add comment explaining why we need to do serialize/deserialize

1 files changed, 3 insertions, 0 deletions

diff --git a/docprocs/src/main/java/com/yahoo/docprocs/indexing/IndexingProcessor.java b/docprocs/src/main/java/com/yahoo/docprocs/indexing/IndexingProcessor.java
index 2561fdc7dc5..a55159cdaa0 100644
--- a/docprocs/src/main/java/com/yahoo/docprocs/indexing/IndexingProcessor.java
+++ b/docprocs/src/main/java/com/yahoo/docprocs/indexing/IndexingProcessor.java
@@ -107,6 +107,9 @@ public class IndexingProcessor extends DocumentProcessor {
         DocumentType wantType = docTypeMgr.getDocumentType(hadType.getName());
         Document prevDoc = prev.getDocument();
         if (hadType != wantType) {
+            // this happens when you have a concrete document; we need to
+            // convert back to a "normal" Document for indexing of complex structures
+            // to work properly.
             GrowableByteBuffer buffer = new GrowableByteBuffer(64 * 1024, 2.0f);
             DocumentSerializer serializer = DocumentSerializerFactory.createHead(buffer);
             serializer.write(prevDoc);