Handle META tag http-equiv charset?

[asfimport]

Sebb (Bug 43678):
The HTML tag:
<META http-equiv="Content-Type" content="text/html; charset=utf-8">

is treated by browsers the same as the Content-Type header.

At present (JMeter 2.3) JMeter does not extract this information, it only uses
headers.

If no charset is present, then the platform default encoding (overridable by
defining sampleresult.default.encoding) is applied when converting the
response to a string.

This means that the conversion of the responseData to a string may not produce
the correct result - for example in the Tree View Listener, and presumably in
Post-Processors.

JMeter could extract the content-type from the data.
It probably needs to check that the data is intended to be used as HTML first -
i.e. the existing Content-Type should specify text/html.

Votes in Bugzilla: 1
Severity: normal
OS: other

[asfimport]

Sebb (migrated from Bugzilla):
Fixed in SVN r587702

[asfimport]

Sebb (migrated from Bugzilla):
On further reflection, perhaps this could be improved:

rather than only processing the meta content-type if the header content-type
did not include a charset/dataEncoding, maybe the meta tag should always be
processed, and used to update the sampler fields? The original headers would
still be available from getResponseHeaders.

[asfimport]

benoit.wiart (migrated from Bugzilla):
the fix in HTTPSampleResult does not work : in the method getDataEncodingWithDefault the test :
getDataEncodingNoDefault() == null
is always false because of the setDataEncoding(DEFAULT_HTTP_ENCODING); in the constructor

please note that there is no workaround to this bug because DEFAULT_HTTP_ENCODING can not be overriden with sampleresult.default.encoding

[asfimport]

benoit.wiart (migrated from Bugzilla):
line 42 in HTTPSampleResult :

setDataEncoding(DEFAULT_HTTP_ENCODING); // default if encoding not provided be the page

should probably be removed

[asfimport]

@pmouawad (migrated from Bugzilla):
Hello,
Attached is a patch that applies Benoit Proposition.

If encoding is set through header, encoding declared in meta will not be taken into account.
Otherwise it will.

Philippe
http://www.ubik-ingenierie.com

Created attachment patch-HTTPSampleResult.patch: Path that removes encoding setting in constructor

patch-HTTPSampleResult.patch

Index: src/protocol/http/org/apache/jmeter/protocol/http/sampler/HTTPSampleResult.java
===================================================================
--- src/protocol/http/org/apache/jmeter/protocol/http/sampler/HTTPSampleResult.java	(revision 816394)
+++ src/protocol/http/org/apache/jmeter/protocol/http/sampler/HTTPSampleResult.java	(working copy)
@@ -39,7 +39,7 @@
 
     public HTTPSampleResult() {
         super();
-        setDataEncoding(DEFAULT_HTTP_ENCODING); // default if encoding not provided be the page
+//        setDataEncoding(DEFAULT_HTTP_ENCODING); // default if encoding not provided be the page
     }
 
     public HTTPSampleResult(long elapsed) {

[asfimport]

Sebb (migrated from Bugzilla):
Thanks for the patch.

Applied with further fix so ensure correct default was used.

URL: http://svn.apache.org/viewvc?rev=929364&view=rev
Log:
#2038 - Handle META tag http-equiv charset?

Modified:
jakarta/jmeter/trunk/src/core/org/apache/jmeter/samplers/SampleResult.java
jakarta/jmeter/trunk/src/protocol/http/org/apache/jmeter/protocol/http/sampler/HTTPSampleResult.java
jakarta/jmeter/trunk/xdocs/changes.xml