[NUTCH-91] empty encoding causes exception

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 0.8
• Fix Version/s: 0.7.2, 0.8
• Assignee: Unassigned
• Reporter: Michael Nebel
• Created: 10/Sep/05 21:35
• Updated: 10/Mar/06 05:17
• Resolved: 10/Mar/06 05:17

Description

I found some sites, where the header says: "Content-Type: text/html; charset=". This causes an exception in the HtmlParser. My suggestion:

Index: src/plugin/parse-html/src/java/org/apache/nutch/parse/html/HtmlParser.java
===================================================================
— src/plugin/parse-html/src/java/org/apache/nutch/parse/html/HtmlParser.java (revision 279397)
+++ src/plugin/parse-html/src/java/org/apache/nutch/parse/html/HtmlParser.java (working copy)
@@ -120,7 +120,7 @@
byte[] contentInOctets = content.getContent();
InputSource input = new InputSource(new ByteArrayInputStream(contentInOctets));
String encoding = StringUtil.parseCharacterEncoding(contentType);

• if (encoding!=null) {
  + if (encoding!=null && !"".equals(encoding)) {
  metadata.put("OriginalCharEncoding", encoding);
  if ((encoding = StringUtil.resolveEncodingAlias(encoding)) != null) {
  metadata.put("CharEncodingForConversion", encoding);