remove DOCTYPE declarations (#1260)

Athou · Jan 10, 2025 · 62d3ed1 · 62d3ed1
1 parent 74f7c48
commit 62d3ed1
Show file tree

Hide file tree

Showing 3 changed files with 26 additions and 0 deletions.
diff --git a/commafeed-server/src/main/java/com/commafeed/backend/feed/parser/FeedCleaner.java b/commafeed-server/src/main/java/com/commafeed/backend/feed/parser/FeedCleaner.java
@@ -1,6 +1,7 @@
 package com.commafeed.backend.feed.parser;
 
 import java.util.Collection;
+import java.util.regex.Pattern;
 
 import org.ahocorasick.trie.Emit;
 import org.ahocorasick.trie.Trie;
@@ -11,6 +12,8 @@
 @Singleton
 class FeedCleaner {
 
+	private static final Pattern DOCTYPE_PATTERN = Pattern.compile("<!DOCTYPE[^>]*>", Pattern.CASE_INSENSITIVE);
+
 	public String trimInvalidXmlCharacters(String xml) {
 		if (StringUtils.isBlank(xml)) {
 			return null;
@@ -60,4 +63,8 @@ public String replaceHtmlEntitiesWithNumericEntities(String source) {
 		return sb.toString();
 	}
 
+	public String removeDoctypeDeclarations(String xml) {
+		return DOCTYPE_PATTERN.matcher(xml).replaceAll("");
+	}
+
 }
diff --git a/commafeed-server/src/main/java/com/commafeed/backend/feed/parser/FeedParser.java b/commafeed-server/src/main/java/com/commafeed/backend/feed/parser/FeedParser.java
@@ -64,6 +64,7 @@ public FeedParserResult parse(String feedUrl, byte[] xml) throws FeedException {
 				throw new FeedException("Input string is null for url " + feedUrl);
 			}
 			xmlString = feedCleaner.replaceHtmlEntitiesWithNumericEntities(xmlString);
+			xmlString = feedCleaner.removeDoctypeDeclarations(xmlString);
 
 			InputSource source = new InputSource(new StringReader(xmlString));
 			SyndFeed feed = new SyndFeedInput().build(source);

diff --git a/commafeed-server/src/test/java/com/commafeed/backend/feed/parser/FeedCleanerTest.java b/commafeed-server/src/test/java/com/commafeed/backend/feed/parser/FeedCleanerTest.java
@@ -13,4 +13,22 @@ void testReplaceHtmlEntitiesWithNumericEntities() {
 		Assertions.assertEquals("<source>T&#180;l&#180;phone &#8242;</source>", feedCleaner.replaceHtmlEntitiesWithNumericEntities(source));
 	}
 
+	@Test
+	void testRemoveDoctype() {
+		String source = "<!DOCTYPE html><html><head></head><body></body></html>";
+		Assertions.assertEquals("<html><head></head><body></body></html>", feedCleaner.removeDoctypeDeclarations(source));
+	}
+
+	@Test
+	void testRemoveMultilineDoctype() {
+		String source = """
+				<!DOCTYPE
+					html
+				>
+				<html><head></head><body></body></html>""";
+		Assertions.assertEquals("""
+
+				<html><head></head><body></body></html>""", feedCleaner.removeDoctypeDeclarations(source));
+	}
+
 }