diff --git a/tika-parsers/src/main/java/org/apache/tika/parser/html/HtmlParser.java b/tika-parsers/src/main/java/org/apache/tika/parser/html/HtmlParser.java
index e0d7ac2..66100f1 100644
--- a/tika-parsers/src/main/java/org/apache/tika/parser/html/HtmlParser.java
+++ b/tika-parsers/src/main/java/org/apache/tika/parser/html/HtmlParser.java
@@ -45,10 +45,11 @@ public class HtmlParser implements Parser {
// Use the widest, most common charset as our default.
private static final String DEFAULT_CHARSET = "windows-1252";
private static final int META_TAG_BUFFER_SIZE = 4096;
- private static final Pattern HTTP_EQUIV_CHARSET_PATTERN = Pattern.compile(
- "(?is)");
+
private static final Pattern CONTENT_TYPE_PATTERN =
Pattern.compile("(?i);\\s*charset\\s*=\\s*(.*)");
@@ -67,12 +68,20 @@ public class HtmlParser implements Parser {
if (bufferSize != -1) {
String metaString = new String(buffer, 0, bufferSize);
- Matcher m = HTTP_EQUIV_CHARSET_PATTERN.matcher(metaString);
+ Matcher m = HTTP_EQUIV_PATTERN.matcher(metaString);
if (m.find()) {
- String charset = m.group(1);
- if (Charset.isSupported(charset)) {
- metadata.set(Metadata.CONTENT_ENCODING, charset);
- return charset;
+ // TIKA-349: flexible handling of attributes
+ // We have one or more x or x=y attributes, separated by ';'
+ String[] attrs = m.group(1).split(";");
+ for (String attr : attrs) {
+ String[] keyValue = attr.trim().split("=");
+ if ((keyValue.length == 2) && keyValue[0].equalsIgnoreCase("charset")) {
+ String charset = keyValue[1];
+ if (Charset.isSupported(charset)) {
+ metadata.set(Metadata.CONTENT_ENCODING, charset);
+ return charset;
+ }
+ }
}
}
}
diff --git a/tika-parsers/src/test/java/org/apache/tika/parser/html/HtmlParserTest.java b/tika-parsers/src/test/java/org/apache/tika/parser/html/HtmlParserTest.java
index fcdc0bc..da90393 100644
--- a/tika-parsers/src/test/java/org/apache/tika/parser/html/HtmlParserTest.java
+++ b/tika-parsers/src/test/java/org/apache/tika/parser/html/HtmlParserTest.java
@@ -267,7 +267,7 @@ public class HtmlParserTest extends TestCase {
/**
* Test case for TIKA-341
- * @see TIKA-XXX
+ * @see TIKA-341
*/
public void testUsingCharsetInContentTypeHeader() throws Exception {
final String test =
@@ -307,5 +307,34 @@ public class HtmlParserTest extends TestCase {
assertEquals("baz", parts[2]);
}
+ /**
+ * Test case for TIKA-349
+ * @see TIKA-349
+ */
+ public void testHttpEquivCharsetFunkyAttributes() throws Exception {
+ String test1 =
+ ""
+ + "the name is \u00e1ndre"
+ + "";
+ Metadata metadata = new Metadata();
+ new HtmlParser().parse (
+ new ByteArrayInputStream(test1.getBytes("UTF-8")),
+ new BodyContentHandler(), metadata, new ParseContext());
+ assertEquals("ISO-8859-1", metadata.get(Metadata.CONTENT_ENCODING));
+
+ // Some HTML pages have errors like ';;' versus '; ' as separator
+ String test2 =
+ ""
+ + "the name is \u00e1ndre"
+ + "";
+ metadata = new Metadata();
+ new HtmlParser().parse (
+ new ByteArrayInputStream(test2.getBytes("UTF-8")),
+ new BodyContentHandler(), metadata, new ParseContext());
+ assertEquals("ISO-8859-1", metadata.get(Metadata.CONTENT_ENCODING));
+ }
+
}