) {
-		// trim leading and trailing space
-		stringList = append(stringList, strings.TrimSpace(s))
-	}
-
-	return stringList
-}