… | |
… | |
64 | |
64 | |
65 | package CBOR::XS; |
65 | package CBOR::XS; |
66 | |
66 | |
67 | use common::sense; |
67 | use common::sense; |
68 | |
68 | |
69 | our $VERSION = 1.41; |
69 | our $VERSION = 1.5; |
70 | our @ISA = qw(Exporter); |
70 | our @ISA = qw(Exporter); |
71 | |
71 | |
72 | our @EXPORT = qw(encode_cbor decode_cbor); |
72 | our @EXPORT = qw(encode_cbor decode_cbor); |
73 | |
73 | |
74 | use Exporter; |
74 | use Exporter; |
… | |
… | |
247 | the standard CBOR way. |
247 | the standard CBOR way. |
248 | |
248 | |
249 | This option does not affect C<decode> in any way - string references will |
249 | This option does not affect C<decode> in any way - string references will |
250 | always be decoded properly if present. |
250 | always be decoded properly if present. |
251 | |
251 | |
|
|
252 | =item $cbor = $cbor->text_keys ([$enable]) |
|
|
253 | |
|
|
254 | =item $enabled = $cbor->get_text_keys |
|
|
255 | |
|
|
256 | If C<$enabled> is true (or missing), then C<encode> will encode all |
|
|
257 | perl hash keys as CBOR text strings/UTF-8 string, upgrading them as needed. |
|
|
258 | |
|
|
259 | If C<$enable> is false (the default), then C<encode> will encode hash keys |
|
|
260 | normally - upgraded perl strings (strings internally encoded as UTF-8) as |
|
|
261 | CBOR text strings, and downgraded perl strings as CBOR byte strings. |
|
|
262 | |
|
|
263 | This option does not affect C<decode> in any way. |
|
|
264 | |
|
|
265 | This option is useful for interoperability with CBOR decoders that don't |
|
|
266 | treat byte strings as a form of text. It is especially useful as Perl |
|
|
267 | gives very little control over hash keys. |
|
|
268 | |
|
|
269 | Enabling this option can be slow, as all downgraded hash keys that are |
|
|
270 | encoded need to be scanned and converted to UTF-8. |
|
|
271 | |
|
|
272 | =item $cbor = $cbor->text_strings ([$enable]) |
|
|
273 | |
|
|
274 | =item $enabled = $cbor->get_text_strings |
|
|
275 | |
|
|
276 | This option works similar to C<text_keys>, above, but works on all strings |
|
|
277 | (including hash keys), so C<text_keys> has no further effect after |
|
|
278 | enabling C<text_strings>. |
|
|
279 | |
|
|
280 | If C<$enabled> is true (or missing), then C<encode> will encode all perl |
|
|
281 | strings as CBOR text strings/UTF-8 strings, upgrading them as needed. |
|
|
282 | |
|
|
283 | If C<$enable> is false (the default), then C<encode> will encode strings |
|
|
284 | normally (but see C<text_keys>) - upgraded perl strings (strings |
|
|
285 | internally encoded as UTF-8) as CBOR text strings, and downgraded perl |
|
|
286 | strings as CBOR byte strings. |
|
|
287 | |
|
|
288 | This option does not affect C<decode> in any way. |
|
|
289 | |
|
|
290 | This option has similar advantages and disadvantages as C<text_keys>. In |
|
|
291 | addition, this option effectively removes the ability to encode byte |
|
|
292 | strings, which might break some C<FREEZE> and C<TO_CBOR> methods that rely |
|
|
293 | on this, such as bignum encoding, so this option is mainly useful for very |
|
|
294 | simple data. |
|
|
295 | |
252 | =item $cbor = $cbor->validate_utf8 ([$enable]) |
296 | =item $cbor = $cbor->validate_utf8 ([$enable]) |
253 | |
297 | |
254 | =item $enabled = $cbor->get_validate_utf8 |
298 | =item $enabled = $cbor->get_validate_utf8 |
255 | |
299 | |
256 | If C<$enable> is true (or missing), then C<decode> will validate that |
300 | If C<$enable> is true (or missing), then C<decode> will validate that |
… | |
… | |
261 | The concept of "valid UTF-8" used is perl's concept, which is a superset |
305 | The concept of "valid UTF-8" used is perl's concept, which is a superset |
262 | of the official UTF-8. |
306 | of the official UTF-8. |
263 | |
307 | |
264 | If C<$enable> is false (the default), then C<decode> will blindly accept |
308 | If C<$enable> is false (the default), then C<decode> will blindly accept |
265 | UTF-8 data, marking them as valid UTF-8 in the resulting data structure |
309 | UTF-8 data, marking them as valid UTF-8 in the resulting data structure |
266 | regardless of whether thats true or not. |
310 | regardless of whether that's true or not. |
267 | |
311 | |
268 | Perl isn't too happy about corrupted UTF-8 in strings, but should |
312 | Perl isn't too happy about corrupted UTF-8 in strings, but should |
269 | generally not crash or do similarly evil things. Extensions might be not |
313 | generally not crash or do similarly evil things. Extensions might be not |
270 | so forgiving, so it's recommended to turn on this setting if you receive |
314 | so forgiving, so it's recommended to turn on this setting if you receive |
271 | untrusted CBOR. |
315 | untrusted CBOR. |
… | |
… | |
481 | |
525 | |
482 | =item hash references |
526 | =item hash references |
483 | |
527 | |
484 | Perl hash references become CBOR maps. As there is no inherent ordering in |
528 | Perl hash references become CBOR maps. As there is no inherent ordering in |
485 | hash keys (or CBOR maps), they will usually be encoded in a pseudo-random |
529 | hash keys (or CBOR maps), they will usually be encoded in a pseudo-random |
486 | order. This order can be different each time a hahs is encoded. |
530 | order. This order can be different each time a hash is encoded. |
487 | |
531 | |
488 | Currently, tied hashes will use the indefinite-length format, while normal |
532 | Currently, tied hashes will use the indefinite-length format, while normal |
489 | hashes will use the fixed-length format. |
533 | hashes will use the fixed-length format. |
490 | |
534 | |
491 | =item array references |
535 | =item array references |
… | |
… | |
544 | my $x = 3.1; # some variable containing a number |
588 | my $x = 3.1; # some variable containing a number |
545 | "$x"; # stringified |
589 | "$x"; # stringified |
546 | $x .= ""; # another, more awkward way to stringify |
590 | $x .= ""; # another, more awkward way to stringify |
547 | print $x; # perl does it for you, too, quite often |
591 | print $x; # perl does it for you, too, quite often |
548 | |
592 | |
549 | You can force whether a string ie encoded as byte or text string by using |
593 | You can force whether a string is encoded as byte or text string by using |
550 | C<utf8::upgrade> and C<utf8::downgrade>): |
594 | C<utf8::upgrade> and C<utf8::downgrade> (if C<text_strings> is disabled): |
551 | |
595 | |
552 | utf8::upgrade $x; # encode $x as text string |
596 | utf8::upgrade $x; # encode $x as text string |
553 | utf8::downgrade $x; # encode $x as byte string |
597 | utf8::downgrade $x; # encode $x as byte string |
554 | |
598 | |
555 | Perl doesn't define what operations up- and downgrade strings, so if the |
599 | Perl doesn't define what operations up- and downgrade strings, so if the |
556 | difference between byte and text is important, you should up- or downgrade |
600 | difference between byte and text is important, you should up- or downgrade |
557 | your string as late as possible before encoding. |
601 | your string as late as possible before encoding. You can also force the |
|
|
602 | use of CBOR text strings by using C<text_keys> or C<text_strings>. |
558 | |
603 | |
559 | You can force the type to be a CBOR number by numifying it: |
604 | You can force the type to be a CBOR number by numifying it: |
560 | |
605 | |
561 | my $x = "3"; # some variable containing a string |
606 | my $x = "3"; # some variable containing a string |
562 | $x += 0; # numify it, ensuring it will be dumped as a number |
607 | $x += 0; # numify it, ensuring it will be dumped as a number |
… | |
… | |
663 | "$self" # encode url string |
708 | "$self" # encode url string |
664 | } |
709 | } |
665 | |
710 | |
666 | sub URI::THAW { |
711 | sub URI::THAW { |
667 | my ($class, $serialiser, $uri) = @_; |
712 | my ($class, $serialiser, $uri) = @_; |
668 | |
|
|
669 | $class->new ($uri) |
713 | $class->new ($uri) |
670 | } |
714 | } |
671 | |
715 | |
672 | Unlike C<TO_CBOR>, multiple values can be returned by C<FREEZE>. For |
716 | Unlike C<TO_CBOR>, multiple values can be returned by C<FREEZE>. For |
673 | example, a C<FREEZE> method that returns "type", "id" and "variant" values |
717 | example, a C<FREEZE> method that returns "type", "id" and "variant" values |
… | |
… | |
804 | additional tags (such as base64url). |
848 | additional tags (such as base64url). |
805 | |
849 | |
806 | =head2 ENFORCED TAGS |
850 | =head2 ENFORCED TAGS |
807 | |
851 | |
808 | These tags are always handled when decoding, and their handling cannot be |
852 | These tags are always handled when decoding, and their handling cannot be |
809 | overriden by the user. |
853 | overridden by the user. |
810 | |
854 | |
811 | =over 4 |
855 | =over 4 |
812 | |
856 | |
813 | =item 26 (perl-object, L<http://cbor.schmorp.de/perl-object>) |
857 | =item 26 (perl-object, L<http://cbor.schmorp.de/perl-object>) |
814 | |
858 | |
… | |
… | |
842 | encoded, however, when C<pack_strings> is enabled. |
886 | encoded, however, when C<pack_strings> is enabled. |
843 | |
887 | |
844 | =item 22098 (indirection, L<http://cbor.schmorp.de/indirection>) |
888 | =item 22098 (indirection, L<http://cbor.schmorp.de/indirection>) |
845 | |
889 | |
846 | This tag is automatically generated when a reference are encountered (with |
890 | This tag is automatically generated when a reference are encountered (with |
847 | the exception of hash and array refernces). It is converted to a reference |
891 | the exception of hash and array references). It is converted to a reference |
848 | when decoding. |
892 | when decoding. |
849 | |
893 | |
850 | =item 55799 (self-describe CBOR, RFC 7049) |
894 | =item 55799 (self-describe CBOR, RFC 7049) |
851 | |
895 | |
852 | This value is not generated on encoding (unless explicitly requested by |
896 | This value is not generated on encoding (unless explicitly requested by |
… | |
… | |
855 | =back |
899 | =back |
856 | |
900 | |
857 | =head2 NON-ENFORCED TAGS |
901 | =head2 NON-ENFORCED TAGS |
858 | |
902 | |
859 | These tags have default filters provided when decoding. Their handling can |
903 | These tags have default filters provided when decoding. Their handling can |
860 | be overriden by changing the C<%CBOR::XS::FILTER> entry for the tag, or by |
904 | be overridden by changing the C<%CBOR::XS::FILTER> entry for the tag, or by |
861 | providing a custom C<filter> callback when decoding. |
905 | providing a custom C<filter> callback when decoding. |
862 | |
906 | |
863 | When they result in decoding into a specific Perl class, the module |
907 | When they result in decoding into a specific Perl class, the module |
864 | usually provides a corresponding C<TO_CBOR> method as well. |
908 | usually provides a corresponding C<TO_CBOR> method as well. |
865 | |
909 | |
… | |
… | |
883 | |
927 | |
884 | These tags are decoded into L<Math::BigInt> objects. The corresponding |
928 | These tags are decoded into L<Math::BigInt> objects. The corresponding |
885 | C<Math::BigInt::TO_CBOR> method encodes "small" bigints into normal CBOR |
929 | C<Math::BigInt::TO_CBOR> method encodes "small" bigints into normal CBOR |
886 | integers, and others into positive/negative CBOR bignums. |
930 | integers, and others into positive/negative CBOR bignums. |
887 | |
931 | |
888 | =item 4, 5 (decimal fraction/bigfloat) |
932 | =item 4, 5, 264, 265 (decimal fraction/bigfloat) |
889 | |
933 | |
890 | Both decimal fractions and bigfloats are decoded into L<Math::BigFloat> |
934 | Both decimal fractions and bigfloats are decoded into L<Math::BigFloat> |
891 | objects. The corresponding C<Math::BigFloat::TO_CBOR> method I<always> |
935 | objects. The corresponding C<Math::BigFloat::TO_CBOR> method I<always> |
892 | encodes into a decimal fraction. |
936 | encodes into a decimal fraction (either tag 4 or 264). |
893 | |
937 | |
894 | CBOR cannot represent bigfloats with I<very> large exponents - conversion |
938 | NaN and infinities are not encoded properly, as they cannot be represented |
895 | of such big float objects is undefined. |
939 | in CBOR. |
896 | |
940 | |
897 | Also, NaN and infinities are not encoded properly. |
941 | See L<BIGNUM SECURITY CONSIDERATIONS> for more info. |
898 | |
942 | |
899 | =item 21, 22, 23 (expected later JSON conversion) |
943 | =item 21, 22, 23 (expected later JSON conversion) |
900 | |
944 | |
901 | CBOR::XS is not a CBOR-to-JSON converter, and will simply ignore these |
945 | CBOR::XS is not a CBOR-to-JSON converter, and will simply ignore these |
902 | tags. |
946 | tags. |
… | |
… | |
907 | C<URI::TO_CBOR> method again results in a CBOR URI value. |
951 | C<URI::TO_CBOR> method again results in a CBOR URI value. |
908 | |
952 | |
909 | =back |
953 | =back |
910 | |
954 | |
911 | =cut |
955 | =cut |
912 | |
|
|
913 | our %FILTER = ( |
|
|
914 | # 0 # rfc4287 datetime, utf-8 |
|
|
915 | # 1 # unix timestamp, any |
|
|
916 | |
|
|
917 | 2 => sub { # pos bigint |
|
|
918 | require Math::BigInt; |
|
|
919 | Math::BigInt->new ("0x" . unpack "H*", pop) |
|
|
920 | }, |
|
|
921 | |
|
|
922 | 3 => sub { # neg bigint |
|
|
923 | require Math::BigInt; |
|
|
924 | -Math::BigInt->new ("0x" . unpack "H*", pop) |
|
|
925 | }, |
|
|
926 | |
|
|
927 | 4 => sub { # decimal fraction, array |
|
|
928 | require Math::BigFloat; |
|
|
929 | Math::BigFloat->new ($_[1][1] . "E" . $_[1][0]) |
|
|
930 | }, |
|
|
931 | |
|
|
932 | 5 => sub { # bigfloat, array |
|
|
933 | require Math::BigFloat; |
|
|
934 | scalar Math::BigFloat->new ($_[1][1])->blsft ($_[1][0], 2) |
|
|
935 | }, |
|
|
936 | |
|
|
937 | 21 => sub { pop }, # expected conversion to base64url encoding |
|
|
938 | 22 => sub { pop }, # expected conversion to base64 encoding |
|
|
939 | 23 => sub { pop }, # expected conversion to base16 encoding |
|
|
940 | |
|
|
941 | # 24 # embedded cbor, byte string |
|
|
942 | |
|
|
943 | 32 => sub { |
|
|
944 | require URI; |
|
|
945 | URI->new (pop) |
|
|
946 | }, |
|
|
947 | |
|
|
948 | # 33 # base64url rfc4648, utf-8 |
|
|
949 | # 34 # base64 rfc46484, utf-8 |
|
|
950 | # 35 # regex pcre/ecma262, utf-8 |
|
|
951 | # 36 # mime message rfc2045, utf-8 |
|
|
952 | ); |
|
|
953 | |
|
|
954 | |
956 | |
955 | =head1 CBOR and JSON |
957 | =head1 CBOR and JSON |
956 | |
958 | |
957 | CBOR is supposed to implement a superset of the JSON data model, and is, |
959 | CBOR is supposed to implement a superset of the JSON data model, and is, |
958 | with some coercion, able to represent all JSON texts (something that other |
960 | with some coercion, able to represent all JSON texts (something that other |
… | |
… | |
999 | |
1001 | |
1000 | Also keep in mind that CBOR::XS might leak contents of your Perl data |
1002 | Also keep in mind that CBOR::XS might leak contents of your Perl data |
1001 | structures in its error messages, so when you serialise sensitive |
1003 | structures in its error messages, so when you serialise sensitive |
1002 | information you might want to make sure that exceptions thrown by CBOR::XS |
1004 | information you might want to make sure that exceptions thrown by CBOR::XS |
1003 | will not end up in front of untrusted eyes. |
1005 | will not end up in front of untrusted eyes. |
|
|
1006 | |
|
|
1007 | |
|
|
1008 | =head1 BIGNUM SECURITY CONSIDERATIONS |
|
|
1009 | |
|
|
1010 | CBOR::XS provides a C<TO_CBOR> method for both L<Math::BigInt> and |
|
|
1011 | L<Math::BigFloat> that tries to encode the number in the simplest possible |
|
|
1012 | way, that is, either a CBOR integer, a CBOR bigint/decimal fraction (tag |
|
|
1013 | 4) or an arbitrary-exponent decimal fraction (tag 264). |
|
|
1014 | |
|
|
1015 | It will also understand base-2 bigfloat or arbitrary-exponent bigfloats |
|
|
1016 | (tags 5 and 265), but it will never generate these on its own. |
|
|
1017 | |
|
|
1018 | Using the built-in L<Math::BigInt::Calc> support, encoding and decoding |
|
|
1019 | decimal fractions is generally fast. Decoding bigints can be slow for very |
|
|
1020 | big numbers, and decoding bigfloats or arbitrary-exponent bigfloats can be |
|
|
1021 | extremely slow (minutes, decades) for large exponents. |
|
|
1022 | |
|
|
1023 | Additionally, L<Math::BigInt> can take advantage of other bignum |
|
|
1024 | libraries, such as L<Math::GMP>, which cannot handle big |
|
|
1025 | floats with large exponents, and might simply abort or crash your program, |
|
|
1026 | due to their code quality. |
|
|
1027 | |
|
|
1028 | This can be a concern if you want to parse untrusted CBOR. If it is, you |
|
|
1029 | need to disable decoding of tag 2 (bigint) and 3 (negative bigint) types, |
|
|
1030 | which will also disable bigfloat support (to be sure, you can also disable |
|
|
1031 | types 4, 5, 264 and 265). |
|
|
1032 | |
1004 | |
1033 | |
1005 | =head1 CBOR IMPLEMENTATION NOTES |
1034 | =head1 CBOR IMPLEMENTATION NOTES |
1006 | |
1035 | |
1007 | This section contains some random implementation notes. They do not |
1036 | This section contains some random implementation notes. They do not |
1008 | describe guaranteed behaviour, but merely behaviour as-is implemented |
1037 | describe guaranteed behaviour, but merely behaviour as-is implemented |
… | |
… | |
1081 | scalar Time::Piece::gmtime (pop) |
1110 | scalar Time::Piece::gmtime (pop) |
1082 | }, |
1111 | }, |
1083 | |
1112 | |
1084 | 2 => sub { # pos bigint |
1113 | 2 => sub { # pos bigint |
1085 | require Math::BigInt; |
1114 | require Math::BigInt; |
1086 | Math::BigInt->new ("0x" . unpack "H*", pop) |
1115 | Math::BigInt->from_hex ("0x" . unpack "H*", pop) |
1087 | }, |
1116 | }, |
1088 | |
1117 | |
1089 | 3 => sub { # neg bigint |
1118 | 3 => sub { # neg bigint |
1090 | require Math::BigInt; |
1119 | require Math::BigInt; |
1091 | -Math::BigInt->new ("0x" . unpack "H*", pop) |
1120 | -Math::BigInt->from_hex ("0x" . unpack "H*", pop) |
1092 | }, |
1121 | }, |
1093 | |
1122 | |
1094 | 4 => sub { # decimal fraction, array |
1123 | 4 => sub { # decimal fraction, array |
1095 | require Math::BigFloat; |
1124 | require Math::BigFloat; |
1096 | Math::BigFloat->new ($_[1][1] . "E" . $_[1][0]) |
1125 | Math::BigFloat->new ($_[1][1] . "E" . $_[1][0]) |
1097 | }, |
1126 | }, |
1098 | |
1127 | |
1099 | 5 => sub { # bigfloat, array |
1128 | 5 => sub { # bigfloat, array |
1100 | require Math::BigFloat; |
1129 | require Math::BigFloat; |
1101 | scalar Math::BigFloat->new ($_[1][1])->blsft ($_[1][0], 2) |
1130 | scalar Math::BigFloat->new ($_[1][1]) * Math::BigFloat->new (2)->bpow ($_[1][0]) |
1102 | }, |
1131 | }, |
1103 | |
1132 | |
1104 | 21 => sub { pop }, # expected conversion to base64url encoding |
1133 | 21 => sub { pop }, # expected conversion to base64url encoding |
1105 | 22 => sub { pop }, # expected conversion to base64 encoding |
1134 | 22 => sub { pop }, # expected conversion to base64 encoding |
1106 | 23 => sub { pop }, # expected conversion to base16 encoding |
1135 | 23 => sub { pop }, # expected conversion to base16 encoding |
… | |
… | |
1114 | |
1143 | |
1115 | # 33 # base64url rfc4648, utf-8 |
1144 | # 33 # base64url rfc4648, utf-8 |
1116 | # 34 # base64 rfc46484, utf-8 |
1145 | # 34 # base64 rfc46484, utf-8 |
1117 | # 35 # regex pcre/ecma262, utf-8 |
1146 | # 35 # regex pcre/ecma262, utf-8 |
1118 | # 36 # mime message rfc2045, utf-8 |
1147 | # 36 # mime message rfc2045, utf-8 |
|
|
1148 | |
|
|
1149 | 264 => sub { # decimal fraction with arbitrary exponent |
|
|
1150 | require Math::BigFloat; |
|
|
1151 | Math::BigFloat->new ($_[1][1] . "E" . $_[1][0]) |
|
|
1152 | }, |
|
|
1153 | |
|
|
1154 | 265 => sub { # bigfloat with arbitrary exponent |
|
|
1155 | require Math::BigFloat; |
|
|
1156 | scalar Math::BigFloat->new ($_[1][1]) * Math::BigFloat->new (2)->bpow ($_[1][0]) |
|
|
1157 | }, |
1119 | ); |
1158 | ); |
1120 | |
1159 | |
1121 | sub CBOR::XS::default_filter { |
1160 | sub CBOR::XS::default_filter { |
1122 | &{ $FILTER{$_[0]} or return } |
1161 | &{ $FILTER{$_[0]} or return } |
1123 | } |
1162 | } |
… | |
… | |
1127 | utf8::upgrade $uri; |
1166 | utf8::upgrade $uri; |
1128 | tag 32, $uri |
1167 | tag 32, $uri |
1129 | } |
1168 | } |
1130 | |
1169 | |
1131 | sub Math::BigInt::TO_CBOR { |
1170 | sub Math::BigInt::TO_CBOR { |
1132 | if ($_[0] >= -2147483648 && $_[0] <= 2147483647) { |
1171 | if (-2147483648 <= $_[0] && $_[0] <= 2147483647) { |
1133 | $_[0]->numify |
1172 | $_[0]->numify |
1134 | } else { |
1173 | } else { |
1135 | my $hex = substr $_[0]->as_hex, 2; |
1174 | my $hex = substr $_[0]->as_hex, 2; |
1136 | $hex = "0$hex" if 1 & length $hex; # sigh |
1175 | $hex = "0$hex" if 1 & length $hex; # sigh |
1137 | tag $_[0] >= 0 ? 2 : 3, pack "H*", $hex |
1176 | tag $_[0] >= 0 ? 2 : 3, pack "H*", $hex |
1138 | } |
1177 | } |
1139 | } |
1178 | } |
1140 | |
1179 | |
1141 | sub Math::BigFloat::TO_CBOR { |
1180 | sub Math::BigFloat::TO_CBOR { |
1142 | my ($m, $e) = $_[0]->parts; |
1181 | my ($m, $e) = $_[0]->parts; |
|
|
1182 | |
|
|
1183 | -9223372036854775808 <= $e && $e <= 18446744073709551615 |
1143 | tag 4, [$e->numify, $m] |
1184 | ? tag 4, [$e->numify, $m] |
|
|
1185 | : tag 264, [$e, $m] |
1144 | } |
1186 | } |
1145 | |
1187 | |
1146 | sub Time::Piece::TO_CBOR { |
1188 | sub Time::Piece::TO_CBOR { |
1147 | tag 1, 0 + $_[0]->epoch |
1189 | tag 1, 0 + $_[0]->epoch |
1148 | } |
1190 | } |