11 YAML is a human readable data serialization language. The full YAML language
12 spec can be read at `yaml.org
13 <http://www.yaml.org/spec/1.2/spec.html#Introduction>`_. The simplest form of
14 yaml is just "scalars", "mappings", and "sequences". A scalar is any number
15 or string. The pound/hash symbol (#) begins a comment line. A mapping is
16 a set of key-value pairs where the key ends with a colon. For example:
24 A sequence is a list of items where each item starts with a leading dash ('-').
34 You can combine mappings and sequences by indenting. For example a sequence
35 of mappings in which one of the mapping values is itself a sequence:
39 # a sequence of mappings with one key's value being a sequence
52 Sometime sequences are known to be short and the one entry per line is too
53 verbose, so YAML offers an alternate syntax for sequences called a "Flow
54 Sequence" in which you put comma separated sequence elements into square
55 brackets. The above example could then be simplified to :
60 # a sequence of mappings with one key's value being a flow sequence
66 cpus: [ PowerPC, x86 ]
69 Introduction to YAML I/O
70 ========================
72 The use of indenting makes the YAML easy for a human to read and understand,
73 but having a program read and write YAML involves a lot of tedious details.
74 The YAML I/O library structures and simplifies reading and writing YAML
77 YAML I/O assumes you have some "native" data structures which you want to be
78 able to dump as YAML and recreate from YAML. The first step is to try
79 writing example YAML for your data structures. You may find after looking at
80 possible YAML representations that a direct mapping of your data structures
81 to YAML is not very readable. Often the fields are not in the order that
82 a human would find readable. Or the same information is replicated in multiple
83 locations, making it hard for a human to write such YAML correctly.
85 In relational database theory there is a design step called normalization in
86 which you reorganize fields and tables. The same considerations need to
87 go into the design of your YAML encoding. But, you may not want to change
88 your exisiting native data structures. Therefore, when writing out YAML
89 there may be a normalization step, and when reading YAML there would be a
90 corresponding denormalization step.
92 YAML I/O uses a non-invasive, traits based design. YAML I/O defines some
93 abstract base templates. You specialize those templates on your data types.
94 For instance, if you have an eumerated type FooBar you could specialize
95 ScalarEnumerationTraits on that type and define the enumeration() method:
99 using llvm::yaml::ScalarEnumerationTraits;
100 using llvm::yaml::IO;
103 struct ScalarEnumerationTraits<FooBar> {
104 static void enumeration(IO &io, FooBar &value) {
110 As with all YAML I/O template specializations, the ScalarEnumerationTraits is used for
111 both reading and writing YAML. That is, the mapping between in-memory enum
112 values and the YAML string representation is only in place.
113 This assures that the code for writing and parsing of YAML stays in sync.
115 To specify a YAML mappings, you define a specialization on
116 llvm::yaml::MapppingTraits.
117 If your native data structure happens to be a struct that is already normalized,
118 then the specialization is simple. For example:
122 using llvm::yaml::MapppingTraits;
123 using llvm::yaml::IO;
126 struct MapppingTraits<Person> {
127 static void mapping(IO &io, Person &info) {
128 io.mapRequired("name", info.name);
129 io.mapOptional("hat-size", info.hatSize);
134 A YAML sequence is automatically infered if you data type has begin()/end()
135 iterators and a push_back() method. Therefore any of the STL containers
136 (such as std::vector<>) will automatically translate to YAML sequences.
138 Once you have defined specializations for your data types, you can
139 programmatically use YAML I/O to write a YAML document:
143 using llvm::yaml::Output;
151 std::vector<Person> persons;
152 persons.push_back(tom);
153 persons.push_back(dan);
155 Output yout(llvm::outs());
158 This would write the following:
167 And you can also read such YAML documents with the following code:
171 using llvm::yaml::Input;
173 typedef std::vector<Person> PersonList;
174 std::vector<PersonList> docs;
176 Input yin(document.getBuffer());
182 // Process read document
183 for ( PersonList &pl : docs ) {
184 for ( Person &person : pl ) {
185 cout << "name=" << person.name;
189 One other feature of YAML is the ability to define multiple documents in a
190 single file. That is why reading YAML produces a vector of your document type.
197 When parsing a YAML document, if the input does not match your schema (as
198 expressed in your XxxTraits<> specializations). YAML I/O
199 will print out an error message and your Input object's error() method will
200 return true. For instance the following document:
209 Has a key (shoe-size) that is not defined in the schema. YAML I/O will
210 automatically generate this error:
214 YAML:2:2: error: unknown key 'shoe-size'
218 Similar errors are produced for other input not conforming to the schema.
224 YAML scalars are just strings (i.e. not a sequence or mapping). The YAML I/O
225 library provides support for translating between YAML scalars and specific
231 The following types have built-in support in YAML I/O:
246 That is, you can use those types in fields of MapppingTraits or as element type
247 in sequence. When reading, YAML I/O will validate that the string found
248 is convertible to that type and error out if not.
253 Given that YAML I/O is trait based, the selection of how to convert your data
254 to YAML is based on the type of your data. But in C++ type matching, typedefs
255 do not generate unique type names. That means if you have two typedefs of
256 unsigned int, to YAML I/O both types look exactly like unsigned int. To
257 facilitate make unique type names, YAML I/O provides a macro which is used
258 like a typedef on built-in types, but expands to create a class with conversion
259 operators to and from the base type. For example:
263 LLVM_YAML_STRONG_TYPEDEF(uint32_t, MyFooFlags)
264 LLVM_YAML_STRONG_TYPEDEF(uint32_t, MyBarFlags)
266 This generates two classes MyFooFlags and MyBarFlags which you can use in your
267 native data structures instead of uint32_t. They are implicitly
268 converted to and from uint32_t. The point of creating these unique types
269 is that you can now specify traits on them to get different YAML conversions.
273 An example use of a unique type is that YAML I/O provides fixed sized unsigned
274 integers that are written with YAML I/O as hexadecimal instead of the decimal
275 format used by the built-in integer types:
282 You can use llvm::yaml::Hex32 instead of uint32_t and the only different will
283 be that when YAML I/O writes out that type it will be formatted in hexadecimal.
286 ScalarEnumerationTraits
287 -----------------------
288 YAML I/O supports translating between in-memory enumerations and a set of string
289 values in YAML documents. This is done by specializing ScalarEnumerationTraits<>
290 on your enumeration type and define a enumeration() method.
291 For instance, suppose you had an enumeration of CPUs and a struct with it as
307 To support reading and writing of this enumeration, you can define a
308 ScalarEnumerationTraits specialization on CPUs, which can then be used
313 using llvm::yaml::ScalarEnumerationTraits;
314 using llvm::yaml::MapppingTraits;
315 using llvm::yaml::IO;
318 struct ScalarEnumerationTraits<CPUs> {
319 static void enumeration(IO &io, CPUs &value) {
320 io.enumCase(value, "x86_64", cpu_x86_64);
321 io.enumCase(value, "x86", cpu_x86);
322 io.enumCase(value, "PowerPC", cpu_PowerPC);
327 struct MapppingTraits<Info> {
328 static void mapping(IO &io, Info &info) {
329 io.mapRequired("cpu", info.cpu);
330 io.mapOptional("flags", info.flags, 0);
334 When reading YAML, if the string found does not match any of the the strings
335 specified by enumCase() methods, an error is automatically generated.
336 When writing YAML, if the value being written does not match any of the values
337 specified by the enumCase() methods, a runtime assertion is triggered.
342 Another common data structure in C++ is a field where each bit has a unique
343 meaning. This is often used in a "flags" field. YAML I/O has support for
344 converting such fields to a flow sequence. For instance suppose you
345 had the following bit flags defined:
356 LLVM_YAML_UNIQUE_TYPE(MyFlags, uint32_t)
358 To support reading and writing of MyFlags, you specialize ScalarBitSetTraits<>
359 on MyFlags and provide the bit values and their names.
363 using llvm::yaml::ScalarBitSetTraits;
364 using llvm::yaml::MapppingTraits;
365 using llvm::yaml::IO;
368 struct ScalarBitSetTraits<MyFlags> {
369 static void bitset(IO &io, MyFlags &value) {
370 io.bitSetCase(value, "hollow", flagHollow);
371 io.bitSetCase(value, "flat", flagFlat);
372 io.bitSetCase(value, "round", flagRound);
373 io.bitSetCase(value, "pointy", flagPointy);
383 struct MapppingTraits<Info> {
384 static void mapping(IO &io, Info& info) {
385 io.mapRequired("name", info.name);
386 io.mapRequired("flags", info.flags);
390 With the above, YAML I/O (when writing) will test mask each value in the
391 bitset trait against the flags field, and each that matches will
392 cause the corresponding string to be added to the flow sequence. The opposite
393 is done when reading and any unknown string values will result in a error. With
394 the above schema, a same valid YAML document is:
399 flags: [ pointy, flat ]
404 Sometimes for readability a scalar needs to be formatted in a custom way. For
405 instance your internal data structure may use a integer for time (seconds since
406 some epoch), but in YAML it would be much nicer to express that integer in
407 some time format (e.g. 4-May-2012 10:30pm). YAML I/O has a way to support
408 custom formatting and parsing of scalar types by specializing ScalarTraits<> on
409 your data type. When writing, YAML I/O will provide the native type and
410 your specialization must create a temporary llvm::StringRef. When reading,
411 YAML I/O will provide a llvm::StringRef of scalar and your specialization
412 must convert that to your native data type. An outline of a custom scalar type
417 using llvm::yaml::ScalarTraits;
418 using llvm::yaml::IO;
421 struct ScalarTraits<MyCustomType> {
422 static void output(const T &value, llvm::raw_ostream &out) {
423 out << value; // do custom formatting here
425 static StringRef input(StringRef scalar, T &value) {
426 // do custom parsing here. Return the empty string on success,
427 // or an error message on failure.
436 To be translated to or from a YAML mapping for your type T you must specialize
437 llvm::yaml::MapppingTraits on T and implement the "void mapping(IO &io, T&)"
438 method. If your native data structures use pointers to a class everywhere,
439 you can specialize on the class pointer. Examples:
443 using llvm::yaml::MapppingTraits;
444 using llvm::yaml::IO;
446 // Example of struct Foo which is used by value
448 struct MapppingTraits<Foo> {
449 static void mapping(IO &io, Foo &foo) {
450 io.mapOptional("size", foo.size);
455 // Example of struct Bar which is natively always a pointer
457 struct MapppingTraits<Bar*> {
458 static void mapping(IO &io, Bar *&bar) {
459 io.mapOptional("size", bar->size);
468 The mapping() method is responsible, if needed, for normalizing and
469 denormalizing. In a simple case where the native data structure requires no
470 normalization, the mapping method just uses mapOptional() or mapRequired() to
471 bind the struct's fields to YAML key names. For example:
475 using llvm::yaml::MapppingTraits;
476 using llvm::yaml::IO;
479 struct MapppingTraits<Person> {
480 static void mapping(IO &io, Person &info) {
481 io.mapRequired("name", info.name);
482 io.mapOptional("hat-size", info.hatSize);
490 When [de]normalization is required, the mapping() method needs a way to access
491 normalized values as fields. To help with this, there is
492 a template MappingNormalization<> which you can then use to automatically
493 do the normalization and denormalization. The template is used to create
494 a local variable in your mapping() method which contains the normalized keys.
496 Suppose you have native data type
497 Polar which specifies a position in polar coordinates (distance, angle):
506 but you've decided the normalized YAML for should be in x,y coordinates. That
507 is, you want the yaml to look like:
514 You can support this by defining a MapppingTraits that normalizes the polar
515 coordinates to x,y coordinates when writing YAML and denormalizes x,y
516 coordindates into polar when reading YAML.
520 using llvm::yaml::MapppingTraits;
521 using llvm::yaml::IO;
524 struct MapppingTraits<Polar> {
526 class NormalizedPolar {
528 NormalizedPolar(IO &io)
531 NormalizedPolar(IO &, Polar &polar)
532 : x(polar.distance * cos(polar.angle)),
533 y(polar.distance * sin(polar.angle)) {
535 Polar denormalize(IO &) {
536 return Polar(sqrt(x*x+y*y, arctan(x,y));
543 static void mapping(IO &io, Polar &polar) {
544 MappingNormalization<NormalizedPolar, Polar> keys(io, polar);
546 io.mapRequired("x", keys->x);
547 io.mapRequired("y", keys->y);
551 When writing YAML, the local variable "keys" will be a stack allocated
552 instance of NormalizedPolar, constructed from the suppled polar object which
553 initializes it x and y fields. The mapRequired() methods then write out the x
554 and y values as key/value pairs.
556 When reading YAML, the local variable "keys" will be a stack allocated instance
557 of NormalizedPolar, constructed by the empty constructor. The mapRequired
558 methods will find the matching key in the YAML document and fill in the x and y
559 fields of the NormalizedPolar object keys. At the end of the mapping() method
560 when the local keys variable goes out of scope, the denormalize() method will
561 automatically be called to convert the read values back to polar coordinates,
562 and then assigned back to the second parameter to mapping().
564 In some cases, the normalized class may be a subclass of the native type and
565 could be returned by the denormalize() method, except that the temporary
566 normalized instance is stack allocated. In these cases, the utility template
567 MappingNormalizationHeap<> can be used instead. It just like
568 MappingNormalization<> except that it heap allocates the normalized object
569 when reading YAML. It never destroyes the normalized object. The denormalize()
570 method can this return "this".
575 Within a mapping() method, calls to io.mapRequired() mean that that key is
576 required to exist when parsing YAML documents, otherwise YAML I/O will issue an
579 On the other hand, keys registered with io.mapOptional() are allowed to not
580 exist in the YAML document being read. So what value is put in the field
581 for those optional keys?
582 There are two steps to how those optional fields are filled in. First, the
583 second parameter to the mapping() method is a reference to a native class. That
584 native class must have a default constructor. Whatever value the default
585 constructor initially sets for an optional field will be that field's value.
586 Second, the mapOptional() method has an optional third parameter. If provided
587 it is the value that mapOptional() should set that field to if the YAML document
588 does not have that key.
590 There is one important difference between those two ways (default constructor
591 and third parameter to mapOptional). When YAML I/O generates a YAML document,
592 if the mapOptional() third parameter is used, if the actual value being written
593 is the same as (using ==) the default value, then that key/value is not written.
599 When writing out a YAML document, the keys are written in the order that the
600 calls to mapRequired()/mapOptional() are made in the mapping() method. This
601 gives you a chance to write the fields in an order that a human reader of
602 the YAML document would find natural. This may be different that the order
603 of the fields in the native class.
605 When reading in a YAML document, the keys in the document can be in any order,
606 but they are processed in the order that the calls to mapRequired()/mapOptional()
607 are made in the mapping() method. That enables some interesting
608 functionality. For instance, if the first field bound is the cpu and the second
609 field bound is flags, and the flags are cpu specific, you can programmatically
610 switch how the flags are converted to and from YAML based on the cpu.
611 This works for both reading and writing. For example:
615 using llvm::yaml::MapppingTraits;
616 using llvm::yaml::IO;
624 struct MapppingTraits<Info> {
625 static void mapping(IO &io, Info &info) {
626 io.mapRequired("cpu", info.cpu);
627 // flags must come after cpu for this to work when reading yaml
628 if ( info.cpu == cpu_x86_64 )
629 io.mapRequired("flags", *(My86_64Flags*)info.flags);
631 io.mapRequired("flags", *(My86Flags*)info.flags);
639 To be translated to or from a YAML sequence for your type T you must specialize
640 llvm::yaml::SequenceTraits on T and implement two methods:
641 ``size_t size(IO &io, T&)`` and
642 ``T::value_type& element(IO &io, T&, size_t indx)``. For example:
647 struct SequenceTraits<MySeq> {
648 static size_t size(IO &io, MySeq &list) { ... }
649 static MySeqEl element(IO &io, MySeq &list, size_t index) { ... }
652 The size() method returns how many elements are currently in your sequence.
653 The element() method returns a reference to the i'th element in the sequence.
654 When parsing YAML, the element() method may be called with an index one bigger
655 than the current size. Your element() method should allocate space for one
656 more element (using default constructor if element is a C++ object) and returns
657 a reference to that new allocated space.
662 A YAML "flow sequence" is a sequence that when written to YAML it uses the
663 inline notation (e.g [ foo, bar ] ). To specify that a sequence type should
664 be written in YAML as a flow sequence, your SequenceTraits specialization should
665 add "static const bool flow = true;". For instance:
670 struct SequenceTraits<MyList> {
671 static size_t size(IO &io, MyList &list) { ... }
672 static MyListEl element(IO &io, MyList &list, size_t index) { ... }
674 // The existence of this member causes YAML I/O to use a flow sequence
675 static const bool flow = true;
678 With the above, if you used MyList as the data type in your native data
679 strucutures, then then when converted to YAML, a flow sequence of integers
680 will be used (e.g. [ 10, -3, 4 ]).
685 Since a common source of sequences is std::vector<>, YAML I/O provids macros:
686 LLVM_YAML_IS_SEQUENCE_VECTOR() and LLVM_YAML_IS_FLOW_SEQUENCE_VECTOR() which
687 can be used to easily specify SequenceTraits<> on a std::vector type. YAML
688 I/O does not partial specialize SequenceTraits on std::vector<> because that
689 would force all vectors to be sequences. An example use of the macros:
693 std::vector<MyType1>;
694 std::vector<MyType2>;
695 LLVM_YAML_IS_SEQUENCE_VECTOR(MyType1)
696 LLVM_YAML_IS_FLOW_SEQUENCE_VECTOR(MyType2)
703 YAML allows you to define multiple "documents" in a single YAML file. Each
704 new document starts with a left aligned "---" token. The end of all documents
705 is denoted with a left aligned "..." token. Many users of YAML will never
706 have need for multiple documents. The top level node in their YAML schema
707 will be a mapping or sequence. For those cases, the following is not needed.
708 But for cases where you do want multiple documents, you can specify a
709 trait for you document list type. The trait has the same methods as
710 SequenceTraits but is named DocumentListTraits. For example:
715 struct DocumentListTraits<MyDocList> {
716 static size_t size(IO &io, MyDocList &list) { ... }
717 static MyDocType element(IO &io, MyDocList &list, size_t index) { ... }
723 When an llvm::yaml::Input or llvm::yaml::Output object is created their
724 constructors take an optional "context" parameter. This is a pointer to
725 whatever state information you might need.
727 For instance, in a previous example we showed how the conversion type for a
728 flags field could be determined at runtime based on the value of another field
729 in the mapping. But what if an inner mapping needs to know some field value
730 of an outer mapping? That is where the "context" parameter comes in. You
731 can set values in the context in the outer map's mapping() method and
732 retrieve those values in the inner map's mapping() method.
734 The context value is just a void*. All your traits which use the context
735 and operate on your native data types, need to agree what the context value
736 actually is. It could be a pointer to an object or struct which your various
737 traits use to shared context sensitive information.
743 The llvm::yaml::Output class is used to generate a YAML document from your
744 in-memory data structures, using traits defined on your data types.
745 To instantiate an Output object you need an llvm::raw_ostream, and optionally
750 class Output : public IO {
752 Output(llvm::raw_ostream &, void *context=NULL);
754 Once you have an Output object, you can use the C++ stream operator on it
755 to write your native data as YAML. One thing to recall is that a YAML file
756 can contain multiple "documents". If the top level data structure you are
757 streaming as YAML is a mapping, scalar, or sequence, then Output assumes you
758 are generating one document and wraps the mapping output
759 with "``---``" and trailing "``...``".
763 using llvm::yaml::Output;
765 void dumpMyMapDoc(const MyMapType &info) {
766 Output yout(llvm::outs());
770 The above could produce output like:
779 On the other hand, if the top level data structure you are streaming as YAML
780 has a DocumentListTraits specialization, then Output walks through each element
781 of your DocumentList and generates a "---" before the start of each element
782 and ends with a "...".
786 using llvm::yaml::Output;
788 void dumpMyMapDoc(const MyDocListType &docList) {
789 Output yout(llvm::outs());
793 The above could produce output like:
808 The llvm::yaml::Input class is used to parse YAML document(s) into your native
809 data structures. To instantiate an Input
810 object you need a StringRef to the entire YAML file, and optionally a context
815 class Input : public IO {
817 Input(StringRef inputContent, void *context=NULL);
819 Once you have an Input object, you can use the C++ stream operator to read
820 the document(s). If you expect there might be multiple YAML documents in
821 one file, you'll need to specialize DocumentListTraits on a list of your
822 document type and stream in that document list type. Otherwise you can
823 just stream in the document type. Also, you can check if there was
824 any syntax errors in the YAML be calling the error() method on the Input
829 // Reading a single document
830 using llvm::yaml::Input;
832 Input yin(mb.getBuffer());
834 // Parse the YAML file
845 // Reading multiple documents in one file
846 using llvm::yaml::Input;
848 LLVM_YAML_IS_DOCUMENT_LIST_VECTOR(std::vector<MyDocType>)
850 Input yin(mb.getBuffer());
852 // Parse the YAML file
853 std::vector<MyDocType> theDocList;