1 files changed, 144 insertions, 0 deletions
diff --git a/reader/sanitizer/sanitizer_test.go b/reader/sanitizer/sanitizer_test.go
new file mode 100644
index 0000000..73862d3
--- /dev/null
+++ b/reader/sanitizer/sanitizer_test.go
@@ -0,0 +1,144 @@
+// Copyright 2017 Frédéric Guillot. All rights reserved.
+// Use of this source code is governed by the Apache 2.0
+// license that can be found in the LICENSE file.
+
+package sanitizer
+
+import "testing"
+
+func TestValidInput(t *testing.T) {
+	input := `<p>This is a <strong>text</strong> with an image: <img src="http://example.org/" alt="Test">.</p>`
+	output := Sanitize("http://example.org/", input)
+
+	if input != output {
+		t.Errorf(`Wrong output: "%s" != "%s"`, input, output)
+	}
+}
+
+func TestSelfClosingTags(t *testing.T) {
+	input := `<p>This <br> is a <strong>text</strong> <br/>with an image: <img src="http://example.org/" alt="Test"/>.</p>`
+	output := Sanitize("http://example.org/", input)
+
+	if input != output {
+		t.Errorf(`Wrong output: "%s" != "%s"`, input, output)
+	}
+}
+
+func TestTable(t *testing.T) {
+	input := `<table><tr><th>A</th><th colspan="2">B</th></tr><tr><td>C</td><td>D</td><td>E</td></tr></table>`
+	output := Sanitize("http://example.org/", input)
+
+	if input != output {
+		t.Errorf(`Wrong output: "%s" != "%s"`, input, output)
+	}
+}
+
+func TestRelativeURL(t *testing.T) {
+	input := `This <a href="/test.html">link is relative</a> and this image: <img src="../folder/image.png"/>`
+	expected := `This <a href="http://example.org/test.html" rel="noopener noreferrer" target="_blank" referrerpolicy="no-referrer">link is relative</a> and this image: <img src="http://example.org/folder/image.png"/>`
+	output := Sanitize("http://example.org/", input)
+
+	if expected != output {
+		t.Errorf(`Wrong output: "%s" != "%s"`, expected, output)
+	}
+}
+
+func TestProtocolRelativeURL(t *testing.T) {
+	input := `This <a href="//static.example.org/index.html">link is relative</a>.`
+	expected := `This <a href="https://static.example.org/index.html" rel="noopener noreferrer" target="_blank" referrerpolicy="no-referrer">link is relative</a>.`
+	output := Sanitize("http://example.org/", input)
+
+	if expected != output {
+		t.Errorf(`Wrong output: "%s" != "%s"`, expected, output)
+	}
+}
+
+func TestInvalidTag(t *testing.T) {
+	input := `<p>My invalid <b>tag</b>.</p>`
+	expected := `<p>My invalid tag.</p>`
+	output := Sanitize("http://example.org/", input)
+
+	if expected != output {
+		t.Errorf(`Wrong output: "%s" != "%s"`, expected, output)
+	}
+}
+
+func TestVideoTag(t *testing.T) {
+	input := `<p>My valid <video src="videofile.webm" autoplay poster="posterimage.jpg">fallback</video>.</p>`
+	expected := `<p>My valid <video src="http://example.org/videofile.webm" poster="http://example.org/posterimage.jpg" controls>fallback</video>.</p>`
+	output := Sanitize("http://example.org/", input)
+
+	if expected != output {
+		t.Errorf(`Wrong output: "%s" != "%s"`, expected, output)
+	}
+}
+
+func TestAudioAndSourceTag(t *testing.T) {
+	input := `<p>My music <audio controls="controls"><source src="foo.wav" type="audio/wav"></audio>.</p>`
+	expected := `<p>My music <audio controls><source src="http://example.org/foo.wav" type="audio/wav"></audio>.</p>`
+	output := Sanitize("http://example.org/", input)
+
+	if expected != output {
+		t.Errorf(`Wrong output: "%s" != "%s"`, expected, output)
+	}
+}
+
+func TestUnknownTag(t *testing.T) {
+	input := `<p>My invalid <unknown>tag</unknown>.</p>`
+	expected := `<p>My invalid tag.</p>`
+	output := Sanitize("http://example.org/", input)
+
+	if expected != output {
+		t.Errorf(`Wrong output: "%s" != "%s"`, expected, output)
+	}
+}
+
+func TestInvalidNestedTag(t *testing.T) {
+	input := `<p>My invalid <b>tag with some <em>valid</em> tag</b>.</p>`
+	expected := `<p>My invalid tag with some <em>valid</em> tag.</p>`
+	output := Sanitize("http://example.org/", input)
+
+	if expected != output {
+		t.Errorf(`Wrong output: "%s" != "%s"`, expected, output)
+	}
+}
+
+func TestInvalidIFrame(t *testing.T) {
+	input := `<iframe src="http://example.org/"></iframe>`
+	expected := ``
+	output := Sanitize("http://example.org/", input)
+
+	if expected != output {
+		t.Errorf(`Wrong output: "%s" != "%s"`, expected, output)
+	}
+}
+
+func TestInvalidURLScheme(t *testing.T) {
+	input := `<p>This link is <a src="file:///etc/passwd">not valid</a></p>`
+	expected := `<p>This link is not valid</p>`
+	output := Sanitize("http://example.org/", input)
+
+	if expected != output {
+		t.Errorf(`Wrong output: "%s" != "%s"`, expected, output)
+	}
+}
+
+func TestBlacklistedLink(t *testing.T) {
+	input := `<p>This image is not valid <img src="https://stats.wordpress.com/some-tracker"></p>`
+	expected := `<p>This image is not valid </p>`
+	output := Sanitize("http://example.org/", input)
+
+	if expected != output {
+		t.Errorf(`Wrong output: "%s" != "%s"`, expected, output)
+	}
+}
+
+func TestPixelTracker(t *testing.T) {
+	input := `<p><img src="https://tracker1.example.org/" height="1" width="1"> and <img src="https://tracker2.example.org/" height="1" width="1"/></p>`
+	expected := `<p> and </p>`
+	output := Sanitize("http://example.org/", input)
+
+	if expected != output {
+		t.Errorf(`Wrong output: "%s" != "%s"`, expected, output)
+	}
+}