From 3e58bcd54f064c48e8fc0e9d3267f5b1c620f375 Mon Sep 17 00:00:00 2001
From: vfilippov <vfilippov@6955db30-a419-402b-8a0d-67ecbb4d7f56>
Date: Tue, 15 Sep 2009 12:33:27 +0000
Subject: [PATCH] Bug 46221 Relevance + AND by default

git-svn-id: svn://svn.office.custis.ru/3rdparty/bugzilla.org/trunk@427 6955db30-a419-402b-8a0d-67ecbb4d7f56
---
 Bugzilla/DB/Mysql.pm | 31 +++++++++++++++++++------------
 Bugzilla/Util.pm     | 14 +++++++++++---
 2 files changed, 30 insertions(+), 15 deletions(-)

diff --git a/Bugzilla/DB/Mysql.pm b/Bugzilla/DB/Mysql.pm
index 20f0035ee..8b2b3a3ad 100644
--- a/Bugzilla/DB/Mysql.pm
+++ b/Bugzilla/DB/Mysql.pm
@@ -165,22 +165,29 @@ sub sql_string_concat {
 sub sql_fulltext_search {
     my ($self, $column, $text) = @_;
 
-    # Add the boolean mode modifier if the search string contains
-    # boolean operators at the start or end of a word.
-    my $mode = '';
-    if ($text =~ /(?:^|\W)[+\-<>~"()]/ || $text =~ /[()"*](?:$|\W)/) {
-        $mode = 'IN BOOLEAN MODE';
-
-        # quote un-quoted compound words
-        my @words = quotewords('[\s()]+', 'delimiters', $text);
-        foreach my $word (@words) {
+    # quote un-quoted compound words
+    my @words = quotewords('[\s()]+', 'delimiters', $text);
+    if ($text =~ /(?:^|\W)[+\-<>~"()]/ || $text =~ /[()"*](?:$|\W)/)
+    {
+        # already a boolean mode search
+        foreach my $word (@words)
+        {
             # match words that have non-word chars in the middle of them
-            if ($word =~ /\w\W+\w/ && $word !~ m/"/) {
+            if ($word =~ /\w\W+\w/ && $word !~ /\"/)
+            {
                 $word = '"' . $word . '"';
             }
         }
-        $text = join('', @words);
     }
+    else
+    {
+        # make search a boolean mode search
+        for (@words)
+        {
+            $_ = "+$_*" if /\w$/;
+        }
+    }
+    $text = join '', @words;
 
     # quote the text for use in the MATCH AGAINST expression
     $text = $self->quote($text);
@@ -188,7 +195,7 @@ sub sql_fulltext_search {
     # untaint the text, since it's safe to use now that we've quoted it
     trick_taint($text);
 
-    return "MATCH($column) AGAINST($text $mode)";
+    return ("MATCH($column) AGAINST($text IN BOOLEAN MODE)", "MATCH($column) AGAINST($text)");
 }
 
 sub sql_istring {
diff --git a/Bugzilla/Util.pm b/Bugzilla/Util.pm
index db576621d..8a707f899 100644
--- a/Bugzilla/Util.pm
+++ b/Bugzilla/Util.pm
@@ -676,6 +676,7 @@ sub disable_utf8 {
 sub stem_text
 {
     my ($text, $allow_verbatim) = @_;
+    Encode::_utf8_on($text) if Bugzilla->params->{utf8};
     $text = [ split /(?<=\w)(?=\W)|(?<=\W)(?=\w)/, $text ];
     my $q = 0;
     for (@$text)
@@ -684,10 +685,17 @@ sub stem_text
         {
             $_ = Lingua::Stem::RuUTF8::stem_word($_) unless $q;
         }
-        elsif ($allow_verbatim)
+        else
         {
-            # If $allow_verbatim is TRUE then text in "double quotes" doesn't stem
-            $q = ($q + tr/\"/\"/) % 2;
+            if ($allow_verbatim)
+            {
+                # If $allow_verbatim is TRUE then text in "double quotes" doesn't stem
+                $q = ($q + tr/\"/\"/) % 2;
+            }
+            if (!/\s$/so)
+            {
+                $_ .= ' ';
+            }
         }
     }
     return join '', @$text;