[SOLR-8366] ConcurrentUpdateSolrClient attempts to use response's content type as charset encoding for parsing exceptions

• Type: Bug
• Status: Closed
• Priority: Minor
• Resolution: Fixed
• Affects Version/s: 5.3, 5.4
• Fix Version/s: 5.5, 6.0
• Component/s: clients - java
• Assignee: Shalin Shekhar Mangar
• Reporter: Shalin Shekhar Mangar
• Created: 03/Dec/15 21:28
• Updated: 09/May/16 18:56
• Resolved: 04/Dec/15 15:18

Description

While debugging the SolrExampleStreamingTest.testUpdateField failures on trunk, I noticed that ConcurrentUpdateSolrClient always logs the following when the server throws a conflict error:

WARN  (concurrentUpdateScheduler-2-thread-1-processing-http:////127.0.0.1:35848//solr//collection1) [    ] o.a.s.c.s.i.ConcurrentUpdateSolrClient Failed to parse error response from http://127.0.0.1:35848/solr/collection1 due to: org.apache.solr.common.SolrException: parsing error

The problem is the following code which uses the response.getEntity().getContentType().getValue() as the charset encoding which is wrong because content type has mime type as well as charset.

try {
              NamedList<Object> resp =
                  client.parser.processResponse(response.getEntity().getContent(),
                      response.getEntity().getContentType().getValue());
              NamedList<Object> error = (NamedList<Object>) resp.get("error");
              if (error != null)
                solrExc.setMetadata((NamedList<String>) error.get("metadata"));
            } catch (Exception exc) {
              // don't want to fail to report error if parsing the response fails
              log.warn("Failed to parse error response from " + client.getBaseURL() + " due to: " + exc);
            }