… | |
… | |
33 | with the added ability of supporting serialisation of Perl objects. |
33 | with the added ability of supporting serialisation of Perl objects. |
34 | (JSON often compresses better than CBOR though, so if you plan to |
34 | (JSON often compresses better than CBOR though, so if you plan to |
35 | compress the data later and speed is less important you might want to |
35 | compress the data later and speed is less important you might want to |
36 | compare both formats first). |
36 | compare both formats first). |
37 | |
37 | |
|
|
38 | The primary goal of this module is to be *correct* and the secondary |
|
|
39 | goal is to be *fast*. To reach the latter goal it was written in C. |
|
|
40 | |
38 | To give you a general idea about speed, with texts in the megabyte |
41 | To give you a general idea about speed, with texts in the megabyte |
39 | range, "CBOR::XS" usually encodes roughly twice as fast as Storable or |
42 | range, "CBOR::XS" usually encodes roughly twice as fast as Storable or |
40 | JSON::XS and decodes about 15%-30% faster than those. The shorter the |
43 | JSON::XS and decodes about 15%-30% faster than those. The shorter the |
41 | data, the worse Storable performs in comparison. |
44 | data, the worse Storable performs in comparison. |
42 | |
45 | |
… | |
… | |
47 | In addition to the core CBOR data format, this module implements a |
50 | In addition to the core CBOR data format, this module implements a |
48 | number of extensions, to support cyclic and shared data structures (see |
51 | number of extensions, to support cyclic and shared data structures (see |
49 | "allow_sharing" and "allow_cycles"), string deduplication (see |
52 | "allow_sharing" and "allow_cycles"), string deduplication (see |
50 | "pack_strings") and scalar references (always enabled). |
53 | "pack_strings") and scalar references (always enabled). |
51 | |
54 | |
52 | The primary goal of this module is to be *correct* and the secondary |
|
|
53 | goal is to be *fast*. To reach the latter goal it was written in C. |
|
|
54 | |
|
|
55 | See MAPPING, below, on how CBOR::XS maps perl values to CBOR values and |
55 | See MAPPING, below, on how CBOR::XS maps perl values to CBOR values and |
56 | vice versa. |
56 | vice versa. |
57 | |
57 | |
58 | FUNCTIONAL INTERFACE |
58 | FUNCTIONAL INTERFACE |
59 | The following convenience methods are provided by this module. They are |
59 | The following convenience methods are provided by this module. They are |
… | |
… | |
86 | but configures the coder object to be safe to use with untrusted |
86 | but configures the coder object to be safe to use with untrusted |
87 | data. Currently, this is equivalent to: |
87 | data. Currently, this is equivalent to: |
88 | |
88 | |
89 | my $cbor = CBOR::XS |
89 | my $cbor = CBOR::XS |
90 | ->new |
90 | ->new |
|
|
91 | ->validate_utf8 |
91 | ->forbid_objects |
92 | ->forbid_objects |
92 | ->filter (\&CBOR::XS::safe_filter) |
93 | ->filter (\&CBOR::XS::safe_filter) |
93 | ->max_size (1e8); |
94 | ->max_size (1e8); |
94 | |
95 | |
95 | But is more future proof (it is better to crash because of a change |
96 | But is more future proof (it is better to crash because of a change |
… | |
… | |
164 | partner supports the value sharing extensions to CBOR |
165 | partner supports the value sharing extensions to CBOR |
165 | (<http://cbor.schmorp.de/value-sharing>), as without decoder |
166 | (<http://cbor.schmorp.de/value-sharing>), as without decoder |
166 | support, the resulting data structure might be unusable. |
167 | support, the resulting data structure might be unusable. |
167 | |
168 | |
168 | Detecting shared values incurs a runtime overhead when values are |
169 | Detecting shared values incurs a runtime overhead when values are |
169 | encoded that have a reference counter large than one, and might |
170 | encoded that have a reference counter larger than one, and might |
170 | unnecessarily increase the encoded size, as potentially shared |
171 | unnecessarily increase the encoded size, as potentially shared |
171 | values are encode as shareable whether or not they are actually |
172 | values are encoded as shareable whether or not they are actually |
172 | shared. |
173 | shared. |
173 | |
174 | |
174 | At the moment, only targets of references can be shared (e.g. |
175 | At the moment, only targets of references can be shared (e.g. |
175 | scalars, arrays or hashes pointed to by a reference). Weirder |
176 | scalars, arrays or hashes pointed to by a reference). Weirder |
176 | constructs, such as an array with multiple "copies" of the *same* |
177 | constructs, such as an array with multiple "copies" of the *same* |
… | |
… | |
192 | code that isn't prepared for this will not leak memory. |
193 | code that isn't prepared for this will not leak memory. |
193 | |
194 | |
194 | If $enable is false (the default), then "decode" will throw an error |
195 | If $enable is false (the default), then "decode" will throw an error |
195 | when it encounters a self-referential/cyclic data structure. |
196 | when it encounters a self-referential/cyclic data structure. |
196 | |
197 | |
197 | FUTURE DIRECTION: the motivation behind this option is to avoid |
198 | This option does not affect "encode" in any way - shared values and |
198 | *real* cycles - future versions of this module might chose to decode |
199 | references will always be encoded properly if present. |
199 | cyclic data structures using weak references when this option is |
200 | |
200 | off, instead of throwing an error. |
201 | $cbor = $cbor->allow_weak_cycles ([$enable]) |
|
|
202 | $enabled = $cbor->get_allow_weak_cycles |
|
|
203 | This works like "allow_cycles" in that it allows the resulting data |
|
|
204 | structures to contain cycles, but unlike "allow_cycles", those |
|
|
205 | cyclic rreferences will be weak. That means that code that |
|
|
206 | recurrsively walks the data structure must be prepared with cycles, |
|
|
207 | but at least not special precautions must be implemented to free |
|
|
208 | these data structures. |
|
|
209 | |
|
|
210 | Only those references leading to actual cycles will be weakened - |
|
|
211 | other references, e.g. when the same hash or arrray is referenced |
|
|
212 | multiple times in an arrray, will be normal references. |
201 | |
213 | |
202 | This option does not affect "encode" in any way - shared values and |
214 | This option does not affect "encode" in any way - shared values and |
203 | references will always be encoded properly if present. |
215 | references will always be encoded properly if present. |
204 | |
216 | |
205 | $cbor = $cbor->forbid_objects ([$enable]) |
217 | $cbor = $cbor->forbid_objects ([$enable]) |
… | |
… | |
276 | perl strings as CBOR byte strings. |
288 | perl strings as CBOR byte strings. |
277 | |
289 | |
278 | This option does not affect "decode" in any way. |
290 | This option does not affect "decode" in any way. |
279 | |
291 | |
280 | This option has similar advantages and disadvantages as "text_keys". |
292 | This option has similar advantages and disadvantages as "text_keys". |
281 | In addition, this option effectively removes the ability to encode |
293 | In addition, this option effectively removes the ability to |
282 | byte strings, which might break some "FREEZE" and "TO_CBOR" methods |
294 | automatically encode byte strings, which might break some "FREEZE" |
283 | that rely on this, such as bignum encoding, so this option is mainly |
295 | and "TO_CBOR" methods that rely on this. |
284 | useful for very simple data. |
296 | |
|
|
297 | A workaround is to use explicit type casts, which are unaffected by |
|
|
298 | this option. |
285 | |
299 | |
286 | $cbor = $cbor->validate_utf8 ([$enable]) |
300 | $cbor = $cbor->validate_utf8 ([$enable]) |
287 | $enabled = $cbor->get_validate_utf8 |
301 | $enabled = $cbor->get_validate_utf8 |
288 | If $enable is true (or missing), then "decode" will validate that |
302 | If $enable is true (or missing), then "decode" will validate that |
289 | elements (text strings) containing UTF-8 data in fact contain valid |
303 | elements (text strings) containing UTF-8 data in fact contain valid |
… | |
… | |
396 | will silently stop parsing there and return the number of characters |
410 | will silently stop parsing there and return the number of characters |
397 | consumed so far. |
411 | consumed so far. |
398 | |
412 | |
399 | This is useful if your CBOR texts are not delimited by an outer |
413 | This is useful if your CBOR texts are not delimited by an outer |
400 | protocol and you need to know where the first CBOR string ends amd |
414 | protocol and you need to know where the first CBOR string ends amd |
401 | the next one starts. |
415 | the next one starts - CBOR strings are self-delimited, so it is |
|
|
416 | possible to concatenate CBOR strings without any delimiters or size |
|
|
417 | fields and recover their data. |
402 | |
418 | |
403 | CBOR::XS->new->decode_prefix ("......") |
419 | CBOR::XS->new->decode_prefix ("......") |
404 | => ("...", 3) |
420 | => ("...", 3) |
405 | |
421 | |
406 | INCREMENTAL PARSING |
422 | INCREMENTAL PARSING |
… | |
… | |
409 | data structure in memory at one time, it does allow you to parse a CBOR |
425 | data structure in memory at one time, it does allow you to parse a CBOR |
410 | stream incrementally, using a similar to using "decode_prefix" to see if |
426 | stream incrementally, using a similar to using "decode_prefix" to see if |
411 | a full CBOR object is available, but is much more efficient. |
427 | a full CBOR object is available, but is much more efficient. |
412 | |
428 | |
413 | It basically works by parsing as much of a CBOR string as possible - if |
429 | It basically works by parsing as much of a CBOR string as possible - if |
414 | the CBOR data is not complete yet, the pasrer will remember where it |
430 | the CBOR data is not complete yet, the parser will remember where it |
415 | was, to be able to restart when more data has been accumulated. Once |
431 | was, to be able to restart when more data has been accumulated. Once |
416 | enough data is available to either decode a complete CBOR value or raise |
432 | enough data is available to either decode a complete CBOR value or raise |
417 | an error, a real decode will be attempted. |
433 | an error, a real decode will be attempted. |
418 | |
434 | |
419 | A typical use case would be a network protocol that consists of sending |
435 | A typical use case would be a network protocol that consists of sending |
… | |
… | |
543 | "CBOR::XS::tag" to create such objects. |
559 | "CBOR::XS::tag" to create such objects. |
544 | |
560 | |
545 | Types::Serialiser::true, Types::Serialiser::false, |
561 | Types::Serialiser::true, Types::Serialiser::false, |
546 | Types::Serialiser::error |
562 | Types::Serialiser::error |
547 | These special values become CBOR true, CBOR false and CBOR undefined |
563 | These special values become CBOR true, CBOR false and CBOR undefined |
548 | values, respectively. You can also use "\1", "\0" and "\undef" |
564 | values, respectively. |
549 | directly if you want. |
|
|
550 | |
565 | |
551 | other blessed objects |
566 | other blessed objects |
552 | Other blessed objects are serialised via "TO_CBOR" or "FREEZE". See |
567 | Other blessed objects are serialised via "TO_CBOR" or "FREEZE". See |
553 | "TAG HANDLING AND EXTENSIONS" for specific classes handled by this |
568 | "TAG HANDLING AND EXTENSIONS" for specific classes handled by this |
554 | module, and "OBJECT SERIALISATION" for generic object serialisation. |
569 | module, and "OBJECT SERIALISATION" for generic object serialisation. |
… | |
… | |
579 | $x .= ""; # another, more awkward way to stringify |
594 | $x .= ""; # another, more awkward way to stringify |
580 | print $x; # perl does it for you, too, quite often |
595 | print $x; # perl does it for you, too, quite often |
581 | |
596 | |
582 | You can force whether a string is encoded as byte or text string by |
597 | You can force whether a string is encoded as byte or text string by |
583 | using "utf8::upgrade" and "utf8::downgrade" (if "text_strings" is |
598 | using "utf8::upgrade" and "utf8::downgrade" (if "text_strings" is |
584 | disabled): |
599 | disabled). |
585 | |
600 | |
586 | utf8::upgrade $x; # encode $x as text string |
601 | utf8::upgrade $x; # encode $x as text string |
587 | utf8::downgrade $x; # encode $x as byte string |
602 | utf8::downgrade $x; # encode $x as byte string |
|
|
603 | |
|
|
604 | More options are available, see "TYPE CASTS", below, and the |
|
|
605 | "text_keys" and "text_strings" options. |
588 | |
606 | |
589 | Perl doesn't define what operations up- and downgrade strings, so if |
607 | Perl doesn't define what operations up- and downgrade strings, so if |
590 | the difference between byte and text is important, you should up- or |
608 | the difference between byte and text is important, you should up- or |
591 | downgrade your string as late as possible before encoding. You can |
609 | downgrade your string as late as possible before encoding. You can |
592 | also force the use of CBOR text strings by using "text_keys" or |
610 | also force the use of CBOR text strings by using "text_keys" or |
… | |
… | |
606 | possible representation. Floating-point values will use either the |
624 | possible representation. Floating-point values will use either the |
607 | IEEE single format if possible without loss of precision, otherwise |
625 | IEEE single format if possible without loss of precision, otherwise |
608 | the IEEE double format will be used. Perls that use formats other |
626 | the IEEE double format will be used. Perls that use formats other |
609 | than IEEE double to represent numerical values are supported, but |
627 | than IEEE double to represent numerical values are supported, but |
610 | might suffer loss of precision. |
628 | might suffer loss of precision. |
|
|
629 | |
|
|
630 | TYPE CASTS |
|
|
631 | EXPERIMENTAL: As an experimental extension, "CBOR::XS" allows you to |
|
|
632 | force specific CBOR types to be used when encoding. That allows you to |
|
|
633 | encode types not normally accessible (e.g. half floats) as well as force |
|
|
634 | string types even when "text_strings" is in effect. |
|
|
635 | |
|
|
636 | Type forcing is done by calling a special "cast" function which keeps a |
|
|
637 | copy of the value and returns a new value that can be handed over to any |
|
|
638 | CBOR encoder function. |
|
|
639 | |
|
|
640 | The following casts are currently available (all of which are unary |
|
|
641 | operators, that is, have a prototype of "$"): |
|
|
642 | |
|
|
643 | CBOR::XS::as_int $value |
|
|
644 | Forces the value to be encoded as some form of (basic, not bignum) |
|
|
645 | integer type. |
|
|
646 | |
|
|
647 | CBOR::XS::as_text $value |
|
|
648 | Forces the value to be encoded as (UTF-8) text values. |
|
|
649 | |
|
|
650 | CBOR::XS::as_bytes $value |
|
|
651 | Forces the value to be encoded as a (binary) string value. |
|
|
652 | |
|
|
653 | Example: encode a perl string as binary even though "text_strings" |
|
|
654 | is in effect. |
|
|
655 | |
|
|
656 | CBOR::XS->new->text_strings->encode ([4, "text", CBOR::XS::bytes "bytevalue"]); |
|
|
657 | |
|
|
658 | CBOR::XS::as_bool $value |
|
|
659 | Converts a Perl boolean (which can be any kind of scalar) into a |
|
|
660 | CBOR boolean. Strictly the same, but shorter to write, than: |
|
|
661 | |
|
|
662 | $value ? Types::Serialiser::true : Types::Serialiser::false |
|
|
663 | |
|
|
664 | CBOR::XS::as_float16 $value |
|
|
665 | Forces half-float (IEEE 754 binary16) encoding of the given value. |
|
|
666 | |
|
|
667 | CBOR::XS::as_float32 $value |
|
|
668 | Forces single-float (IEEE 754 binary32) encoding of the given value. |
|
|
669 | |
|
|
670 | CBOR::XS::as_float64 $value |
|
|
671 | Forces double-float (IEEE 754 binary64) encoding of the given value. |
|
|
672 | |
|
|
673 | CBOR::XS::as_cbor $cbor_text |
|
|
674 | Not a type cast per-se, this type cast forces the argument to be |
|
|
675 | encoded as-is. This can be used to embed pre-encoded CBOR data. |
|
|
676 | |
|
|
677 | Note that no checking on the validity of the $cbor_text is done - |
|
|
678 | it's the callers responsibility to correctly encode values. |
|
|
679 | |
|
|
680 | CBOR::XS::as_map [key => value...] |
|
|
681 | Treat the array reference as key value pairs and output a CBOR map. |
|
|
682 | This allows you to generate CBOR maps with arbitrary key types (or, |
|
|
683 | if you don't care about semantics, duplicate keys or pairs in a |
|
|
684 | custom order), which is otherwise hard to do with Perl. |
|
|
685 | |
|
|
686 | The single argument must be an array reference with an even number |
|
|
687 | of elements. |
|
|
688 | |
|
|
689 | Note that only the reference to the array is copied, the array |
|
|
690 | itself is not. Modifications done to the array before calling an |
|
|
691 | encoding function will be reflected in the encoded output. |
|
|
692 | |
|
|
693 | Example: encode a CBOR map with a string and an integer as keys. |
|
|
694 | |
|
|
695 | encode_cbor CBOR::XS::as_map [string => "value", 5 => "value"] |
611 | |
696 | |
612 | OBJECT SERIALISATION |
697 | OBJECT SERIALISATION |
613 | This module implements both a CBOR-specific and the generic |
698 | This module implements both a CBOR-specific and the generic |
614 | Types::Serialier object serialisation protocol. The following |
699 | Types::Serialier object serialisation protocol. The following |
615 | subsections explain both methods. |
700 | subsections explain both methods. |
… | |
… | |
907 | ensure that decoded JSON data will round-trip encoding and decoding to |
992 | ensure that decoded JSON data will round-trip encoding and decoding to |
908 | CBOR intact. |
993 | CBOR intact. |
909 | |
994 | |
910 | SECURITY CONSIDERATIONS |
995 | SECURITY CONSIDERATIONS |
911 | Tl;dr... if you want to decode or encode CBOR from untrusted sources, |
996 | Tl;dr... if you want to decode or encode CBOR from untrusted sources, |
912 | you should start with a coder object created via "new_safe": |
997 | you should start with a coder object created via "new_safe" (which |
|
|
998 | implements the mitigations explained below): |
913 | |
999 | |
914 | my $coder = CBOR::XS->new_safe; |
1000 | my $coder = CBOR::XS->new_safe; |
915 | |
1001 | |
916 | my $data = $coder->decode ($cbor_text); |
1002 | my $data = $coder->decode ($cbor_text); |
917 | my $cbor = $coder->encode ($data); |
1003 | my $cbor = $coder->encode ($data); |
… | |
… | |
936 | even if all your "THAW" methods are secure, encoding data structures |
1022 | even if all your "THAW" methods are secure, encoding data structures |
937 | from untrusted sources can invoke those and trigger bugs in those. |
1023 | from untrusted sources can invoke those and trigger bugs in those. |
938 | |
1024 | |
939 | So, if you are not sure about the security of all the modules you |
1025 | So, if you are not sure about the security of all the modules you |
940 | have loaded (you shouldn't), you should disable this part using |
1026 | have loaded (you shouldn't), you should disable this part using |
941 | "forbid_objects". |
1027 | "forbid_objects" or using "new_safe". |
942 | |
1028 | |
943 | CBOR can be extended with tags that call library code |
1029 | CBOR can be extended with tags that call library code |
944 | CBOR can be extended with tags, and "CBOR::XS" has a registry of |
1030 | CBOR can be extended with tags, and "CBOR::XS" has a registry of |
945 | conversion functions for many existing tags that can be extended via |
1031 | conversion functions for many existing tags that can be extended via |
946 | third-party modules (see the "filter" method). |
1032 | third-party modules (see the "filter" method). |
947 | |
1033 | |
948 | If you don't trust these, you should configure the "safe" filter |
1034 | If you don't trust these, you should configure the "safe" filter |
949 | function, "CBOR::XS::safe_filter", which by default only includes |
1035 | function, "CBOR::XS::safe_filter" ("new_safe" does this), which by |
950 | conversion functions that are considered "safe" by the author (but |
1036 | default only includes conversion functions that are considered |
951 | again, they can be extended by third party modules). |
1037 | "safe" by the author (but again, they can be extended by third party |
|
|
1038 | modules). |
952 | |
1039 | |
953 | Depending on your level of paranoia, you can use the "safe" filter: |
1040 | Depending on your level of paranoia, you can use the "safe" filter: |
954 | |
1041 | |
955 | $cbor->filter (\&CBOR::XS::safe_filter); |
1042 | $cbor->filter (\&CBOR::XS::safe_filter); |
956 | |
1043 | |
… | |
… | |
970 | limit the size of CBOR data you accept, or make sure then when your |
1057 | limit the size of CBOR data you accept, or make sure then when your |
971 | resources run out, that's just fine (e.g. by using a separate |
1058 | resources run out, that's just fine (e.g. by using a separate |
972 | process that can crash safely). The size of a CBOR string in octets |
1059 | process that can crash safely). The size of a CBOR string in octets |
973 | is usually a good indication of the size of the resources required |
1060 | is usually a good indication of the size of the resources required |
974 | to decode it into a Perl structure. While CBOR::XS can check the |
1061 | to decode it into a Perl structure. While CBOR::XS can check the |
975 | size of the CBOR text (using "max_size"), it might be too late when |
1062 | size of the CBOR text (using "max_size" - done by "new_safe"), it |
976 | you already have it in memory, so you might want to check the size |
1063 | might be too late when you already have it in memory, so you might |
977 | before you accept the string. |
1064 | want to check the size before you accept the string. |
978 | |
1065 | |
979 | As for encoding, it is possible to construct data structures that |
1066 | As for encoding, it is possible to construct data structures that |
980 | are relatively small but result in large CBOR texts (for example by |
1067 | are relatively small but result in large CBOR texts (for example by |
981 | having an array full of references to the same big data structure, |
1068 | having an array full of references to the same big data structure, |
982 | which will all be deep-cloned during encoding by default). This is |
1069 | which will all be deep-cloned during encoding by default). This is |
… | |
… | |
996 | |
1083 | |
997 | Resource-starving attacks: CPU en-/decoding complexity |
1084 | Resource-starving attacks: CPU en-/decoding complexity |
998 | CBOR::XS will use the Math::BigInt, Math::BigFloat and Math::BigRat |
1085 | CBOR::XS will use the Math::BigInt, Math::BigFloat and Math::BigRat |
999 | libraries to represent encode/decode bignums. These can be very slow |
1086 | libraries to represent encode/decode bignums. These can be very slow |
1000 | (as in, centuries of CPU time) and can even crash your program (and |
1087 | (as in, centuries of CPU time) and can even crash your program (and |
1001 | are generally not very trustworthy). See the next section for |
1088 | are generally not very trustworthy). See the next section on bignum |
1002 | details. |
1089 | security for details. |
1003 | |
1090 | |
1004 | Data breaches: leaking information in error messages |
1091 | Data breaches: leaking information in error messages |
1005 | CBOR::XS might leak contents of your Perl data structures in its |
1092 | CBOR::XS might leak contents of your Perl data structures in its |
1006 | error messages, so when you serialise sensitive information you |
1093 | error messages, so when you serialise sensitive information you |
1007 | might want to make sure that exceptions thrown by CBOR::XS will not |
1094 | might want to make sure that exceptions thrown by CBOR::XS will not |
… | |
… | |
1060 | |
1147 | |
1061 | LIMITATIONS ON PERLS WITHOUT 64-BIT INTEGER SUPPORT |
1148 | LIMITATIONS ON PERLS WITHOUT 64-BIT INTEGER SUPPORT |
1062 | On perls that were built without 64 bit integer support (these are rare |
1149 | On perls that were built without 64 bit integer support (these are rare |
1063 | nowadays, even on 32 bit architectures, as all major Perl distributions |
1150 | nowadays, even on 32 bit architectures, as all major Perl distributions |
1064 | are built with 64 bit integer support), support for any kind of 64 bit |
1151 | are built with 64 bit integer support), support for any kind of 64 bit |
1065 | integer in CBOR is very limited - most likely, these 64 bit values will |
1152 | value in CBOR is very limited - most likely, these 64 bit values will be |
1066 | be truncated, corrupted, or otherwise not decoded correctly. This also |
1153 | truncated, corrupted, or otherwise not decoded correctly. This also |
1067 | includes string, array and map sizes that are stored as 64 bit integers. |
1154 | includes string, float, array and map sizes that are stored as 64 bit |
|
|
1155 | integers. |
1068 | |
1156 | |
1069 | THREADS |
1157 | THREADS |
1070 | This module is *not* guaranteed to be thread safe and there are no plans |
1158 | This module is *not* guaranteed to be thread safe and there are no plans |
1071 | to change this until Perl gets thread support (as opposed to the |
1159 | to change this until Perl gets thread support (as opposed to the |
1072 | horribly slow so-called "threads" which are simply slow and bloated |
1160 | horribly slow so-called "threads" which are simply slow and bloated |