const string cleanup

src/rcldb/rcldb.cpp

@@ -94,7 +94,6 @@ static const string cstr_syntAbs("?!#@");
 // Only ONE field name inside the index data record differs from the
 // Rcl::Doc ones: caption<->title, for a remnant of compatibility with
 // omega
-static const string cstr_keycap("caption");
 
 // Static/Default table for field->prefix/weight translation. 
 // This is logically const after initialization. Can't use a
@@ -125,7 +124,7 @@ static void initFldToTraits()
     fldToTraits["ext"] = FieldTraits("XE");
     fldToTraits[Doc::keyfn] = FieldTraits("XSFN");
 
-    fldToTraits[cstr_keycap] = FieldTraits("S");
+    fldToTraits[cstr_caption] = FieldTraits("S");
     fldToTraits[Doc::keytt] = FieldTraits("S");
     fldToTraits["subject"] = FieldTraits("S");
 
@@ -198,7 +197,7 @@ bool Db::Native::dbDataToRclDoc(Xapian::docid docid, std::string &data,
     parms.get(Doc::keyfmt, doc.fmtime);
     parms.get(Doc::keydmt, doc.dmtime);
     parms.get(Doc::keyoc, doc.origcharset);
-    parms.get(cstr_keycap, doc.meta[Doc::keytt]);
+    parms.get(cstr_caption, doc.meta[Doc::keytt]);
     parms.get(Doc::keykw, doc.meta[Doc::keykw]);
     parms.get(Doc::keyabs, doc.meta[Doc::keyabs]);
     // Possibly remove synthetic abstract indicator (if it's there, we
@@ -1179,7 +1178,7 @@ bool Db::addOrUpdate(const string &udi, const string &parent_udi,
     doc.meta[Doc::keytt] = 
 	neutchars(truncate_to_word(doc.meta[Doc::keytt], 150), cstr_nc);
     if (!doc.meta[Doc::keytt].empty())
-	RECORD_APPEND(record, cstr_keycap, doc.meta[Doc::keytt]);
+	RECORD_APPEND(record, cstr_caption, doc.meta[Doc::keytt]);
 
     trimstring(doc.meta[Doc::keykw], " \t\r\n");
     doc.meta[Doc::keykw] = 
@@ -1526,7 +1525,7 @@ bool Db::filenameWildExp(const string& fnexp, list<string>& names)
     // each end: match any substring
     if (pattern[0] == '"' && pattern[pattern.size()-1] == '"') {
 	pattern = pattern.substr(1, pattern.size() -2);
-    } else if (pattern.find_first_of("*?[") == string::npos && 
+    } else if (pattern.find_first_of(cstr_minwilds) == string::npos && 
 	       !unaciscapital(pattern)) {
 	pattern = "*" + pattern + "*";
     } // else let it be

src/rcldb/rcldb.h

@@ -21,6 +21,7 @@
 #include <list>
 #include <vector>
 
+#include "cstr.h"
 #include "refcntr.h"
 #include "rcldoc.h"
 #include "stoplist.h"
@@ -186,7 +187,7 @@ class Db {
     enum MatchType {ET_WILD, ET_REGEXP, ET_STEM};
     bool termMatch(MatchType typ, const string &lang, const string &s, 
 		   TermMatchResult& result, int max = -1, 
-		   const string& field = "",
+		   const string& field = cstr_null,
                    string *prefix = 0
         );
     /** Return min and max years for doc mod times in db */

src/rcldb/rclquery.cpp

@@ -24,6 +24,7 @@
 
 #include "xapian.h"
 
+#include "cstr.h"
 #include "rcldb.h"
 #include "rcldb_p.h"
 #include "rclquery.h"
@@ -39,17 +40,15 @@
 namespace Rcl {
 #endif
 
-static const string cstr_keycap("caption");
-static const string cstr_keydmtime("dmtime");
 
 // Field names inside the index data record may differ from the rcldoc ones
 // (esp.: caption / title)
 static const string& docfToDatf(const string& df)
 {
     if (!df.compare(Doc::keytt)) {
-	return cstr_keycap;
+	return cstr_caption;
     } else if (!df.compare(Doc::keymt)) {
-	return cstr_keydmtime;
+	return cstr_dmtime;
     } else {
 	return df;
     }

src/rcldb/searchdata.cpp

@@ -25,6 +25,7 @@
 
 #include "xapian.h"
 
+#include "cstr.h"
 #include "rcldb.h"
 #include "searchdata.h"
 #include "debuglog.h"
@@ -346,8 +347,10 @@ bool SearchData::maybeAddAutoPhrase()
 	words +=  clp->gettext();
     }
 
-    if (words.find_first_of("\"*[]?") != string::npos && // has wildcards
-	TextSplit::countWords(words) <= 1) { // Just one word.
+    // If there are wildcards or quotes in there, or this is a single word,
+    // bail out
+    if (words.find_first_of("\"*[?") != string::npos &&
+	TextSplit::countWords(words) <= 1) { 
 	LOGDEB2(("SearchData::maybeAddAutoPhrase: wildcards or single word\n"));
 	return false;
     }
@@ -561,7 +564,7 @@ void StringToXapianQ::expandTerm(bool nostemexp,
 	return;
     }
 
-    bool haswild = term.find_first_of("*?[") != string::npos;
+    bool haswild = term.find_first_of(cstr_minwilds) != string::npos;
 
     // No stemming if there are wildcards or prevented globally.
     if (haswild || m_stemlang.empty())
@@ -857,8 +860,6 @@ bool StringToXapianQ::processUserString(const string &iq,
     return true;
 }
 
-static const string cstr_null;
-
 // Translate a simple OR, AND, or EXCL search clause. 
 bool SearchDataClauseSimple::toNativeQuery(Rcl::Db &db, void *p, 
 					   const string& stemlang)
@@ -964,10 +965,10 @@ bool SearchDataClauseDist::toNativeQuery(Rcl::Db &db, void *p,
     // stringToXapianQueries() to lowercase and simplify the phrase
     // terms etc. This will result into a single (complex)
     // Xapian::Query.
-    if (m_text.find_first_of("\"") != string::npos) {
+    if (m_text.find('\"') != string::npos) {
 	m_text = neutchars(m_text, "\"");
     }
-    string s = string("\"") + m_text + string("\"");
+    string s = cstr_dquote + m_text + cstr_dquote;
     bool useNear = (m_tp == SCLT_NEAR);
     StringToXapianQ tr(db, m_field, l_stemlang, doBoostUserTerm);
     if (!tr.processUserString(s, m_reason, pqueries, db.getStopList(),

src/rcldb/searchdata.h

@@ -30,6 +30,7 @@
 #include "rcldb.h"
 #include "refcntr.h"
 #include "smallut.h"
+#include "cstr.h"
 
 class RclConfig;
 
@@ -215,7 +216,7 @@ public:
     SearchDataClauseSimple(SClType tp, const string& txt, 
 			   const string& fld = string())
 	: SearchDataClause(tp), m_text(txt), m_field(fld), m_slack(0) {
-	m_haveWildCards = (txt.find_first_of("*?[") != string::npos);
+	m_haveWildCards = (txt.find_first_of(cstr_minwilds) != string::npos);
     }
 
     virtual ~SearchDataClauseSimple() {}

src/rcldb/stemdb.cpp

@@ -291,13 +291,14 @@ static bool stemExpandOne(const std::string& dbdir,
 		// Build expansion list from database data No need for
 		// a conftree, but we need to massage the data a
 		// little
-		string::size_type pos = data.find_first_of("=");
-		++pos;
-		string::size_type pos1 = data.find_last_of("\n");
+		string::size_type pos = data.find('=');
+		string::size_type pos1 = data.rfind('\n');
 		if (pos == string::npos || pos1 == string::npos || 
-		    pos1 <= pos) {
-		    // ??
+		    pos1 <= pos+1) {
+		    LOGERR(("stemExpand: bad data in db: [%s]\n", 
+			    data.c_str()));
 		} else {
+		    ++pos;
 		    stringToStrings(data.substr(pos, pos1-pos), result);
 		}
 	    }