--- CBOR-XS/XS.pm 2013/11/20 16:29:02 1.21 +++ CBOR-XS/XS.pm 2016/02/08 04:11:11 1.46 @@ -28,38 +28,31 @@ =head1 DESCRIPTION -WARNING! This module is very new, and not very well tested (that's up -to you to do). Furthermore, details of the implementation might change -freely before version 1.0. And lastly, most extensions depend on an IANA -assignment, and until that assignment is official, this implementation is -not interoperable with other implementations (even future versions of this -module) until the assignment is done. - -You are still invited to try out CBOR, and this module. - This module converts Perl data structures to the Concise Binary Object Representation (CBOR) and vice versa. CBOR is a fast binary serialisation -format that aims to use a superset of the JSON data model, i.e. when you -can represent something in JSON, you should be able to represent it in -CBOR. +format that aims to use an (almost) superset of the JSON data model, i.e. +when you can represent something useful in JSON, you should be able to +represent it in CBOR. -In short, CBOR is a faster and very compact binary alternative to JSON, +In short, CBOR is a faster and quite compact binary alternative to JSON, with the added ability of supporting serialisation of Perl objects. (JSON often compresses better than CBOR though, so if you plan to compress the -data later you might want to compare both formats first). +data later and speed is less important you might want to compare both +formats first). To give you a general idea about speed, with texts in the megabyte range, C usually encodes roughly twice as fast as L or L and decodes about 15%-30% faster than those. The shorter the data, the worse L performs in comparison. -As for compactness, C encoded data structures are usually about -20% smaller than the same data encoded as (compact) JSON or L. - -In addition to the core CBOR data format, this module implements a number -of extensions, to support cyclic and self-referencing data structures -(see C), string deduplication (see C) and -scalar references (always enabled). +Regarding compactness, C-encoded data structures are usually +about 20% smaller than the same data encoded as (compact) JSON or +L. + +In addition to the core CBOR data format, this module implements a +number of extensions, to support cyclic and shared data structures +(see C and C), string deduplication (see +C) and scalar references (always enabled). The primary goal of this module is to be I and the secondary goal is to be I. To reach the latter goal it was written in C. @@ -73,7 +66,7 @@ use common::sense; -our $VERSION = 0.08; +our $VERSION = 1.4; our @ISA = qw(Exporter); our @EXPORT = qw(encode_cbor decode_cbor); @@ -120,7 +113,6 @@ The mutators for flags all return the CBOR object again and thus calls can be chained: -#TODO my $cbor = CBOR::XS->new->encode ({a => [1,2]}); =item $cbor = $cbor->max_depth ([$maximum_nesting_depth]) @@ -189,49 +181,140 @@ This means that such values will only be encoded once, and will not result in a deep cloning of the value on decode, in decoders supporting the value -sharing extension. +sharing extension. This also makes it possible to encode cyclic data +structures (which need C to ne enabled to be decoded by this +module). It is recommended to leave it off unless you know your communication partner supports the value sharing extensions to CBOR -(http://cbor.schmorp.de/value-sharing). +(L), as without decoder support, the +resulting data structure might be unusable. Detecting shared values incurs a runtime overhead when values are encoded that have a reference counter large than one, and might unnecessarily increase the encoded size, as potentially shared values are encode as -sharable whether or not they are actually shared. +shareable whether or not they are actually shared. At the moment, only targets of references can be shared (e.g. scalars, arrays or hashes pointed to by a reference). Weirder constructs, such as an array with multiple "copies" of the I string, which are hard but not impossible to create in Perl, are not supported (this is the same as -for L). +with L). -If C<$enable> is false (the default), then C will encode -exception when it encounters anything it cannot encode as CBOR. +If C<$enable> is false (the default), then C will encode shared +data structures repeatedly, unsharing them in the process. Cyclic data +structures cannot be encoded in this mode. This option does not affect C in any way - shared values and references will always be decoded properly if present. -=item $cbor = $cbor->allow_stringref ([$enable]) +=item $cbor = $cbor->allow_cycles ([$enable]) + +=item $enabled = $cbor->get_allow_cycles + +If C<$enable> is true (or missing), then C will happily decode +self-referential (cyclic) data structures. By default these will not be +decoded, as they need manual cleanup to avoid memory leaks, so code that +isn't prepared for this will not leak memory. -=item $enabled = $cbor->get_allow_stringref +If C<$enable> is false (the default), then C will throw an error +when it encounters a self-referential/cyclic data structure. + +FUTURE DIRECTION: the motivation behind this option is to avoid I +cycles - future versions of this module might chose to decode cyclic data +structures using weak references when this option is off, instead of +throwing an error. + +This option does not affect C in any way - shared values and +references will always be encoded properly if present. + +=item $cbor = $cbor->pack_strings ([$enable]) + +=item $enabled = $cbor->get_pack_strings If C<$enable> is true (or missing), then C will try not to encode the same string twice, but will instead encode a reference to the string -instead. Depending on your data format. this can save a lot of space, but +instead. Depending on your data format, this can save a lot of space, but also results in a very large runtime overhead (expect encoding times to be 2-4 times as high as without). It is recommended to leave it off unless you know your communications partner supports the stringref extension to CBOR -(http://cbor.schmorp.de/stringref). +(L), as without decoder support, the +resulting data structure might not be usable. -If C<$enable> is false (the default), then C will encode -exception when it encounters anything it cannot encode as CBOR. +If C<$enable> is false (the default), then C will encode strings +the standard CBOR way. This option does not affect C in any way - string references will always be decoded properly if present. +=item $cbor = $cbor->validate_utf8 ([$enable]) + +=item $enabled = $cbor->get_validate_utf8 + +If C<$enable> is true (or missing), then C will validate that +elements (text strings) containing UTF-8 data in fact contain valid UTF-8 +data (instead of blindly accepting it). This validation obviously takes +extra time during decoding. + +The concept of "valid UTF-8" used is perl's concept, which is a superset +of the official UTF-8. + +If C<$enable> is false (the default), then C will blindly accept +UTF-8 data, marking them as valid UTF-8 in the resulting data structure +regardless of whether thats true or not. + +Perl isn't too happy about corrupted UTF-8 in strings, but should +generally not crash or do similarly evil things. Extensions might be not +so forgiving, so it's recommended to turn on this setting if you receive +untrusted CBOR. + +This option does not affect C in any way - strings that are +supposedly valid UTF-8 will simply be dumped into the resulting CBOR +string without checking whether that is, in fact, true or not. + +=item $cbor = $cbor->filter ([$cb->($tag, $value)]) + +=item $cb_or_undef = $cbor->get_filter + +Sets or replaces the tagged value decoding filter (when C<$cb> is +specified) or clears the filter (if no argument or C is provided). + +The filter callback is called only during decoding, when a non-enforced +tagged value has been decoded (see L for a +list of enforced tags). For specific tags, it's often better to provide a +default converter using the C<%CBOR::XS::FILTER> hash (see below). + +The first argument is the numerical tag, the second is the (decoded) value +that has been tagged. + +The filter function should return either exactly one value, which will +replace the tagged value in the decoded data structure, or no values, +which will result in default handling, which currently means the decoder +creates a C object to hold the tag and the value. + +When the filter is cleared (the default state), the default filter +function, C, is used. This function simply looks +up the tag in the C<%CBOR::XS::FILTER> hash. If an entry exists it must be +a code reference that is called with tag and value, and is responsible for +decoding the value. If no entry exists, it returns no values. + +Example: decode all tags not handled internally into C +objects, with no other special handling (useful when working with +potentially "unsafe" CBOR data). + + CBOR::XS->new->filter (sub { })->decode ($cbor_data); + +Example: provide a global filter for tag 1347375694, converting the value +into some string form. + + $CBOR::XS::FILTER{1347375694} = sub { + my ($tag, $value); + + "tag 1347375694 value $value" + }; + =item $cbor_data = $cbor->encode ($perl_scalar) Converts the given Perl data structure (a scalar value) to its CBOR @@ -257,6 +340,70 @@ =back +=head2 INCREMENTAL PARSING + +In some cases, there is the need for incremental parsing of JSON +texts. While this module always has to keep both CBOR text and resulting +Perl data structure in memory at one time, it does allow you to parse a +CBOR stream incrementally, using a similar to using "decode_prefix" to see +if a full CBOR object is available, but is much more efficient. + +It basically works by parsing as much of a CBOR string as possible - if +the CBOR data is not complete yet, the pasrer will remember where it was, +to be able to restart when more data has been accumulated. Once enough +data is available to either decode a complete CBOR value or raise an +error, a real decode will be attempted. + +A typical use case would be a network protocol that consists of sending +and receiving CBOR-encoded messages. The solution that works with CBOR and +about anything else is by prepending a length to every CBOR value, so the +receiver knows how many octets to read. More compact (and slightly slower) +would be to just send CBOR values back-to-back, as C knows where +a CBOR value ends, and doesn't need an explicit length. + +The following methods help with this: + +=over 4 + +=item @decoded = $cbor->incr_parse ($buffer) + +This method attempts to decode exactly one CBOR value from the beginning +of the given C<$buffer>. The value is removed from the C<$buffer> on +success. When C<$buffer> doesn't contain a complete value yet, it returns +nothing. Finally, when the C<$buffer> doesn't start with something +that could ever be a valid CBOR value, it raises an exception, just as +C would. In the latter case the decoder state is undefined and +must be reset before being able to parse further. + +This method modifies the C<$buffer> in place. When no CBOR value can be +decoded, the decoder stores the current string offset. On the next call, +continues decoding at the place where it stopped before. For this to make +sense, the C<$buffer> must begin with the same octets as on previous +unsuccessful calls. + +You can call this method in scalar context, in which case it either +returns a decoded value or C. This makes it impossible to +distinguish between CBOR null values (which decode to C) and an +unsuccessful decode, which is often acceptable. + +=item @decoded = $cbor->incr_parse_multiple ($buffer) + +Same as C, but attempts to decode as many CBOR values as +possible in one go, instead of at most one. Calls to C and +C can be interleaved. + +=item $cbor->incr_reset + +Resets the incremental decoder. This throws away any saved state, so that +subsequent calls to C or C start to parse +a new CBOR value from the beginning of the C<$buffer> again. + +This method can be caled at any time, but it I be called if you want +to change your C<$buffer> or there was a decoding error and you want to +reuse the C<$cbor> object for future incremental parsings. + +=back + =head1 MAPPING @@ -281,7 +428,7 @@ =item byte strings -Byte strings will become octet strings in Perl (the byte values 0..255 +Byte strings will become octet strings in Perl (the Byte values 0..255 will simply become characters of the same value in Perl). =item UTF-8 strings @@ -309,24 +456,12 @@ C<1> and C<0> (for true and false) or to throw an exception on access (for error). See the L manpage for details. -=item CBOR tag 256 (perl object) - -The tag value C<256> (TODO: pending iana registration) will be used -to deserialise a Perl object serialised with C. See L, below, for details. +=item tagged values -=item CBOR tag 55799 (magic header) +Tagged items consists of a numeric tag and another CBOR value. -The tag 55799 is ignored (this tag implements the magic header). - -=item other CBOR tags - -Tagged items consists of a numeric tag and another CBOR value. Tags not -handled internally are currently converted into a L -object, which is simply a blessed array reference consisting of the -numeric tag value followed by the (decoded) CBOR value. - -In the future, support for user-supplied conversions might get added. +See L and the description of C<< ->filter >> +for details on which tags are handled how. =item anything else @@ -339,8 +474,8 @@ =head2 PERL -> CBOR The mapping from Perl to CBOR is slightly more difficult, as Perl is a -truly typeless language, so we can only guess which CBOR type is meant by -a Perl value. +typeless language. That means this module can only guess which CBOR type +is meant by a perl value. =over 4 @@ -348,7 +483,7 @@ Perl hash references become CBOR maps. As there is no inherent ordering in hash keys (or CBOR maps), they will usually be encoded in a pseudo-random -order. +order. This order can be different each time a hahs is encoded. Currently, tied hashes will use the indefinite-length format, while normal hashes will use the fixed-length format. @@ -359,15 +494,18 @@ =item other references -Other unblessed references are generally not allowed and will cause an -exception to be thrown, except for references to the integers C<0> and -C<1>, which get turned into false and true in CBOR. +Other unblessed references will be represented using +the indirection tag extension (tag value C<22098>, +L). CBOR decoders are guaranteed +to be able to decode these values somehow, by either "doing the right +thing", decoding into a generic tagged object, simply ignoring the tag, or +something else. =item CBOR::XS::Tagged objects Objects of this type must be arrays consisting of a single C<[tag, value]> pair. The (numerical) tag will be encoded as a CBOR tag, the value will -be encoded as appropriate for the value. You cna use C to +be encoded as appropriate for the value. You must use C to create such objects. =item Types::Serialiser::true, Types::Serialiser::false, Types::Serialiser::error @@ -379,11 +517,11 @@ =item other blessed objects Other blessed objects are serialised via C or C. See -L, below, for details. +L for specific classes handled by this +module, and L for generic object serialisation. =item simple scalars -TODO Simple Perl scalars (any scalar that is not a reference) are the most difficult objects to encode: CBOR::XS will encode undefined scalars as CBOR null values, scalars that have last been used in a string context @@ -394,7 +532,7 @@ encode_cbor [-3.0e17] # yields [-3e+17] my $value = 5; encode_cbor [$value] # yields [5] - # used as string, so dump as string + # used as string, so dump as string (either byte or text) print $value; encode_cbor [$value] # yields ["5"] @@ -408,6 +546,16 @@ $x .= ""; # another, more awkward way to stringify print $x; # perl does it for you, too, quite often +You can force whether a string ie encoded as byte or text string by using +C and C): + + utf8::upgrade $x; # encode $x as text string + utf8::downgrade $x; # encode $x as byte string + +Perl doesn't define what operations up- and downgrade strings, so if the +difference between byte and text is important, you should up- or downgrade +your string as late as possible before encoding. + You can force the type to be a CBOR number by numifying it: my $x = "3"; # some variable containing a string @@ -429,10 +577,16 @@ =head2 OBJECT SERIALISATION +This module implements both a CBOR-specific and the generic +L object serialisation protocol. The following +subsections explain both methods. + +=head3 ENCODING + This module knows two way to serialise a Perl object: The CBOR-specific way, and the generic way. -Whenever the encoder encounters a Perl object that it cnanot serialise +Whenever the encoder encounters a Perl object that it cannot serialise directly (most of them), it will first look up the C method on it. @@ -448,11 +602,18 @@ more). These will be encoded as CBOR perl object, together with the classname. +These methods I change the data structure that is being +serialised. Failure to comply to this can result in memory corruption - +and worse. + If an object supports neither C nor C, encoding will fail with an error. -Objects encoded via C cannot be automatically decoded, but -objects encoded via C can be decoded using the following protocol: +=head3 DECODING + +Objects encoded via C cannot (normally) be automatically decoded, +but objects encoded via C can be decoded using the following +protocol: When an encoded CBOR perl object is encountered by the decoder, it will look up the C method, by using the stored classname, and will fail @@ -462,7 +623,7 @@ as first argument, the constant string C as second argument, and all values returned by C as remaining arguments. -=head4 EXAMPLES +=head3 EXAMPLES Here is an example C method: @@ -483,7 +644,7 @@ my ($self) = @_; my $uri = "$self"; # stringify uri utf8::upgrade $uri; # make sure it will be encoded as UTF-8 string - CBOR::XS::tagged 32, "$_[0]" + CBOR::XS::tag 32, "$_[0]" } This will encode URIs as a UTF-8 string with tag 32, which indicates an @@ -629,32 +790,56 @@ =head1 TAG HANDLING AND EXTENSIONS -This section describes how this module handles specific tagged values and -extensions. If a tag is not mentioned here, then the default handling -applies (creating a CBOR::XS::Tagged object on decoding, and only encoding -the tag when explicitly requested). +This section describes how this module handles specific tagged values +and extensions. If a tag is not mentioned here and no additional filters +are provided for it, then the default handling applies (creating a +CBOR::XS::Tagged object on decoding, and only encoding the tag when +explicitly requested). + +Tags not handled specifically are currently converted into a +L object, which is simply a blessed array reference +consisting of the numeric tag value followed by the (decoded) CBOR value. Future versions of this module reserve the right to special case -additional tags (such as bigfloat or base64url). +additional tags (such as base64url). + +=head2 ENFORCED TAGS + +These tags are always handled when decoding, and their handling cannot be +overriden by the user. =over 4 -=item (perl-object, L) +=item 26 (perl-object, L) -These tags are automatically created for serialisable objects using the -C methods (the L object serialisation -protocol). +These tags are automatically created (and decoded) for serialisable +objects using the C methods (the L object +serialisation protocol). See L for details. -=item , (sharable, sharedref, L ) +=item 28, 29 (shareable, sharedref, L) -These tags are automatically decoded when encountered, resulting in +These tags are automatically decoded when encountered (and they do not +result in a cyclic data structure, see C), resulting in shared values in the decoded object. They are only encoded, however, when -C is enabled. +C is enabled. + +Not all shared values can be successfully decoded: values that reference +themselves will I decode as C (this is not the same +as a reference pointing to itself, which will be represented as a value +that contains an indirect reference to itself - these will be decoded +properly). + +Note that considerably more shared value data structures can be decoded +than will be encoded - currently, only values pointed to by references +will be shared, others will not. While non-reference shared values can be +generated in Perl with some effort, they were considered too unimportant +to be supported in the encoder. The decoder, however, will decode these +values as shared values. -=item , (stringref-namespace, stringref, L ) +=item 256, 25 (stringref-namespace, stringref, L) These tags are automatically decoded when encountered. They are only -encoded, however, when C is enabled. +encoded, however, when C is enabled. =item 22098 (indirection, L) @@ -669,6 +854,103 @@ =back +=head2 NON-ENFORCED TAGS + +These tags have default filters provided when decoding. Their handling can +be overriden by changing the C<%CBOR::XS::FILTER> entry for the tag, or by +providing a custom C callback when decoding. + +When they result in decoding into a specific Perl class, the module +usually provides a corresponding C method as well. + +When any of these need to load additional modules that are not part of the +perl core distribution (e.g. L), it is (currently) up to the user to +provide these modules. The decoding usually fails with an exception if the +required module cannot be loaded. + +=over 4 + +=item 0, 1 (date/time string, seconds since the epoch) + +These tags are decoded into L objects. The corresponding +C method always encodes into tag 1 values currently. + +The L API is generally surprisingly bad, and fractional +seconds are only accidentally kept intact, so watch out. On the plus side, +the module comes with perl since 5.10, which has to count for something. + +=item 2, 3 (positive/negative bignum) + +These tags are decoded into L objects. The corresponding +C method encodes "small" bigints into normal CBOR +integers, and others into positive/negative CBOR bignums. + +=item 4, 5 (decimal fraction/bigfloat) + +Both decimal fractions and bigfloats are decoded into L +objects. The corresponding C method I +encodes into a decimal fraction. + +CBOR cannot represent bigfloats with I large exponents - conversion +of such big float objects is undefined. + +Also, NaN and infinities are not encoded properly. + +=item 21, 22, 23 (expected later JSON conversion) + +CBOR::XS is not a CBOR-to-JSON converter, and will simply ignore these +tags. + +=item 32 (URI) + +These objects decode into L objects. The corresponding +C method again results in a CBOR URI value. + +=back + +=cut + +our %FILTER = ( + # 0 # rfc4287 datetime, utf-8 + # 1 # unix timestamp, any + + 2 => sub { # pos bigint + require Math::BigInt; + Math::BigInt->new ("0x" . unpack "H*", pop) + }, + + 3 => sub { # neg bigint + require Math::BigInt; + -Math::BigInt->new ("0x" . unpack "H*", pop) + }, + + 4 => sub { # decimal fraction, array + require Math::BigFloat; + Math::BigFloat->new ($_[1][1] . "E" . $_[1][0]) + }, + + 5 => sub { # bigfloat, array + require Math::BigFloat; + scalar Math::BigFloat->new ($_[1][1])->blsft ($_[1][0], 2) + }, + + 21 => sub { pop }, # expected conversion to base64url encoding + 22 => sub { pop }, # expected conversion to base64 encoding + 23 => sub { pop }, # expected conversion to base16 encoding + + # 24 # embedded cbor, byte string + + 32 => sub { + require URI; + URI->new (pop) + }, + + # 33 # base64url rfc4648, utf-8 + # 34 # base64 rfc46484, utf-8 + # 35 # regex pcre/ecma262, utf-8 + # 36 # mime message rfc2045, utf-8 +); + =head1 CBOR and JSON @@ -739,6 +1021,16 @@ Strict mode and canonical mode are not implemented. +=head1 LIMITATIONS ON PERLS WITHOUT 64-BIT INTEGER SUPPORT + +On perls that were built without 64 bit integer support (these are rare +nowadays, even on 32 bit architectures, as all major Perl distributions +are built with 64 bit integer support), support for any kind of 64 bit +integer in CBOR is very limited - most likely, these 64 bit values will +be truncated, corrupted, or otherwise not decoded correctly. This also +includes string, array and map sizes that are stored as 64 bit integers. + + =head1 THREADS This module is I guaranteed to be thread safe and there are no @@ -760,6 +1052,101 @@ =cut +our %FILTER = ( + 0 => sub { # rfc4287 datetime, utf-8 + require Time::Piece; + # Time::Piece::Strptime uses the "incredibly flexible date parsing routine" + # from FreeBSD, which can't parse ISO 8601, RFC3339, RFC4287 or much of anything + # else either. Whats incredibe over standard strptime totally escapes me. + # doesn't do fractional times, either. sigh. + # In fact, it's all a lie, it uses whatever strptime it wants, and of course, + # they are all incomptible. The openbsd one simply ignores %z (but according to the + # docs, it would be much more incredibly flexible indeed. If it worked, that is.). + scalar eval { + my $s = $_[1]; + + $s =~ s/Z$/+00:00/; + $s =~ s/(\.[0-9]+)?([+-][0-9][0-9]):([0-9][0-9])$// + or die; + + my $b = $1 - ($2 * 60 + $3) * 60; # fractional part + offset. hopefully + my $d = Time::Piece->strptime ($s, "%Y-%m-%dT%H:%M:%S"); + + Time::Piece::gmtime ($d->epoch + $b) + } || die "corrupted CBOR date/time string ($_[0])"; + }, + + 1 => sub { # seconds since the epoch, possibly fractional + require Time::Piece; + scalar Time::Piece::gmtime (pop) + }, + + 2 => sub { # pos bigint + require Math::BigInt; + Math::BigInt->new ("0x" . unpack "H*", pop) + }, + + 3 => sub { # neg bigint + require Math::BigInt; + -Math::BigInt->new ("0x" . unpack "H*", pop) + }, + + 4 => sub { # decimal fraction, array + require Math::BigFloat; + Math::BigFloat->new ($_[1][1] . "E" . $_[1][0]) + }, + + 5 => sub { # bigfloat, array + require Math::BigFloat; + scalar Math::BigFloat->new ($_[1][1])->blsft ($_[1][0], 2) + }, + + 21 => sub { pop }, # expected conversion to base64url encoding + 22 => sub { pop }, # expected conversion to base64 encoding + 23 => sub { pop }, # expected conversion to base16 encoding + + # 24 # embedded cbor, byte string + + 32 => sub { + require URI; + URI->new (pop) + }, + + # 33 # base64url rfc4648, utf-8 + # 34 # base64 rfc46484, utf-8 + # 35 # regex pcre/ecma262, utf-8 + # 36 # mime message rfc2045, utf-8 +); + +sub CBOR::XS::default_filter { + &{ $FILTER{$_[0]} or return } +} + +sub URI::TO_CBOR { + my $uri = $_[0]->as_string; + utf8::upgrade $uri; + tag 32, $uri +} + +sub Math::BigInt::TO_CBOR { + if ($_[0] >= -2147483648 && $_[0] <= 2147483647) { + $_[0]->numify + } else { + my $hex = substr $_[0]->as_hex, 2; + $hex = "0$hex" if 1 & length $hex; # sigh + tag $_[0] >= 0 ? 2 : 3, pack "H*", $hex + } +} + +sub Math::BigFloat::TO_CBOR { + my ($m, $e) = $_[0]->parts; + tag 4, [$e->numify, $m] +} + +sub Time::Piece::TO_CBOR { + tag 1, 0 + $_[0]->epoch +} + XSLoader::load "CBOR::XS", $VERSION; =head1 SEE ALSO