Merge pull request #1168 from metacpan/haarg/fix-author-indexing

fix author indexing
metacpan · Feb 23, 2024 · 333322d · 333322d
2 parents 256111b + bdf34ef
commit 333322d
Show file tree

Hide file tree

Showing 2 changed files with 29 additions and 13 deletions.
diff --git a/lib/MetaCPAN/Script/Author.pm b/lib/MetaCPAN/Script/Author.pm
@@ -103,6 +103,7 @@ sub index_authors {
 
     my $scroll = $self->es->scroll_helper(
         index       => $self->index->name,
+        type        => 'author',
         search_type => 'scan',
         size        => 500,
         body        => {
@@ -224,8 +225,19 @@ sub update_author {
         return;
     }
 
-    return
-        unless diff_struct( $current_data, $data );
+    if ( my $diff = diff_struct( $current_data, $data ) ) {
+
+        # log a sampling of differences
+        if ( $self->has_surrogate_keys_to_purge % 10 == 9 ) {
+            Dlog_debug {
+                "Found difference in $pauseid: $_"
+            }
+            $diff;
+        }
+    }
+    else {
+        return;
+    }
 
     $data->{updated} = DateTime->now( time_zone => 'UTC' )->iso8601;
 

diff --git a/lib/MetaCPAN/Util.pm b/lib/MetaCPAN/Util.pm
@@ -163,34 +163,38 @@ sub single_valued_arrayref_to_scalar {
 }
 
 sub diff_struct {
-    my (@queue) = [@_];
+    my ( $old_root, $new_root, $allow_extra ) = @_;
+    my (@queue) = [ $old_root, $new_root, '', $allow_extra ];
 
     while ( my $check = shift @queue ) {
-        my ( $old, $new, $allow_extra ) = @$check;
+        my ( $old, $new, $path, $allow_extra ) = @$check;
         if ( !defined $new ) {
-            return !!1
+            return [ $path, $old, $new ]
                 if defined $old;
         }
         elsif ( !is_ref($new) ) {
-            return !!1
+            return [ $path, $old, $new ]
                 if is_ref($old)
                 or $new ne $old;
         }
         elsif ( is_plain_arrayref($new) ) {
-            return !!1
+            return [ $path, $old, $new ]
                 if !is_plain_arrayref($old) || @$new != @$old;
-            push @queue, map [ $old->[$_], $new->[$_] ], 0 .. $#$new;
+            push @queue, map [ $old->[$_], $new->[$_], "$path/$_" ],
+                0 .. $#$new;
         }
         elsif ( is_plain_hashref($new) ) {
-            return !!1
-                if !is_plain_hashref($old) || keys %$new != keys %$old;
-            push @queue, map [ $old->{$_}, $new->{$_} ], keys %$new;
+            return [ $path, $old, $new ]
+                if !is_plain_hashref($old)
+                || !$allow_extra && keys %$new != keys %$old;
+            push @queue, map [ $old->{$_}, $new->{$_}, "$path/$_" ],
+                keys %$new;
         }
         else {
-            die "can't compare $new type data";
+            die "can't compare $new type data at $path";
         }
     }
-    return !!0;
+    return undef;
 }
 
 1;