From 7035d8063c00c7c109b458f75c857ee5871065ea Mon Sep 17 00:00:00 2001
From: Torsten Grote <t@grobox.de>
Date: Tue, 6 Sep 2016 19:08:14 -0300
Subject: [PATCH] Clean HTML from RSS feeds with Jsoup and show Link Warning

---
 .../res/layout/fragment_link_dialog.xml       | 73 ++++++++++++++++++
 briar-android/res/values/strings.xml          |  6 ++
 .../android/blogs/BlogPostViewHolder.java     | 10 ++-
 .../android/util/AndroidUtils.java            | 37 +++++++++
 .../android/widget/LinkDialogFragment.java    | 77 +++++++++++++++++++
 briar-core/build.gradle                       |  2 +
 .../briarproject/feed/FeedManagerImpl.java    | 38 ++++-----
 .../src/org/briarproject/util/HtmlUtils.java  | 16 ++++
 .../org/briarproject/util/PrivacyUtils.java   |  1 -
 9 files changed, 238 insertions(+), 22 deletions(-)
 create mode 100644 briar-android/res/layout/fragment_link_dialog.xml
 create mode 100644 briar-android/src/org/briarproject/android/widget/LinkDialogFragment.java
 create mode 100644 briar-core/src/org/briarproject/util/HtmlUtils.java

diff --git a/briar-android/res/layout/fragment_link_dialog.xml b/briar-android/res/layout/fragment_link_dialog.xml
new file mode 100644
index 000000000..307b7110e
--- /dev/null
+++ b/briar-android/res/layout/fragment_link_dialog.xml
@@ -0,0 +1,73 @@
+<?xml version="1.0" encoding="utf-8"?>
+<LinearLayout
+	xmlns:android="http://schemas.android.com/apk/res/android"
+	xmlns:tools="http://schemas.android.com/tools"
+	android:layout_width="match_parent"
+	android:layout_height="match_parent"
+	android:orientation="vertical">
+
+	<LinearLayout
+		android:layout_width="wrap_content"
+		android:layout_height="wrap_content"
+		android:orientation="vertical"
+		android:padding="@dimen/margin_large">
+
+		<TextView
+			android:layout_width="match_parent"
+			android:layout_height="wrap_content"
+			android:text="@string/link_warning_title"
+			android:textColor="@color/briar_primary"
+			android:textSize="@dimen/text_size_large"
+			android:textStyle="bold"/>
+
+		<TextView
+			android:layout_width="match_parent"
+			android:layout_height="wrap_content"
+			android:layout_marginTop="@dimen/margin_large"
+			android:text="@string/link_warning_intro"
+			android:textColor="@color/briar_primary"
+			android:textSize="@dimen/text_size_medium"/>
+
+		<TextView
+			android:id="@+id/urlView"
+			android:layout_width="match_parent"
+			android:layout_height="wrap_content"
+			android:layout_marginTop="@dimen/margin_large"
+			android:textIsSelectable="true"
+			android:typeface="monospace"
+			tools:text="http://very.bad.site.com"/>
+
+		<TextView
+			android:layout_width="match_parent"
+			android:layout_height="wrap_content"
+			android:layout_marginTop="@dimen/margin_large"
+			android:text="@string/link_warning_text"
+			android:textColor="@color/briar_primary"
+			android:textSize="@dimen/text_size_medium"/>
+
+	</LinearLayout>
+
+	<LinearLayout
+		android:layout_width="match_parent"
+		android:layout_height="wrap_content"
+		android:orientation="horizontal">
+
+		<Button
+			android:id="@+id/cancelButton"
+			style="@style/BriarButtonFlat.Positive"
+			android:layout_width="wrap_content"
+			android:layout_height="wrap_content"
+			android:layout_weight="0.5"
+			android:text="@string/cancel"/>
+
+		<Button
+			android:id="@+id/openButton"
+			style="@style/BriarButtonFlat.Negative"
+			android:layout_width="wrap_content"
+			android:layout_height="wrap_content"
+			android:layout_weight="0.5"
+			android:text="@string/link_warning_open_link"/>
+
+	</LinearLayout>
+
+</LinearLayout>
\ No newline at end of file
diff --git a/briar-android/res/values/strings.xml b/briar-android/res/values/strings.xml
index e86654509..626f3d8a8 100644
--- a/briar-android/res/values/strings.xml
+++ b/briar-android/res/values/strings.xml
@@ -308,6 +308,12 @@
 	<string name="feedback_settings_title">Feedback</string>
 	<string name="send_feedback">Send feedback</string>
 
+	<!-- Link Warning -->
+	<string name="link_warning_title">Link Warning</string>
+	<string name="link_warning_intro">You are about to open the following link with an external app.</string>
+	<string name="link_warning_text">This can be used to identify you. Think about whether you trust the person that sent you this link and consider opening it with Orfox.</string>
+	<string name="link_warning_open_link">Open Link</string>
+
 	<!-- Multiple Identities -->
 	<string name="anonymous">Anonymous</string>
 	<string name="new_identity_title">New Identity</string>
diff --git a/briar-android/src/org/briarproject/android/blogs/BlogPostViewHolder.java b/briar-android/src/org/briarproject/android/blogs/BlogPostViewHolder.java
index 13a1b3f17..2fb84671e 100644
--- a/briar-android/src/org/briarproject/android/blogs/BlogPostViewHolder.java
+++ b/briar-android/src/org/briarproject/android/blogs/BlogPostViewHolder.java
@@ -30,6 +30,8 @@ import static org.briarproject.android.BriarActivity.GROUP_ID;
 import static org.briarproject.android.blogs.BasePostPagerFragment.POST_ID;
 import static org.briarproject.android.util.AndroidUtils.TEASER_LENGTH;
 import static org.briarproject.android.util.AndroidUtils.getTeaser;
+import static org.briarproject.android.util.AndroidUtils.getSpanned;
+import static org.briarproject.android.util.AndroidUtils.makeLinksClickable;
 import static org.briarproject.api.blogs.MessageType.POST;
 
 @UiThread
@@ -108,15 +110,17 @@ class BlogPostViewHolder extends RecyclerView.ViewHolder {
 		}
 
 		// post body
-		CharSequence bodyText = item.getBody();
 		if (listener == null) {
+			body.setText(getSpanned(item.getBody()));
+			makeLinksClickable(body);
 			body.setTextIsSelectable(true);
 		} else {
 			body.setTextIsSelectable(false);
 			if (item.getBody().length() > TEASER_LENGTH)
-				bodyText = getTeaser(ctx, item.getBody());
+				body.setText(getTeaser(ctx, item.getBody()));
+			else
+				body.setText(item.getBody());
 		}
-		body.setText(bodyText);
 
 		// reblog button
 		reblogButton.setOnClickListener(new OnClickListener() {
diff --git a/briar-android/src/org/briarproject/android/util/AndroidUtils.java b/briar-android/src/org/briarproject/android/util/AndroidUtils.java
index 66593d0f8..6a591f1dc 100644
--- a/briar-android/src/org/briarproject/android/util/AndroidUtils.java
+++ b/briar-android/src/org/briarproject/android/util/AndroidUtils.java
@@ -6,14 +6,24 @@ import android.content.Context;
 import android.os.Build;
 import android.provider.Settings;
 import android.support.design.widget.TextInputLayout;
+import android.support.v4.app.FragmentManager;
 import android.support.v4.content.ContextCompat;
+import android.support.v7.app.AppCompatActivity;
+import android.text.Html;
 import android.text.Spannable;
 import android.text.SpannableString;
 import android.text.SpannableStringBuilder;
+import android.text.Spanned;
 import android.text.format.DateUtils;
+import android.text.method.LinkMovementMethod;
+import android.text.style.ClickableSpan;
 import android.text.style.ForegroundColorSpan;
+import android.text.style.URLSpan;
+import android.view.View;
+import android.widget.TextView;
 
 import org.briarproject.R;
+import org.briarproject.android.widget.LinkDialogFragment;
 import org.briarproject.util.IoUtils;
 import org.briarproject.util.StringUtils;
 
@@ -142,4 +152,31 @@ public class AndroidUtils {
 		return builder;
 	}
 
+	public static Spanned getSpanned(String s) {
+		return Html.fromHtml(s);
+	}
+
+	public static void makeLinksClickable(TextView v) {
+		SpannableStringBuilder ssb = new SpannableStringBuilder(v.getText());
+		URLSpan[] spans = ssb.getSpans(0, ssb.length(), URLSpan.class);
+		for (URLSpan span : spans) {
+			int start = ssb.getSpanStart(span);
+			int end = ssb.getSpanEnd(span);
+			final String url = span.getURL();
+			ssb.removeSpan(span);
+			ClickableSpan cSpan = new ClickableSpan() {
+				@Override
+				public void onClick(View v2) {
+					LinkDialogFragment f = LinkDialogFragment.newInstance(url);
+					FragmentManager fm = ((AppCompatActivity) v2.getContext())
+							.getSupportFragmentManager();
+					f.show(fm, f.getUniqueTag());
+				}
+			};
+			ssb.setSpan(cSpan, start, end, 0);
+		}
+		v.setText(ssb);
+		v.setMovementMethod(LinkMovementMethod.getInstance());
+	}
+
 }
diff --git a/briar-android/src/org/briarproject/android/widget/LinkDialogFragment.java b/briar-android/src/org/briarproject/android/widget/LinkDialogFragment.java
new file mode 100644
index 000000000..1a6e636ee
--- /dev/null
+++ b/briar-android/src/org/briarproject/android/widget/LinkDialogFragment.java
@@ -0,0 +1,77 @@
+package org.briarproject.android.widget;
+
+
+import android.content.Intent;
+import android.net.Uri;
+import android.os.Bundle;
+import android.support.v4.app.DialogFragment;
+import android.view.LayoutInflater;
+import android.view.View;
+import android.view.View.OnClickListener;
+import android.view.ViewGroup;
+import android.widget.Button;
+import android.widget.TextView;
+
+import org.briarproject.R;
+
+
+public class LinkDialogFragment extends DialogFragment {
+
+	private static final String TAG = LinkDialogFragment.class.getName();
+
+	private String url;
+
+	public static LinkDialogFragment newInstance(String url) {
+		LinkDialogFragment f = new LinkDialogFragment();
+
+		Bundle args = new Bundle();
+		args.putString("url", url);
+		f.setArguments(args);
+
+		return f;
+	}
+
+	@Override
+	public void onCreate(Bundle savedInstanceState) {
+		super.onCreate(savedInstanceState);
+
+		url = getArguments().getString("url");
+
+		setStyle(STYLE_NO_TITLE, R.style.BriarDialogTheme);
+	}
+
+	@Override
+	public View onCreateView(LayoutInflater inflater, ViewGroup container,
+			Bundle savedInstanceState) {
+
+		View v = inflater.inflate(R.layout.fragment_link_dialog, container,
+				false);
+
+		TextView urlView = (TextView) v.findViewById(R.id.urlView);
+		urlView.setText(url);
+
+		Button openButton = (Button) v.findViewById(R.id.openButton);
+		openButton.setOnClickListener(new OnClickListener() {
+			@Override
+			public void onClick(View v) {
+				Intent i = new Intent(Intent.ACTION_VIEW, Uri.parse(url));
+				startActivity(i);
+			}
+		});
+
+		Button cancelButton = (Button) v.findViewById(R.id.cancelButton);
+		cancelButton.setOnClickListener(new OnClickListener() {
+			@Override
+			public void onClick(View v) {
+				getDialog().cancel();
+			}
+		});
+
+		return v;
+	}
+
+	public String getUniqueTag() {
+		return TAG;
+	}
+
+}
diff --git a/briar-core/build.gradle b/briar-core/build.gradle
index ae5071be7..9ffd0987f 100644
--- a/briar-core/build.gradle
+++ b/briar-core/build.gradle
@@ -13,6 +13,7 @@ dependencies {
 	compile 'org.jdom:jdom2:2.0.6'
 	compile 'org.slf4j:slf4j-api:1.7.21'
 	compile 'com.squareup.okhttp3:okhttp:3.3.1'
+	compile 'org.jsoup:jsoup:1.9.2'
 }
 
 dependencyVerification {
@@ -25,6 +26,7 @@ dependencyVerification {
 			'com.squareup.okhttp3:okhttp:a47f4efa166551cd5acc04f1071d82dafbf05638c21f9ca13068bc6633e3bff6',
 			'com.rometools:rome-utils:2be18a1edc601c31fe49c2000bb5484dd75182309270c2a2561d71888d81587a',
 			'com.squareup.okio:okio:5cfea5afe6c6e441a4dbf6053a07a733b1249d1009382eb44ac2255ccedd0c15',
+			'org.jsoup:jsoup:9c1885f1b182256e06f1e30b8451caed0c0dee96299d6348f968d18b54d0a46a',
 	]
 }
 
diff --git a/briar-core/src/org/briarproject/feed/FeedManagerImpl.java b/briar-core/src/org/briarproject/feed/FeedManagerImpl.java
index 6384f06c0..b9263c062 100644
--- a/briar-core/src/org/briarproject/feed/FeedManagerImpl.java
+++ b/briar-core/src/org/briarproject/feed/FeedManagerImpl.java
@@ -66,6 +66,9 @@ import static org.briarproject.api.feed.FeedConstants.FETCH_DELAY_INITIAL;
 import static org.briarproject.api.feed.FeedConstants.FETCH_INTERVAL;
 import static org.briarproject.api.feed.FeedConstants.FETCH_UNIT;
 import static org.briarproject.api.feed.FeedConstants.KEY_FEEDS;
+import static org.briarproject.util.HtmlUtils.article;
+import static org.briarproject.util.HtmlUtils.clean;
+import static org.briarproject.util.HtmlUtils.stripAll;
 
 class FeedManagerImpl implements FeedManager, Client, EventListener {
 
@@ -337,13 +340,13 @@ class FeedManagerImpl implements FeedManager, Client, EventListener {
 
 		SyndFeed f = getSyndFeed(getFeedInputStream(feed.getUrl()));
 		title = StringUtils.isNullOrEmpty(f.getTitle()) ? null : f.getTitle();
-		if (title != null) title = stripHTML(title);
+		if (title != null) title = clean(title, stripAll);
 		description = StringUtils.isNullOrEmpty(f.getDescription()) ? null :
 				f.getDescription();
-		if (description != null) description = stripHTML(description);
+		if (description != null) description = clean(description, stripAll);
 		author =
 				StringUtils.isNullOrEmpty(f.getAuthor()) ? null : f.getAuthor();
-		if (author != null) author = stripHTML(author);
+		if (author != null) author = clean(author, stripAll);
 
 		if (f.getEntries().size() == 0)
 			throw new FeedException("Feed has no entries");
@@ -418,23 +421,23 @@ class FeedManagerImpl implements FeedManager, Client, EventListener {
 		// build post body
 		StringBuilder b = new StringBuilder();
 		if (feed.getTitle() != null) {
-			// HTML in feed title was already stripped
-			b.append(feed.getTitle()).append("\n\n");
+			b.append("<h3>").append(feed.getTitle()).append("</h3>");
 		}
 		if (!StringUtils.isNullOrEmpty(entry.getTitle())) {
-			b.append(stripHTML(entry.getTitle())).append("\n\n");
+			b.append("<h1>").append(entry.getTitle()).append("</h1>");
 		}
 		for (SyndContent content : entry.getContents()) {
-			// extract content and do a very simple HTML tag stripping
 			if (content.getValue() != null)
-				b.append(stripHTML(content.getValue()));
+				b.append(content.getValue());
 		}
 		if (entry.getContents().size() == 0) {
-			if (entry.getDescription().getValue() != null)
-				b.append(stripHTML(entry.getDescription().getValue()));
+			if (entry.getDescription() != null &&
+					entry.getDescription().getValue() != null)
+				b.append(entry.getDescription().getValue());
 		}
+		b.append("<p>");
 		if (!StringUtils.isNullOrEmpty(entry.getAuthor())) {
-			b.append("\n\n-- ").append(stripHTML(entry.getAuthor()));
+			b.append("\n\n-- ").append(clean(entry.getAuthor(), stripAll));
 		}
 		if (entry.getPublishedDate() != null) {
 			b.append(" (").append(entry.getPublishedDate().toString())
@@ -443,8 +446,11 @@ class FeedManagerImpl implements FeedManager, Client, EventListener {
 			b.append(" (").append(entry.getUpdatedDate().toString())
 					.append(")");
 		}
-		if (!StringUtils.isNullOrEmpty(entry.getLink())) {
-			b.append("\n\n").append(stripHTML(entry.getLink()));
+		b.append("</p>");
+		String link = entry.getLink();
+		if (!StringUtils.isNullOrEmpty(link)) {
+			b.append("<a href=\"").append(link).append("\">").append(link)
+					.append("</a>");
 		}
 
 		// get other information for post
@@ -476,12 +482,8 @@ class FeedManagerImpl implements FeedManager, Client, EventListener {
 		}
 	}
 
-	private String stripHTML(String s) {
-		s = s.replaceAll("<script.*?>(?s).*?</script>", "");
-		return StringUtils.trim(s.replaceAll("<(?s).*?>", ""));
-	}
-
 	private String getPostBody(String text) {
+		text = clean(text, article);
 		if (text.length() <= MAX_BLOG_POST_BODY_LENGTH) return text;
 		else return text.substring(0, MAX_BLOG_POST_BODY_LENGTH);
 	}
diff --git a/briar-core/src/org/briarproject/util/HtmlUtils.java b/briar-core/src/org/briarproject/util/HtmlUtils.java
new file mode 100644
index 000000000..e9c338401
--- /dev/null
+++ b/briar-core/src/org/briarproject/util/HtmlUtils.java
@@ -0,0 +1,16 @@
+package org.briarproject.util;
+
+import org.jsoup.Jsoup;
+import org.jsoup.safety.Whitelist;
+
+public class HtmlUtils {
+
+	public static Whitelist stripAll = Whitelist.none();
+	public static Whitelist article =
+			Whitelist.basic().addTags("h1", "h2", "h3", "h4", "h5", "h6");
+
+	public static String clean(String s, Whitelist list) {
+		return Jsoup.clean(s, list);
+	}
+
+}
diff --git a/briar-core/src/org/briarproject/util/PrivacyUtils.java b/briar-core/src/org/briarproject/util/PrivacyUtils.java
index 247ccda3f..c2bf42628 100644
--- a/briar-core/src/org/briarproject/util/PrivacyUtils.java
+++ b/briar-core/src/org/briarproject/util/PrivacyUtils.java
@@ -1,6 +1,5 @@
 package org.briarproject.util;
 
-import java.net.Inet4Address;
 import java.net.Inet6Address;
 import java.net.InetAddress;
 import java.net.InetSocketAddress;