Date: 05 May 2003 16:15:06 +0200
From: Gabriele Bartolini <g.bartol@comune.prato.it>
To: "ht://Dig - Dev" <htdig-dev@lists.sourceforge.net>
Subject: [htdig-dev] Regex patch for 3.1.6 - ready

Hi guys,

   I got to make it work today ... If some of you could please put in
the patch archive ... :-)

   Basically it is possibile to specify a regular expression in the
restrict and exclude attribute, as:

restrict: [^http://www.comune.prato.it/\$]

Ciao ciao
-Gabriele
-- 
Gabriele Bartolini - Web Programmer
Comune di Prato - Prato - Tuscany - Italy
g.bartol@comune.prato.it | http://www.comune.prato.it
> find bin/laden -name osama -exec rm {} ;


diff -3upr ../htdig-3.1.6/htsearch/Display.cc ./htsearch/Display.cc
--- ../htdig-3.1.6/htsearch/Display.cc	Fri Feb  1 00:47:18 2002
+++ ./htsearch/Display.cc	Mon May  5 16:06:34 2003
@@ -212,15 +212,14 @@ Display::display(int pageNumber)
 int
 Display::includeURL(char *url)
 {
-    if (limitTo && limitTo->FindFirst(url) < 0)
+    if (limitTo && limitTo->match(url, 1, 0) == 0)
     {
 	return 0;
     }
     else
     {
 	if (excludeFrom &&
-            excludeFrom->hasPattern() &&
-            excludeFrom->FindFirst(url) >= 0)
+            excludeFrom->match(url, 0, 0) != 0)
 	    return 0;
 	else
 	    return 1;
Only in ./htsearch: Display.cc.orig
diff -3upr ../htdig-3.1.6/htsearch/Display.h ./htsearch/Display.h
--- ../htdig-3.1.6/htsearch/Display.h	Fri Feb  1 00:47:18 2002
+++ ./htsearch/Display.h	Mon May  5 16:06:34 2003
@@ -18,6 +18,7 @@
 #include "DocumentDB.h"
 #include "Database.h"
 #include "Dictionary.h"
+#include "HtRegex.h"
 
 class Display : public Object
 {
@@ -34,8 +35,8 @@ public:
 	
     void		setResults(ResultList *results);
     void		setSearchWords(List *searchWords);
-    void		setLimit(StringMatch *);
-    void		setExclude(StringMatch *);
+    void		setLimit(HtRegex *);
+    void		setExclude(HtRegex *);
     void		setAllWordsPattern(StringMatch *);
     void		setLogicalWords(char *);
     void		setOriginalWords(char *);
@@ -82,8 +83,8 @@ protected:
     //
     // Pattern that all result URLs must match or exclude
     //
-    StringMatch		*limitTo;
-    StringMatch		*excludeFrom;
+    HtRegex		*limitTo;
+    HtRegex		*excludeFrom;
 
     //
     // Pattern of all the words
@@ -173,13 +174,13 @@ protected:
 
 //*****************************************************************************
 inline void
-Display::setLimit(StringMatch *limit)
+Display::setLimit(HtRegex *limit)
 {
     limitTo = limit;
 }
 
 inline void
-Display::setExclude(StringMatch *exclude)
+Display::setExclude(HtRegex *exclude)
 {
     excludeFrom = exclude;
 }
Only in ./htsearch: Display.h.orig
diff -3upr ../htdig-3.1.6/htsearch/htsearch.cc ./htsearch/htsearch.cc
--- ../htdig-3.1.6/htsearch/htsearch.cc	Fri Feb  1 00:47:18 2002
+++ ./htsearch/htsearch.cc	Mon May  5 16:07:26 2003
@@ -27,6 +27,7 @@ static char RCSid[] = "$Id: htsearch.cc,
 #include "HtURLCodec.h"
 #include "HtURLRewriter.h"
 #include "HtWordType.h"
+#include "HtRegex.h"
 
 // If we have this, we probably want it.
 #ifdef HAVE_GETOPT_H
@@ -62,8 +63,8 @@ main(int ac, char **av)
     List		searchWords;
     String		configFile = DEFAULT_CONFIG_FILE;
     int			pageNumber = 1;
-    StringMatch		limit_to;
-    StringMatch		exclude_these;
+    HtRegex		limit_to;
+    HtRegex		exclude_these;
     String		logicalWords;
     String              origPattern;
     String              logicalPattern;
@@ -215,13 +216,13 @@ main(int ac, char **av)
 
     if (strlen(config["restrict"]))
     {
-        // Create a temporary list from either the configuration
-        // file or the input parameter
-        urllist.Create(config["restrict"], "| \t\r\n\001");
-        urlpat = urllist.Join('|');
-        urllist.Release();  // release the temporary list of URLs
-        config.Add("restrict", urlpat);  // re-create the config attribute
-	limit_to.Pattern(urlpat);  // Set the new limit pattern
+	// Create a temporary list from either the configuration
+	// file or the input parameter
+	urllist.Create(config["restrict"], "| \t\r\n\001");
+	limit_to.setEscaped(urllist);  // Set the new limit pattern
+	urlpat = urllist.Join('|');
+	config.Add("restrict", urlpat);  // re-create the config attribute
+	urllist.Release();  // release the temporary list of URLs
     }
 
     if (strlen(config["exclude"]))
@@ -229,10 +230,10 @@ main(int ac, char **av)
         // Create a temporary list from either the configuration
         // file or the input parameter
         urllist.Create(config["exclude"], "| \t\r\n\001");
+        exclude_these.setEscaped(urllist);
         urlpat = urllist.Join('|');
-        urllist.Release();  // release the temporary list of URLs
         config.Add("exclude", urlpat);  // re-create the config attribute
-        exclude_these.Pattern(urlpat);
+        urllist.Release();  // release the temporary list of URLs
     }
 
     // Ctype-like functions for what constitutes a word.