From c7d9d524f9761230be40e1c86469854870eda1bd Mon Sep 17 00:00:00 2001
From: Jean-Francois Dockes <jfd@recoll.org>
Date: Tue, 16 Apr 2013 15:56:49 +0200
Subject: [PATCH] add terms for doc md5s

---
 src/rcldb/rcldb.cpp   | 12 +++++++-----
 src/sampleconf/fields |  1 +
 2 files changed, 8 insertions(+), 5 deletions(-)

diff --git a/src/rcldb/rcldb.cpp b/src/rcldb/rcldb.cpp
index 37013a58..a4a54b4b 100644
--- a/src/rcldb/rcldb.cpp
+++ b/src/rcldb/rcldb.cpp
@@ -1129,13 +1129,15 @@ bool Db::addOrUpdate(const string &udi, const string &parent_udi, Doc &doc)
 	RECORD_APPEND(record, string(cstr_mbreaks), multibreaks.str());
     }
     
-    // If the file's md5 was computed, add value. This is optionally
-    // used for query result duplicate elimination.
-    string& md5 = doc.meta[Doc::keymd5];
-    if (!md5.empty()) {
+    // If the file's md5 was computed, add value and term. 
+    // The value is optionally used for query result duplicate elimination, 
+    // and the term to find the duplicates.
+    const string *md5;
+    if (doc.peekmeta(Doc::keymd5, &md5) && !md5->empty()) {
 	string digest;
-	MD5HexScan(md5, digest);
+	MD5HexScan(*md5, digest);
 	newdocument.add_value(VALUE_MD5, digest);
+	newdocument.add_boolean_term(wrap_prefix("XM") + *md5);
     }
 
     LOGDEB0(("Rcl::Db::add: new doc record:\n%s\n", record.c_str()));
diff --git a/src/sampleconf/fields b/src/sampleconf/fields
index 2d013b87..0ec4d846 100644
--- a/src/sampleconf/fields
+++ b/src/sampleconf/fields
@@ -40,6 +40,7 @@ xapyearmon = M
 title = S ; wdfinc = 10
 mtype = T
 ext = XE
+rclmd5 = XM
 dir = XP
 abstract = XS
 filename = XSFN