From dd42a9f51e377580d012a12f30192d927a01c937 Mon Sep 17 00:00:00 2001
From: Jean-Francois Dockes <jfd@recoll.org>
Date: Mon, 1 Apr 2013 16:01:21 +0200
Subject: [PATCH] allow wild chars in dir clauses

---
 src/rcldb/rcldb.cpp      |  4 ++++
 src/rcldb/searchdata.cpp | 35 +++++++++++++++++++++++++++++------
 src/rcldb/searchdata.h   | 10 ++--------
 src/sampleconf/fields    | 26 ++++++++++++--------------
 4 files changed, 47 insertions(+), 28 deletions(-)

diff --git a/src/rcldb/rcldb.cpp b/src/rcldb/rcldb.cpp
index 2d5804b5..b97d3ce7 100644
--- a/src/rcldb/rcldb.cpp
+++ b/src/rcldb/rcldb.cpp
@@ -917,6 +917,10 @@ bool Db::addOrUpdate(const string &udi, const string &parent_udi, Doc &doc)
 	string path = url_gpath(doc.url);
 	vector<string> vpath;
 	stringToTokens(path, vpath, "/");
+	// If vpath is not /, the last elt is the file/dir name, not a
+	// part of the path.
+	if (vpath.size())
+	    vpath.resize(vpath.size()-1);
 	splitter.curpos = 0;
 	newdocument.add_posting(wrap_prefix(pathelt_prefix),
 				splitter.basepos + splitter.curpos++);
diff --git a/src/rcldb/searchdata.cpp b/src/rcldb/searchdata.cpp
index 7581fb19..18c084b4 100644
--- a/src/rcldb/searchdata.cpp
+++ b/src/rcldb/searchdata.cpp
@@ -517,7 +517,7 @@ static void listVector(const string& what, const vector<string>&l)
     for (vector<string>::const_iterator it = l.begin(); it != l.end(); it++) {
         a = a + *it + " ";
     }
-    LOGDEB(("%s: %s\n", what.c_str(), a.c_str()));
+    LOGDEB0(("%s: %s\n", what.c_str(), a.c_str()));
 }
 #endif
 
@@ -1056,19 +1056,42 @@ bool SearchDataClausePath::toNativeQuery(Rcl::Db &db, void *p)
 
     if (m_text.empty()) {
 	LOGERR(("SearchDataClausePath: empty path??\n"));
+	m_reason = "Empty path ?";
 	return false;
     }
+
+    vector<Xapian::Query> orqueries;
+
+    if (m_text[0] == '/')
+	orqueries.push_back(Xapian::Query(wrap_prefix(pathelt_prefix)));
+
     vector<string> vpath;
     stringToTokens(m_text, vpath, "/");
-    vector<string> pvpath;
-    if (m_text[0] == '/')
-	pvpath.push_back(wrap_prefix(pathelt_prefix));
+
     for (vector<string>::const_iterator pit = vpath.begin(); 
 	 pit != vpath.end(); pit++){
-	pvpath.push_back(wrap_prefix(pathelt_prefix) + *pit);
+
+	string sterm;
+	vector<string> exp;
+	if (!expandTerm(db, m_reason, 
+			SDCM_NOSTEMMING|SDCM_CASESENS|SDCM_DIACSENS,
+			*pit, exp, sterm, wrap_prefix(pathelt_prefix))) {
+	    return false;
+	}
+	LOGDEB0(("SDataPath::toNative: exp size %d\n", exp.size()));
+	listVector("", exp);
+	if (exp.size() == 1)
+	    orqueries.push_back(Xapian::Query(exp[0]));
+	else 
+	    orqueries.push_back(Xapian::Query(Xapian::Query::OP_OR, 
+					      exp.begin(), exp.end()));
+	m_curcl += exp.size();
+	if (m_curcl >= getMaxCl())
+	    return false;
     }
+
     *qp = Xapian::Query(Xapian::Query::OP_PHRASE, 
-			pvpath.begin(), pvpath.end());
+			orqueries.begin(), orqueries.end());
 
     if (m_weight != 1.0) {
 	*qp = Xapian::Query(Xapian::Query::OP_SCALE_WEIGHT, *qp, m_weight);
diff --git a/src/rcldb/searchdata.h b/src/rcldb/searchdata.h
index 24fd38a3..4c10f3d0 100644
--- a/src/rcldb/searchdata.h
+++ b/src/rcldb/searchdata.h
@@ -409,10 +409,10 @@ protected:
  * change but is no important enough to warrant it, this has to wait for
  * the next format change.
  */
-class SearchDataClausePath : public SearchDataClause {
+class SearchDataClausePath : public SearchDataClauseSimple {
 public:
     SearchDataClausePath(const std::string& txt, bool excl = false)
-	: SearchDataClause(SCLT_PATH), m_text(txt)
+	: SearchDataClauseSimple(SCLT_PATH, txt, "dir")
     {
 	m_exclude = excl;
 	m_haveWildCards = false;
@@ -427,13 +427,7 @@ public:
     }
 
     virtual bool toNativeQuery(Rcl::Db &, void *);
-    virtual const std::string& gettext() const
-    {
-	return m_text;
-    }
 
-protected:
-    std::string m_text;
 };
 
 /** 
diff --git a/src/sampleconf/fields b/src/sampleconf/fields
index 8b154fa0..2d013b87 100644
--- a/src/sampleconf/fields
+++ b/src/sampleconf/fields
@@ -33,20 +33,18 @@
 # The following ones are probably hard-coded in the c code, can't change at
 # all.
 # Also reserved: F(parentid), Q(uniqueid)
-title = S ; wdfinc = 10
 author = A
-abstract = XS
-caption = S
-title = S
-subject = S
-keywords= K
-xapyear = Y
-xapyearmon = M
 xapdate = D
+keywords= K
+xapyearmon = M
+title = S ; wdfinc = 10
 mtype = T
-rclUnsplitFN = XSFS
-filename = XSFN
 ext = XE
+dir = XP
+abstract = XS
+filename = XSFN
+rclUnsplitFN = XSFS
+xapyear = Y
 
 # Extension examples. These are actually used by default by Recoll, you can
 # add your own to search for fields produced by the filters and not handled
@@ -72,13 +70,13 @@ recipient = XTO
 # "rclaptg" is used for viewer specialization (depending on local config)
 # "rclbes" defines the backend type (ie normal fs, firefox cache). Should
 #   probably be hardcoded, don't remove it
+abstract=
 author=
-recipient=
-rclaptg=
-rclbes=
 filename=
 keywords=
-abstract=
+rclaptg=
+rclbes=
+recipient=
 
 [aliases]
 ##########################