Cannot use non ASCII characters in the blog

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 3.2.0
• Fix Version/s: 4.1.0
• Component/s: None
• Labels: None
• Environment: Red Hat el3, Liferay enterprise 3.2.0
• Similar Issues: Show 5 results ; ;

  LEP-279	When passing a parameter containing a non-ascii character using the portlet:param tag the non-ascii characters are corrupted.
  LEP-516	Renaming og mail folders with non-ASCII characters fails
  LEP-341	Non-ASCII characters in folders names
  LEP-222	Non-ASCII characters in the documents names
  LEP-327	Non ASCII characters get corrupted in rich text format in mail portlet

When using non ASCII characters like the Norwegian characters Ã?Â?Ã?Â?Ã?Â? the rss feed generated from the Blog portlet does not work. The message says that invalid character(s) is found in the text.

[Brett Randall]

Hi Anders - do you have a stack-trace for this, or is an Exception not thrown?

[Anders Brynilsen]

Yes i got a stack trace:

09:45:50,019 INFO [STDOUT] com.liferay.util.ConverterException: http://localhost/blogs/xml?blogs_user_id=liferay.com.1&type=rss2.0 de.nava.informa.core.ParseException: ParseException: Error in building: Invalid byte 2 of 3-byte UTF-8 sequence.
09:45:50,022 INFO [STDOUT] at com.liferay.portlet.rss.util.RSSConverter.convert(RSSConverter.java:71)
09:45:50,022 INFO [STDOUT] at com.liferay.portal.util.WebCachePool.get(WebCachePool.java:70)
09:45:50,022 INFO [STDOUT] at com.liferay.portlet.rss.util.RSSUtil.getChannel(RSSUtil.java:42)
09:45:50,022 INFO [STDOUT] at org.apache.jsp.html.portlet.rss.view_jsp._jspService(view_jsp.java:647)
09:45:50,032 INFO [STDOUT] at org.apache.jasper.runtime.HttpJspBase.service(HttpJspBase.java:94)

This is the stack trace from the rss portlet when trying to show the feed.
I also have tried to show som enorwegian newspaper feeds who frequently uses non-ascii characters. The do not cause an exception.

[Anders Brynilsen]

Here is the text it chockes on: Test with norwegian characters æøå (from the blog view, the blog in itself got no problems showing the non-ascii characters)

[Brian Chan]

we use informa rss to process our feeds. I wonder if it's a bug on their part?

[Anders Brynilsen]

Do you guys at Liferay report it to them?

[Brian Chan]

We don't. Do you mind checking their bug list and seeing if this is an issue? We use 6.0 I believe.

[Anders Brynilsen]

It is not listed in their bug list.

[Rafael Martí­nez]

In my installation, using characters non-ASCII of the spanish language, this doesn't work.

In my case, the problem is the encoding of the generated XML, UTF-8 instead of ISO-8859-1. What function of Informa RSS set this encoding?

In any case, in BlogsEntryManagerImpl the encoding UTF-8 of the RSS___Exporter function is hardcode, but I don't know if this is the problem.

[Rafael Martí­nez]

I think that the problem is in RSS___Exporter of Informa RSS.

The method 'writer' of this class use:

XMLOutputter outputter = new XMLOutputter(Format.getPrettyFormat());

then, encoding=UTF-8 is set in the XML.

This is an inconvenience. This is a very critical issue for me, I will open a bug on Informa's issue tracker, and I will modify the code of RSS_Exporter in order to patch this problem.

[Anders Brynilsen]

Good work Rafael!

Can you submit the patch here when you have tested it?

[Rafael Martí­nez]

Here is the patch. (all changes marked in the code between //RMS-I and //RMS-F

1. I use JDM 1.0 (jdom.zip or download from www.jdom.org). Compile informa.jar with this version and update in /lib o LPortal.

2. I have modified 3 java classes of Informa (exporters.zip, informa.zip).

3. I have modified system.properties of LPortal adding:

1. 1. RMS-I
      ##
   2. XML Encoding for Blogs
      ##

#

1. The encoding that is set by InformaRSS in the RSS of blogs.
   #
   com.liferay.portlet.blogs.ejb.BlogsEntryManagerImpl.encoding=ISO-8859-1
   1. RMS-F

Set here the correct value, for spanish language, ISO-8859-1.

4. I have modified BlogsEntryManagerImpl.java (BlogsEntryManagerImpl.zip).

That's all

[Anders Brynilsen]

Thanks Rafael, i'll test it out next week.

[Brett Randall]

This issue appears to be resolved.