Ticket #350: 0001-Add-set_max_wildcard_expansion-method-to-the-queryp4.patch

search-xapian/XS/QueryParser.xs

@@ -49 +49 @@
     CODE:
         THIS->set_database(*database);
 
+void
+QueryParser::set_max_wildcard_expansion(max)
+    termcount max
+    CODE:
+        THIS->set_max_wildcard_expansion(max);
+
 Query *
 QueryParser::parse_query(q, flags = QueryParser::FLAG_DEFAULT)
     string q

search-xapian/Xapian/QueryParser.pm

@@ -166 +166 @@
 
 If there were no corrections, an empty string is returned.
 
+=item set_max_wildcard_expansion <max>
+
+If the queryparser expands a wildcard to more than max terms, an
+exception will be thrown.
+
+=cut
+
 =back
 
 =head1 REFERENCE

xapian-core/docs/queryparser.html

@@ -117 +117 @@
 to enable it, and tell the QueryParser which database to expand wildcards from
 using the <code>QueryParser::set_database(<i>database</i>)</code> method.
 
+<P>You can limit the number of terms a wildcard will expand to by
+calling <code>Xapian::QueryParser::set_max_wildcard_expansion()</code>. If a
+wildcard expands to more terms than that number, an exception will be
+thrown. The exception may be thrown by the QueryParser, or later when
+Enquire handles the query. The default is not to limit the expansion.
+
 <H3>Partially entered query matching</H3>
 
 <P>

xapian-core/include/xapian/queryparser.h

@@ -446 +446 @@
      */
     void set_database(const Database &db);
 
+    /// Specify the maximum expansion of a wildcard term.
+    void set_max_wildcard_expansion(Xapian::termcount);
+
     /** Parse a query.
      *
      *  @param query_string  A free-text query as entered by a user

xapian-core/queryparser/queryparser.cc

@@ -103 +103 @@
     internal->db = db;
 }
 
+void
+QueryParser::set_max_wildcard_expansion(Xapian::termcount max)
+{
+    internal->max_wildcard_expansion = max;
+}
+
 Query
 QueryParser::parse_query(const string &query_string, unsigned flags,
                          const string &default_prefix)

xapian-core/queryparser/queryparser.lemony

@@ -27 +27 @@
 #include <xapian/error.h>
 #include <xapian/unicode.h>
 #include "stringutils.h"
+#include "utils.h"
 
 // Include the list of token values lemon generates.
 #include "queryparser_token.h"
@@ -251 +252 @@
     void stoplist_resize(size_t s) {
         qpi->stoplist.resize(s);
     }
+
+    Xapian::termcount get_max_wildcard_expansion() {
+        return qpi->max_wildcard_expansion;
+    }
 };
 
 string
@@ -364 +369 @@
 
     const list<string> & prefixes = prefix_info->prefixes;
     list<string>::const_iterator piter;
+    Xapian::termcount expansion_count = 0;
+    Xapian::termcount max = state_->get_max_wildcard_expansion();
     for (piter = prefixes.begin(); piter != prefixes.end(); ++piter) {
         string root = *piter;
         root += name;
@@ -371 +378 @@
         while (t != db.allterms_end(root)) {
             subqs.push_back(Query(*t, 1, pos));
             ++t;
+            if (max != 0 && ++expansion_count > max) {
+                throw Xapian::QueryParserError("Wildcard "+unstemmed+"* expands too much (exceeding "+om_tostring(max)+" terms)");
+            }
         }
     }
     delete this;

xapian-core/queryparser/queryparser_internal.h

@@ -79 +79 @@
     void add_prefix(const string &field, const string &prefix,
                     filter_type type);
 
+    Xapian::termcount max_wildcard_expansion;
+
     std::string parse_term(Utf8Iterator &it, const Utf8Iterator &end,
                            bool &was_acronym);
 
   public:
     Internal() : stem_action(STEM_NONE), stopper(NULL),
-        default_op(Query::OP_OR), errmsg(NULL) { }
+        default_op(Query::OP_OR), errmsg(NULL) { max_wildcard_expansion(0); }
     Query parse_query(const string & query_string, unsigned int flags, const string & default_prefix);
 };