small fixups and compilation issues

src/rcldb/rcldb.cpp

@@ -1697,7 +1697,7 @@ static void addPrefix(vector<TermMatchEntry>& terms, const string& prefix)
 // If field is set, we return a list of appropriately prefixed terms (which 
 // are going to be used to build a Xapian query).
 bool Db::termMatch(MatchType typ, const string &lang,
-		   const string &root, 
+		   const string &_root,
 		   TermMatchResult& res,
 		   int max, 
 		   const string& field)
@@ -1714,15 +1714,14 @@ bool Db::termMatch(MatchType typ, const string &lang,
     if (!m_reason.empty())
         return false;
 
-    // Get rid of capitals and accents
+    string droot = _root;
-
-    string droot = root;
 
+    // If index is stripped, get rid of capitals and accents
 #ifndef RCL_INDEX_STRIPCHARS
     if (o_index_stripchars)
 #endif
-	if (!unacmaybefold(root, droot, "UTF-8", UNACOP_UNACFOLD)) {
+	if (!unacmaybefold(_root, droot, "UTF-8", UNACOP_UNACFOLD)) {
-	    LOGERR(("Db::termMatch: unac failed for [%s]\n", root.c_str()));
+	    LOGERR(("Db::termMatch: unac failed for [%s]\n", _root.c_str()));
 	    return false;
 	}
 
@@ -1742,7 +1741,7 @@ bool Db::termMatch(MatchType typ, const string &lang,
     res.prefix = prefix;
 
     if (typ == ET_STEM) {
-	if (!stemExpand(lang, root, res))
+	if (!stemExpand(lang, droot, res))
 	    return false;
 	for (vector<TermMatchEntry>::iterator it = res.entries.begin(); 
 	     it != res.entries.end(); it++) {
@@ -1759,8 +1758,7 @@ bool Db::termMatch(MatchType typ, const string &lang,
 	regex_t reg;
 	int errcode;
 	if (typ == ET_REGEXP) {
-	    string mroot = droot;
+	    if ((errcode = regcomp(&reg, droot.c_str(), 
-	    if ((errcode = regcomp(&reg, mroot.c_str(), 
 				   REG_EXTENDED|REG_NOSUB))) {
 		char errbuf[200];
 		regerror(errcode, &reg, errbuf, 199);

src/rcldb/searchdata.cpp

@@ -606,10 +606,13 @@ bool SearchDataClauseSimple::expandTerm(Rcl::Db &db,
 	return true;
     } 
 
+#ifndef RCL_INDEX_STRIPCHARS
     // The case/diac expansion db
     SynTermTransUnac unacfoldtrans(UNACOP_UNACFOLD);
     XapComputableSynFamMember synac(db.m_ndb->xrdb, synFamDiCa, "all", 
 				    &unacfoldtrans);
+#endif // RCL_INDEX_STRIPCHARS
+
     TermMatchResult res;
 
     if (haswild) {
@@ -969,10 +972,11 @@ static int stringToMods(string& s)
  *   count)
  */
 bool SearchDataClauseSimple::processUserString(Rcl::Db &db, const string &iq,
-					int mods, string &ermsg,
+					       string &ermsg, void *pq, 
-					void *pq, int slack, bool useNear)
+					       int slack, bool useNear)
 {
     vector<Xapian::Query> &pqueries(*(vector<Xapian::Query>*)pq);
+    int mods = m_modifiers;
 
     LOGDEB(("StringToXapianQ:pUS:: qstr [%s] fld [%s] mods 0x%x "
 	    "slack %d near %d\n", 
@@ -1094,7 +1098,7 @@ bool SearchDataClauseSimple::toNativeQuery(Rcl::Db &db, void *p)
     }
 
     vector<Xapian::Query> pqueries;
-    if (!processUserString(db, m_text, getModifiers(), m_reason, &pqueries))
+    if (!processUserString(db, m_text, m_reason, &pqueries))
 	return false;
     if (pqueries.empty()) {
 	LOGERR(("SearchDataClauseSimple: resolved to null query\n"));
@@ -1185,8 +1189,7 @@ bool SearchDataClauseDist::toNativeQuery(Rcl::Db &db, void *p)
     }
     string s = cstr_dquote + m_text + cstr_dquote;
     bool useNear = (m_tp == SCLT_NEAR);
-    if (!processUserString(db, s, getModifiers(), m_reason, &pqueries, 
+    if (!processUserString(db, s, m_reason, &pqueries, m_slack, useNear))
-			   m_slack, useNear))
 	return false;
     if (pqueries.empty()) {
 	LOGERR(("SearchDataClauseDist: resolved to null query\n"));

src/rcldb/searchdata.h

@@ -265,15 +265,9 @@ public:
     {
 	m_modifiers = mod;
     }
-    virtual int getModifiers() 
-    {
-	return m_modifiers;
-    }
     virtual void addModifier(Modifier mod) 
     {
-	int imod = getModifiers();
+	m_modifiers = Modifier(m_modifiers | mod);
-	imod |= mod;
-	setModifiers(Modifier(imod));
     }
     virtual void setWeight(float w) 
     {
@@ -316,7 +310,7 @@ class SearchDataClauseSimple : public SearchDataClause {
 public:
     SearchDataClauseSimple(SClType tp, const std::string& txt, 
 			   const std::string& fld = std::string())
-	: SearchDataClause(tp), m_text(txt), m_field(fld)
+	: SearchDataClause(tp), m_text(txt), m_field(fld), m_curcl(0)
     {
 	m_haveWildCards = 
 	    (txt.find_first_of(cstr_minwilds) != std::string::npos);
@@ -345,9 +339,10 @@ protected:
     std::string  m_text;  // Raw user entry text.
     std::string  m_field; // Field specification if any
     HighlightData m_hldata;
+    // Current count of Xapian clauses, to check against expansion limit
     int  m_curcl;
 
-    bool processUserString(Rcl::Db &db, const string &iq, int mods,  
+    bool processUserString(Rcl::Db &db, const string &iq,
 			   std::string &ermsg,
 			   void* pq, int slack = 0, bool useNear = false);
     bool expandTerm(Rcl::Db &db, std::string& ermsg, int mods, 

src/rcldb/stemdb.cpp

@@ -43,11 +43,18 @@ namespace Rcl {
 /**
  * Expand for one or several languages
  */
-bool StemDb::stemExpand(const std::string& langs, const std::string& term,
+bool StemDb::stemExpand(const std::string& langs, const std::string& _term,
 			vector<string>& result)
 {
     vector<string> llangs;
     stringToStrings(langs, llangs);
+    
+    // The stemdb keys may have kept their diacritics or not but they
+    // are always lower-case. It would be more logical for the term
+    // transformers to perform before doing the stemming, but this
+    // would be inefficient when there are several stemming languages
+    string term;
+    unacmaybefold(_term, term, "UTF-8", UNACOP_FOLD);
 
     for (vector<string>::const_iterator it = llangs.begin();
 	 it != llangs.end(); it++) {

src/rcldb/synfamily.h

@@ -210,6 +210,7 @@ private:
 // Lowercase accented stem to expansion. Family member name: language
 static const std::string synFamStem("Stm");
 
+#ifndef RCL_INDEX_STRIPCHARS
 // Lowercase unaccented stem to expansion. Family member name: language
 static const std::string synFamStemUnac("StU");
 
@@ -217,6 +218,7 @@ static const std::string synFamStemUnac("StU");
 // member, named "all". This set is used for separate case/diac
 // expansion by post-filtering the results of dual expansion.
 static const std::string synFamDiCa("DCa");
+#endif // !RCL_INDEX_STRIPCHARS
 
 } // end namespace Rcl