Merge branch 'devel' of github.com:peczenyj/GDPR-IAB-TCFv2 into devel

peczenyj · Dec 13, 2023 · 5c62481 · 5c62481
2 parents 89bda65 + e0ef785
commit 5c62481
Show file tree

Hide file tree

Showing 8 changed files with 140 additions and 112 deletions.
diff --git a/Changes b/Changes
@@ -1,3 +1,5 @@
+  - refactor on Publisher Restriction parsing.
+  - small fixes about data and offset.
   - performance improvement: when we parse a range-based consent string now the Parse method is 23% faster, TO_JSON is 9% faster and check vendor consent or legitimate interest is between 122% and 137% faster than the previous version
   - remove GDPR::IAB::TCFv2::RangeConsent package
 

diff --git a/lib/GDPR/IAB/TCFv2.pm b/lib/GDPR/IAB/TCFv2.pm
@@ -120,15 +120,6 @@ sub Parse {
 
     croak 'invalid vendor list version' if $self->vendor_list_version == 0;
 
-    # TODO parse special feature opt in
-
-    #_parse_bitfield()
-
-    # TODO parse purpose section
-    # TODO parse purpose consent
-
-    # TODO parse purpose legitimate interest
-
     # parse vendor section
     # parse vendor consent
 
@@ -330,7 +321,7 @@ sub check_publisher_restriction {
     my ( $self, $purpose_id, $restrict_type, $vendor ) = @_;
 
     return $self->{publisher_restrictions}
-      ->check_publisher_restriction( $purpose_id, $restrict_type, $vendor );
+      ->contains( $purpose_id, $restrict_type, $vendor );
 }
 
 sub _format_date {
@@ -495,37 +486,20 @@ sub _parse_vendor_legitimate_interests {
 sub _parse_publisher_restrictions {
     my ( $self, $pub_restrict_offset ) = @_;
 
-    my ( $num_restrictions, $next_offset ) =
-      get_uint12( $self->{data}, $pub_restrict_offset );
-
-    my %restrictions;
-
-    for ( 1 .. $num_restrictions ) {
-        my ( $purpose_id, $restriction_type, $vendor_restrictions );
-
-        ( $purpose_id, $next_offset ) =
-          get_uint6( $self->{data}, $next_offset );
-
-        ( $restriction_type, $next_offset ) =
-          get_uint2( $self->{data}, $next_offset );
-
-        ( $vendor_restrictions, $next_offset ) = $self->_parse_range_section(
-            ASSUMED_MAX_VENDOR_ID,
-            $next_offset
-        );
+    my $data =
+      substr( $self->{data}, $pub_restrict_offset, ASSUMED_MAX_VENDOR_ID );
 
-        $restrictions{$purpose_id} ||= {};
-
-        $restrictions{$purpose_id}->{$restriction_type} = $vendor_restrictions;
-    }
-
-    my $publisher_restrictions = GDPR::IAB::TCFv2::PublisherRestrictions->new(
-        restrictions => \%restrictions,
-    );
+    my ( $publisher_restrictions, $relative_next_offset ) =
+      GDPR::IAB::TCFv2::PublisherRestrictions->Parse(
+        data      => $data,
+        data_size => length( $self->{data} ),
+        max_id    => ASSUMED_MAX_VENDOR_ID,
+        options   => $self->{options},
+      );
 
     $self->{publisher_restrictions} = $publisher_restrictions;
 
-    return $next_offset;
+    return $pub_restrict_offset + $relative_next_offset;
 }
 
 sub _get_core_tc_string {
@@ -571,30 +545,40 @@ sub _is_vendor_consent_range_encoding {
 }
 
 sub _parse_range_section {
-    my ( $self, $max_id, $offset ) = @_;
+    my ( $self, $max_id, $range_section_start_offset ) = @_;
+
+    my $data = substr( $self->{data}, $range_section_start_offset, $max_id );
 
     my ( $range_section, $next_offset ) =
       GDPR::IAB::TCFv2::RangeSection->Parse(
-        data    => $self->{data},
-        offset  => $offset,
-        max_id  => $max_id,
-        options => $self->{options},
+        data      => $data,
+        data_size => length( $self->{data} ),
+        offset    => 0,
+        max_id    => $max_id,
+        options   => $self->{options},
       );
 
-    return ( $range_section, $next_offset );
+    return
+      wantarray
+      ? ( $range_section, $range_section_start_offset + $next_offset )
+      : $range_section;
 }
 
 sub _parse_bitfield {
-    my ( $self, $max_id, $offset ) = @_;
+    my ( $self, $max_id, $bitfield_start_offset ) = @_;
+
+    my $data = substr( $self->{data}, $bitfield_start_offset, $max_id );
 
     my ( $bitfield, $next_offset ) = GDPR::IAB::TCFv2::BitField->Parse(
-        data    => $self->{data},
-        offset  => $offset,
-        max_id  => $max_id,
-        options => $self->{options},
+        data      => $data,
+        data_size => length( $self->{data} ),
+        max_id    => $max_id,
+        options   => $self->{options},
     );
 
-    return ( $bitfield, $next_offset );
+    return wantarray
+      ? ( $bitfield, $bitfield_start_offset + $next_offset )
+      : $bitfield;
 }
 
 sub looksLikeIsConsentVersion2 {

diff --git a/lib/GDPR/IAB/TCFv2/BitField.pm b/lib/GDPR/IAB/TCFv2/BitField.pm
@@ -10,31 +10,27 @@ use Carp                       qw<croak>;
 sub Parse {
     my ( $klass, %args ) = @_;
 
-    croak "missing 'data'"   unless defined $args{data};
-    croak "missing 'offset'" unless defined $args{offset};
+    croak "missing 'data'" unless defined $args{data};
     croak "missing 'max_id'"
       unless defined $args{max_id};
 
     croak "missing 'options'"      unless defined $args{options};
     croak "missing 'options.json'" unless defined $args{options}->{json};
 
-    my $data    = $args{data};
-    my $offset  = $args{offset};
-    my $max_id  = $args{max_id};
-    my $options = $args{options};
-
-    my $data_size = length($data);
+    my $data      = $args{data};
+    my $data_size = $args{data_size};
+    my $offset    = 0;
+    my $max_id    = $args{max_id};
+    my $options   = $args{options};
 
     # add 7 to force rounding to next integer value
-    my $bytes_required = ( $max_id + $offset + 7 ) / 8;
+    my $bytes_required = ( $max_id + 7 ) / 8;
 
     croak
       "a BitField for $max_id requires a consent string of $bytes_required bytes. This consent string had $data_size"
       if $data_size < $bytes_required;
 
     my $self = {
-
-        # TODO consider store data as arrayref of bits
         data    => substr( $data, $offset, $max_id ),
         max_id  => $max_id,
         options => $options,
@@ -56,14 +52,6 @@ sub contains {
     return is_set( $self->{data}, $id - 1 );
 }
 
-sub all {
-    my $self = shift;
-
-    my @data = split //, $self->{data};
-
-    return [ grep { $data[ $_ - 1 ] } 1 .. $self->{max_id} ];
-}
-
 sub TO_JSON {
     my $self = shift;
 

diff --git a/lib/GDPR/IAB/TCFv2/BitUtils.pm b/lib/GDPR/IAB/TCFv2/BitUtils.pm
@@ -32,8 +32,11 @@ our @EXPORT_OK = qw<is_set
 sub is_set {
     my ( $data, $offset ) = @_;
 
-    croak "index out of bounds on offset $offset"
-      if $offset + 1 > length($data);
+    my $data_size = length($data);
+
+    croak
+      "index out of bounds on offset $offset: can't read 1, only has: $data_size"
+      if $offset + 1 > $data_size;
 
     my $r = substr( $data, $offset, 1 ) == 1;
 
@@ -139,8 +142,6 @@ sub get_uint36 {
 sub _get_bits_with_padding {
     my ( $data, $bits, $offset, $nbits ) = @_;
 
-    # TODO check if offset is in range of $data ?
-
     my ( $data_with_padding, $next_offset ) =
       _add_padding( $data, $bits, $offset, $nbits );
 
@@ -152,8 +153,11 @@ sub _get_bits_with_padding {
 sub _add_padding {
     my ( $data, $bits, $offset, $nbits ) = @_;
 
-    croak "index out of bounds on offset $offset"
-      if $offset + $nbits > length($data);
+    my $data_size = length($data);
+
+    croak
+      "index out of bounds on offset $offset: can't read $nbits, only has: $data_size"
+      if $offset + $nbits > $data_size;
 
     my $padding = "0" x ( $bits - $nbits );
 

diff --git a/lib/GDPR/IAB/TCFv2/PublisherRestrictions.pm b/lib/GDPR/IAB/TCFv2/PublisherRestrictions.pm
@@ -4,22 +4,67 @@ use warnings;
 
 use Carp qw<croak>;
 
-sub new {
+use GDPR::IAB::TCFv2::BitUtils qw<is_set
+  get_uint2
+  get_uint6
+  get_uint12
+  get_uint16
+  get_uint36
+  get_char6_pair
+>;
+
+sub Parse {
     my ( $klass, %args ) = @_;
 
-    my $restrictions = $args{restrictions}
-      or croak "missing field 'restrictions'";
+    croak "missing 'data'"      unless defined $args{data};
+    croak "missing 'data_size'" unless defined $args{data_size};
+    croak "missing 'max_id'"
+      unless defined $args{max_id};
+
+    croak "missing 'options'"      unless defined $args{options};
+    croak "missing 'options.json'" unless defined $args{options}->{json};
+
+    my $data      = $args{data};
+    my $data_size = $args{data_size};
+    my $offset    = 0;
+    my $max_id    = $args{max_id};
+    my $options   = $args{options};
+
+    my ( $num_restrictions, $next_offset ) = get_uint12( $data, $offset );
+
+    my %restrictions;
+
+    for ( 1 .. $num_restrictions ) {
+        my ( $purpose_id, $restriction_type, $vendor_restrictions );
+
+        ( $purpose_id, $next_offset ) = get_uint6( $data, $next_offset );
+
+        ( $restriction_type, $next_offset ) = get_uint2( $data, $next_offset );
+
+        ( $vendor_restrictions, $next_offset ) =
+          GDPR::IAB::TCFv2::RangeSection->Parse(
+            data      => $data,
+            data_size => $data_size,
+            offset    => $next_offset,
+            max_id    => $max_id,
+            options   => $options,
+          );
+
+        $restrictions{$purpose_id} ||= {};
+
+        $restrictions{$purpose_id}->{$restriction_type} = $vendor_restrictions;
+    }
 
     my $self = {
-        restrictions => $restrictions,
+        restrictions => \%restrictions,
     };
 
     bless $self, $klass;
 
-    return $self;
+    return wantarray ? ( $self, $next_offset ) : $self;
 }
 
-sub check_publisher_restriction {
+sub contains {
     my ( $self, $purpose_id, $restrict_type, $vendor ) = @_;
 
     return 0
@@ -62,12 +107,11 @@ GDPR::IAB::TCFv2::PublisherRestrictions - Transparency & Consent String version
 
 =head1 SYNOPSIS
 
-    my $range = GDPR::IAB::TCFv2::PublisherRestrictions->new(
-        restrictions => {
-            purpose id => {
-                restriction type => instance of GDPR::IAB::TCFv2::RangeSection
-            },
-        },
+    my ($publisher_restrictions, $next_offset) = GDPR::IAB::TCFv2::PublisherRestrictions->Parse(
+        data => $self->{data},
+        offset => $pub_restrict_offset,
+        max_id =>ASSUMED_MAX_VENDOR_ID,
+        options => $self->{options},
     );
 
     die "there is publisher restriction on purpose id 1, type 0 on vendor 284"