JCR-2873: Add a way to locate full text extraction problems

Use a special "TextExtractionError" token to make these cases searchable.
Also reduce the log level extraction failures to avoid confusing users.

git-svn-id: https://svn.apache.org/repos/asf/jackrabbit/trunk@1085038 13f79535-47bb-0310-9956-ffa450edef68

jackrabbit-core/src/main/java/org/apache/jackrabbit/core/query/lucene/LazyTextExtractorField.java

@@ -178,7 +178,11 @@ public void run() {
                 }
             } catch (Throwable t) {
                 if (t != STOP) {
-                    log.warn("Failed to extract text from a binary property", t);
+                    log.debug("Failed to extract text from a binary property."
+                            + " This is a fairly common case, and nothing to"
+                            + " worry about. The stack trace is included to"
+                            + " help improve the text extraction feature.", t);
+                    builder.replace(0, builder.length(), "TextExtractionError");
                 }
             } finally {
                 value.discard();