Fix output parser to show more results

Fix #63

The git output parser was pretty naive and buggy when multiple
chunks were found from the same file.

This was resulting in skipping results we should have displayed to begin with.
That changeset is fixing that issue and will now show more results.

The summary was accurate but the output truncated due to poor man parsing reason.

Author: atoomic

src/lib/GrepCpan/Grep.pm

@@ -282,7 +282,7 @@ sub _do_search ( $self, %opts ) {
     $page = 0 if $page < 0;
 
     #
-    my $cache = $self->get_match_cache( $search, $search_distro, $filetype,
+    my $cache = $self->_get_match_cache( $search, $search_distro, $filetype,
         $caseinsensitive, $ignore_files );
 
     my $is_a_known_distro
@@ -362,41 +362,55 @@ sub _do_search ( $self, %opts ) {
         return;
     };
 
+    my $previous_file;
+    my $qr_match_line = qr{^([0-9]+)([-:])};
+
     foreach my $line (@$matches) {
-        if ( !defined $current_file && $line !~ qr{^([0-9]+)[-:]} )
-        { # when more than one block match we are just going to have a -- separator
-            $current_file = $line;
+        if ( !defined $current_file ) {
+
+     # when more than one block match we are just going to have a -- separator
+            if ( $line =~ m{^distros/} ) {
+                $previous_file = $current_file = $line;
+                next;
+            }
+            $current_file //= $previous_file;
         }
-        elsif ( $line eq '--' ) {
+
+        if ( $line eq '--' ) {
+
+        # we found a new block, it's either from the current file or a new one
             $process_file->();
-            undef $current_file;
+            undef $current_file;    # reset: could use previous or next file
             $diff       = '';
             @diffblocks = ();
             undef $start_line;
             undef $line_number;
             @matching_lines = ();
+            next;
         }
-        elsif ( $line =~ s{^([0-9]+)([-:])}{} ) {    # mainly else
-            my $new_line = $1;
-            my $prefix   = $2;
-            $start_line //= $new_line;
-            if ( length($line) > 250 )
-            {    # max length autorized ( js minified & co )
-                $line = substr( $line, 0, 250 ) . '...';
-            }
-            if ( !defined $line_number || $new_line == $line_number + 1 ) {
 
-                # same block
-                push @matching_lines, $new_line if $prefix eq ':';
-                $diff .= $line . "\n";
-            }
-            else {
-                # new block
-                $add_block->();
-                $diff = $line . "\n";    # reset the block
-            }
-            $line_number = $new_line;
+        # matching the main part
+        next unless $line =~ s/$qr_match_line//;
+        my ( $new_line, $prefix ) = ( $1, $2 );
+
+        $start_line //= $new_line;
+        if ( length($line) > 250 )
+        {    # max length autorized ( js minified & co )
+            $line = substr( $line, 0, 250 ) . '...';
         }
+        if ( !defined $line_number || $new_line == $line_number + 1 ) {
+
+            # same block
+            push @matching_lines, $new_line if $prefix eq ':';
+            $diff .= $line . "\n";
+        }
+        else {
+            # new block
+            $add_block->();
+            $diff = $line . "\n";    # reset the block
+        }
+        $line_number = $new_line;
+
     }
     $process_file->();    # process the last block
 
@@ -592,7 +606,7 @@ sub _parse_ignore_files ( $self, $ignore_files ) {
     return \@rules;
 }
 
-sub get_match_cache(
+sub _get_match_cache(
     $self, $search, $search_distro, $query_filetype,
     $caseinsensitive = 0,
     $ignore_files = undef
@@ -601,6 +615,11 @@ sub get_match_cache(
 
     $caseinsensitive //= 0;
 
+    my @keys_for_cache = (
+        $search,          $search_distro, $query_filetype,
+        $caseinsensitive, $ignore_files // ''
+    );
+
     my $gitdir = $self->git()->work_tree;
     my $limit  = $self->config()->{limit}->{files_per_search} or die;
 
@@ -614,10 +633,8 @@ sub get_match_cache(
     }
 
     # use the full cache when available -- need to filter it later
-    my $request_cache_file = $self->_get_cache_file(
-        [ $search, $search_distro, $query_filetype, $caseinsensitive, $ignore_files // '' ] );
-    {
-        my $load = $self->_load_cache($request_cache_file);
+    my $request_cache_file = $self->_get_cache_file( \@keys_for_cache );
+    if ( my $load = $self->_load_cache($request_cache_file) ) {
         return $load if $load;
     }
 
@@ -647,8 +664,7 @@ sub get_match_cache(
 
     # fallback to a shorter search ( and a different cache )
     my $cache_file = $self->_get_cache_file( [@git_cmd] );
-    {
-        my $load = $self->_load_cache($cache_file);
+    if ( my $load = $self->_load_cache($cache_file) ) {
         return $load if $load;
     }
 
@@ -699,7 +715,7 @@ sub get_match_cache(
 
     $cache->{match} = {
         files   => $match_files,
-        distros => scalar keys %{ $cache->{distros} }
+        distros => scalar keys $cache->{distros}->%*,
     };
 
     if ( !$search_in_progress ) {