BountyBoard/internal/chat/sanitize_test.go

package chat

import (
	"strings"
	"testing"
)

func TestSanitizeHTML(t *testing.T) {
	tests := []struct {
		name string
		in   string
		want string
	}{
		{"plain text escaped", `hello <world> & "friends"`, "hello  &amp; &#34;friends&#34;"},
		{"allowed formatting kept", "<b>bold</b> <i>it</i> <u>u</u> <s>s</s>", "<b>bold</b> <i>it</i> <u>u</u> <s>s</s>"},
		{"paragraph and lists", "<p>a</p><ul><li>x</li><li>y</li></ul>", "<p>a</p><ul><li>x</li><li>y</li></ul>"},
		{"code and pre", "<pre><code>x := 1</code></pre>", "<pre><code>x := 1</code></pre>"},
		{"blockquote", "<blockquote>q</blockquote>", "<blockquote>q</blockquote>"},
		{"br void", "a<br>b<br/>c", "a<br>b<br>c"},
		{"script stripped", `<script>alert(1)</script>hi`, "alert(1)hi"},
		{"img dropped", `<img src=x onerror=alert(1)>safe`, "safe"},
		{"event handlers stripped from allowed tag", `<b onclick="evil()">x</b>`, "<b>x</b>"},
		{"style attr stripped", `<p style="position:fixed">x</p>`, "<p>x</p>"},
		{"https link kept with rel", `<a href="https://x.y/z">l</a>`,
			`<a href="https://x.y/z" rel="noopener noreferrer" target="_blank">l</a>`},
		{"http link kept", `<a href="http://x.y">l</a>`,
			`<a href="http://x.y" rel="noopener noreferrer" target="_blank">l</a>`},
		{"javascript href removed", `<a href="javascript:alert(1)">l</a>`, "<a>l</a>"},
		{"data href removed", `<a href="data:text/html,x">l</a>`, "<a>l</a>"},
		{"single-quoted js href removed", `<a href='javascript:x'>l</a>`, "<a>l</a>"},
		{"unquoted href", `<a href=https://ok.example>l</a>`,
			`<a href="https://ok.example" rel="noopener noreferrer" target="_blank">l</a>`},
		{"unclosed tags balanced", "<b>bold <i>both", "<b>bold <i>both</i></b>"},
		{"stray close ignored", "</b>text", "text"},
		{"mismatched nesting closed in order", "<b><i>x</b></i>", "<b><i>x</i></b>"},
		{"case insensitive tags", "<B>x</B><SCRIPT>y</SCRIPT>", "<b>x</b>y"},
		{"nested quotes attack", `<a href="https://x" "><script>alert(1)</script>">x</a>`,
			`<a href="https://x" rel="noopener noreferrer" target="_blank">alert(1)&#34;&gt;x</a>`},
		{"incomplete tag escaped", "a <b", "a &lt;b"},
		{"empty input", "", ""},
	}
	for _, tt := range tests {
		t.Run(tt.name, func(t *testing.T) {
			if got := SanitizeHTML(tt.in); got != tt.want {
				t.Errorf("SanitizeHTML(%q)\n got  %q\n want %q", tt.in, got, tt.want)
			}
		})
	}
}

func TestSanitizeHTMLNeverEmitsDangerousOutput(t *testing.T) {
	vectors := []string{
		`<script>alert(1)</script>`,
		`<img src=x onerror=alert(1)>`,
		`<svg onload=alert(1)>`,
		`<iframe src="https://evil"></iframe>`,
		`<a href="javascript:alert(1)">x</a>`,
		`<A HREF="JAVASCRIPT:alert(1)">x</A>`,
		`<b onmouseover="alert(1)">x</b>`,
		`<<script>script>alert(1)<</script>/script>`,
		`<p><style>*{}</style></p>`,
		`<form action="https://evil"><input></form>`,
	}
	for _, v := range vectors {
		got := strings.ToLower(SanitizeHTML(v))
		for _, bad := range []string{"<script", "onerror", "onload", "onmouseover",
			"javascript:", "<iframe", "<svg", "<img", "<form", "<style", "<input"} {
			if strings.Contains(got, bad) {
				t.Errorf("vector %q produced dangerous output %q (contains %q)", v, got, bad)
			}
		}
	}
}

func TestSanitizePlain(t *testing.T) {
	if got := SanitizePlain(`<b>x</b> & <script>y</script>`); got != "x &amp; y" {
		t.Errorf("SanitizePlain = %q", got)
	}
}