searchidx: fix -Lmedium for IDs and filenames

This fixes "m:", "l:", "f:", "t:", "c:", "dfn:", and "n:" search
prefixes under indexlevel=medium when mixed with indexlevel=full
inboxish.  We need positional data for Message-IDs, List-Id,
email addresses and filenames for exact matches, though we still
want to support wildcards.

Fortunately the storage cost is still small as these prefixes
tend to be small compared to message bodies.  These are NOT
boolean terms since wildcard support and partial matching is
desired.

1 files changed, 37 insertions, 10 deletions

diff --git a/lib/PublicInbox/SearchIdx.pm b/lib/PublicInbox/SearchIdx.pm
index 3372bea5..772f5a64 100644
--- a/lib/PublicInbox/SearchIdx.pm
+++ b/lib/PublicInbox/SearchIdx.pm
@@ -22,6 +22,7 @@ use PublicInbox::OverIdx;
 use PublicInbox::Spawn qw(spawn nodatacow_dir);
 use PublicInbox::Git qw(git_unquote);
 use PublicInbox::MsgTime qw(msg_timestamp msg_datestamp);
+use PublicInbox::Address;
 our @EXPORT_OK = qw(log2stack is_ancestor check_size prepare_stack
         index_text term_generator add_val is_bad_blob);
 my $X = \%PublicInbox::Search::X;
@@ -158,22 +159,44 @@ sub term_generator ($) { # write-only
         }
 }
 
+sub index_phrase ($$$$) {
+        my ($self, $text, $wdf_inc, $prefix) = @_;
+
+        my $tg = term_generator($self);
+        $tg->index_text($text, $wdf_inc, $prefix);
+        $tg->increase_termpos;
+}
+
 sub index_text ($$$$) {
         my ($self, $text, $wdf_inc, $prefix) = @_;
-        my $tg = term_generator($self); # man Search::Xapian::TermGenerator
 
         if ($self->{indexlevel} eq 'full') {
-                $tg->index_text($text, $wdf_inc, $prefix);
-                $tg->increase_termpos;
+                index_phrase($self, $text, $wdf_inc, $prefix);
         } else {
+                my $tg = term_generator($self);
                 $tg->index_text_without_positions($text, $wdf_inc, $prefix);
         }
 }
 
 sub index_headers ($$) {
         my ($self, $smsg) = @_;
-        my @x = (from => 'A', # Author
-                subject => 'S', to => 'XTO', cc => 'XCC');
+        my @x = (from => 'A', to => 'XTO', cc => 'XCC'); # A: Author
+        while (my ($field, $pfx) = splice(@x, 0, 2)) {
+                my $val = $smsg->{$field};
+                next if $val eq '';
+                # include "(comments)" after the address, too, so not using
+                # PublicInbox::Address::names or pairs
+                index_text($self, $val, 1, $pfx);
+
+                # we need positional info for email addresses since they
+                # can be considered phrases
+                if ($self->{indexlevel} eq 'medium') {
+                        for my $addr (PublicInbox::Address::emails($val)) {
+                                index_phrase($self, $addr, 1, $pfx);
+                        }
+                }
+        }
+        @x = (subject => 'S');
         while (my ($field, $pfx) = splice(@x, 0, 2)) {
                 my $val = $smsg->{$field};
                 index_text($self, $val, 1, $pfx) if $val ne '';
@@ -186,7 +209,11 @@ sub index_diff_inc ($$$$) {
                 index_text($self, join("\n", @$xnq), 1, 'XNQ');
                 @$xnq = ();
         }
-        index_text($self, $text, 1, $pfx);
+        if ($pfx eq 'XDFN') {
+                index_phrase($self, $text, 1, $pfx);
+        } else {
+                index_text($self, $text, 1, $pfx);
+        }
 }
 
 sub index_old_diff_fn {
@@ -292,7 +319,7 @@ sub index_xapian { # msg_iter callback
         my $ct = $part->content_type || 'text/plain';
         my $fn = $part->filename;
         if (defined $fn && $fn ne '') {
-                index_text($self, $fn, 1, 'XFN');
+                index_phrase($self, $fn, 1, 'XFN');
         }
         if ($part->{is_submsg}) {
                 my $mids = mids_for_index($part);
@@ -330,20 +357,20 @@ sub index_list_id ($$$) {
                 $l =~ /<([^>]+)>/ or next;
                 my $lid = lc $1;
                 $doc->add_boolean_term('G' . $lid);
-                index_text($self, $lid, 1, 'XL'); # probabilistic
+                index_phrase($self, $lid, 1, 'XL'); # probabilistic
         }
 }
 
 sub index_ids ($$$$) {
         my ($self, $doc, $hdr, $mids) = @_;
         for my $mid (@$mids) {
-                index_text($self, $mid, 1, 'XM');
+                index_phrase($self, $mid, 1, 'XM');
 
                 # because too many Message-IDs are prefixed with
                 # "Pine.LNX."...
                 if ($mid =~ /\w{12,}/) {
                         my @long = ($mid =~ /(\w{3,}+)/g);
-                        index_text($self, join(' ', @long), 1, 'XM');
+                        index_phrase($self, join(' ', @long), 1, 'XM');
                 }
         }
         $doc->add_boolean_term('Q' . $_) for @$mids;