minimal experimental stopword functionality

src/common/rclconfig.cpp

@@ -1,5 +1,5 @@
 #ifndef lint
-static char rcsid[] = "@(#$Id: rclconfig.cpp,v 1.43 2007-02-07 17:17:11 dockes Exp $ (C) 2004 J.F.Dockes";
+static char rcsid[] = "@(#$Id: rclconfig.cpp,v 1.44 2007-06-02 08:30:41 dockes Exp $ (C) 2004 J.F.Dockes";
 #endif
 /*
  *   This program is free software; you can redistribute it and/or modify
@@ -461,6 +461,11 @@ string RclConfig::getDbDir()
     return path_canon(dbdir);
 }
 
+string RclConfig::getStopfile()
+{
+    return path_cat(getConfDir(), "stoplist.txt");
+}
+
 list<string> RclConfig::getSkippedNames()
 {
     list<string> skpl;

src/common/rclconfig.h

@@ -16,7 +16,7 @@
  */
 #ifndef _RCLCONFIG_H_INCLUDED_
 #define _RCLCONFIG_H_INCLUDED_
-/* @(#$Id: rclconfig.h,v 1.31 2007-02-02 10:12:58 dockes Exp $  (C) 2004 J.F.Dockes */
+/* @(#$Id: rclconfig.h,v 1.32 2007-06-02 08:30:41 dockes Exp $  (C) 2004 J.F.Dockes */
 
 #include <list>
 #include <string>
@@ -78,8 +78,10 @@ class RclConfig {
 
     /** Get database directory */
     string getDbDir();
+    /** Get stoplist file name */
+    string getStopfile();
 
-    /** Get list of skipped names for current keydir */
+    /** Get list of skipped file names for current keydir */
     list<string> getSkippedNames();
 
     /** Get list of skipped paths patterns. Doesn't depend on the keydir */

src/index/indexer.cpp

@@ -1,5 +1,5 @@
 #ifndef lint
-static char rcsid[] = "@(#$Id: indexer.cpp,v 1.56 2007-05-30 12:31:19 dockes Exp $ (C) 2004 J.F.Dockes";
+static char rcsid[] = "@(#$Id: indexer.cpp,v 1.57 2007-06-02 08:30:41 dockes Exp $ (C) 2004 J.F.Dockes";
 #endif
 /*
  *   This program is free software; you can redistribute it and/or modify
@@ -181,7 +181,7 @@ bool DbIndexer::init(bool resetbefore, bool rdonly)
     }
     Rcl::Db::OpenMode mode = rdonly ? Rcl::Db::DbRO :
 	resetbefore ? Rcl::Db::DbTrunc : Rcl::Db::DbUpd;
-    if (!m_db.open(m_dbdir, mode)) {
+    if (!m_db.open(m_dbdir, m_config->getStopfile(), mode)) {
 	LOGERR(("DbIndexer: error opening database in %s\n", m_dbdir.c_str()));
 	return false;
     }

src/lib/Makefile

@@ -8,8 +8,8 @@ LIBS = librcl.a
 
 all: $(LIBS)
 
-OBJS =  rclaspell.o rclconfig.o rclinit.o textsplit.o unacpp.o csguess.o indexer.o mimetype.o htmlparse.o myhtmlparse.o mimehandler.o internfile.o mh_exec.o mh_html.o mh_mail.o mh_mbox.o mh_text.o docseq.o docseqdb.o docseqhist.o history.o sortseq.o wasastringtoquery.o wasatorcl.o pathhash.o rcldb.o searchdata.o stemdb.o base64.o conftree.o copyfile.o debuglog.o execmd.o fstreewalk.o idfile.o md5.o mimeparse.o pathut.o readfile.o smallut.o transcode.o wipedir.o x11mon.o
+OBJS =  rclaspell.o rclconfig.o rclinit.o textsplit.o unacpp.o csguess.o indexer.o mimetype.o htmlparse.o myhtmlparse.o mimehandler.o internfile.o mh_exec.o mh_html.o mh_mail.o mh_mbox.o mh_text.o docseq.o docseqdb.o docseqhist.o history.o sortseq.o wasastringtoquery.o wasatorcl.o pathhash.o rcldb.o searchdata.o stemdb.o stoplist.o base64.o conftree.o copyfile.o debuglog.o execmd.o fstreewalk.o idfile.o md5.o mimeparse.o pathut.o readfile.o smallut.o transcode.o wipedir.o x11mon.o
-DEPS =  rclaspell.dep.stamp rclconfig.dep.stamp rclinit.dep.stamp textsplit.dep.stamp unacpp.dep.stamp csguess.dep.stamp indexer.dep.stamp mimetype.dep.stamp htmlparse.dep.stamp myhtmlparse.dep.stamp mimehandler.dep.stamp internfile.dep.stamp mh_exec.dep.stamp mh_html.dep.stamp mh_mail.dep.stamp mh_mbox.dep.stamp mh_text.dep.stamp docseq.dep.stamp docseqdb.dep.stamp docseqhist.dep.stamp history.dep.stamp sortseq.dep.stamp wasastringtoquery.dep.stamp wasatorcl.dep.stamp pathhash.dep.stamp rcldb.dep.stamp searchdata.dep.stamp stemdb.dep.stamp base64.dep.stamp conftree.dep.stamp copyfile.dep.stamp debuglog.dep.stamp execmd.dep.stamp fstreewalk.dep.stamp idfile.dep.stamp md5.dep.stamp mimeparse.dep.stamp pathut.dep.stamp readfile.dep.stamp smallut.dep.stamp transcode.dep.stamp wipedir.dep.stamp x11mon.dep.stamp
+DEPS =  rclaspell.dep.stamp rclconfig.dep.stamp rclinit.dep.stamp textsplit.dep.stamp unacpp.dep.stamp csguess.dep.stamp indexer.dep.stamp mimetype.dep.stamp htmlparse.dep.stamp myhtmlparse.dep.stamp mimehandler.dep.stamp internfile.dep.stamp mh_exec.dep.stamp mh_html.dep.stamp mh_mail.dep.stamp mh_mbox.dep.stamp mh_text.dep.stamp docseq.dep.stamp docseqdb.dep.stamp docseqhist.dep.stamp history.dep.stamp sortseq.dep.stamp wasastringtoquery.dep.stamp wasatorcl.dep.stamp pathhash.dep.stamp rcldb.dep.stamp searchdata.dep.stamp stemdb.dep.stamp stoplist.dep.stamp base64.dep.stamp conftree.dep.stamp copyfile.dep.stamp debuglog.dep.stamp execmd.dep.stamp fstreewalk.dep.stamp idfile.dep.stamp md5.dep.stamp mimeparse.dep.stamp pathut.dep.stamp readfile.dep.stamp smallut.dep.stamp transcode.dep.stamp wipedir.dep.stamp x11mon.dep.stamp
 
 librcl.a : $(DEPS) $(OBJS) unac.o
 	ar ru librcl.a $(OBJS) unac.o
@@ -73,6 +73,8 @@ searchdata.o : ../rcldb/searchdata.cpp
 	$(CXX) $(ALL_CXXFLAGS) -c ../rcldb/searchdata.cpp
 stemdb.o : ../rcldb/stemdb.cpp
 	$(CXX) $(ALL_CXXFLAGS) -c ../rcldb/stemdb.cpp
+stoplist.o : ../rcldb/stoplist.cpp
+	$(CXX) $(ALL_CXXFLAGS) -c ../rcldb/stoplist.cpp
 base64.o : ../utils/base64.cpp
 	$(CXX) $(ALL_CXXFLAGS) -c ../utils/base64.cpp
 conftree.o : ../utils/conftree.cpp
@@ -193,6 +195,9 @@ searchdata.dep.stamp : ../rcldb/searchdata.cpp
 stemdb.dep.stamp : ../rcldb/stemdb.cpp
 	$(CXX) -M $(ALL_CXXFLAGS) ../rcldb/stemdb.cpp > stemdb.dep
 	touch stemdb.dep.stamp
+stoplist.dep.stamp : ../rcldb/stoplist.cpp
+	$(CXX) -M $(ALL_CXXFLAGS) ../rcldb/stoplist.cpp > stoplist.dep
+	touch stoplist.dep.stamp
 base64.dep.stamp : ../utils/base64.cpp
 	$(CXX) -M $(ALL_CXXFLAGS) ../utils/base64.cpp > base64.dep
 	touch base64.dep.stamp
@@ -266,6 +271,7 @@ include pathhash.dep
 include rcldb.dep
 include searchdata.dep
 include stemdb.dep
+include stoplist.dep
 include base64.dep
 include conftree.dep
 include copyfile.dep

src/lib/mkMake

@@ -32,6 +32,7 @@ ${depth}/rcldb/pathhash.cpp \
 ${depth}/rcldb/rcldb.cpp \
 ${depth}/rcldb/searchdata.cpp \
 ${depth}/rcldb/stemdb.cpp \
+${depth}/rcldb/stoplist.cpp \
 ${depth}/utils/base64.cpp \
 ${depth}/utils/conftree.cpp \
 ${depth}/utils/copyfile.cpp \

src/qtgui/main.cpp

@@ -1,5 +1,5 @@
 #ifndef lint
-static char rcsid[] = "@(#$Id: main.cpp,v 1.59 2007-05-21 13:30:21 dockes Exp $ (C) 2005 J.F.Dockes";
+static char rcsid[] = "@(#$Id: main.cpp,v 1.60 2007-06-02 08:30:41 dockes Exp $ (C) 2005 J.F.Dockes";
 #endif
 /*
  *   This program is free software; you can redistribute it and/or modify
@@ -102,7 +102,8 @@ bool maybeOpenDb(string &reason, bool force)
 	LOGDEB(("main: adding [%s]\n", it->c_str()));
 	rcldb->addQueryDb(*it);
     }
-    if (!rcldb->isopen() && !rcldb->open(dbdir, Rcl::Db::DbRO, qopts)) {
+    if (!rcldb->isopen() && !rcldb->open(dbdir, rclconfig->getStopfile(),
+					 Rcl::Db::DbRO, qopts)) {
 	reason = "Could not open database in " + 
 	    dbdir + " wait for indexing to complete?";
 	return false;

src/rcldb/Makefile

@@ -0,0 +1,24 @@
+# @(#$Id: Makefile,v 1.1 2007-06-02 08:30:42 dockes Exp $  (C) 2005 J.F.Dockes
+depth = ..
+include $(depth)/mk/sysconf
+
+# Only test executables get build in here
+PROGS = stoplist
+
+all: $(BIGLIB) $(PROGS) 
+
+$(BIGLIB): force
+	cd $(depth)/lib;$(MAKE)
+force:
+
+STOPLIST_OBJS= trstoplist.o  $(BIGLIB)
+stoplist : $(STOPLIST_OBJS)
+	$(CXX) $(ALL_CXXFLAGS) -o stoplist $(STOPLIST_OBJS) \
+	   $(LIBICONV) $(LIBSYS)
+trstoplist.o : stoplist.cpp 
+	$(CXX) $(ALL_CXXFLAGS) -DTEST_STOPLIST -c -o trstoplist.o \
+	       stoplist.cpp
+
+clean:
+	rm -f *.o $(PROGS)
+

src/rcldb/rcldb.cpp

@@ -1,5 +1,5 @@
 #ifndef lint
-static char rcsid[] = "@(#$Id: rcldb.cpp,v 1.110 2007-05-30 12:30:38 dockes Exp $ (C) 2004 J.F.Dockes";
+static char rcsid[] = "@(#$Id: rcldb.cpp,v 1.111 2007-06-02 08:30:42 dockes Exp $ (C) 2004 J.F.Dockes";
 #endif
 /*
  *   This program is free software; you can redistribute it and/or modify
@@ -528,7 +528,7 @@ Db::~Db()
     LOGERR(("Db::~Db: got exception: %s\n", ermsg));
 }
 
-bool Db::open(const string& dir, OpenMode mode, int qops)
+bool Db::open(const string& dir, const string &stops, OpenMode mode, int qops)
 {
     bool keep_updated = (qops & QO_KEEP_UPDATED) != 0;
     qops &= ~QO_KEEP_UPDATED;
@@ -543,6 +543,9 @@ bool Db::open(const string& dir, OpenMode mode, int qops)
 	if (!close())
 	    return false;
     }
+    if (!stops.empty())
+	m_stops.setFile(stops);
+
     const char *ermsg = "Unknown";
     try {
 	switch (mode) {
@@ -652,7 +655,7 @@ bool Db::reOpen()
     if (m_ndb && m_ndb->m_isopen) {
 	if (!close())
 	    return false;
-	if (!open(m_basedir, m_mode, m_qOpts | QO_KEEP_UPDATED)) {
+	if (!open(m_basedir, "", m_mode, m_qOpts | QO_KEEP_UPDATED)) {
 	    return false;
 	}
     }
@@ -737,8 +740,9 @@ class mySplitterCB : public TextSplitCB {
     Xapian::termpos basepos; // Base for document section
     Xapian::termpos curpos;  // Current position. Used to set basepos for the
                              // following section
-    mySplitterCB(Xapian::Document &d) 
+    StopList &stops;
-	: doc(d), basepos(1), curpos(0)
+    mySplitterCB(Xapian::Document &d, StopList &_stops) 
+	: doc(d), basepos(1), curpos(0), stops(_stops)
     {}
     bool takeword(const std::string &term, int pos, int, int);
     void setprefix(const string& pref) {prefix = pref;}
@@ -762,6 +766,10 @@ bool mySplitterCB::takeword(const std::string &term, int pos, int, int)
 
     const char *ermsg;
     try {
+	if (stops.hasStops() && stops.isStop(term)) {
+	    LOGDEB1(("Db: takeword [%s] in stop list\n", term.c_str()));
+	    return true;
+	}
 	// Note: 1 is the within document frequency increment. It would 
 	// be possible to assign different weigths to doc parts (ie title)
 	// by using a higher value
@@ -849,7 +857,7 @@ bool Db::add(const string &fn, const Doc &idoc,
 
     Xapian::Document newdocument;
 
-    mySplitterCB splitData(newdocument);
+    mySplitterCB splitData(newdocument, m_stops);
 
     TextSplit splitter(&splitData);
 

src/rcldb/rcldb.h

@@ -16,7 +16,7 @@
  */
 #ifndef _DB_H_INCLUDED_
 #define _DB_H_INCLUDED_
-/* @(#$Id: rcldb.h,v 1.47 2007-05-22 07:40:00 dockes Exp $  (C) 2004 J.F.Dockes */
+/* @(#$Id: rcldb.h,v 1.48 2007-06-02 08:30:42 dockes Exp $  (C) 2004 J.F.Dockes */
 
 #include <string>
 #include <list>
@@ -24,6 +24,7 @@
 
 #include "refcntr.h"
 #include "rcldoc.h"
+#include "stoplist.h"
 
 #ifndef NO_NAMESPACES
 using std::string;
@@ -77,7 +78,8 @@ class Db {
     // KEEP_UPDATED is internal use by reOpen() only
     enum QueryOpts {QO_NONE=0, QO_STEM = 1, QO_KEEP_UPDATED = 8};
 
-    bool open(const string &dbdir, OpenMode mode, int qops = QO_NONE);
+    bool open(const string &dbdir, const string &stoplistfn, 
+	      OpenMode mode, int qops = QO_NONE);
     bool close();
     bool isopen();
 
@@ -172,11 +174,13 @@ class Db {
     
     /** Filename wildcard expansion */
     bool filenameWildExp(const string& exp, list<string>& names);
-    string getReason(){return m_reason;}
+    string getReason() const {return m_reason;}
 
     /** Adjust flush threshold */
     void setFlushMb(int mb) {m_flushmb = mb;}
 
+    const StopList& getStopList() const {return m_stops;}
+
 private:
 
     string m_filterTopDir; // Current query filter on subtree top directory 
@@ -217,6 +221,8 @@ private:
 
     vector<bool> updated;
 
+    StopList m_stops;
+
     bool reOpen(); // Close/open, same mode/opts
     bool stemExpand(const string &lang, const string &s, 
 		    list<TermMatchEntry>& result, int max = -1);

src/rcldb/searchdata.cpp

@@ -1,5 +1,5 @@
 #ifndef lint
-static char rcsid[] = "@(#$Id: searchdata.cpp,v 1.13 2007-02-13 10:58:31 dockes Exp $ (C) 2006 J.F.Dockes";
+static char rcsid[] = "@(#$Id: searchdata.cpp,v 1.14 2007-06-02 08:30:42 dockes Exp $ (C) 2006 J.F.Dockes";
 #endif
 /*
  *   This program is free software; you can redistribute it and/or modify
@@ -32,6 +32,7 @@ static char rcsid[] = "@(#$Id: searchdata.cpp,v 1.13 2007-02-13 10:58:31 dockes
 #include "textsplit.h"
 #include "unacpp.h"
 #include "utf8iter.h"
+#include "stoplist.h"
 
 #ifndef NO_NAMESPACES
 using namespace std;
@@ -136,12 +137,23 @@ bool SearchData::getTerms(vector<string>& terms,
 // terms and phrases. 
 class wsQData : public TextSplitCB {
  public:
+    wsQData(const StopList &_stops) 
+	: stops(_stops), alltermcount(0)
+    {}
     vector<string> terms;
     bool takeword(const std::string &term, int , int, int) {
+	alltermcount++;
 	LOGDEB1(("wsQData::takeword: %s\n", term.c_str()));
+	if (stops.hasStops() && stops.isStop(term)) {
+	    LOGDEB1(("wsQData::takeword [%s] in stop list\n", term.c_str()));
+	    return true;
+	}
 	terms.push_back(term);
 	return true;
     }
+    const StopList &stops;
+    int alltermcount; // Count of terms including stopwords: this is
+		      // for adjusting phrase/near slack
 };
 
 /** 
@@ -158,10 +170,11 @@ public:
     { }
 
     bool processUserString(const string &iq,
-		   const string &prefix,
+			   const string &prefix,
-		   string &ermsg,
+			   string &ermsg,
-		   list<Xapian::Query> &pqueries,
+			   list<Xapian::Query> &pqueries, 
-		   int slack = 0, bool useNear = false);
+			   const StopList &stops,
+			   int slack = 0, bool useNear = false);
 
     bool getTerms(vector<string>& terms, 
 		  vector<vector<string> >& groups) 
@@ -313,7 +326,9 @@ bool StringToXapianQ::processUserString(const string &iq,
 					const string &prefix,
 					string &ermsg,
 					list<Xapian::Query> &pqueries,
-					int slack, bool useNear)
+					const StopList& stops,
+					int slack, bool useNear
+					)
 {
     LOGDEB(("StringToXapianQ:: query string: [%s]\n", iq.c_str()));
     ermsg.erase();
@@ -339,7 +354,7 @@ bool StringToXapianQ::processUserString(const string &iq,
 	    // we need to use a word split, else a phrase query including
 	    // a span would fail if we didn't adjust the proximity to
 	    // account for the additional span term which is complicated.
-	    wsQData splitDataS, splitDataW;
+	    wsQData splitDataS(stops), splitDataW(stops);
 	    TextSplit splitterS(&splitDataS, (TextSplit::Flags)
 				(TextSplit::TXTS_ONLYSPANS | 
 				 TextSplit::TXTS_KEEPWILD));
@@ -418,7 +433,8 @@ bool StringToXapianQ::processUserString(const string &iq,
 		pqueries.push_back(Xapian::Query(op,
 						 orqueries.begin(),
 						 orqueries.end(),
-					 splitData->terms.size() + slack));
+						 splitData->alltermcount 
+						 + slack));
 		// Add NEAR/PHRASE groups to the highlighting data. Must
 		// push all combinations
 		vector<vector<string> > allcombs;
@@ -508,7 +524,8 @@ bool SearchDataClauseSimple::toNativeQuery(Rcl::Db &db, void *p,
 	(m_parentSearch == 0 && !m_haveWildCards);
 
     StringToXapianQ tr(db, l_stemlang, doBoostUserTerm);
-    if (!tr.processUserString(m_text, prefix, m_reason, pqueries))
+    if (!tr.processUserString(m_text, prefix, m_reason, pqueries, 
+			      db.getStopList()))
 	return false;
     if (pqueries.empty()) {
 	LOGERR(("SearchDataClauseSimple: resolved to null query\n"));
@@ -570,7 +587,8 @@ bool SearchDataClauseDist::toNativeQuery(Rcl::Db &db, void *p,
     string s = string("\"") + m_text + string("\"");
     bool useNear = (m_tp == SCLT_NEAR);
     StringToXapianQ tr(db, l_stemlang, doBoostUserTerm);
-    if (!tr.processUserString(s, prefix, m_reason, pqueries, m_slack, useNear))
+    if (!tr.processUserString(s, prefix, m_reason, pqueries, db.getStopList(),
+			      m_slack, useNear))
 	return false;
     if (pqueries.empty()) {
 	LOGERR(("SearchDataClauseDist: resolved to null query\n"));

src/rcldb/stoplist.cpp

@@ -0,0 +1,104 @@
+#ifndef lint
+static char rcsid[] = "@(#$Id: stoplist.cpp,v 1.1 2007-06-02 08:30:42 dockes Exp $ (C) 2007 J.F.Dockes";
+#endif
+#ifndef TEST_STOPLIST
+#include "debuglog.h"
+#include "readfile.h"
+#include "unacpp.h"
+#include "stoplist.h"
+
+#ifndef NO_NAMESPACES
+namespace Rcl 
+{
+#endif
+
+bool StopList::setFile(const string &filename)
+{
+    m_hasStops = false;
+    m_stops.clear();
+    string stoptext, reason;
+    if (!file_to_string(filename, stoptext, &reason)) {
+	LOGDEB(("StopList::StopList: file_to_string(%s) failed: %s\n", 
+		filename.c_str(), reason.c_str()));
+	return false;
+    }
+    TextSplit ts(this, TextSplit::TXTS_ONLYSPANS);
+    ts.text_to_words(stoptext);
+    return true;
+}
+
+bool StopList::takeword(const string& term, int, int, int)
+{
+    string dterm;
+    unacmaybefold(term, dterm, "UTF-8", true);
+    LOGDEB2(("StopList::takeword: inserting [%s]\n", dterm.c_str()));
+    m_hasStops = true;
+    m_stops.insert(dterm);
+    return true;
+}
+
+bool StopList::isStop(const string &term) const
+{
+    return m_hasStops ? m_stops.find(term) != m_stops.end() : false;
+}
+
+
+#ifndef NO_NAMESPACES
+}
+#endif
+
+#else // TEST_STOPLIST
+
+#include <stdio.h>
+#include <stdlib.h>
+#include <unistd.h>
+#include <errno.h>
+#include <string.h>
+
+#include <string>
+#include <iostream>
+
+#include "stoplist.h"
+
+using namespace std;
+using namespace Rcl;
+
+static char *thisprog;
+
+static char usage [] =
+"trstoplist stopstermsfile\n\n"
+;
+static void
+Usage(void)
+{
+    fprintf(stderr, "%s: usage:\n%s", thisprog, usage);
+    exit(1);
+}
+
+const string tstwords[] = {
+    "the", "is", "xweird"
+};
+const int tstsz = sizeof(tstwords) / sizeof(string);
+
+int main(int argc, char **argv)
+{
+  int count = 10;
+    
+  thisprog = argv[0];
+  argc--; argv++;
+
+  if (argc != 1)
+    Usage();
+  string filename = argv[0]; argc--;
+
+  StopList sl(filename);
+
+  for (int i = 0; i < tstsz; i++) {
+      const string &tst = tstwords[i];
+      cout << "[" << tst << "] " << 
+	  (sl.isStop(tst) ? "in stop list" : "not in stop list") << endl;
+  }
+  exit(0);
+}
+
+#endif // TEST_STOPLIST

src/rcldb/stoplist.h

@@ -0,0 +1,37 @@
+#ifndef _STOPLIST_H_INCLUDED_
+#define _STOPLIST_H_INCLUDED_
+/* @(#$Id: stoplist.h,v 1.1 2007-06-02 08:30:42 dockes Exp $  (C) 2006 J.F.Dockes */
+
+#include <set>
+#include <string>
+
+#include "textsplit.h"
+
+#ifndef NO_NAMESPACES
+using std::set;
+using std::string;
+namespace Rcl 
+{
+#endif
+
+class StopList : public TextSplitCB {
+public:
+    StopList() : m_hasStops(false) {}
+    StopList(const string &filename) {setFile(filename);}
+    virtual ~StopList() {}
+
+    bool setFile(const string &filename);
+    bool isStop(const string &term) const;
+    bool hasStops() const {return m_hasStops;}
+    virtual bool takeword(const string& term, int pos, int bts, int bte); 
+
+private:
+    bool m_hasStops;
+    set<string> m_stops;
+};
+
+#ifndef NO_NAMESPACES
+}
+#endif
+
+#endif /* _STOPLIST_H_INCLUDED_ */

src/utils/readfile.cpp

@@ -1,5 +1,5 @@
 #ifndef lint
-static char rcsid[] = "@(#$Id: readfile.cpp,v 1.3 2006-01-23 13:32:28 dockes Exp $ (C) 2004 J.F.Dockes";
+static char rcsid[] = "@(#$Id: readfile.cpp,v 1.4 2007-06-02 08:30:42 dockes Exp $ (C) 2004 J.F.Dockes";
 #endif
 /*
  *   This program is free software; you can redistribute it and/or modify
@@ -32,20 +32,28 @@ using std::string;
 
 #include "readfile.h"
 
-bool file_to_string(const string &fn, string &data)
+bool file_to_string(const string &fn, string &data, string *reason)
 {
+#define ERRBUFSZ 200    
+    char errbuf[ERRBUFSZ];
     bool ret = false;
 
     int fd = open(fn.c_str(), O_RDONLY|O_STREAMING);
     if (fd < 0) {
-	// perror("open");
+	if (reason) {
+	    strerror_r(errno, errbuf, ERRBUFSZ);
+	    *reason += string("file_to_string: open failed: ") + errbuf;
+	}
 	return false;
     }
     char buf[4096];
     for (;;) {
 	int n = read(fd, buf, 4096);
 	if (n < 0) {
-	    // perror("read");
+	    if (reason) {
+		strerror_r(errno, errbuf, ERRBUFSZ);
+		*reason += string("file_to_string: read failed: ") + errbuf;
+	    }
 	    goto out;
 	}
 	if (n == 0)
@@ -54,7 +62,10 @@ bool file_to_string(const string &fn, string &data)
 	try {
 	    data.append(buf, n);
 	} catch (...) {
-	    //	    fprintf(stderr, "file_to_string: out of memory\n");
+	    if (reason) {
+		strerror_r(errno, errbuf, ERRBUFSZ);
+		*reason += string("file_to_string: out of memory? : ") +errbuf;
+	    }
 	    goto out;
 	}
     }

src/utils/readfile.h

@@ -16,7 +16,7 @@
  */
 #ifndef _READFILE_H_INCLUDED_
 #define _READFILE_H_INCLUDED_
-/* @(#$Id: readfile.h,v 1.2 2006-01-30 11:15:28 dockes Exp $  (C) 2004 J.F.Dockes */
+/* @(#$Id: readfile.h,v 1.3 2007-06-02 08:30:42 dockes Exp $  (C) 2004 J.F.Dockes */
 
 #include <string>
 
@@ -24,6 +24,7 @@
  * Read whole file into string. 
  * @return true for ok, false else
  */
-bool file_to_string(const std::string &filename, std::string &data);
+bool file_to_string(const std::string &filename, std::string &data, 
+		    std::string *reason = 0);
 
 #endif /* _READFILE_H_INCLUDED_ */