Quickstart

Example code for every feature. More...

Topics
	Sample helpers
	Helper utilities used in the sample.

Functions
void	sample_lightning_overview ()
	a lightning tour over most features see sample_quick_overview
void	sample_quick_overview ()
	a brief tour over most features
void	sample_substr ()
	demonstrate usage of ryml::substr and ryml::csubstr
void	sample_parse_file ()
	demonstrate how to load a YAML file from disk to parse with ryml.
void	sample_parse_in_place ()
	demonstrate in-place parsing of a mutable YAML source buffer.
void	sample_parse_in_arena ()
	demonstrate parsing of a read-only YAML source buffer
void	sample_parse_reuse_tree ()
	demonstrate reuse/modification of tree when parsing
void	sample_parse_reuse_parser ()
	Demonstrates reuse of an existing parser.
void	sample_parse_reuse_tree_and_parser ()
	for ultimate speed when parsing multiple times, reuse both the tree and parser
void	sample_iterate_trees ()
	shows how to programatically iterate through trees
void	sample_create_trees ()
	shows how to programatically create trees
void	sample_tree_arena ()
	demonstrates explicit and implicit interaction with the tree's string arena.
void	sample_fundamental_types ()
	ryml provides facilities for serializing and deserializing the C++ fundamental types, including boolean and null values; this is provided by the several overloads in to_chars: generalized chars to value and from_chars: generalized chars to value.
void	sample_empty_null_values ()
	Shows how to deal with empty/null values.
void	sample_formatting ()
	ryml provides facilities for formatting/deformatting (imported from c4core into the ryml namespace).
void	sample_base64 ()
	demonstrates how to read and write base64-encoded blobs.
void	sample_user_scalar_types ()
	to add scalar types (ie leaf types converting to/from string), define the functions above for those types.
void	sample_user_container_types ()
	shows how to serialize/deserialize container types.
void	sample_std_types ()
	demonstrates usage with the std implementations provided by ryml in the ryml_std.hpp header
void	sample_float_precision ()
	control precision of serialized floats
void	sample_emit_to_container ()
	demonstrates how to emit to a linear container of char
void	sample_emit_to_stream ()
	demonstrates how to emit to a stream-like structure
void	sample_emit_to_file ()
	demonstrates how to emit to a FILE*
void	sample_emit_nested_node ()
	just like parsing into a nested node, you can also emit from a nested node.
void	sample_style ()
	query/set/modify node style to control formatting of emitted YAML code.
void	sample_style_flow_formatting ()
	Shows how to control formatting of flow styles.
void	sample_style_flow_ml_indent ()
	control the indentation of emitted flow multiline containers
void	sample_json ()
	shows how to parse and emit JSON.
void	sample_anchors_and_aliases ()
	demonstrates usage with anchors and alias references.
void	sample_anchors_and_aliases_create ()
	demonstrates how to use the API to programatically create anchors and aliases
void	sample_tags ()
void	sample_tag_directives ()
void	sample_docs ()
void	sample_error_handler ()
	demonstrates how to set a custom error handler for ryml
void	sample_error_basic ()
void	sample_error_parse ()
void	sample_error_visit ()
	Visit errors happen when an error is triggered while reading from a node.
void	sample_error_visit_location ()
	It is possible to obtain the YAML location from a visit error: when the tree is obtained from parsing YAML, the messages may be enriched by using a parser set to track the locations.
void	sample_global_allocator ()
	demonstrates how to set the global allocator for ryml
void	sample_per_tree_allocator ()
void	sample_static_trees ()
	shows how to work around the static initialization order fiasco when using a static-duration ryml tree
void	sample_location_tracking ()
	demonstrates how to obtain the (zero-based) location of a node from a recently parsed tree

Detailed Description

Example code for every feature.

This file does a quick tour of ryml.

It has multiple self-contained and well-commented samples that illustrate how to use ryml, and how it works.

Although this is not a unit test, the samples are written as a sequence of actions and predicate checks to better convey what is the expected result at any stage. And to ensure the code here is correct and up to date, it's also run as part of the CI tests.

If something is unclear, please open an issue or send a pull request at https://github.com/biojppm/rapidyaml . If you have an issue while using ryml, it is also encouraged to try to reproduce the issue here, or look first through the relevant section.

Happy ryml'ing!

Note on use of raw strings

This sample uses multiline C strings instead of C++11 R"(raw strings)" because the doxygen parser is badly broken when handling raw strings.

Some guidance on building

The directories that exist side-by-side with this file contain several examples on how to build this with cmake, such that you can hit the ground running. See the relevant section of the main README for an overview of the different choices. I suggest starting first with the add_subdirectory example, treating it just like any other self-contained cmake project.

Or very quickly, to build and run this sample on your PC, start by creating this CMakeLists.txt:

cmake_minimum_required(VERSION 3.13)
project(ryml-quickstart LANGUAGES CXX)
include(FetchContent)
FetchContent_Declare(ryml
    GIT_REPOSITORY https://github.com/biojppm/rapidyaml.git
    GIT_TAG v0.15.1
    GIT_SHALLOW FALSE  # ensure submodules are checked out
)
FetchContent_MakeAvailable(ryml)
add_executable(ryml-quickstart ${ryml_SOURCE_DIR}/samples/quickstart.cpp)
target_link_libraries(ryml-quickstart ryml::ryml)
add_custom_target(run ryml-quickstart
    COMMAND $<TARGET_FILE:ryml-quickstart>
    DEPENDS ryml-quickstart
    COMMENT "running: $<TARGET_FILE:ryml-quickstart>")

Now run the following commands in the same folder:

# configure the project
cmake -S . -B build
# build and run
cmake --build build --target ryml-quickstart -j
# optionally, open in your IDE
cmake --open build

Function Documentation

sample_lightning_overview()

void sample_lightning_overview

(

)

a lightning tour over most features see sample_quick_overview

Definition at line 337 of file quickstart.cpp.

{
    // Parse YAML code in place, potentially mutating the buffer:
    char yml_buf[] = "{foo: 1, bar: [2, 3], john: doe}";
    ryml::Tree tree = ryml::parse_in_place(yml_buf);
 
    // read from the tree:
    ryml::NodeRef bar = tree["bar"];
    CHECK(bar[0].val() == "2");
    CHECK(bar[1].val() == "3");
    CHECK(bar[0].val().str == yml_buf + 15); // points at the source buffer
    CHECK(bar[1].val().str == yml_buf + 18);
 
    // deserializing:
    int bar0 = 0, bar1 = 0;
    bar[0] >> bar0;
    bar[1] >> bar1;
    CHECK(bar0 == 2);
    CHECK(bar1 == 3);
 
    // serializing:
    bar[0] << 10; // creates a string in the tree's arena
    bar[1] << 11;
    CHECK(bar[0].val() == "10");
    CHECK(bar[1].val() == "11");
 
    // add nodes
    bar.append_child() << 12; // see also operator= (explanation below)
    CHECK(bar[2].val() == "12");
 
    // emit tree
    std::string expected = "{foo: 1,bar: [10,11,12],john: doe}";
    // emit tree to std::string
    CHECK(ryml::emitrs_yaml<std::string>(tree) == expected);
    // emit tree to FILE*
    ryml::emit_yaml(tree, stdout); printf("\n");
    // emit tree to ostream
    std::cout << tree << "\n";
 
    // emit node
    ryml::ConstNodeRef foo = tree["foo"];
    expected = "foo: 1\n";
    // emit node to std::string
    CHECK(ryml::emitrs_yaml<std::string>(foo) == expected);
    // emit node to FILE*
    ryml::emit_yaml(foo, stdout);
    // emit node to ostream
    std::cout << foo;
}

sample_quick_overview()

void sample_quick_overview

(

)

a brief tour over most features

Definition at line 391 of file quickstart.cpp.

{
    // Parse YAML code in place, potentially mutating the buffer:
    char yml_buf[] = ""
        "foo: 1"         "\n"
        "bar: [2, 3]"    "\n"
        "john: doe"      "\n"
        "";
    ryml::Tree tree = ryml::parse_in_place(yml_buf);
 
    // The resulting tree contains only views to the parsed string. If
    // the string was parsed in place, then the string must outlive
    // the tree! This works here because both `yml_buf` and `tree`
    // live on the same scope, so have the same lifetime.
 
    // It is also possible to:
    //
    //   - parse a read-only buffer using parse_in_arena(). This
    //     copies the YAML buffer to the tree's arena, and spares the
    //     headache of the string's lifetime.
    //
    //   - reuse an existing tree (advised)
    //
    //   - reuse an existing parser (advised)
    //
    //   - parse into an existing node deep in a tree
    //
    // Note: it will always be significantly faster to parse in place
    // and reuse tree+parser.
    //
    // Below you will find samples that show how to achieve reuse; but
    // please note that for brevity and clarity, many of the examples
    // here are parsing in the arena, and not reusing tree or parser.
 
 
    //------------------------------------------------------------------
    // API overview
 
    // ryml has a two-level API:
    //
    // The lower level index API is based on the indices of nodes,
    // where the node's id is the node's position in the tree's data
    // array. This API is very efficient, but somewhat difficult to use:
    ryml::id_type root_id = tree.root_id();
    ryml::id_type bar_id = tree.find_child(root_id, "bar"); // need to get the index right
    CHECK(tree.is_map(root_id)); // all of the index methods are in the tree
    CHECK(tree.is_seq(bar_id));  // ... and receive the subject index
 
    // The node API is a lightweight abstraction sitting on top of the
    // index API, but offering a much more convenient interaction:
    ryml::ConstNodeRef root = tree.rootref();  // a const node reference
    ryml::ConstNodeRef bar = tree["bar"];
    CHECK(root.is_map());
    CHECK(bar.is_seq());
 
    // A node ref is a lightweight handle to the tree and associated id:
    CHECK(root.tree() == &tree); // a node ref points at its tree, WITHOUT refcount
    CHECK(root.id() == root_id); // a node ref's id is the index of the node
    CHECK(bar.id() == bar_id);   // a node ref's id is the index of the node
 
    // The node API translates very cleanly to the index API, so most
    // of the code examples below are using the node API.
 
    // WARNING. A node ref holds a raw pointer to the tree. Care must
    // be taken to ensure the lifetimes match, so that a node will
    // never access the tree after the goes out of scope.
 
 
    //------------------------------------------------------------------
    // To read the parsed tree
 
    // ConstNodeRef::operator[] does a lookup, is O(num_children[node]).
    CHECK(tree["foo"].is_keyval());
    CHECK(tree["foo"].val() == "1");   // get the val of a node (must be leaf node, otherwise it is a container and has no val)
    CHECK(tree["foo"].key() == "foo"); // get the key of a node (must be child of a map, otherwise it has no key)
    CHECK(tree["bar"].is_seq());
    CHECK(tree["bar"].has_key());
    CHECK(tree["bar"].key() == "bar");
    // maps use string keys, seqs use index keys:
    CHECK(tree["bar"][0].val() == "2");
    CHECK(tree["bar"][1].val() == "3");
    CHECK(tree["john"].val() == "doe");
    // An index key is the position of the child within its parent,
    // so even maps can also use int keys, if the key position is
    // known.
    CHECK(tree[0].id() == tree["foo"].id());
    CHECK(tree[1].id() == tree["bar"].id());
    CHECK(tree[2].id() == tree["john"].id());
    // Tree::operator[](int) searches a ***root*** child by its position.
    CHECK(tree[0].id() == tree["foo"].id());  // 0: first child of root
    CHECK(tree[1].id() == tree["bar"].id());  // 1: second child of root
    CHECK(tree[2].id() == tree["john"].id()); // 2: third child of root
    // NodeRef::operator[](int) searches a ***node*** child by its position:
    CHECK(bar[0].val() == "2"); // 0 means first child of bar
    CHECK(bar[1].val() == "3"); // 1 means second child of bar
    // NodeRef::operator[](string):
    // A string key is the key of the node: lookup is by name. So it
    // is only available for maps, and it is NOT available for seqs,
    // since seq members do not have keys.
    CHECK(tree["foo"].key() == "foo");
    CHECK(tree["bar"].key() == "bar");
    CHECK(tree["john"].key() == "john");
    CHECK(bar.is_seq());
    // CHECK(bar["BOOM!"].is_seed()); // error, seqs do not have key lookup
 
    // Note that maps can also use index keys as well as string keys:
    CHECK(root["foo"].id() == root[0].id());
    CHECK(root["bar"].id() == root[1].id());
    CHECK(root["john"].id() == root[2].id());
 
    // IMPORTANT. The ryml tree uses an index-based linked list for
    // storing children, so the complexity of
    // `Tree::operator[csubstr]` and `Tree::operator[id_type]` is O(n),
    // linear on the number of root children. If you use
    // `Tree::operator[]` with a large tree where the root has many
    // children, you will see a performance hit.
    //
    // To avoid this hit, you can create your own accelerator
    // structure. For example, before doing a lookup, do a single
    // traverse at the root level to fill an `map<csubstr,id_type>`
    // mapping key names to node indices; with a node index, a lookup
    // (via `Tree::get()`) is O(1), so this way you can get O(log n)
    // lookup from a key. (But please do not use `std::map` if you
    // care about performance; use something else like a flat map or
    // sorted vector).
    //
    // As for node refs, the difference from `NodeRef::operator[]` and
    // `ConstNodeRef::operator[]` to `Tree::operator[]` is that the
    // latter refers to the root node, whereas the former are invoked
    // on their target node. But the lookup process works the same for
    // both and their algorithmic complexity is the same: they are
    // both linear in the number of direct children. But of course,
    // depending on the data, that number may be very different from
    // one to another.
 
 
    //------------------------------------------------------------------
    // Hierarchy:
 
    {
        ryml::ConstNodeRef foo = root.first_child();
        ryml::ConstNodeRef john = root.last_child();
        CHECK(tree.size() == 6); // O(1) number of nodes in the tree
        CHECK(root.num_children() == 3); // O(num_children[root])
        CHECK(foo.num_siblings() == 3); // O(num_children[parent(foo)])
        CHECK(foo.parent().id() == root.id()); // parent() is O(1)
        CHECK(root.first_child().id() == root["foo"].id()); // first_child() is O(1)
        CHECK(root.last_child().id() == root["john"].id()); // last_child() is O(1)
        CHECK(john.first_sibling().id() == foo.id());
        CHECK(foo.last_sibling().id() == john.id());
        // prev_sibling(), next_sibling(): (both are O(1))
        CHECK(foo.num_siblings() == root.num_children());
        CHECK(foo.prev_sibling().id() == ryml::NONE); // foo is the first_child()
        CHECK(foo.next_sibling().key() == "bar");
        CHECK(foo.next_sibling().next_sibling().key() == "john");
        CHECK(foo.next_sibling().next_sibling().next_sibling().id() == ryml::NONE); // john is the last_child()
    }
 
 
    //------------------------------------------------------------------
    // Iterating:
    {
        ryml::csubstr expected_keys[] = {"foo", "bar", "john"};
        // iterate children using the high-level node API:
        {
            ryml::id_type count = 0;
            for(ryml::ConstNodeRef const& child : root.children())
                CHECK(child.key() == expected_keys[count++]);
        }
        // iterate siblings using the high-level node API:
        {
            ryml::id_type count = 0;
            for(ryml::ConstNodeRef const& child : root["foo"].siblings())
                CHECK(child.key() == expected_keys[count++]);
        }
        // iterate children using the lower-level tree index API:
        {
            ryml::id_type count = 0;
            for(ryml::id_type child_id = tree.first_child(root_id); child_id != ryml::NONE; child_id = tree.next_sibling(child_id))
                CHECK(tree.key(child_id) == expected_keys[count++]);
        }
        // iterate siblings using the lower-level tree index API:
        // (notice the only difference from above is in the loop
        // preamble, which calls tree.first_sibling(bar_id) instead of
        // tree.first_child(root_id))
        {
            ryml::id_type count = 0;
            for(ryml::id_type child_id = tree.first_sibling(bar_id); child_id != ryml::NONE; child_id = tree.next_sibling(child_id))
                CHECK(tree.key(child_id) == expected_keys[count++]);
        }
    }
 
 
    //------------------------------------------------------------------
    // Gotchas:
 
    // ryml uses assertions to prevent you from trying to obtain
    // things that do not exist. For example:
 
    {
        ryml::ConstNodeRef seq_node = tree["bar"];
        ryml::ConstNodeRef val_node = seq_node[0];
 
        CHECK(seq_node.is_seq());          // seq is a container
        CHECK(!seq_node.has_val());        // ... so it has no val
        //CHECK(seq_node.val() == BOOM!);  // ... so attempting to get a val is undefined behavior
 
        CHECK(val_node.parent() == seq_node); // belongs to a seq
        CHECK(!val_node.has_key());           // ... so it has no key
        //CHECK(val_node.key() == BOOM!);     // ... so attempting to get a key is undefined behavior
 
        CHECK(val_node.is_val());         // this node is a val
        //CHECK(val_node.first_child() == BOOM!); // ... so attempting to get a child is undefined behavior
 
        // assertions are also present in methods that /may/ read the val:
        CHECK(seq_node.is_seq());              // seq is a container
        //CHECK(seq_node.val_is_null() BOOM!); // so cannot get the val to check
    }
 
    // By default, assertions are enabled unless the NDEBUG macro is
    // defined (which happens in release builds).
    //
    // This adheres to the pay-only-for-what-you-use philosophy: if
    // you are sure that your intent is correct, why would you need to
    // pay the runtime cost for the assertions?
    //
    // The downside, of course, is that you may be unsure, or your
    // code may be wrong; and then release builds may end up doing
    // something wrong.
    //
    // So in that case, you can use the appropriate ryml predicates to
    // check your intent (as in the examples above), or you can
    // explicitly enable/disable assertions, by defining the macro
    // RYML_USE_ASSERT to a proper value (see c4/yml/common.hpp). This
    // will make problematic code trigger a call to the appropriate
    // error callback.
    //
    // Also, to be clear, this does not apply to parse errors
    // occurring when the YAML is parsed. Parse errors are always
    // detected by the parser: this is not done through assertions,
    // but through the appropriate error checking mechanism. So
    // checking for these errors is always enabled and cannot be
    // switched off.
 
 
    //------------------------------------------------------------------
    // Deserializing: use operator>>
    {
        int foo = 0, bar0 = 0, bar1 = 0;
        std::string john_str;
        std::string bar_str;
        root["foo"] >> foo;
        root["bar"][0] >> bar0;
        root["bar"][1] >> bar1;
        root["john"] >> john_str; // requires from_chars(std::string). see serialization samples below.
        root["bar"] >> ryml::key(bar_str); // to deserialize the key, use the tag function ryml::key()
        CHECK(foo == 1);
        CHECK(bar0 == 2);
        CHECK(bar1 == 3);
        CHECK(john_str == "doe");
        CHECK(bar_str == "bar");
    }
 
 
    //------------------------------------------------------------------
    // Modifying existing nodes: operator= vs operator<<
 
    // As implied by its name, ConstNodeRef is a reference to a const
    // node. It can be used to read from the node, but not write to it
    // or modify the hierarchy of the node. If any modification is
    // desired then a NodeRef must be used instead:
    ryml::NodeRef wroot = tree.rootref(); // writeable root
 
    // operator= assigns an existing string to the receiving node.
    // The contents are NOT copied, and the string pointer will be in
    // effect until the tree goes out of scope! So BEWARE to only
    // assign from strings outliving the tree.
    wroot["foo"] = "says you";
    wroot["bar"][0] = "-2";
    wroot["bar"][1] = "-3";
    wroot["john"] = "ron";
    // Now the tree is _pointing_ at the memory of the strings above.
    // In this case it is OK because those are static strings, located
    // in the executable's static section, and will outlive the tree.
    CHECK(root["foo"].val() == "says you");
    CHECK(root["bar"][0].val() == "-2");
    CHECK(root["bar"][1].val() == "-3");
    CHECK(root["john"].val() == "ron");
    // But WATCHOUT: do not assign from temporary objects:
    // {
    //     std::string crash("will dangle");
    //     root["john"] = ryml::to_csubstr(crash);
    // }
    // CHECK(root["john"] == "dangling"); // CRASH! the string was deallocated
 
    // operator<<: for cases where the lifetime of the string is
    // problematic WRT the tree, you can create and save a string in
    // the tree using operator<<. It first serializes values to a
    // string arena owned by the tree, then assigns the serialized
    // string to the receiving node. This avoids constraints with the
    // lifetime, since the arena lives with the tree.
    CHECK(tree.arena().empty());
    wroot["foo"] << "says who";  // requires to_chars(). see serialization samples below.
    wroot["bar"][0] << 20;
    wroot["bar"][1] << 30;
    wroot["john"] << "deere";
    CHECK(root["foo"].val() == "says who");
    CHECK(root["bar"][0].val() == "20");
    CHECK(root["bar"][1].val() == "30");
    CHECK(root["john"].val() == "deere");
    CHECK(tree.arena() == "says who2030deere"); // the result of serializations to the tree arena
    // using operator<< instead of operator=, the crash above is avoided:
    {
        std::string ok("in_scope");
        // root["john"] = ryml::to_csubstr(ok); // don't, will dangle
        wroot["john"] << ryml::to_csubstr(ok); // OK, copy to the tree's arena
    }
    CHECK(root["john"].val() == "in_scope"); // OK! val is now in the tree's arena
    // serializing floating points:
    wroot["float"] << 2.4f;
    // to force a particular precision or float format:
    // (see sample_float_precision() and sample_formatting())
    wroot["digits"] << ryml::fmt::real(2.4, /*num_digits*/6, ryml::FTOA_FLOAT);
    CHECK(tree.arena() == "says who2030deerein_scope2.42.400000"); // the result of serializations to the tree arena
 
 
    //------------------------------------------------------------------
    // Adding new nodes:
 
    // adding a keyval node to a map:
    CHECK(root.num_children() == 5);
    wroot["newkeyval"] = "shiny and new"; // using these strings
    wroot.append_child() << ryml::key("newkeyval (serialized)") << "shiny and new (serialized)"; // serializes and assigns the serialization
    CHECK(root.num_children() == 7);
    CHECK(root["newkeyval"].key() == "newkeyval");
    CHECK(root["newkeyval"].val() == "shiny and new");
    CHECK(root["newkeyval (serialized)"].key() == "newkeyval (serialized)");
    CHECK(root["newkeyval (serialized)"].val() == "shiny and new (serialized)");
    CHECK( ! tree.in_arena(root["newkeyval"].key())); // it's using directly the static string above
    CHECK( ! tree.in_arena(root["newkeyval"].val())); // it's using directly the static string above
    CHECK(   tree.in_arena(root["newkeyval (serialized)"].key())); // it's using a serialization of the string above
    CHECK(   tree.in_arena(root["newkeyval (serialized)"].val())); // it's using a serialization of the string above
    // adding a val node to a seq:
    CHECK(root["bar"].num_children() == 2);
    wroot["bar"][2] = "oh so nice";
    wroot["bar"][3] << "oh so nice (serialized)";
    CHECK(root["bar"].num_children() == 4);
    CHECK(root["bar"][2].val() == "oh so nice");
    CHECK(root["bar"][3].val() == "oh so nice (serialized)");
    // adding a seq node:
    CHECK(root.num_children() == 7);
    wroot["newseq"] |= ryml::SEQ;
    wroot.append_child() << ryml::key("newseq (serialized)") |= ryml::SEQ;
    CHECK(root.num_children() == 9);
    CHECK(root["newseq"].num_children() == 0);
    CHECK(root["newseq"].is_seq());
    CHECK(root["newseq (serialized)"].num_children() == 0);
    CHECK(root["newseq (serialized)"].is_seq());
    // adding a map node:
    CHECK(root.num_children() == 9);
    wroot["newmap"] |= ryml::MAP;
    wroot.append_child() << ryml::key("newmap (serialized)") |= ryml::MAP;
    CHECK(root.num_children() == 11);
    CHECK(root["newmap"].num_children() == 0);
    CHECK(root["newmap"].is_map());
    CHECK(root["newmap (serialized)"].num_children() == 0);
    CHECK(root["newmap (serialized)"].is_map());
    //
    // When the tree is const, operator[] searches the tree for a
    // matching key/position, and returns it, or raises an error if
    // none is found.
    //
    // When the tree is mutable, operator[] will not mutate the tree
    // until the returned node is written to. operator[] first
    // searches the tree for a node matching the key/position and
    // returns it if it exists; otherwise, the result is a NodeRef
    // object which keeps in itself the required information to write
    // to the proper place in the tree. This is called being in a
    // "seed" state.
    //
    // This means that passing a key/position which does not exist
    // will not mutate the tree, but will instead store (in the
    // returned NodeRef) the proper place of the tree to be able to do
    // so, if and when it is required. This is why the node is said to
    // be in "seed" state - it allows creating the entry in the tree
    // in the future.
    //
    // This is a significant difference from eg, the behavior of
    // std::map, which mutates the map immediately within the call to
    // operator[].
    //
    // All of the points above apply only if the tree is mutable. If
    // the tree is const, then a NodeRef cannot be obtained from it;
    // only a ConstNodeRef, which can never be used to mutate the
    // tree.
    //
    CHECK(!root.has_child("I am not nothing"));
    ryml::NodeRef nothing;
    CHECK(nothing.invalid());    // invalid because it points at nothing
    nothing = wroot["I am nothing"];
    CHECK(!nothing.invalid());   // points at the tree, and a specific place in the tree
    CHECK(nothing.is_seed()); // ... but nothing is there yet.
    CHECK(!root.has_child("I am nothing")); // same as above
    CHECK(!nothing.readable()); // ... and this node cannot be used to
                                // read anything from the tree
    ryml::NodeRef something = wroot["I am something"];
    ryml::ConstNodeRef constsomething = wroot["I am something"];
    CHECK(!root.has_child("I am something")); // same as above
    CHECK(!something.invalid());
    CHECK(something.is_seed()); // same as above
    CHECK(!something.readable()); // same as above
    CHECK(constsomething.invalid()); // NOTE: because a ConstNodeRef cannot be
                                     // used to mutate a tree, it is only valid()
                                     // if it is pointing at an existing node.
    something = "indeed";  // this will commit the seed to the tree, mutating at the proper place
    CHECK(root.has_child("I am something"));
    CHECK(root["I am something"].val() == "indeed");
    CHECK(!something.invalid()); // it was already valid
    CHECK(!something.is_seed()); // now the tree has this node, so the
                                 // ref is no longer a seed
    CHECK(something.readable()); // and it is now readable
    //
    // now the constref is also valid (but it needs to be reassigned):
    ryml::ConstNodeRef constsomethingnew = wroot["I am something"];
    CHECK(!constsomethingnew.invalid());
    CHECK(constsomethingnew.readable());
    // note that the old constref is now stale, because it only keeps
    // the state at creation:
    CHECK(constsomething.invalid());
    CHECK(!constsomething.readable());
    //
    // -----------------------------------------------------------
    // Remember: a seed node cannot be used to read from the tree!
    // -----------------------------------------------------------
    //
    // The seed node needs to be created and become readable first.
    //
    // Trying to invoke any tree-reading method on a node that is not
    // readable will cause an assertion (see RYML_USE_ASSERT).
    //
    // It is your responsibility to verify that the preconditions are
    // met. If you are not sure about the structure of your data,
    // write your code defensively to signify your full intent:
    //
    ryml::NodeRef wbar = wroot["bar"];
    if(wbar.readable() && wbar.is_seq()) // .is_seq() requires .readable()
    {
        CHECK(wbar[0].readable() && wbar[0].val() == "20");
        CHECK( ! wbar[100].readable());
        CHECK( ! wbar[100].readable() || wbar[100].val() == "100"); // <- no crash because it is not .readable(), so never tries to call .val()
        // this would work as well:
        CHECK( ! wbar[0].is_seed() && wbar[0].val() == "20");
        CHECK(wbar[100].is_seed() || wbar[100].val() == "100");
    }
 
 
    //------------------------------------------------------------------
    // .operator[]() vs .at()
 
    // (Const)NodeRef::operator[] is an analogue to std::vector::operator[].
    // (Const)NodeRef::at() is an analogue to std::vector::at()
    //
    // at() will always check the subject node is .readable().
    //
    // [] is meant for the happy path, and unverified in Release
    // builds.
    {
        // in this example we will be checking errors, so set up a
        // temporary error handler to catch them:
        ScopedErrorHandlerExample errh; // calls ryml::set_callbacks()
        // instantiate the tree after errh
        ryml::Tree err_tree = ryml::parse_in_arena("{foo: bar}");
        // ... so that the tree uses the current callbacks:
        CHECK(err_tree.callbacks() == errh.callbacks());
        // node does not exist, only a seed node
        ryml::NodeRef seed_node = err_tree["this"];
        // ... therefore not .readable()
        CHECK(!seed_node.readable());
        // using .at() reliably produces an error:
        CHECK(errh.check_error_occurs([&]{
            return seed_node.at("is").at("an").at("invalid").at("operation");
            //               ^
            //               error occurs here because it is unreadable
        }));
        // ... but using [] fails only when RYML_USE_ASSERT is
        // defined. otherwise, it's the dreaded Undefined Behavior:
        CHECK(errh.check_assertion_occurs([&]{
            return seed_node["is"]["an"]["invalid"]["operation"];
            //              ^
            //              assertion occurs here because it is unreadable
        }));
    }
 
 
    //------------------------------------------------------------------
    // Emitting:
 
    ryml::csubstr expected_result =
        "foo: says who"                                         "\n"
        "bar: [20,30,oh so nice,oh so nice (serialized)]"       "\n"
        "john: in_scope"                                        "\n"
        "float: 2.4"                                            "\n"
        "digits: 2.400000"                                      "\n"
        "newkeyval: shiny and new"                              "\n"
        "newkeyval (serialized): shiny and new (serialized)"    "\n"
        "newseq: []"                                            "\n"
        "newseq (serialized): []"                               "\n"
        "newmap: {}"                                            "\n"
        "newmap (serialized): {}"                               "\n"
        "I am something: indeed"                                "\n"
        "";
 
    // emit to a FILE*
    ryml::emit_yaml(tree, stdout);
    // emit to a stream
    // (this is templated on the stream, so it works both with
    // standard streams like std::ostream/std::stringstream, or with
    // user-defined streams not inheriting from ostream)
    std::stringstream ss;
    ss << tree;
    std::string stream_result = ss.str();
    // emit to a buffer:
    std::string str_result = ryml::emitrs_yaml<std::string>(tree);
    // can emit to any given buffer:
    char buf[1024];
    ryml::csubstr buf_result = ryml::emit_yaml(tree, buf);
    // now check
    CHECK(buf_result == expected_result);
    CHECK(str_result == expected_result);
    CHECK(stream_result == expected_result);
    // There are many possibilities to emit to buffer;
    // please look at the emit sample functions below.
 
 
    //------------------------------------------------------------------
    // ConstNodeRef vs NodeRef
 
    ryml::NodeRef noderef = tree["bar"][0];
    ryml::ConstNodeRef constnoderef = tree["bar"][0];
 
    // ConstNodeRef cannot be used to mutate the tree:
    //constnoderef = "21";  // compile error
    //constnoderef << "22"; // compile error
    // ... but a NodeRef can:
    noderef = "21";         // ok, can assign because it's not const
    CHECK(tree["bar"][0].val() == "21");
    noderef << "22";        // ok, can serialize and assign because it's not const
    CHECK(tree["bar"][0].val() == "22");
 
    // it is not possible to obtain a NodeRef from a ConstNodeRef:
    // noderef = constnoderef; // compile error
 
    // it is always possible to obtain a ConstNodeRef from a NodeRef:
    constnoderef = noderef;    // ok can assign const <- nonconst
 
    // If a tree is const, then only ConstNodeRef's can be
    // obtained from that tree:
    ryml::Tree const& consttree = tree;
    //noderef = consttree["bar"][0];    // compile error
    noderef = tree["bar"][0];           // ok
    constnoderef = consttree["bar"][0]; // ok
 
    // ConstNodeRef and NodeRef can be compared for equality.
    // Equality means they point at the same node.
    CHECK(constnoderef == noderef);
    CHECK(!(constnoderef != noderef));
 
 
    //------------------------------------------------------------------
    // Getting the location of nodes in the source:
    //
    // Location tracking is opt-in:
    ryml::EventHandlerTree evt_handler = {};
    ryml::Parser parser(&evt_handler, ryml::ParserOptions().locations(true));
    // Now the parser will start by building the accelerator structure:
    ryml::Tree tree2 = parse_in_arena(&parser, "expected.yml", expected_result);
    // ... and use it when querying
    ryml::ConstNodeRef subject_node = tree2["bar"][1];
    CHECK(subject_node.val() == "30");
    ryml::Location loc = subject_node.location(parser);
    CHECK(parser.location_contents(loc).begins_with("30"));
    CHECK(loc.line == 1u);
    CHECK(loc.col == 9u);
    // For further details in location tracking,
    // refer to the sample function below.
 
    //------------------------------------------------------------------
    // Dealing with UTF8
    ryml::Tree langs = ryml::parse_in_arena(""
        "#"                                                       "\n"
        "# UTF8/16/32 can be placed directly in any scalar:"      "\n"
        "#"                                                       "\n"
        "en: Planet (Gas)"                                        "\n"
        "fr: Planète (Gazeuse)"                                   "\n"
        "ru: Планета (Газ)"                                      "\n"
        "ja: 惑星（ガス）"                                        "\n"
        "zh: 行星（气体）"                                        "\n"
        "#"                                                       "\n"
        "# UTF8 decoding only happens in double-quoted strings,"  "\n"
        "# as per the YAML standard"                              "\n"
        "#"                                                       "\n"
        "decode this: \"\\u263A c\\x61f\\xE9\""                   "\n"
        "and this as well: \"\\u2705 \\U0001D11E\""               "\n"
        "not decoded: '\\u263A \\xE2\\x98\\xBA'"                  "\n"
        "neither this: '\\u2705 \\U0001D11E'"                     "\n");
    // in-place UTF8 just works:
    CHECK(langs["en"].val() == "Planet (Gas)");
    CHECK(langs["fr"].val() == "Planète (Gazeuse)");
    CHECK(langs["ru"].val() == "Планета (Газ)");
    CHECK(langs["ja"].val() == "惑星（ガス）");
    CHECK(langs["zh"].val() == "行星（气体）");
    // and \x \u \U codepoints are decoded, but only when they appear
    // inside double-quoted strings, as dictated by the YAML
    // standard:
    CHECK(langs["decode this"].val() == "☺ café");
    CHECK(langs["and this as well"].val() == "✅ 𝄞");
    CHECK(langs["not decoded"].val() == "\\u263A \\xE2\\x98\\xBA");
    CHECK(langs["neither this"].val() == "\\u2705 \\U0001D11E");
}

sample_substr()

void sample_substr

(

)

demonstrate usage of ryml::substr and ryml::csubstr

These types are imported from the c4core library into the ryml namespace You may have noticed above the use of a csubstr class. This class is defined in another library, c4core, which is imported by ryml. This is a library I use with my projects consisting of multiplatform low-level utilities. One of these is c4::csubstr (the name comes from "constant substring") which is a non-owning read-only string view, with many methods that make it practical to use (I would certainly argue more practical than std::string). In fact, c4::csubstr and its writeable counterpart c4::substr are the workhorses of the ryml parsing and serialization code.

See also

Substring: read/write string views

Definition at line 1029 of file quickstart.cpp.

{
    // substr is a mutable view: pointer and length to a string in memory.
    // csubstr is a const-substr (immutable).
 
    // construct from explicit args
    {
        const char foobar_str[] = "foobar";
        auto s = ryml::csubstr(foobar_str, strlen(foobar_str));
        CHECK(s == "foobar");
        CHECK(s.size() == 6);
        CHECK(s.data() == foobar_str);
        CHECK(s.size() == s.len);
        CHECK(s.data() == s.str);
    }
 
    // construct from a string array
    {
        const char foobar_str[] = "foobar";
        ryml::csubstr s = foobar_str;
        CHECK(s == "foobar");
        CHECK(s != "foobar0");
        CHECK(s.size() == 6); // does not include the terminating \0
        CHECK(s.data() == foobar_str);
        CHECK(s.size() == s.len);
        CHECK(s.data() == s.str);
    }
    // you can also declare directly in-place from an array:
    {
        ryml::csubstr s = "foobar";
        CHECK(s == "foobar");
        CHECK(s != "foobar0");
        CHECK(s.size() == 6);
        CHECK(s.size() == s.len);
        CHECK(s.data() == s.str);
    }
 
    // construct from a C-string:
    //
    // Since the input is only a pointer, the string length can only
    // be found with a call to strlen(). To make this cost evident, we
    // require calling to_csubstr():
    {
        const char *foobar_str = "foobar";
        ryml::csubstr s = ryml::to_csubstr(foobar_str);
        CHECK(s == "foobar");
        CHECK(s != "foobar0");
        CHECK(s.size() == 6);
        CHECK(s.size() == s.len);
        CHECK(s.data() == s.str);
    }
 
    // construct from a std::string: same approach as above.
    // requires inclusion of the <ryml/std/string.hpp> header
    // or of the umbrella header <ryml_std.hpp>.
    //
    // not including <string> in the default header is a deliberate
    // design choice to avoid including the heavy std:: allocation
    // machinery
    {
        std::string foobar_str = "foobar";
        ryml::csubstr s = ryml::to_csubstr(foobar_str); // defined in <ryml/std/string.hpp>
        CHECK(s == "foobar");
        CHECK(s != "foobar0");
        CHECK(s.size() == 6);
        CHECK(s.size() == s.len);
        CHECK(s.data() == s.str);
    }
 
    // convert substr -> csubstr
    {
        char buf[] = "foo";
        ryml::substr foo = buf;
        CHECK(foo.len == 3);
        CHECK(foo.data() == buf);
        ryml::csubstr cfoo = foo;
        CHECK(cfoo.data() == buf);
    }
    // cannot convert csubstr -> substr:
    {
        // ryml::substr foo2 = cfoo; // compile error: cannot write to csubstr
    }
 
    // construct from char[]/const char[]: mutable vs immutable memory
    {
        char const foobar_str_ro[] = "foobar"; // ro := read-only
        char       foobar_str_rw[] = "foobar"; // rw := read-write
        static_assert(std::is_array<decltype(foobar_str_ro)>::value, "this is an array");
        static_assert(std::is_array<decltype(foobar_str_rw)>::value, "this is an array");
        // csubstr <- read-only memory
        {
            ryml::csubstr foobar = foobar_str_ro;
            CHECK(foobar.data() == foobar_str_ro);
            CHECK(foobar.size() == strlen(foobar_str_ro));
            CHECK(foobar == "foobar"); // AKA strcmp
        }
        // csubstr <- read-write memory: you can create an immutable csubstr from mutable memory
        {
            ryml::csubstr foobar = foobar_str_rw;
            CHECK(foobar.data() == foobar_str_rw);
            CHECK(foobar.size() == strlen(foobar_str_rw));
            CHECK(foobar == "foobar"); // AKA strcmp
        }
        // substr <- read-write memory.
        {
            ryml::substr foobar = foobar_str_rw;
            CHECK(foobar.data() == foobar_str_rw);
            CHECK(foobar.size() == strlen(foobar_str_rw));
            CHECK(foobar == "foobar"); // AKA strcmp
        }
        // substr <- ro is impossible.
        {
            //ryml::substr foobar = foobar_str_ro; // compile error!
        }
    }
 
    // construct from char*/const char*: mutable vs immutable memory.
    // use to_substr()/to_csubstr()
    {
        char const* foobar_str_ro = "foobar";       // ro := read-only
        char        foobar_str_rw_[] = "foobar";    // rw := read-write
        char      * foobar_str_rw = foobar_str_rw_; // rw := read-write
        static_assert(!std::is_array<decltype(foobar_str_ro)>::value, "this is a decayed pointer");
        static_assert(!std::is_array<decltype(foobar_str_rw)>::value, "this is a decayed pointer");
        // csubstr <- read-only memory
        {
            //ryml::csubstr foobar = foobar_str_ro; // compile error: length is not known
            ryml::csubstr foobar = ryml::to_csubstr(foobar_str_ro);
            CHECK(foobar.data() == foobar_str_ro);
            CHECK(foobar.size() == strlen(foobar_str_ro));
            CHECK(foobar == "foobar"); // AKA strcmp
        }
        // csubstr <- read-write memory: you can create an immutable csubstr from mutable memory
        {
            ryml::csubstr foobar = ryml::to_csubstr(foobar_str_rw);
            CHECK(foobar.data() == foobar_str_rw);
            CHECK(foobar.size() == strlen(foobar_str_rw));
            CHECK(foobar == "foobar"); // AKA strcmp
        }
        // substr <- read-write memory.
        {
            ryml::substr foobar = ryml::to_substr(foobar_str_rw);
            CHECK(foobar.data() == foobar_str_rw);
            CHECK(foobar.size() == strlen(foobar_str_rw));
            CHECK(foobar == "foobar"); // AKA strcmp
        }
        // substr <- read-only is impossible.
        {
            //ryml::substr foobar = ryml::to_substr(foobar_str_ro); // compile error!
        }
    }
 
    // substr is mutable, without changing the size:
    {
        char buf[] = "foobar";
        ryml::substr foobar = buf;
        CHECK(foobar == "foobar");
        foobar[0] = 'F';            CHECK(foobar == "Foobar");
        foobar.back() = 'R';        CHECK(foobar == "FoobaR");
        foobar.reverse();           CHECK(foobar == "RabooF");
        foobar.reverse();           CHECK(foobar == "FoobaR");
        foobar.reverse_sub(1, 4);   CHECK(foobar == "FabooR");
        foobar.reverse_sub(1, 4);   CHECK(foobar == "FoobaR");
        foobar.reverse_range(2, 5); CHECK(foobar == "FoaboR");
        foobar.reverse_range(2, 5); CHECK(foobar == "FoobaR");
        foobar.replace('o', '0');   CHECK(foobar == "F00baR");
        foobar.replace('a', '_');   CHECK(foobar == "F00b_R");
        foobar.replace("_0b", 'a'); CHECK(foobar == "FaaaaR");
        foobar.toupper();           CHECK(foobar == "FAAAAR");
        foobar.tolower();           CHECK(foobar == "faaaar");
        foobar.fill('.');           CHECK(foobar == "......");
        // see also:
        // - .erase()
        // - .replace_all()
    }
 
    // sub-views
    {
        ryml::csubstr s = "fooFOObarBAR";
        CHECK(s.len == 12u);
        // sub(): <- first,[num]
        CHECK(s.sub(0)     == "fooFOObarBAR");
        CHECK(s.sub(0, 12) == "fooFOObarBAR");
        CHECK(s.sub(0,  3) == "foo"         );
        CHECK(s.sub(3)     ==    "FOObarBAR");
        CHECK(s.sub(3,  3) ==    "FOO"      );
        CHECK(s.sub(6)     ==       "barBAR");
        CHECK(s.sub(6,  3) ==       "bar"   );
        CHECK(s.sub(9)     ==          "BAR");
        CHECK(s.sub(9,  3) ==          "BAR");
        // first(): <- length
        CHECK(s.first(0) == ""   );
        CHECK(s.first(1) == "f"  );
        CHECK(s.first(2) != "f"  );
        CHECK(s.first(2) == "fo" );
        CHECK(s.first(3) == "foo");
        // last(): <- length
        CHECK(s.last(0) ==    "");
        CHECK(s.last(1) ==   "R");
        CHECK(s.last(2) ==  "AR");
        CHECK(s.last(3) == "BAR");
        // range(): <- first, last
        CHECK(s.range(0, 12) == "fooFOObarBAR");
        CHECK(s.range(1, 12) ==  "ooFOObarBAR");
        CHECK(s.range(1, 11) ==  "ooFOObarBA" );
        CHECK(s.range(2, 10) ==   "oFOObarB"  );
        CHECK(s.range(3,  9) ==    "FOObar"   );
        // offs(): offset from beginning, end
        CHECK(s.offs(0, 0) == "fooFOObarBAR");
        CHECK(s.offs(1, 0) ==  "ooFOObarBAR");
        CHECK(s.offs(1, 1) ==  "ooFOObarBA" );
        CHECK(s.offs(2, 1) ==   "oFOObarBA" );
        CHECK(s.offs(2, 2) ==   "oFOObarB"  );
        CHECK(s.offs(3, 3) ==    "FOObar"  );
        // right_of(): <- pos, include_pos
        CHECK(s.right_of(0,  true) == "fooFOObarBAR");
        CHECK(s.right_of(0, false) ==  "ooFOObarBAR");
        CHECK(s.right_of(1,  true) ==  "ooFOObarBAR");
        CHECK(s.right_of(1, false) ==   "oFOObarBAR");
        CHECK(s.right_of(2,  true) ==   "oFOObarBAR");
        CHECK(s.right_of(2, false) ==    "FOObarBAR");
        CHECK(s.right_of(3,  true) ==    "FOObarBAR");
        CHECK(s.right_of(3, false) ==     "OObarBAR");
        // left_of() <- pos, include_pos
        CHECK(s.left_of(12, false) == "fooFOObarBAR");
        CHECK(s.left_of(11,  true) == "fooFOObarBAR");
        CHECK(s.left_of(11, false) == "fooFOObarBA" );
        CHECK(s.left_of(10,  true) == "fooFOObarBA" );
        CHECK(s.left_of(10, false) == "fooFOObarB"  );
        CHECK(s.left_of( 9,  true) == "fooFOObarB"  );
        CHECK(s.left_of( 9, false) == "fooFOObar"   );
        // left_of(),right_of() <- substr
        ryml::csubstr FOO = s.sub(3, 3);
        CHECK(s.is_super(FOO)); // required for the following
        CHECK(s.left_of(FOO) == "foo");
        CHECK(s.right_of(FOO) == "barBAR");
    }
 
    // printing a substr/csubstr using printf-like
    {
        ryml::csubstr s = "some substring";
        ryml::csubstr some = s.first(4);
        CHECK(some == "some");
        CHECK(s == "some substring");
        // To print a csubstr using printf(), use the %.*s format specifier:
        {
            char result[32] = {0};
            std::snprintf(result, sizeof(result), "%.*s", (int)some.len, some.str);
            printf("~~~%s~~~\n", result);
            CHECK(ryml::to_csubstr((const char*)result) == "some");
            CHECK(ryml::to_csubstr((const char*)result) == some);
        }
        // But NOTE: because this is a string view type, in general
        // the C-string is NOT zero terminated.  So NEVER print it
        // directly, or it will overflow past the end of the given
        // substr, with a potential unbounded access.  For example,
        // this is bad:
        {
            char result[32] = {0};
            std::snprintf(result, sizeof(result), "%s", some.str); // ERROR! do not print the c-string directly
            CHECK(ryml::to_csubstr((const char*)result) == "some substring");
            CHECK(ryml::to_csubstr((const char*)result) == s);
        }
    }
 
    // printing a substr/csubstr using ostreams
    {
        ryml::csubstr s = "some substring";
        ryml::csubstr some = s.first(4);
        CHECK(some == "some");
        CHECK(s == "some substring");
        // simple! just use plain operator<<
        {
            std::stringstream ss;
            ss << s;
            CHECK(ss.str() == "some substring"); // as expected
            CHECK(ss.str() == s); // as expected
        }
        // But NOTE: because this is a string view type, in general
        // the C-string is NOT zero terminated.  So NEVER print it
        // directly, or it will overflow past the end of the given
        // substr, with a potential unbounded access.  For example,
        // this is bad:
        {
            std::stringstream ss;
            ss << some.str; // ERROR! do not print the c-string directly
            CHECK(ss.str() == "some substring"); // NOT "some"
            CHECK(ss.str() == s); // NOT some
        }
        // this is also bad (the same)
        {
            std::stringstream ss;
            ss << some.data(); // ERROR! do not print the c-string directly
            CHECK(ss.str() == "some substring"); // NOT "some"
            CHECK(ss.str() == s); // NOT some
        }
        // this is ok:
        {
            std::stringstream ss;
            ss << some;
            CHECK(ss.str() == "some"); // ok
            CHECK(ss.str() == some); // ok
        }
    }
 
    // is_sub(),is_super()
    {
        ryml::csubstr foobar = "foobar";
        ryml::csubstr foo = foobar.first(3);
        CHECK(foo.is_sub(foobar));
        CHECK(foo.is_sub(foo));
        CHECK(!foo.is_super(foobar));
        CHECK(!foobar.is_sub(foo));
        // identity comparison is true:
        CHECK(foo.is_super(foo));
        CHECK(foo.is_sub(foo));
        CHECK(foobar.is_sub(foobar));
        CHECK(foobar.is_super(foobar));
    }
 
    // overlaps()
    {
        ryml::csubstr foobar = "foobar";
        ryml::csubstr foo = foobar.first(3);
        ryml::csubstr oba = foobar.offs(2, 1);
        ryml::csubstr abc = "abc";
        CHECK(foobar.overlaps(foo));
        CHECK(foobar.overlaps(oba));
        CHECK(foo.overlaps(foobar));
        CHECK(foo.overlaps(oba));
        CHECK(!foo.overlaps(abc));
        CHECK(!abc.overlaps(foo));
    }
 
    // triml(): trim characters from the left
    // trimr(): trim characters from the right
    // trim(): trim characters from left AND right
    {
        CHECK(ryml::csubstr(" \t\n\rcontents without whitespace\t \n\r").trim("\t \n\r") == "contents without whitespace");
        ryml::csubstr aaabbb = "aaabbb";
        ryml::csubstr aaa___bbb = "aaa___bbb";
        // trim a character:
        CHECK(aaabbb.triml('a') == aaabbb.last(3)); // bbb
        CHECK(aaabbb.trimr('a') == aaabbb);
        CHECK(aaabbb.trim ('a') == aaabbb.last(3)); // bbb
        CHECK(aaabbb.triml('b') == aaabbb);
        CHECK(aaabbb.trimr('b') == aaabbb.first(3)); // aaa
        CHECK(aaabbb.trim ('b') == aaabbb.first(3)); // aaa
        CHECK(aaabbb.triml('c') == aaabbb);
        CHECK(aaabbb.trimr('c') == aaabbb);
        CHECK(aaabbb.trim ('c') == aaabbb);
        CHECK(aaa___bbb.triml('a') == aaa___bbb.last(6)); // ___bbb
        CHECK(aaa___bbb.trimr('a') == aaa___bbb);
        CHECK(aaa___bbb.trim ('a') == aaa___bbb.last(6)); // ___bbb
        CHECK(aaa___bbb.triml('b') == aaa___bbb);
        CHECK(aaa___bbb.trimr('b') == aaa___bbb.first(6)); // aaa___
        CHECK(aaa___bbb.trim ('b') == aaa___bbb.first(6)); // aaa___
        CHECK(aaa___bbb.triml('c') == aaa___bbb);
        CHECK(aaa___bbb.trimr('c') == aaa___bbb);
        CHECK(aaa___bbb.trim ('c') == aaa___bbb);
        // trim ANY of the characters:
        CHECK(aaabbb.triml("ab") == "");
        CHECK(aaabbb.trimr("ab") == "");
        CHECK(aaabbb.trim ("ab") == "");
        CHECK(aaabbb.triml("ba") == "");
        CHECK(aaabbb.trimr("ba") == "");
        CHECK(aaabbb.trim ("ba") == "");
        CHECK(aaabbb.triml("cd") == aaabbb);
        CHECK(aaabbb.trimr("cd") == aaabbb);
        CHECK(aaabbb.trim ("cd") == aaabbb);
        CHECK(aaa___bbb.triml("ab") == aaa___bbb.last(6)); // ___bbb
        CHECK(aaa___bbb.triml("ba") == aaa___bbb.last(6)); // ___bbb
        CHECK(aaa___bbb.triml("cd") == aaa___bbb);
        CHECK(aaa___bbb.trimr("ab") == aaa___bbb.first(6)); // aaa___
        CHECK(aaa___bbb.trimr("ba") == aaa___bbb.first(6)); // aaa___
        CHECK(aaa___bbb.trimr("cd") == aaa___bbb);
        CHECK(aaa___bbb.trim ("ab") == aaa___bbb.range(3, 6)); // ___
        CHECK(aaa___bbb.trim ("ba") == aaa___bbb.range(3, 6)); // ___
        CHECK(aaa___bbb.trim ("cd") == aaa___bbb);
    }
 
    // unquoted():
    {
        CHECK(ryml::csubstr( "'this is is single quoted'" ).unquoted() == "this is is single quoted");
        CHECK(ryml::csubstr("\"this is is double quoted\"").unquoted() == "this is is double quoted");
    }
 
    // stripl(): remove pattern from the left
    // stripr(): remove pattern from the right
    {
        ryml::csubstr abc___cba = "abc___cba";
        ryml::csubstr abc___abc = "abc___abc";
        CHECK(abc___cba.stripl("abc") == abc___cba.last(6)); // ___cba
        CHECK(abc___cba.stripr("abc") == abc___cba);
        CHECK(abc___cba.stripl("ab") == abc___cba.last(7)); // c___cba
        CHECK(abc___cba.stripr("ab") == abc___cba);
        CHECK(abc___cba.stripl("a") == abc___cba.last(8)); // bc___cba, same as triml('a')
        CHECK(abc___cba.stripr("a") == abc___cba.first(8));
        CHECK(abc___abc.stripl("abc") == abc___abc.last(6)); // ___abc
        CHECK(abc___abc.stripr("abc") == abc___abc.first(6)); // abc___
        CHECK(abc___abc.stripl("ab") == abc___abc.last(7)); // c___cba
        CHECK(abc___abc.stripr("ab") == abc___abc);
        CHECK(abc___abc.stripl("a") == abc___abc.last(8)); // bc___cba, same as triml('a')
        CHECK(abc___abc.stripr("a") == abc___abc);
    }
 
    // begins_with()/ends_with()
    // begins_with_any()/ends_with_any()
    {
        ryml::csubstr s = "foobar123";
        // char overloads
        CHECK(s.begins_with('f'));
        CHECK(s.ends_with('3'));
        CHECK(!s.ends_with('2'));
        CHECK(!s.ends_with('o'));
        // char[] overloads
        CHECK(s.begins_with("foobar"));
        CHECK(s.begins_with("foo"));
        CHECK(s.begins_with_any("foo"));
        CHECK(!s.begins_with("oof"));
        CHECK(s.begins_with_any("oof"));
        CHECK(s.ends_with("23"));
        CHECK(s.ends_with("123"));
        CHECK(s.ends_with_any("123"));
        CHECK(!s.ends_with("321"));
        CHECK(s.ends_with_any("231"));
    }
 
    // select()
    {
        ryml::csubstr s = "0123456789";
        CHECK(s.select('0') == s.sub(0, 1));
        CHECK(s.select('1') == s.sub(1, 1));
        CHECK(s.select('2') == s.sub(2, 1));
        CHECK(s.select('8') == s.sub(8, 1));
        CHECK(s.select('9') == s.sub(9, 1));
        CHECK(s.select("0123") == s.range(0, 4));
        CHECK(s.select("012" ) == s.range(0, 3));
        CHECK(s.select("01"  ) == s.range(0, 2));
        CHECK(s.select("0"   ) == s.range(0, 1));
        CHECK(s.select( "123") == s.range(1, 4));
        CHECK(s.select(  "23") == s.range(2, 4));
        CHECK(s.select(   "3") == s.range(3, 4));
    }
 
    // find()
    {
        ryml::csubstr s012345 = "012345";
        // find single characters:
        CHECK(s012345.find('a') == ryml::npos);
        CHECK(s012345.find('0'    ) == 0u);
        CHECK(s012345.find('0', 1u) == ryml::npos);
        CHECK(s012345.find('1'    ) == 1u);
        CHECK(s012345.find('1', 2u) == ryml::npos);
        CHECK(s012345.find('2'    ) == 2u);
        CHECK(s012345.find('2', 3u) == ryml::npos);
        CHECK(s012345.find('3'    ) == 3u);
        CHECK(s012345.find('3', 4u) == ryml::npos);
        // find patterns
        CHECK(s012345.find("ab"    ) == ryml::npos);
        CHECK(s012345.find("01"    ) == 0u);
        CHECK(s012345.find("01", 1u) == ryml::npos);
        CHECK(s012345.find("12"    ) == 1u);
        CHECK(s012345.find("12", 2u) == ryml::npos);
        CHECK(s012345.find("23"    ) == 2u);
        CHECK(s012345.find("23", 3u) == ryml::npos);
    }
 
    // count(): count the number of occurrences of a character
    {
        ryml::csubstr buf = "00110022003300440055";
        CHECK(buf.count('1'     ) ==  2u);
        CHECK(buf.count('1',  0u) ==  2u);
        CHECK(buf.count('1',  1u) ==  2u);
        CHECK(buf.count('1',  2u) ==  2u);
        CHECK(buf.count('1',  3u) ==  1u);
        CHECK(buf.count('1',  4u) ==  0u);
        CHECK(buf.count('1',  5u) ==  0u);
        CHECK(buf.count('0'     ) == 10u);
        CHECK(buf.count('0',  0u) == 10u);
        CHECK(buf.count('0',  1u) ==  9u);
        CHECK(buf.count('0',  2u) ==  8u);
        CHECK(buf.count('0',  3u) ==  8u);
        CHECK(buf.count('0',  4u) ==  8u);
        CHECK(buf.count('0',  5u) ==  7u);
        CHECK(buf.count('0',  6u) ==  6u);
        CHECK(buf.count('0',  7u) ==  6u);
        CHECK(buf.count('0',  8u) ==  6u);
        CHECK(buf.count('0',  9u) ==  5u);
        CHECK(buf.count('0', 10u) ==  4u);
        CHECK(buf.count('0', 11u) ==  4u);
        CHECK(buf.count('0', 12u) ==  4u);
        CHECK(buf.count('0', 13u) ==  3u);
        CHECK(buf.count('0', 14u) ==  2u);
        CHECK(buf.count('0', 15u) ==  2u);
        CHECK(buf.count('0', 16u) ==  2u);
        CHECK(buf.count('0', 17u) ==  1u);
        CHECK(buf.count('0', 18u) ==  0u);
        CHECK(buf.count('0', 19u) ==  0u);
        CHECK(buf.count('0', 20u) ==  0u);
    }
 
    // first_of(),last_of()
    {
        ryml::csubstr s012345 = "012345";
        CHECK(s012345.first_of('a') == ryml::npos);
        CHECK(s012345.first_of("ab") == ryml::npos);
        CHECK(s012345.first_of('0') == 0u);
        CHECK(s012345.first_of("0") == 0u);
        CHECK(s012345.first_of("01") == 0u);
        CHECK(s012345.first_of("10") == 0u);
        CHECK(s012345.first_of("012") == 0u);
        CHECK(s012345.first_of("210") == 0u);
        CHECK(s012345.first_of("0123") == 0u);
        CHECK(s012345.first_of("3210") == 0u);
        CHECK(s012345.first_of("01234") == 0u);
        CHECK(s012345.first_of("43210") == 0u);
        CHECK(s012345.first_of("012345") == 0u);
        CHECK(s012345.first_of("543210") == 0u);
        CHECK(s012345.first_of('5') == 5u);
        CHECK(s012345.first_of("5") == 5u);
        CHECK(s012345.first_of("45") == 4u);
        CHECK(s012345.first_of("54") == 4u);
        CHECK(s012345.first_of("345") == 3u);
        CHECK(s012345.first_of("543") == 3u);
        CHECK(s012345.first_of("2345") == 2u);
        CHECK(s012345.first_of("5432") == 2u);
        CHECK(s012345.first_of("12345") == 1u);
        CHECK(s012345.first_of("54321") == 1u);
        CHECK(s012345.first_of("012345") == 0u);
        CHECK(s012345.first_of("543210") == 0u);
        CHECK(s012345.first_of('0', 6u) == ryml::npos);
        CHECK(s012345.first_of('5', 6u) == ryml::npos);
        CHECK(s012345.first_of("012345", 6u) == ryml::npos);
        //
        CHECK(s012345.last_of('a') == ryml::npos);
        CHECK(s012345.last_of("ab") == ryml::npos);
        CHECK(s012345.last_of('0') == 0u);
        CHECK(s012345.last_of("0") == 0u);
        CHECK(s012345.last_of("01") == 1u);
        CHECK(s012345.last_of("10") == 1u);
        CHECK(s012345.last_of("012") == 2u);
        CHECK(s012345.last_of("210") == 2u);
        CHECK(s012345.last_of("0123") == 3u);
        CHECK(s012345.last_of("3210") == 3u);
        CHECK(s012345.last_of("01234") == 4u);
        CHECK(s012345.last_of("43210") == 4u);
        CHECK(s012345.last_of("012345") == 5u);
        CHECK(s012345.last_of("543210") == 5u);
        CHECK(s012345.last_of('5') == 5u);
        CHECK(s012345.last_of("5") == 5u);
        CHECK(s012345.last_of("45") == 5u);
        CHECK(s012345.last_of("54") == 5u);
        CHECK(s012345.last_of("345") == 5u);
        CHECK(s012345.last_of("543") == 5u);
        CHECK(s012345.last_of("2345") == 5u);
        CHECK(s012345.last_of("5432") == 5u);
        CHECK(s012345.last_of("12345") == 5u);
        CHECK(s012345.last_of("54321") == 5u);
        CHECK(s012345.last_of("012345") == 5u);
        CHECK(s012345.last_of("543210") == 5u);
        CHECK(s012345.last_of('0', 6u) == 0u);
        CHECK(s012345.last_of('5', 6u) == 5u);
        CHECK(s012345.last_of("012345", 6u) == 5u);
    }
 
    // first_not_of(), last_not_of()
    {
        ryml::csubstr s012345 = "012345";
        CHECK(s012345.first_not_of('a') == 0u);
        CHECK(s012345.first_not_of("ab") == 0u);
        CHECK(s012345.first_not_of('0') == 1u);
        CHECK(s012345.first_not_of("0") == 1u);
        CHECK(s012345.first_not_of("01") == 2u);
        CHECK(s012345.first_not_of("10") == 2u);
        CHECK(s012345.first_not_of("012") == 3u);
        CHECK(s012345.first_not_of("210") == 3u);
        CHECK(s012345.first_not_of("0123") == 4u);
        CHECK(s012345.first_not_of("3210") == 4u);
        CHECK(s012345.first_not_of("01234") == 5u);
        CHECK(s012345.first_not_of("43210") == 5u);
        CHECK(s012345.first_not_of("012345") == ryml::npos);
        CHECK(s012345.first_not_of("543210") == ryml::npos);
        CHECK(s012345.first_not_of('5') == 0u);
        CHECK(s012345.first_not_of("5") == 0u);
        CHECK(s012345.first_not_of("45") == 0u);
        CHECK(s012345.first_not_of("54") == 0u);
        CHECK(s012345.first_not_of("345") == 0u);
        CHECK(s012345.first_not_of("543") == 0u);
        CHECK(s012345.first_not_of("2345") == 0u);
        CHECK(s012345.first_not_of("5432") == 0u);
        CHECK(s012345.first_not_of("12345") == 0u);
        CHECK(s012345.first_not_of("54321") == 0u);
        CHECK(s012345.first_not_of("012345") == ryml::npos);
        CHECK(s012345.first_not_of("543210") == ryml::npos);
        CHECK(s012345.last_not_of('a') == 5u);
        CHECK(s012345.last_not_of("ab") == 5u);
        CHECK(s012345.last_not_of('5') == 4u);
        CHECK(s012345.last_not_of("5") == 4u);
        CHECK(s012345.last_not_of("45") == 3u);
        CHECK(s012345.last_not_of("54") == 3u);
        CHECK(s012345.last_not_of("345") == 2u);
        CHECK(s012345.last_not_of("543") == 2u);
        CHECK(s012345.last_not_of("2345") == 1u);
        CHECK(s012345.last_not_of("5432") == 1u);
        CHECK(s012345.last_not_of("12345") == 0u);
        CHECK(s012345.last_not_of("54321") == 0u);
        CHECK(s012345.last_not_of("012345") == ryml::npos);
        CHECK(s012345.last_not_of("543210") == ryml::npos);
        CHECK(s012345.last_not_of('0') == 5u);
        CHECK(s012345.last_not_of("0") == 5u);
        CHECK(s012345.last_not_of("01") == 5u);
        CHECK(s012345.last_not_of("10") == 5u);
        CHECK(s012345.last_not_of("012") == 5u);
        CHECK(s012345.last_not_of("210") == 5u);
        CHECK(s012345.last_not_of("0123") == 5u);
        CHECK(s012345.last_not_of("3210") == 5u);
        CHECK(s012345.last_not_of("01234") == 5u);
        CHECK(s012345.last_not_of("43210") == 5u);
        CHECK(s012345.last_not_of("012345") == ryml::npos);
        CHECK(s012345.last_not_of("543210") == ryml::npos);
    }
 
    // first_non_empty_span()
    {
        CHECK(ryml::csubstr("foo bar").first_non_empty_span() == "foo");
        CHECK(ryml::csubstr("       foo bar").first_non_empty_span() == "foo");
        CHECK(ryml::csubstr("\n   \r  \t  foo bar").first_non_empty_span() == "foo");
        CHECK(ryml::csubstr("\n   \r  \t  foo\n\r\t bar").first_non_empty_span() == "foo");
        CHECK(ryml::csubstr("\n   \r  \t  foo\n\r\t bar").first_non_empty_span() == "foo");
        CHECK(ryml::csubstr(",\n   \r  \t  foo\n\r\t bar").first_non_empty_span() == ",");
    }
    // first_uint_span()
    {
        CHECK(ryml::csubstr("1234 asdkjh").first_uint_span() == "1234");
        CHECK(ryml::csubstr("1234\rasdkjh").first_uint_span() == "1234");
        CHECK(ryml::csubstr("1234\tasdkjh").first_uint_span() == "1234");
        CHECK(ryml::csubstr("1234\nasdkjh").first_uint_span() == "1234");
        CHECK(ryml::csubstr("1234]asdkjh").first_uint_span() == "1234");
        CHECK(ryml::csubstr("1234)asdkjh").first_uint_span() == "1234");
        CHECK(ryml::csubstr("1234gasdkjh").first_uint_span() == "");
    }
    // first_int_span()
    {
        CHECK(ryml::csubstr("-1234 asdkjh").first_int_span() == "-1234");
        CHECK(ryml::csubstr("-1234\rasdkjh").first_int_span() == "-1234");
        CHECK(ryml::csubstr("-1234\tasdkjh").first_int_span() == "-1234");
        CHECK(ryml::csubstr("-1234\nasdkjh").first_int_span() == "-1234");
        CHECK(ryml::csubstr("-1234]asdkjh").first_int_span() == "-1234");
        CHECK(ryml::csubstr("-1234)asdkjh").first_int_span() == "-1234");
        CHECK(ryml::csubstr("-1234gasdkjh").first_int_span() == "");
    }
    // first_real_span()
    {
        CHECK(ryml::csubstr("-1234 asdkjh").first_real_span() == "-1234");
        CHECK(ryml::csubstr("-1234\rasdkjh").first_real_span() == "-1234");
        CHECK(ryml::csubstr("-1234\tasdkjh").first_real_span() == "-1234");
        CHECK(ryml::csubstr("-1234\nasdkjh").first_real_span() == "-1234");
        CHECK(ryml::csubstr("-1234]asdkjh").first_real_span() == "-1234");
        CHECK(ryml::csubstr("-1234)asdkjh").first_real_span() == "-1234");
        CHECK(ryml::csubstr("-1234gasdkjh").first_real_span() == "");
        CHECK(ryml::csubstr("1.234 asdkjh").first_real_span() == "1.234");
        CHECK(ryml::csubstr("1.234e+5 asdkjh").first_real_span() == "1.234e+5");
        CHECK(ryml::csubstr("1.234e-5 asdkjh").first_real_span() == "1.234e-5");
        CHECK(ryml::csubstr("1.234 asdkjh").first_real_span() == "1.234");
        CHECK(ryml::csubstr("1.234e+5 asdkjh").first_real_span() == "1.234e+5");
        CHECK(ryml::csubstr("1.234e-5 asdkjh").first_real_span() == "1.234e-5");
        CHECK(ryml::csubstr("-1.234 asdkjh").first_real_span() == "-1.234");
        CHECK(ryml::csubstr("-1.234e+5 asdkjh").first_real_span() == "-1.234e+5");
        CHECK(ryml::csubstr("-1.234e-5 asdkjh").first_real_span() == "-1.234e-5");
        // hexadecimal real numbers
        CHECK(ryml::csubstr("0x1.e8480p+19 asdkjh").first_real_span() == "0x1.e8480p+19");
        CHECK(ryml::csubstr("0x1.e8480p-19 asdkjh").first_real_span() == "0x1.e8480p-19");
        CHECK(ryml::csubstr("-0x1.e8480p+19 asdkjh").first_real_span() == "-0x1.e8480p+19");
        CHECK(ryml::csubstr("-0x1.e8480p-19 asdkjh").first_real_span() == "-0x1.e8480p-19");
        CHECK(ryml::csubstr("+0x1.e8480p+19 asdkjh").first_real_span() == "+0x1.e8480p+19");
        CHECK(ryml::csubstr("+0x1.e8480p-19 asdkjh").first_real_span() == "+0x1.e8480p-19");
        // binary real numbers
        CHECK(ryml::csubstr("0b101.011p+19 asdkjh").first_real_span() == "0b101.011p+19");
        CHECK(ryml::csubstr("0b101.011p-19 asdkjh").first_real_span() == "0b101.011p-19");
        CHECK(ryml::csubstr("-0b101.011p+19 asdkjh").first_real_span() == "-0b101.011p+19");
        CHECK(ryml::csubstr("-0b101.011p-19 asdkjh").first_real_span() == "-0b101.011p-19");
        CHECK(ryml::csubstr("+0b101.011p+19 asdkjh").first_real_span() == "+0b101.011p+19");
        CHECK(ryml::csubstr("+0b101.011p-19 asdkjh").first_real_span() == "+0b101.011p-19");
        // octal real numbers
        CHECK(ryml::csubstr("0o173.045p+19 asdkjh").first_real_span() == "0o173.045p+19");
        CHECK(ryml::csubstr("0o173.045p-19 asdkjh").first_real_span() == "0o173.045p-19");
        CHECK(ryml::csubstr("-0o173.045p+19 asdkjh").first_real_span() == "-0o173.045p+19");
        CHECK(ryml::csubstr("-0o173.045p-19 asdkjh").first_real_span() == "-0o173.045p-19");
        CHECK(ryml::csubstr("+0o173.045p+19 asdkjh").first_real_span() == "+0o173.045p+19");
        CHECK(ryml::csubstr("+0o173.045p-19 asdkjh").first_real_span() == "+0o173.045p-19");
    }
    // see also is_number()
 
    // basename(), dirname(), extshort(), extlong()
    {
        CHECK(ryml::csubstr("/path/to/file.tar.gz").basename() == "file.tar.gz");
        CHECK(ryml::csubstr("/path/to/file.tar.gz").dirname() == "/path/to/");
        CHECK(ryml::csubstr("C:\\path\\to\\file.tar.gz").basename('\\') == "file.tar.gz");
        CHECK(ryml::csubstr("C:\\path\\to\\file.tar.gz").dirname('\\') == "C:\\path\\to\\");
        CHECK(ryml::csubstr("/path/to/file.tar.gz").extshort() == "gz");
        CHECK(ryml::csubstr("/path/to/file.tar.gz").extlong() == "tar.gz");
        CHECK(ryml::csubstr("/path/to/file.tar.gz").name_wo_extshort() == "/path/to/file.tar");
        CHECK(ryml::csubstr("/path/to/file.tar.gz").name_wo_extlong() == "/path/to/file");
    }
 
    // split()
    {
        using namespace ryml;
        csubstr parts[] = {"aa", "bb", "cc", "dd", "ee", "ff"};
        {
            size_t count = 0;
            for(csubstr part : csubstr("aa/bb/cc/dd/ee/ff").split('/'))
                CHECK(part == parts[count++]);
            CHECK(count == 6u);
        }
        {
            size_t count = 0;
            for(csubstr part : csubstr("aa.bb.cc.dd.ee.ff").split('.'))
                CHECK(part == parts[count++]);
            CHECK(count == 6u);
        }
        {
            size_t count = 0;
            for(csubstr part : csubstr("aa-bb-cc-dd-ee-ff").split('-'))
                CHECK(part == parts[count++]);
            CHECK(count == 6u);
        }
        // see also next_split()
    }
 
    //  pop_left(),  pop_right() --- non-greedy version
    // gpop_left(), gpop_right() --- greedy version
    {
        const bool skip_empty = true;
        // pop_left(): pop the last element from the left
        CHECK(ryml::csubstr(  "0/1/2"   ). pop_left('/'            ) ==   "0"    );
        CHECK(ryml::csubstr( "/0/1/2"   ). pop_left('/'            ) == ""       );
        CHECK(ryml::csubstr("//0/1/2"   ). pop_left('/'            ) == ""       );
        CHECK(ryml::csubstr(  "0/1/2"   ). pop_left('/', skip_empty) ==   "0"    );
        CHECK(ryml::csubstr( "/0/1/2"   ). pop_left('/', skip_empty) ==  "/0"    );
        CHECK(ryml::csubstr("//0/1/2"   ). pop_left('/', skip_empty) == "//0"    );
        // gpop_left(): pop all but the first element (greedy pop)
        CHECK(ryml::csubstr(  "0/1/2"   ).gpop_left('/'            ) ==   "0/1"  );
        CHECK(ryml::csubstr( "/0/1/2"   ).gpop_left('/'            ) ==  "/0/1"  );
        CHECK(ryml::csubstr("//0/1/2"   ).gpop_left('/'            ) == "//0/1"  );
        CHECK(ryml::csubstr(  "0/1/2/"  ).gpop_left('/'            ) ==   "0/1/2");
        CHECK(ryml::csubstr( "/0/1/2/"  ).gpop_left('/'            ) ==  "/0/1/2");
        CHECK(ryml::csubstr("//0/1/2/"  ).gpop_left('/'            ) == "//0/1/2");
        CHECK(ryml::csubstr(  "0/1/2//" ).gpop_left('/'            ) ==   "0/1/2/");
        CHECK(ryml::csubstr( "/0/1/2//" ).gpop_left('/'            ) ==  "/0/1/2/");
        CHECK(ryml::csubstr("//0/1/2//" ).gpop_left('/'            ) == "//0/1/2/");
        CHECK(ryml::csubstr(  "0/1/2"   ).gpop_left('/', skip_empty) ==   "0/1"  );
        CHECK(ryml::csubstr( "/0/1/2"   ).gpop_left('/', skip_empty) ==  "/0/1"  );
        CHECK(ryml::csubstr("//0/1/2"   ).gpop_left('/', skip_empty) == "//0/1"  );
        CHECK(ryml::csubstr(  "0/1/2/"  ).gpop_left('/', skip_empty) ==   "0/1"  );
        CHECK(ryml::csubstr( "/0/1/2/"  ).gpop_left('/', skip_empty) ==  "/0/1"  );
        CHECK(ryml::csubstr("//0/1/2/"  ).gpop_left('/', skip_empty) == "//0/1"  );
        CHECK(ryml::csubstr(  "0/1/2//" ).gpop_left('/', skip_empty) ==   "0/1"  );
        CHECK(ryml::csubstr( "/0/1/2//" ).gpop_left('/', skip_empty) ==  "/0/1"  );
        CHECK(ryml::csubstr("//0/1/2//" ).gpop_left('/', skip_empty) == "//0/1"  );
        // pop_right(): pop the last element from the right
        CHECK(ryml::csubstr(  "0/1/2"   ). pop_right('/'            ) ==   "2"    );
        CHECK(ryml::csubstr(  "0/1/2/"  ). pop_right('/'            ) == ""       );
        CHECK(ryml::csubstr(  "0/1/2//" ). pop_right('/'            ) == ""       );
        CHECK(ryml::csubstr(  "0/1/2"   ). pop_right('/', skip_empty) ==   "2"    );
        CHECK(ryml::csubstr(  "0/1/2/"  ). pop_right('/', skip_empty) ==   "2/"   );
        CHECK(ryml::csubstr(  "0/1/2//" ). pop_right('/', skip_empty) ==   "2//"  );
        // gpop_right(): pop all but the first element (greedy pop)
        CHECK(ryml::csubstr(  "0/1/2"   ).gpop_right('/'            ) ==     "1/2"  );
        CHECK(ryml::csubstr(  "0/1/2/"  ).gpop_right('/'            ) ==     "1/2/" );
        CHECK(ryml::csubstr(  "0/1/2//" ).gpop_right('/'            ) ==     "1/2//");
        CHECK(ryml::csubstr( "/0/1/2"   ).gpop_right('/'            ) ==   "0/1/2"  );
        CHECK(ryml::csubstr( "/0/1/2/"  ).gpop_right('/'            ) ==   "0/1/2/" );
        CHECK(ryml::csubstr( "/0/1/2//" ).gpop_right('/'            ) ==   "0/1/2//");
        CHECK(ryml::csubstr("//0/1/2"   ).gpop_right('/'            ) ==  "/0/1/2"  );
        CHECK(ryml::csubstr("//0/1/2/"  ).gpop_right('/'            ) ==  "/0/1/2/" );
        CHECK(ryml::csubstr("//0/1/2//" ).gpop_right('/'            ) ==  "/0/1/2//");
        CHECK(ryml::csubstr(  "0/1/2"   ).gpop_right('/', skip_empty) ==     "1/2"  );
        CHECK(ryml::csubstr(  "0/1/2/"  ).gpop_right('/', skip_empty) ==     "1/2/" );
        CHECK(ryml::csubstr(  "0/1/2//" ).gpop_right('/', skip_empty) ==     "1/2//");
        CHECK(ryml::csubstr( "/0/1/2"   ).gpop_right('/', skip_empty) ==     "1/2"  );
        CHECK(ryml::csubstr( "/0/1/2/"  ).gpop_right('/', skip_empty) ==     "1/2/" );
        CHECK(ryml::csubstr( "/0/1/2//" ).gpop_right('/', skip_empty) ==     "1/2//");
        CHECK(ryml::csubstr("//0/1/2"   ).gpop_right('/', skip_empty) ==     "1/2"  );
        CHECK(ryml::csubstr("//0/1/2/"  ).gpop_right('/', skip_empty) ==     "1/2/" );
        CHECK(ryml::csubstr("//0/1/2//" ).gpop_right('/', skip_empty) ==     "1/2//");
    }
}

sample_parse_file()

void sample_parse_file

(

)

demonstrate how to load a YAML file from disk to parse with ryml.

ryml offers no overload to directly parse files from disk; it only parses source buffers (which may be mutable or immutable). It is up to the caller to first load the file contents into a buffer before parsing with ryml. To help with this you can use the (efficient) helper [file_get_contents](c4::yml::file_get_contents()). See also the analogous file_put_contents

See also

Parse utilities

Definition at line 1834 of file quickstart.cpp.

{
    const char filename[] = "ryml_example.yml";
    std::string yaml = ""
        "foo: 1"  "\n"
        "bar:"    "\n"
        "- 2"     "\n"
        "- 3"     "\n";
    // because this is a minimal sample, it assumes nothing on the
    // environment/OS (other than that it can read/write files). So we
    // create the file on the fly:
    ryml::file_put_contents(yaml, filename);
 
    // now we can load it into a std::string (for example):
    {
        std::string contents = ryml::file_get_contents<std::string>(filename);
        ryml::Tree tree = ryml::parse_in_arena(ryml::to_csubstr(contents)); // immutable (csubstr) overload
        CHECK(tree["foo"].val() == "1");
        CHECK(tree["bar"][0].val() == "2");
        CHECK(tree["bar"][1].val() == "3");
    }
 
    // or we can use a vector<char> instead:
    {
        std::vector<char> contents = ryml::file_get_contents<std::vector<char>>(filename);
        ryml::Tree tree = ryml::parse_in_place(ryml::to_substr(contents)); // mutable (substr) overload
        CHECK(tree["foo"].val() == "1");
        CHECK(tree["bar"][0].val() == "2");
        CHECK(tree["bar"][1].val() == "3");
    }
 
    // generally, any contiguous char container can be used with ryml,
    // provided that the ryml::substr/ryml::csubstr view can be
    // created out of it.
    //
    // ryml provides the overloads above for these two containers, but
    // if you are using another container it should be very easy (only
    // requires pointer and length).
}

sample_parse_in_place()

void sample_parse_in_place

(

)

demonstrate in-place parsing of a mutable YAML source buffer.

See also

Parse utilities

Definition at line 1879 of file quickstart.cpp.

{
    // Like the name suggests, parse_in_place() directly mutates the
    // source buffer in place
    char src[] = "{foo: 1, bar: [2, 3]}"; // ryml can parse in situ
    ryml::substr srcview = src; // a mutable view to the source buffer
    ryml::Tree tree = ryml::parse_in_place(srcview); // you can also reuse the tree and/or parser
    ryml::ConstNodeRef root = tree.crootref(); // get a constant reference to the root
 
    CHECK(root.is_map());
    CHECK(root["foo"].is_keyval());
    CHECK(root["foo"].key() == "foo");
    CHECK(root["foo"].val() == "1");
    CHECK(root["bar"].is_seq());
    CHECK(root["bar"].has_key());
    CHECK(root["bar"].key() == "bar");
    CHECK(root["bar"][0].val() == "2");
    CHECK(root["bar"][1].val() == "3");
 
    // deserializing:
    int foo = 0, bar0 = 0, bar1 = 0;
    root["foo"] >> foo;
    root["bar"][0] >> bar0;
    root["bar"][1] >> bar1;
    CHECK(foo == 1);
    CHECK(bar0 == 2);
    CHECK(bar1 == 3);
 
    // after parsing, the tree holds views to the source buffer:
    CHECK(root["foo"].val().data()     == src + strlen("{foo: "));
    CHECK(root["foo"].val().begin()    == src + strlen("{foo: "));
    CHECK(root["foo"].val().end()      == src + strlen("{foo: 1"));
    CHECK(root["foo"].val().is_sub(srcview)); // equivalent to the previous three assertions
    CHECK(root["bar"][0].val().data()  == src + strlen("{foo: 1, bar: ["));
    CHECK(root["bar"][0].val().begin() == src + strlen("{foo: 1, bar: ["));
    CHECK(root["bar"][0].val().end()   == src + strlen("{foo: 1, bar: [2"));
    CHECK(root["bar"][0].val().is_sub(srcview)); // equivalent to the previous three assertions
    CHECK(root["bar"][1].val().data()  == src + strlen("{foo: 1, bar: [2, "));
    CHECK(root["bar"][1].val().begin() == src + strlen("{foo: 1, bar: [2, "));
    CHECK(root["bar"][1].val().end()   == src + strlen("{foo: 1, bar: [2, 3"));
    CHECK(root["bar"][1].val().is_sub(srcview)); // equivalent to the previous three assertions
 
    // NOTE. parse_in_place() cannot accept ryml::csubstr
    // so this will cause a /compile/ error:
    ryml::csubstr csrcview = srcview; // ok, can assign from mutable to immutable
    //tree = ryml::parse_in_place(csrcview); // compile error, cannot mutate an immutable view
    (void)csrcview;
}

sample_parse_in_arena()

void sample_parse_in_arena

(

)

demonstrate parsing of a read-only YAML source buffer

See also

Parse utilities

Definition at line 1933 of file quickstart.cpp.

{
    // to parse read-only memory, ryml will copy first to the tree's
    // arena, and then parse the copied buffer:
    ryml::Tree tree = ryml::parse_in_arena("{foo: 1, bar: [2, 3]}");
    ryml::ConstNodeRef root = tree.crootref(); // get a const reference to the root
 
    CHECK(root.is_map());
    CHECK(root["foo"].is_keyval());
    CHECK(root["foo"].key() == "foo");
    CHECK(root["foo"].val() == "1");
    CHECK(root["bar"].is_seq());
    CHECK(root["bar"].has_key());
    CHECK(root["bar"].key() == "bar");
    CHECK(root["bar"][0].val() == "2");
    CHECK(root["bar"][1].val() == "3");
 
    // deserializing:
    int foo = 0, bar0 = 0, bar1 = 0;
    root["foo"] >> foo;
    root["bar"][0] >> bar0;
    root["bar"][1] >> bar1;
    CHECK(foo == 1);
    CHECK(bar0 == 2);
    CHECK(bar1 == 3);
 
    // NOTE. parse_in_arena() cannot accept ryml::substr. Overloads
    // receiving substr buffers are declared, but intentionally left
    // undefined, so this will cause a /linker/ error
    char src[] = "{foo: is it really true}";
    ryml::substr srcview = src;
    //tree = ryml::parse_in_place(srcview); // linker error, overload intentionally undefined
 
    // If you really intend to parse a mutable buffer in the arena,
    // then simply convert it to immutable prior to calling
    // parse_in_arena():
    ryml::csubstr csrcview = srcview; // assigning from src also works
    tree = ryml::parse_in_arena(csrcview); // OK! csrcview is immutable
    CHECK(tree["foo"].val() == "is it really true");
}

sample_parse_reuse_tree()

void sample_parse_reuse_tree

(

)

demonstrate reuse/modification of tree when parsing

See also

Parse utilities

Definition at line 1979 of file quickstart.cpp.

{
    ryml::Tree tree;
 
    // it will always be faster if the tree's size is conveniently reserved:
    tree.reserve(30); // reserve 30 nodes (good enough for this sample)
    // if you are using the tree's arena to serialize data,
    // then reserve also the arena's size:
    tree.reserve_arena(256); // reserve 256 characters (good enough for this sample)
 
    // now parse into the tree:
    ryml::csubstr yaml = ""
        "foo: 1"        "\n"
        "bar: [2,3]"    "\n"
        "";
    ryml::parse_in_arena(yaml, &tree);
 
    ryml::ConstNodeRef root = tree.crootref();
    CHECK(root.num_children() == 2);
    CHECK(root.is_map());
    CHECK(root["foo"].is_keyval());
    CHECK(root["foo"].key() == "foo");
    CHECK(root["foo"].val() == "1");
    CHECK(root["bar"].is_seq());
    CHECK(root["bar"].has_key());
    CHECK(root["bar"].key() == "bar");
    CHECK(root["bar"][0].val() == "2");
    CHECK(root["bar"][1].val() == "3");
    CHECK(ryml::emitrs_yaml<std::string>(tree) == yaml);
 
    // WATCHOUT: parsing into an existing tree will APPEND to it:
    ryml::parse_in_arena("{foo2: 12, bar2: [22, 32]}", &tree);
    CHECK(ryml::emitrs_yaml<std::string>(tree) == ""
          "foo: 1"         "\n"
          "bar: [2,3]"     "\n"
          "foo2: 12"       "\n"
          "bar2: [22,32]"  "\n"
          "");
    CHECK(root.num_children() == 4);
    CHECK(root["foo2"].is_keyval());
    CHECK(root["foo2"].key() == "foo2");
    CHECK(root["foo2"].val() == "12");
    CHECK(root["bar2"].is_seq());
    CHECK(root["bar2"].has_key());
    CHECK(root["bar2"].key() == "bar2");
    CHECK(root["bar2"][0].val() == "22");
    CHECK(root["bar2"][1].val() == "32");
 
    // if you want to fully replace the tree, you need to clear first
    // before parsing into an existing tree:
    tree.clear();
    tree.clear_arena();  // you may or may not want to clear the arena
    ryml::parse_in_arena("- a\n- b\n- {x0: 1, x1: 2}", &tree);
    CHECK(ryml::emitrs_yaml<std::string>(tree) == "- a\n- b\n- {x0: 1,x1: 2}\n");
    CHECK(root.is_seq());
    CHECK(root[0].val() == "a");
    CHECK(root[1].val() == "b");
    CHECK(root[2].is_map());
    CHECK(root[2]["x0"].val() == "1");
    CHECK(root[2]["x1"].val() == "2");
 
    // we can parse directly into a node nested deep in an existing tree:
    ryml::NodeRef mroot = tree.rootref(); // modifiable root
    ryml::parse_in_arena("champagne: Dom Perignon\ncoffee: Arabica", mroot.append_child());
    CHECK(ryml::emitrs_yaml<std::string>(tree) == ""
          "- a"                              "\n"
          "- b"                              "\n"
          "- {x0: 1,x1: 2}"                  "\n"
          "- champagne: Dom Perignon"        "\n"
          "  coffee: Arabica"                "\n"
          "");
    CHECK(root.is_seq());
    CHECK(root[0].val() == "a");
    CHECK(root[1].val() == "b");
    CHECK(root[2].is_map());
    CHECK(root[2]["x0"].val() == "1");
    CHECK(root[2]["x1"].val() == "2");
    CHECK(root[3].is_map());
    CHECK(root[3]["champagne"].val() == "Dom Perignon");
    CHECK(root[3]["coffee"].val() == "Arabica");
 
    // watchout: to add to an existing node within a map, the node's
    // key must be separately set first:
    ryml::NodeRef more = mroot[3].append_child({ryml::KEYMAP, "more"});
    ryml::NodeRef beer = mroot[3].append_child({ryml::KEYSEQ, "beer"});
    ryml::NodeRef always = mroot[3].append_child({ryml::KEY, "always"});
    ryml::parse_in_arena("{vinho verde: Soalheiro, vinho tinto: Redoma 2017}", more);
    ryml::parse_in_arena("- Rochefort 10\n- Busch\n- Leffe Rituel", beer);
    ryml::parse_in_arena("lots\nof\nwater", always);
    CHECK(ryml::emitrs_yaml<std::string>(tree) == ""
          "- a"                              "\n"
          "- b"                              "\n"
          "- {x0: 1,x1: 2}"                  "\n"
          "- champagne: Dom Perignon"        "\n"
          "  coffee: Arabica"                "\n"
          // note these were added:
          "  more:"                          "\n"
          "    vinho verde: Soalheiro"       "\n"
          "    vinho tinto: Redoma 2017"     "\n"
          "  beer:"                          "\n"
          "    - Rochefort 10"               "\n"
          "    - Busch"                      "\n"
          "    - Leffe Rituel"               "\n"
          "  always: lots of water"          "\n"
          "");
 
    // can append at the top:
    ryml::parse_in_arena("- foo\n- bar\n- baz\n- bat", mroot);
    CHECK(ryml::emitrs_yaml<std::string>(tree) == ""
          "- a"                              "\n"
          "- b"                              "\n"
          "- {x0: 1,x1: 2}"                  "\n"
          "- champagne: Dom Perignon"        "\n"
          "  coffee: Arabica"                "\n"
          "  more:"                          "\n"
          "    vinho verde: Soalheiro"       "\n"
          "    vinho tinto: Redoma 2017"     "\n"
          "  beer:"                          "\n"
          "    - Rochefort 10"               "\n"
          "    - Busch"                      "\n"
          "    - Leffe Rituel"               "\n"
          "  always: lots of water"          "\n"
          // note these were added
          "- foo"                            "\n"
          "- bar"                            "\n"
          "- baz"                            "\n"
          "- bat"                            "\n"
          "");
 
    // or nested:
    ryml::parse_in_arena("[Kasteel Donker]", beer);
    CHECK(ryml::emitrs_yaml<std::string>(tree) ==
          "- a"                              "\n"
          "- b"                              "\n"
          "- {x0: 1,x1: 2}"                  "\n"
          "- champagne: Dom Perignon"        "\n"
          "  coffee: Arabica"                "\n"
          "  more:"                          "\n"
          "    vinho verde: Soalheiro"       "\n"
          "    vinho tinto: Redoma 2017"     "\n"
          "  beer:"                          "\n"
          "    - Rochefort 10"               "\n"
          "    - Busch"                      "\n"
          "    - Leffe Rituel"               "\n"
          "    - Kasteel Donker"             "\n"  // added this
          "  always: lots of water"          "\n"
          "- foo"                            "\n"
          "- bar"                            "\n"
          "- baz"                            "\n"
          "- bat"                            "\n"
          "");
}

sample_parse_reuse_parser()

void sample_parse_reuse_parser

(

)

Demonstrates reuse of an existing parser.

Doing this is recommended when multiple files are parsed.

See also

Parse utilities

Definition at line 2138 of file quickstart.cpp.

{
    ryml::EventHandlerTree evt_handler = {};
    ryml::Parser parser(&evt_handler);
 
    // it is also advised to reserve the parser depth
    // to the expected depth of the data tree:
    parser.reserve_stack(10); // uses small storage optimization
                              // defaulting to 16 depth, so this
                              // instruction is a no-op, and the stack
                              // will located in the parser object.
    parser.reserve_stack(20); // But this will cause an allocation
                              // because it is above 16.
 
    ryml::csubstr yaml = "[Dom Perignon,Gosset Grande Reserve,Jacquesson 742]";
    ryml::Tree champagnes = parse_in_arena(&parser, "champagnes.yml", yaml);
    CHECK(ryml::emitrs_yaml<std::string>(champagnes) == yaml);
 
    yaml = "[Rochefort 10,Busch,Leffe Rituel,Kasteel Donker]";
    ryml::Tree beers = parse_in_arena(&parser, "beers.yml", yaml);
    CHECK(ryml::emitrs_yaml<std::string>(beers) == yaml);
}

sample_parse_reuse_tree_and_parser()

void sample_parse_reuse_tree_and_parser

(

)

for ultimate speed when parsing multiple times, reuse both the tree and parser

See also

Parse utilities

Definition at line 2167 of file quickstart.cpp.

{
    ryml::Tree tree;
    // it will always be faster if the tree's size is conveniently reserved:
    tree.reserve(30); // reserve 30 nodes (good enough for this sample)
    // if you are using the tree's arena to serialize data,
    // then reserve also the arena's size:
    tree.reserve(256); // reserve 256 characters (good enough for this sample)
 
    ryml::EventHandlerTree evt_handler;
    ryml::Parser parser(&evt_handler);
    // it is also advised to reserve the parser depth
    // to the expected depth of the data tree:
    parser.reserve_stack(10); // the parser uses small storage
                              // optimization defaulting to 16 depth,
                              // so this instruction is a no-op, and
                              // the stack will be located in the
                              // parser object.
    parser.reserve_stack(20); // But this will cause an allocation
                              // because it is above 16.
 
    ryml::csubstr champagnes = ""
        "- Dom Perignon\n"
        "- Gosset Grande Reserve\n"
        "- Jacquesson 742\n"
        "";
    ryml::csubstr beers = ""
        "- Rochefort 10\n"
        "- Busch\n"
        "- Leffe Rituel\n"
        "- Kasteel Donker\n"
        "";
    ryml::csubstr wines = ""
        "- Soalheiro\n"
        "- Niepoort Redoma 2017\n"
        "- Vina Esmeralda\n"
        "";
 
    parse_in_arena(&parser, "champagnes.yml", champagnes, &tree);
    CHECK(ryml::emitrs_yaml<std::string>(tree) == champagnes);
 
    // watchout: this will APPEND to the given tree:
    parse_in_arena(&parser, "beers.yml", beers, &tree);
    CHECK(ryml::emitrs_yaml<std::string>(tree) == ""
          "- Dom Perignon\n"
          "- Gosset Grande Reserve\n"
          "- Jacquesson 742\n"
          "- Rochefort 10\n"
          "- Busch\n"
          "- Leffe Rituel\n"
          "- Kasteel Donker\n"
          "");
 
    // if you don't wish to append, clear the tree first:
    tree.clear();
    parse_in_arena(&parser, "wines.yml", wines, &tree);
    CHECK(ryml::emitrs_yaml<std::string>(tree) == ""
        "- Soalheiro\n"
        "- Niepoort Redoma 2017\n"
        "- Vina Esmeralda\n"
        "");
}

sample_iterate_trees()

void sample_iterate_trees

(

)

shows how to programatically iterate through trees

See also

Tree utilities

Node classes

Definition at line 2237 of file quickstart.cpp.

{
    const ryml::Tree tree = ryml::parse_in_arena(
        "doe: a deer, a female deer"      "\n"
        "ray: a drop of golden sun"       "\n"
        "pi: 3.14159"                     "\n"
        "xmas: true"                      "\n"
        "french-hens: 3"                  "\n"
        "calling-birds:"                  "\n"
        "   - huey"                       "\n"
        "   - dewey"                      "\n"
        "   - louie"                      "\n"
        "   - fred"                       "\n"
        "xmas-fifth-day:"                 "\n"
        "  calling-birds: four"           "\n"
        "  french-hens: 3"                "\n"
        "  golden-rings: 5"               "\n"
        "  partridges:"                   "\n"
        "    count: 1"                    "\n"
        "    location: a pear tree"       "\n"
        "  turtle-doves: two"             "\n"
        "cars: GTO"                       "\n"
        "");
    ryml::ConstNodeRef root = tree.crootref();
 
    // iterate children
    {
        std::vector<ryml::csubstr> keys, vals; // to store all the root-level keys, vals
        for(ryml::ConstNodeRef n : root.children())
        {
            keys.emplace_back(n.key());
            vals.emplace_back(n.has_val() ? n.val() : ryml::csubstr{});
        }
        CHECK(keys.size() >= 6);
        CHECK(vals.size() >= 6);
        if(keys.size() >= 6 && vals.size() >= 6)
        {
            CHECK(keys[0] == "doe");
            CHECK(vals[0] == "a deer, a female deer");
            CHECK(keys[1] == "ray");
            CHECK(vals[1] == "a drop of golden sun");
            CHECK(keys[2] == "pi");
            CHECK(vals[2] == "3.14159");
            CHECK(keys[3] == "xmas");
            CHECK(vals[3] == "true");
            CHECK(root[5].has_key());
            CHECK(root[5].is_seq());
            CHECK(root[5].key() == "calling-birds");
            CHECK(!root[5].has_val()); // it is a map, so not a val
            //CHECK(root[5].val() == ""); // ERROR! node does not have a val.
            CHECK(keys[5] == "calling-birds");
            CHECK(vals[5] == "");
        }
    }
 
    // iterate siblings
    {
        size_t count = 0;
        ryml::csubstr calling_birds[] = {"huey", "dewey", "louie", "fred"};
        for(ryml::ConstNodeRef n : root["calling-birds"][2].siblings())
            CHECK(n.val() == calling_birds[count++]);
        CHECK(count == 4u);
    }
}

sample_create_trees()

void sample_create_trees

(

)

shows how to programatically create trees

See also

Tree utilities

Node classes

Definition at line 2309 of file quickstart.cpp.

{
    ryml::NodeRef doe;
    CHECK(doe.invalid()); // it's pointing at nowhere
 
    ryml::Tree tree;
    ryml::NodeRef root = tree.rootref();
    root |= ryml::MAP; // mark root as a map
    doe = root["doe"];
    CHECK(!doe.invalid()); // it's now pointing at the tree
    CHECK(doe.is_seed()); // but the tree has nothing there, so this is only a seed
 
    // set the value of the node
    const char a_deer[] = "a deer, a female deer";
    doe = a_deer;
    // now the node really exists in the tree, and this ref is no
    // longer a seed:
    CHECK(!doe.is_seed());
    // WATCHOUT for lifetimes:
    CHECK(doe.val().str == a_deer); // it is pointing at the initial string
    // If you need to avoid lifetime dependency, serialize the data:
    {
        std::string a_drop = "a drop of golden sun";
        // this will copy the string to the tree's arena:
        // (see the serialization samples below)
        root["ray"] << a_drop;
        // and now you can modify the original string without changing
        // the tree:
        a_drop[0] = 'Z';
        a_drop[1] = 'Z';
    }
    CHECK(root["ray"].val() == "a drop of golden sun");
 
    // etc.
    root["pi"] << ryml::fmt::real(3.141592654, 5);
    root["xmas"] << ryml::fmt::boolalpha(true);
    root["french-hens"] << 3;
    ryml::NodeRef calling_birds = root["calling-birds"];
    calling_birds |= ryml::SEQ;
    calling_birds.append_child() = "huey";
    calling_birds.append_child() = "dewey";
    calling_birds.append_child() = "louie";
    calling_birds.append_child() = "fred";
    ryml::NodeRef xmas5 = root["xmas-fifth-day"];
    xmas5 |= ryml::MAP;
    xmas5["calling-birds"] = "four";
    xmas5["french-hens"] << 3;
    xmas5["golden-rings"] << 5;
    xmas5["partridges"] |= ryml::MAP;
    xmas5["partridges"]["count"] << 1;
    xmas5["partridges"]["location"] = "a pear tree";
    xmas5["turtle-doves"] = "two";
    root["cars"] = "GTO";
 
    CHECK(ryml::emitrs_yaml<std::string>(tree) == ""
          "doe: a deer, a female deer"       "\n"
          "ray: a drop of golden sun"        "\n"
          "pi: 3.14159"                      "\n"
          "xmas: true"                       "\n"
          "french-hens: 3"                   "\n"
          "calling-birds:"                   "\n"
          "  - huey"                         "\n"
          "  - dewey"                        "\n"
          "  - louie"                        "\n"
          "  - fred"                         "\n"
          "xmas-fifth-day:"                  "\n"
          "  calling-birds: four"            "\n"
          "  french-hens: 3"                 "\n"
          "  golden-rings: 5"                "\n"
          "  partridges:"                    "\n"
          "    count: 1"                     "\n"
          "    location: a pear tree"        "\n"
          "  turtle-doves: two"              "\n"
          "cars: GTO"                        "\n"
          "");
}

sample_tree_arena()

void sample_tree_arena

(

)

demonstrates explicit and implicit interaction with the tree's string arena.

Notice that ryml only holds strings in the tree's nodes.

Definition at line 2391 of file quickstart.cpp.

{
    // mutable buffers are parsed in situ:
    {
        char buf[] = "[a, b, c, d]";
        ryml::substr yml = buf;
        ryml::Tree tree = ryml::parse_in_place(yml);
        // notice the arena is empty:
        CHECK(tree.arena().empty());
        // and the tree is pointing at the original buffer:
        ryml::NodeRef root = tree.rootref();
        CHECK(root[0].val().is_sub(yml));
        CHECK(root[1].val().is_sub(yml));
        CHECK(root[2].val().is_sub(yml));
        CHECK(root[3].val().is_sub(yml));
        CHECK(yml.is_super(root[0].val()));
        CHECK(yml.is_super(root[1].val()));
        CHECK(yml.is_super(root[2].val()));
        CHECK(yml.is_super(root[3].val()));
    }
 
    // when parsing immutable buffers, the buffer is first copied to the
    // tree's arena; the copy in the arena is then the buffer which is
    // actually parsed
    {
        ryml::csubstr yml = "[a, b, c, d]";
        ryml::Tree tree = ryml::parse_in_arena(yml);
        // notice the buffer was copied to the arena:
        CHECK(tree.arena().data() != yml.data());
        CHECK(tree.arena() == yml);
        // and the tree is pointing at the arena instead of to the
        // original buffer:
        ryml::NodeRef root = tree.rootref();
        ryml::csubstr arena = tree.arena();
        CHECK(root[0].val().is_sub(arena));
        CHECK(root[1].val().is_sub(arena));
        CHECK(root[2].val().is_sub(arena));
        CHECK(root[3].val().is_sub(arena));
        CHECK(arena.is_super(root[0].val()));
        CHECK(arena.is_super(root[1].val()));
        CHECK(arena.is_super(root[2].val()));
        CHECK(arena.is_super(root[3].val()));
    }
 
    // the arena is also used when the data is serialized to string
    // with NodeRef::operator<<(): mutable buffer
    {
        char buf[] = "[a, b, c, d]"; // mutable
        ryml::substr yml = buf;
        ryml::Tree tree = ryml::parse_in_place(yml);
        // notice the arena is empty:
        CHECK(tree.arena().empty());
        ryml::NodeRef root = tree.rootref();
 
        // serialize an integer, and mutate the tree
        CHECK(root[2].val() == "c");
        CHECK(root[2].val().is_sub(yml)); // val is first pointing at the buffer
        root[2] << 12345;
        CHECK(root[2].val() == "12345");
        CHECK(root[2].val().is_sub(tree.arena())); // now val is pointing at the arena
        // notice the serialized string was appended to the tree's arena:
        CHECK(tree.arena() == "12345");
 
        // serialize an integer, and mutate the tree
        CHECK(root[3].val() == "d");
        CHECK(root[3].val().is_sub(yml)); // val is first pointing at the buffer
        root[3] << 67890;
        CHECK(root[3].val() == "67890");
        CHECK(root[3].val().is_sub(tree.arena())); // now val is pointing at the arena
        // notice the serialized string was appended to the tree's arena:
        CHECK(tree.arena() == "1234567890");
    }
    // the arena is also used when the data is serialized to string
    // with NodeRef::operator<<(): immutable buffer
    {
        ryml::csubstr yml = "[a, b, c, d]"; // immutable
        ryml::Tree tree = ryml::parse_in_arena(yml);
        // notice the buffer was copied to the arena:
        CHECK(tree.arena().data() != yml.data());
        CHECK(tree.arena() == yml);
        ryml::NodeRef root = tree.rootref();
 
        // serialize an integer, and mutate the tree
        CHECK(root[2].val() == "c");
        root[2] << 12345; // serialize an integer
        CHECK(root[2].val() == "12345");
        // notice the serialized string was appended to the tree's arena:
        // notice also the previous values remain there.
        // RYML DOES NOT KEEP TRACK OF REFERENCES TO THE ARENA.
        CHECK(tree.arena() == "[a, b, c, d]12345");
        // old values:  --------------^
 
        // serialize an integer, and mutate the tree
        root[3] << 67890;
        CHECK(root[3].val() == "67890");
        // notice the serialized string was appended to the tree's arena:
        // notice also the previous values remain there.
        // RYML DOES NOT KEEP TRACK OF REFERENCES TO THE ARENA.
        CHECK(tree.arena() == "[a, b, c, d]1234567890");
        // old values:  --------------^ ---^^^^^
    }
 
    // to_arena(): directly serialize values to the arena:
    {
        ryml::Tree tree = ryml::parse_in_arena("{a: b}");
        ryml::csubstr c10 = tree.to_arena(10101010);
        CHECK(c10 == "10101010");
        CHECK(c10.is_sub(tree.arena()));
        CHECK(tree.arena() == "{a: b}10101010");
        CHECK(tree.key(1) == "a");
        CHECK(tree.val(1) == "b");
        tree.set_val(1, c10);
        CHECK(tree.val(1) == c10);
        // and you can also do it through a node:
        ryml::NodeRef root = tree.rootref();
        root["a"].set_val_serialized(2222);
        CHECK(root["a"].val() == "2222");
        CHECK(tree.arena() == "{a: b}101010102222");
    }
 
    // copy_to_arena(): manually copy a string to the arena:
    {
        ryml::Tree tree = ryml::parse_in_arena("{a: b}");
        ryml::csubstr mystr = "Gosset Grande Reserve";
        ryml::csubstr copied = tree.copy_to_arena(mystr);
        CHECK(!copied.overlaps(mystr));
        CHECK(copied == mystr);
        CHECK(tree.arena() == "{a: b}Gosset Grande Reserve");
    }
 
    // alloc_arena(): allocate a buffer from the arena:
    {
        ryml::Tree tree = ryml::parse_in_arena("{a: b}");
        ryml::csubstr mystr = "Gosset Grande Reserve";
        ryml::substr copied = tree.alloc_arena(mystr.size());
        CHECK(!copied.overlaps(mystr));
        memcpy(copied.str, mystr.str, mystr.len);
        CHECK(copied == mystr);
        CHECK(tree.arena() == "{a: b}Gosset Grande Reserve");
    }
 
    // reserve_arena(): ensure the arena has a certain size to avoid reallocations
    {
        ryml::Tree tree = ryml::parse_in_arena("{a: b}");
        CHECK(tree.arena().size() == strlen("{a: b}"));
        tree.reserve_arena(100);
        CHECK(tree.arena_capacity() >= 100);
        CHECK(tree.arena().size() == strlen("{a: b}"));
        tree.to_arena(123456);
        CHECK(tree.arena().first(12) == "{a: b}123456");
    }
}

sample_fundamental_types()

void sample_fundamental_types

(

)

ryml provides facilities for serializing and deserializing the C++ fundamental types, including boolean and null values; this is provided by the several overloads in to_chars: generalized chars to value and from_chars: generalized chars to value.

To add serialization for user scalar types (ie, those types that should be serialized as strings in leaf nodes), you just need to define the appropriate overloads of to_chars and from_chars for those types; see sample_user_scalar_types for an example on how to achieve this, and see Serialization/deserialization for more information on serialization.

Definition at line 2556 of file quickstart.cpp.

{
    ryml::Tree tree;
    CHECK(tree.arena().empty());
    CHECK(tree.to_arena('a') == "a");               CHECK(tree.arena() == "a");
    CHECK(tree.to_arena("bcde") == "bcde");         CHECK(tree.arena() == "abcde");
    CHECK(tree.to_arena(unsigned(0)) == "0");       CHECK(tree.arena() == "abcde0");
    CHECK(tree.to_arena(int(1)) == "1");            CHECK(tree.arena() == "abcde01");
    CHECK(tree.to_arena(uint8_t(0)) == "0");        CHECK(tree.arena() == "abcde010");
    CHECK(tree.to_arena(uint16_t(1)) ==  "1");      CHECK(tree.arena() == "abcde0101");
    CHECK(tree.to_arena(uint32_t(2)) ==  "2");      CHECK(tree.arena() == "abcde01012");
    CHECK(tree.to_arena(uint64_t(3)) ==  "3");      CHECK(tree.arena() == "abcde010123");
    CHECK(tree.to_arena(int8_t( 4))  ==  "4");      CHECK(tree.arena() == "abcde0101234");
    CHECK(tree.to_arena(int8_t(-4))  == "-4");      CHECK(tree.arena() == "abcde0101234-4");
    CHECK(tree.to_arena(int16_t( 5)) ==  "5");      CHECK(tree.arena() == "abcde0101234-45");
    CHECK(tree.to_arena(int16_t(-5)) == "-5");      CHECK(tree.arena() == "abcde0101234-45-5");
    CHECK(tree.to_arena(int32_t( 6)) ==  "6");      CHECK(tree.arena() == "abcde0101234-45-56");
    CHECK(tree.to_arena(int32_t(-6)) == "-6");      CHECK(tree.arena() == "abcde0101234-45-56-6");
    CHECK(tree.to_arena(int64_t( 7)) ==  "7");      CHECK(tree.arena() == "abcde0101234-45-56-67");
    CHECK(tree.to_arena(int64_t(-7)) == "-7");      CHECK(tree.arena() == "abcde0101234-45-56-67-7");
    CHECK(tree.to_arena((void*)1) == "0x1");        CHECK(tree.arena() == "abcde0101234-45-56-67-70x1");
    CHECK(tree.to_arena(float(0.124)) == "0.124");  CHECK(tree.arena() == "abcde0101234-45-56-67-70x10.124");
    CHECK(tree.to_arena(double(0.234)) == "0.234"); CHECK(tree.arena() == "abcde0101234-45-56-67-70x10.1240.234");
 
    // write boolean values - see also sample_formatting()
    CHECK(tree.to_arena(bool(true)) == "1");                    CHECK(tree.arena() == "abcde0101234-45-56-67-70x10.1240.2341");
    CHECK(tree.to_arena(bool(false)) == "0");                   CHECK(tree.arena() == "abcde0101234-45-56-67-70x10.1240.23410");
    CHECK(tree.to_arena(c4::fmt::boolalpha(true)) == "true");   CHECK(tree.arena() == "abcde0101234-45-56-67-70x10.1240.23410true");
    CHECK(tree.to_arena(c4::fmt::boolalpha(false)) == "false"); CHECK(tree.arena() == "abcde0101234-45-56-67-70x10.1240.23410truefalse");
 
    // write special float values
    // see also sample_float_precision()
    const float  fnan = std::numeric_limits<float >::quiet_NaN();
    const double dnan = std::numeric_limits<double>::quiet_NaN();
    const float  finf = std::numeric_limits<float >::infinity();
    const double dinf = std::numeric_limits<double>::infinity();
    CHECK(tree.to_arena( finf) ==  ".inf"); CHECK(tree.arena() == "abcde0101234-45-56-67-70x10.1240.23410truefalse.inf");
    CHECK(tree.to_arena( dinf) ==  ".inf"); CHECK(tree.arena() == "abcde0101234-45-56-67-70x10.1240.23410truefalse.inf.inf");
    CHECK(tree.to_arena(-finf) == "-.inf"); CHECK(tree.arena() == "abcde0101234-45-56-67-70x10.1240.23410truefalse.inf.inf-.inf");
    CHECK(tree.to_arena(-dinf) == "-.inf"); CHECK(tree.arena() == "abcde0101234-45-56-67-70x10.1240.23410truefalse.inf.inf-.inf-.inf");
    CHECK(tree.to_arena( fnan) ==  ".nan"); CHECK(tree.arena() == "abcde0101234-45-56-67-70x10.1240.23410truefalse.inf.inf-.inf-.inf.nan");
    CHECK(tree.to_arena( dnan) ==  ".nan"); CHECK(tree.arena() == "abcde0101234-45-56-67-70x10.1240.23410truefalse.inf.inf-.inf-.inf.nan.nan");
 
    // read special float values
    // see also sample_float_precision()
    C4_SUPPRESS_WARNING_GCC_CLANG_WITH_PUSH("-Wfloat-equal");
    tree = ryml::parse_in_arena(R"({ninf: -.inf, pinf: .inf, nan: .nan})");
    float f = 0.f;
    double d = 0.;
    CHECK(f == 0.f);
    CHECK(d == 0.);
    tree["ninf"] >> f; CHECK(f == -finf);
    tree["ninf"] >> d; CHECK(d == -dinf);
    tree["pinf"] >> f; CHECK(f ==  finf);
    tree["pinf"] >> d; CHECK(d ==  dinf);
    tree["nan" ] >> f; CHECK(std::isnan(f));
    tree["nan" ] >> d; CHECK(std::isnan(d));
    C4_SUPPRESS_WARNING_GCC_CLANG_POP
 
    // value overflow detection:
    // (for integral types only)
    {
        // we will be detecting errors below, so we use this sample helper
        ScopedErrorHandlerExample err = {};
        ryml::Tree t(err.callbacks()); // instantiate with the error-detecting callbacks
        // create a simple tree with an int value
        ryml::parse_in_arena(R"({val: 258})", &t);
        // by default, overflow is not detected:
        uint8_t valu8 = 0;
        int8_t vali8 = 0;
        t["val"] >> valu8; CHECK(valu8 == 2); // not 257; it wrapped around
        t["val"] >> vali8; CHECK(vali8 == 2); // not 257; it wrapped around
        // ...but there are facilities to detect overflow
        CHECK(ryml::overflows<uint8_t>(t["val"].val()));
        CHECK(ryml::overflows<int8_t>(t["val"].val()));
        CHECK( ! ryml::overflows<int16_t>(t["val"].val()));
        // and there is a format helper
        CHECK(err.check_error_occurs([&]{
            auto checku8 = ryml::fmt::overflow_checked(valu8); // need to declare the wrapper type before using it with >>
            t["val"] >> checku8; // this will cause an error
        }));
        CHECK(err.check_error_occurs([&]{
            auto checki8 = ryml::fmt::overflow_checked(vali8); // need to declare the wrapper type before using it with >>
            t["val"] >> checki8; // this will cause an error
        }));
    }
}

sample_empty_null_values()

void sample_empty_null_values

(

)

Shows how to deal with empty/null values.

See also c4::yml::Tree::val_is_null

Definition at line 2649 of file quickstart.cpp.

{
    // reading empty/null values - see also sample_formatting()
    ryml::Tree tree = ryml::parse_in_arena(""
        "plain:"                                                  "\n"
        "squoted: ''"                                             "\n"
        "dquoted: \"\""                                           "\n"
        "literal: |"                                              "\n"
        "folded: >"                                               "\n"
        "all_null: [~, null, Null, NULL]"                         "\n"
        "non_null: [nULL, non_null, non null, null it is not]"    "\n"
        "");
    // first, remember that .has_val() is a structural predicate
    // indicating the node is a leaf, and not a container.
    CHECK(tree["plain"].has_val()); // has a val, even if it's empty!
    CHECK(tree["squoted"].has_val());
    CHECK(tree["dquoted"].has_val());
    CHECK(tree["literal"].has_val());
    CHECK(tree["folded"].has_val());
    CHECK( ! tree["all_null"].has_val());
    CHECK( ! tree["non_null"].has_val());
    // In essence, has_val() is the logical opposite of is_container()
    CHECK( ! tree["plain"].is_container());
    CHECK( ! tree["squoted"].is_container());
    CHECK( ! tree["dquoted"].is_container());
    CHECK( ! tree["literal"].is_container());
    CHECK( ! tree["folded"].is_container());
    CHECK(tree["all_null"].is_container());
    CHECK(tree["non_null"].is_container());
    //
    // Right. How about the contents of each val?
    //
    // all of these scalars have zero-length:
    CHECK(tree["plain"].val().len == 0);
    CHECK(tree["squoted"].val().len == 0);
    CHECK(tree["dquoted"].val().len == 0);
    CHECK(tree["literal"].val().len == 0);
    CHECK(tree["folded"].val().len == 0);
    // but only the empty scalar has null string:
    CHECK(tree["plain"].val().str == nullptr);
    CHECK(tree["squoted"].val().str != nullptr);
    CHECK(tree["dquoted"].val().str != nullptr);
    CHECK(tree["literal"].val().str != nullptr);
    CHECK(tree["folded"].val().str != nullptr);
    // likewise, scalar comparison to nullptr has the same results:
    // (remember that .val() gives you the scalar value, node must
    // have a val, ie must be a leaf node, not a container)
    CHECK(tree["plain"].val() == nullptr);
    CHECK(tree["squoted"].val() != nullptr);
    CHECK(tree["dquoted"].val() != nullptr);
    CHECK(tree["literal"].val() != nullptr);
    CHECK(tree["folded"].val() != nullptr);
    // the tree and node classes provide the corresponding predicate
    // functions .key_is_null() and .val_is_null().
    // (note that these functions have the same preconditions as .val(),
    // because they need get the val to look into its contents)
    CHECK(tree["plain"].val_is_null());
    CHECK( ! tree["squoted"].val_is_null());
    CHECK( ! tree["dquoted"].val_is_null());
    CHECK( ! tree["literal"].val_is_null());
    CHECK( ! tree["folded"].val_is_null());
    // matching to null is case-sensitive. only the cases shown here
    // match to null:
    for(ryml::ConstNodeRef child : tree["all_null"].children())
    {
        CHECK(child.val() != nullptr); // it is pointing at a string, so it is not nullptr!
        CHECK(child.val_is_null());
    }
    for(ryml::ConstNodeRef child : tree["non_null"].children())
    {
        CHECK(child.val() != nullptr);
        CHECK( ! child.val_is_null());
    }
    //
    //
    // Because the meaning of null/~/empty will vary from application
    // to application, ryml makes no assumption on what should be
    // serialized as null. It leaves this decision to the user. But
    // it also provides the proper toolbox for the user to implement
    // its intended solution.
    //
    // writing/disambiguating null values:
    ryml::csubstr null = {};
    ryml::csubstr nonnull = "";
    ryml::csubstr strnull = "null";
    ryml::csubstr tilde = "~";
    CHECK(null   .len == 0); CHECK(null   .str == nullptr); CHECK(null    == nullptr);
    CHECK(nonnull.len == 0); CHECK(nonnull.str != nullptr); CHECK(nonnull != nullptr);
    CHECK(strnull.len != 0); CHECK(strnull.str != nullptr); CHECK(strnull != nullptr);
    CHECK(tilde  .len != 0); CHECK(tilde  .str != nullptr); CHECK(tilde   != nullptr);
    tree.clear();
    tree.clear_arena();
    tree.rootref() |= ryml::MAP;
    // serializes as an empty plain scalar:
    tree["empty_null"] << null; CHECK(tree.arena() == "");
    // serializes as an empty quoted scalar:
    tree["empty_nonnull"] << nonnull; CHECK(tree.arena() == "");
    // serializes as the normal 'null' string:
    tree["str_null"] << strnull; CHECK(tree.arena() == "null");
    // serializes as the normal '~' string:
    tree["str_tilde"] << tilde; CHECK(tree.arena() == "null~");
    // this is the resulting yaml:
    CHECK(ryml::emitrs_yaml<std::string>(tree) == ""
          "empty_null: "         "\n"
          "empty_nonnull: ''"    "\n"
          "str_null: null"       "\n"
          "str_tilde: ~"         "\n"
          "");
    // To enforce a particular concept of what is a null string, you
    // can use the appropriate condition based on pointer nulity or
    // other appropriate criteria.
    //
    // As an example, proper comparison to nullptr:
    auto null_if_nullptr = [](ryml::csubstr s) {
        return s.str == nullptr ? "null" : s;
    };
    tree["empty_null"] << null_if_nullptr(null);
    tree["empty_nonnull"] << null_if_nullptr(nonnull);
    tree["str_null"] << null_if_nullptr(strnull);
    tree["str_tilde"] << null_if_nullptr(tilde);
    // this is the resulting yaml:
    CHECK(ryml::emitrs_yaml<std::string>(tree) == ""
          "empty_null: null"    "\n"
          "empty_nonnull: ''"   "\n"
          "str_null: null"      "\n"
          "str_tilde: ~"        "\n"
          "");
    //
    // As another example, nulity check based on the YAML nulity
    // predicate:
    auto null_if_predicate = [](ryml::csubstr s) {
        return ryml::scalar_is_null(s) ? "null" : s;
    };
    tree["empty_null"] << null_if_predicate(null);
    tree["empty_nonnull"] << null_if_predicate(nonnull);
    tree["str_null"] << null_if_predicate(strnull);
    tree["str_tilde"] << null_if_predicate(tilde);
    // this is the resulting yaml:
    CHECK(ryml::emitrs_yaml<std::string>(tree) == ""
          "empty_null: null"    "\n"
          "empty_nonnull: ''"   "\n"
          "str_null: null"      "\n"
          "str_tilde: null"     "\n"
          "");
    //
    // As another example, nulity check based on the YAML nulity
    // predicate, but returning "~" to simbolize nulity:
    auto tilde_if_predicate = [](ryml::csubstr s) {
        return ryml::scalar_is_null(s) ? "~" : s;
    };
    tree["empty_null"] << tilde_if_predicate(null);
    tree["empty_nonnull"] << tilde_if_predicate(nonnull);
    tree["str_null"] << tilde_if_predicate(strnull);
    tree["str_tilde"] << tilde_if_predicate(tilde);
    // this is the resulting yaml:
    CHECK(ryml::emitrs_yaml<std::string>(tree) == ""
          "empty_null: ~"         "\n"
          "empty_nonnull: ''"     "\n"
          "str_null: ~"           "\n"
          "str_tilde: ~"          "\n"
          "");
}

sample_formatting()

void sample_formatting

(

)

ryml provides facilities for formatting/deformatting (imported from c4core into the ryml namespace).

See Format utilities . These functions are very useful to serialize and deserialize scalar types; see Serialization/deserialization .

Definition at line 2820 of file quickstart.cpp.

{
    // format(), format_sub(), formatrs(): format arguments
    {
        char buf_[256] = {};
        ryml::substr buf = buf_;
        size_t size = ryml::format(buf, "a={} foo {} {} bar {}", 0.1, 10, 11, 12);
        CHECK(size == strlen("a=0.1 foo 10 11 bar 12"));
        CHECK(buf.first(size) == "a=0.1 foo 10 11 bar 12");
        // it is safe to call on an empty buffer:
        // returns the size needed for the result, and no overflow occurs:
        size = ryml::format({} , "a={} foo {} {} bar {}", "this_is_a", 10, 11, 12);
        CHECK(size == ryml::format(buf, "a={} foo {} {} bar {}", "this_is_a", 10, 11, 12));
        CHECK(size == strlen("a=this_is_a foo 10 11 bar 12"));
        // it is also safe to call on an insufficient buffer:
        char smallbuf[8] = {};
        size = ryml::format(smallbuf, "{} is too large {}", "this", "for the buffer");
        CHECK(size == strlen("this is too large for the buffer"));
        // ... and the result is truncated at the buffer size:
        CHECK(ryml::substr(smallbuf, sizeof(smallbuf)) == "this is\0");
 
        // format_sub() directly returns the written string:
        ryml::csubstr result = ryml::format_sub(buf, "b={}, damn it.", 1);
        CHECK(result == "b=1, damn it.");
        CHECK(result.is_sub(buf));
 
        // formatrs() means FORMAT & ReSize:
        //
        // Instead of a substr, it receives any owning linear char container
        // for which to_substr() is defined (using ADL).
        // <ryml_std.hpp> has to_substr() definitions for std::string and
        // std::vector<char>.
        //
        // formatrs() starts by calling format(), and if needed, resizes the container
        // and calls format() again.
        //
        // Note that unless the container is previously sized, this
        // may cause an allocation, which will make your code slower.
        // Make sure to call .reserve() on the container for real
        // production code.
        std::string sbuf;
        ryml::formatrs(&sbuf, "and c={} seems about right", 2);
        CHECK(sbuf == "and c=2 seems about right");
        std::vector<char> vbuf; // works with any linear char container
        ryml::formatrs(&vbuf, "and c={} seems about right", 2);
        CHECK(sbuf == "and c=2 seems about right");
        // with formatrs() it is also possible to append:
        ryml::formatrs_append(&sbuf, ", and finally d={} - done", 3);
        CHECK(sbuf == "and c=2 seems about right, and finally d=3 - done");
    }
 
    // unformat(): read arguments - opposite of format()
    {
        char buf_[256];
 
        int a = 0, b = 1, c = 2;
        ryml::csubstr result = ryml::format_sub(buf_, "{} and {} and {}", a, b, c);
        CHECK(result == "0 and 1 and 2");
        int aa = -1, bb = -2, cc = -3;
        size_t num_characters = ryml::unformat(result, "{} and {} and {}", aa, bb, cc);
        CHECK(num_characters != ryml::csubstr::npos); // if a conversion fails, returns ryml::csubstr::npos
        CHECK(num_characters == result.size());
        CHECK(aa == a);
        CHECK(bb == b);
        CHECK(cc == c);
 
        result = ryml::format_sub(buf_, "{} and {} and {}", 10, 20, 30);
        CHECK(result == "10 and 20 and 30");
        num_characters = ryml::unformat(result, "{} and {} and {}", aa, bb, cc);
        CHECK(num_characters != ryml::csubstr::npos); // if a conversion fails, returns ryml::csubstr::npos
        CHECK(num_characters == result.size());
        CHECK(aa == 10);
        CHECK(bb == 20);
        CHECK(cc == 30);
    }
 
    // cat(), cat_sub(), catrs(): concatenate arguments
    {
        char buf_[256] = {};
        ryml::substr buf = buf_;
        size_t size = ryml::cat(buf, "a=", 0.1, "foo", 10, 11, "bar", 12);
        CHECK(size == strlen("a=0.1foo1011bar12"));
        CHECK(buf.first(size) == "a=0.1foo1011bar12");
        // it is safe to call on an empty buffer:
        // returns the size needed for the result, and no overflow occurs:
        CHECK(ryml::cat({}, "a=", 0) == 3);
        // it is also safe to call on an insufficient buffer:
        char smallbuf[8] = {};
        size = ryml::cat(smallbuf, "this", " is too large ", "for the buffer");
        CHECK(size == strlen("this is too large for the buffer"));
        // ... and the result is truncated at the buffer size:
        CHECK(ryml::substr(smallbuf, sizeof(smallbuf)) == "this is\0");
 
        // cat_sub() directly returns the written string:
        ryml::csubstr result = ryml::cat_sub(buf, "b=", 1, ", damn it.");
        CHECK(result == "b=1, damn it.");
        CHECK(result.is_sub(buf));
 
        // catrs() means CAT & ReSize:
        //
        // Instead of a substr, it receives any owning linear char container
        // for which to_substr() is defined (using ADL).
        // <ryml_std.hpp> has to_substr() definitions for std::string and
        // std::vector<char>.
        //
        // catrs() starts by calling cat(), and if needed, resizes the container
        // and calls cat() again.
        //
        // Note that unless the container is previously sized, this
        // may cause an allocation, which will make your code slower.
        // Make sure to call .reserve() on the container for real
        // production code.
        std::string sbuf;
        ryml::catrs(&sbuf, "and c=", 2, " seems about right");
        CHECK(sbuf == "and c=2 seems about right");
        std::vector<char> vbuf; // works with any linear char container
        ryml::catrs(&vbuf, "and c=", 2, " seems about right");
        CHECK(sbuf == "and c=2 seems about right");
        // with catrs() it is also possible to append:
        ryml::catrs_append(&sbuf, ", and finally d=", 3, " - done");
        CHECK(sbuf == "and c=2 seems about right, and finally d=3 - done");
    }
 
    // uncat(): read arguments - opposite of cat()
    {
        char buf_[256];
 
        int a = 0, b = 1, c = 2;
        ryml::csubstr result = ryml::cat_sub(buf_, a, ' ', b, ' ', c);
        CHECK(result == "0 1 2");
        int aa = -1, bb = -2, cc = -3;
        char sep1 = 'a', sep2 = 'b';
        size_t num_characters = ryml::uncat(result, aa, sep1, bb, sep2, cc);
        CHECK(num_characters == result.size());
        CHECK(aa == a);
        CHECK(bb == b);
        CHECK(cc == c);
        CHECK(sep1 == ' ');
        CHECK(sep2 == ' ');
 
        result = ryml::cat_sub(buf_, 10, ' ', 20, ' ', 30);
        CHECK(result == "10 20 30");
        num_characters = ryml::uncat(result, aa, sep1, bb, sep2, cc);
        CHECK(num_characters == result.size());
        CHECK(aa == 10);
        CHECK(bb == 20);
        CHECK(cc == 30);
        CHECK(sep1 == ' ');
        CHECK(sep2 == ' ');
    }
 
    // catsep(), catsep_sub(), catseprs(): concatenate arguments, with a separator
    {
        char buf_[256] = {};
        ryml::substr buf = buf_;
        // use ' ' as a separator
        size_t size = ryml::catsep(buf, ' ', "a=", 0, "b=", 1, "c=", 2, 45, 67);
        CHECK(buf.first(size) == "a= 0 b= 1 c= 2 45 67");
        // any separator may be used
        // use " and " as a separator
        size = ryml::catsep(buf, " and ", "a=0", "b=1", "c=2", 45, 67);
        CHECK(buf.first(size) == "a=0 and b=1 and c=2 and 45 and 67");
        // use " ... " as a separator
        size = ryml::catsep(buf, " ... ", "a=0", "b=1", "c=2", 45, 67);
        CHECK(buf.first(size) == "a=0 ... b=1 ... c=2 ... 45 ... 67");
        // use '/' as a separator
        size = ryml::catsep(buf, '/', "a=", 0, "b=", 1, "c=", 2, 45, 67);
        CHECK(buf.first(size) == "a=/0/b=/1/c=/2/45/67");
        // use 888 as a separator
        size = ryml::catsep(buf, 888, "a=0", "b=1", "c=2", 45, 67);
        CHECK(buf.first(size) == "a=0888b=1888c=28884588867");
 
        // it is safe to call on an empty buffer:
        // returns the size needed for the result, and no overflow occurs:
        CHECK(size == ryml::catsep({}, 888, "a=0", "b=1", "c=2", 45, 67));
        // it is also safe to call on an insufficient buffer:
        char smallbuf[8] = {};
        CHECK(size == ryml::catsep(smallbuf, 888, "a=0", "b=1", "c=2", 45, 67));
        CHECK(size == strlen("a=0888b=1888c=28884588867"));
        // ... and the result is truncated:
        CHECK(ryml::substr(smallbuf, sizeof(smallbuf)) == "a=0888b\0");
 
        // catsep_sub() directly returns the written substr:
        ryml::csubstr result = ryml::catsep_sub(buf, " and ", "a=0", "b=1", "c=2", 45, 67);
        CHECK(result == "a=0 and b=1 and c=2 and 45 and 67");
        CHECK(result.is_sub(buf));
 
        // catseprs() means CATSEP & ReSize:
        //
        // Instead of a substr, it receives any owning linear char container
        // for which to_substr() is defined (using ADL).
        // <ryml_std.hpp> has to_substr() definitions for std::string and
        // std::vector<char>.
        //
        // catseprs() starts by calling catsep(), and if needed, resizes the container
        // and calls catsep() again.
        //
        // Note that unless the container is previously sized, this
        // may cause an allocation, which will make your code slower.
        // Make sure to call .reserve() on the container for real
        // production code.
        std::string sbuf;
        ryml::catseprs(&sbuf, " and ", "a=0", "b=1", "c=2", 45, 67);
        CHECK(sbuf == "a=0 and b=1 and c=2 and 45 and 67");
        std::vector<char> vbuf; // works with any linear char container
        ryml::catseprs(&vbuf, " and ", "a=0", "b=1", "c=2", 45, 67);
        CHECK(ryml::to_csubstr(vbuf) == "a=0 and b=1 and c=2 and 45 and 67");
 
        // with catseprs() it is also possible to append:
        ryml::catseprs_append(&sbuf, " well ", " --- a=0", "b=11", "c=12", 145, 167);
        CHECK(sbuf == "a=0 and b=1 and c=2 and 45 and 67 --- a=0 well b=11 well c=12 well 145 well 167");
    }
 
    // uncatsep(): read arguments with a separator - opposite of catsep()
    {
        char buf_[256] = {};
 
        int a = 0, b = 1, c = 2;
        ryml::csubstr result = ryml::catsep_sub(buf_, ' ', a, b, c);
        CHECK(result == "0 1 2");
        int aa = -1, bb = -2, cc = -3;
        size_t num_characters = ryml::uncatsep(result, " ", aa, bb, cc);
        CHECK(num_characters == result.size());
        CHECK(aa == a);
        CHECK(bb == b);
        CHECK(cc == c);
 
        result = ryml::catsep_sub(buf_, "--", 10, 20, 30);
        CHECK(result == "10--20--30");
        num_characters = ryml::uncatsep(result, "--", aa, bb, cc);
        CHECK(num_characters == result.size());
        CHECK(aa == 10);
        CHECK(bb == 20);
        CHECK(cc == 30);
    }
 
    // formatting individual arguments
    {
        using namespace ryml;  // all the symbols below are in the ryml namespace.
        char buf_[256] = {};   // all the results below are written in this buffer
        substr buf = buf_;
        // --------------------------------------
        // fmt::boolalpha(): format as true/false
        // --------------------------------------
        // just as with std streams, printing a bool will output the integer value:
        CHECK("0" == cat_sub(buf, false));
        CHECK("1" == cat_sub(buf, true));
        // to force a "true"/"false", use fmt::boolalpha:
        CHECK("false" == cat_sub(buf, fmt::boolalpha(false)));
        CHECK("true" == cat_sub(buf, fmt::boolalpha(true)));
 
        // ---------------------------------
        // fmt::hex(): format as hexadecimal
        // ---------------------------------
        CHECK("0xff"  == cat_sub(buf, fmt::hex(255)));
        CHECK("0x100" == cat_sub(buf, fmt::hex(256)));
        CHECK("-0xff"  == cat_sub(buf, fmt::hex(-255)));
        CHECK("-0x100" == cat_sub(buf, fmt::hex(-256)));
        CHECK("3735928559" == cat_sub(buf,          UINT32_C(0xdeadbeef)));
        CHECK("0xdeadbeef" == cat_sub(buf, fmt::hex(UINT32_C(0xdeadbeef))));
        // ----------------------------
        // fmt::bin(): format as binary
        // ----------------------------
        CHECK("0b1000" == cat_sub(buf, fmt::bin(8)));
        CHECK("0b1001" == cat_sub(buf, fmt::bin(9)));
        CHECK("0b10001" == cat_sub(buf, fmt::bin(17)));
        CHECK("0b11001" == cat_sub(buf, fmt::bin(25)));
        CHECK("-0b1000" == cat_sub(buf, fmt::bin(-8)));
        CHECK("-0b1001" == cat_sub(buf, fmt::bin(-9)));
        CHECK("-0b10001" == cat_sub(buf, fmt::bin(-17)));
        CHECK("-0b11001" == cat_sub(buf, fmt::bin(-25)));
        // ---------------------------
        // fmt::bin(): format as octal
        // ---------------------------
        CHECK("0o77"    == cat_sub(buf, fmt::oct(63)));
        CHECK("0o100"   == cat_sub(buf, fmt::oct(64)));
        CHECK("0o377"   == cat_sub(buf, fmt::oct(255)));
        CHECK("0o400"   == cat_sub(buf, fmt::oct(256)));
        CHECK("0o1000"  == cat_sub(buf, fmt::oct(512)));
        CHECK("-0o77"   == cat_sub(buf, fmt::oct(-63)));
        CHECK("-0o100"  == cat_sub(buf, fmt::oct(-64)));
        CHECK("-0o377"  == cat_sub(buf, fmt::oct(-255)));
        CHECK("-0o400"  == cat_sub(buf, fmt::oct(-256)));
        CHECK("-0o1000" == cat_sub(buf, fmt::oct(-512)));
        // ---------------------------
        // fmt::zpad(): pad with zeros
        // ---------------------------
        CHECK("000063" == cat_sub(buf, fmt::zpad(63, 6)));
        CHECK( "00063" == cat_sub(buf, fmt::zpad(63, 5)));
        CHECK(  "0063" == cat_sub(buf, fmt::zpad(63, 4)));
        CHECK(   "063" == cat_sub(buf, fmt::zpad(63, 3)));
        CHECK(    "63" == cat_sub(buf, fmt::zpad(63, 2)));
        CHECK(    "63" == cat_sub(buf, fmt::zpad(63, 1))); // will never trim the result
        CHECK(    "63" == cat_sub(buf, fmt::zpad(63, 0))); // will never trim the result
        CHECK("0x00003f" == cat_sub(buf, fmt::zpad(fmt::hex(63), 6)));
        CHECK("0o000077" == cat_sub(buf, fmt::zpad(fmt::oct(63), 6)));
        CHECK("0b00011001" == cat_sub(buf, fmt::zpad(fmt::bin(25), 8)));
        // ------------------------------------------------
        // fmt::left(): align left with a given field width
        // ------------------------------------------------
        CHECK("63    " == cat_sub(buf, fmt::left(63, 6)));
        CHECK("63   "  == cat_sub(buf, fmt::left(63, 5)));
        CHECK("63  "   == cat_sub(buf, fmt::left(63, 4)));
        CHECK("63 "    == cat_sub(buf, fmt::left(63, 3)));
        CHECK("63"     == cat_sub(buf, fmt::left(63, 2)));
        CHECK("63"     == cat_sub(buf, fmt::left(63, 1))); // will never trim the result
        CHECK("63"     == cat_sub(buf, fmt::left(63, 0))); // will never trim the result
        // the fill character can be specified (defaults to ' '):
        CHECK("63----" == cat_sub(buf, fmt::left(63, 6, '-')));
        CHECK("63++++" == cat_sub(buf, fmt::left(63, 6, '+')));
        CHECK("63////" == cat_sub(buf, fmt::left(63, 6, '/')));
        CHECK("630000" == cat_sub(buf, fmt::left(63, 6, '0')));
        CHECK("63@@@@" == cat_sub(buf, fmt::left(63, 6, '@')));
        CHECK("0x003f    " == cat_sub(buf, fmt::left(fmt::zpad(fmt::hex(63), 4), 10)));
        // --------------------------------------------------
        // fmt::right(): align right with a given field width
        // --------------------------------------------------
        CHECK("    63" == cat_sub(buf, fmt::right(63, 6)));
        CHECK("   63"  == cat_sub(buf, fmt::right(63, 5)));
        CHECK("  63"   == cat_sub(buf, fmt::right(63, 4)));
        CHECK(" 63"    == cat_sub(buf, fmt::right(63, 3)));
        CHECK("63"     == cat_sub(buf, fmt::right(63, 2)));
        CHECK("63"     == cat_sub(buf, fmt::right(63, 1))); // will never trim the result
        CHECK("63"     == cat_sub(buf, fmt::right(63, 0))); // will never trim the result
        // the fill character can be specified (defaults to ' '):
        CHECK("----63" == cat_sub(buf, fmt::right(63, 6, '-')));
        CHECK("++++63" == cat_sub(buf, fmt::right(63, 6, '+')));
        CHECK("////63" == cat_sub(buf, fmt::right(63, 6, '/')));
        CHECK("000063" == cat_sub(buf, fmt::right(63, 6, '0')));
        CHECK("@@@@63" == cat_sub(buf, fmt::right(63, 6, '@')));
        CHECK("    0x003f" == cat_sub(buf, fmt::right(fmt::zpad(fmt::hex(63), 4), 10)));
 
        // ------------------------------------------
        // fmt::real(): format floating point numbers
        // ------------------------------------------
        // see also sample_float_precision()
        CHECK("0"       == cat_sub(buf, fmt::real(0.01f, 0)));
        CHECK("0.0"     == cat_sub(buf, fmt::real(0.01f, 1)));
        CHECK("0.01"    == cat_sub(buf, fmt::real(0.01f, 2)));
        CHECK("0.010"   == cat_sub(buf, fmt::real(0.01f, 3)));
        CHECK("0.0100"  == cat_sub(buf, fmt::real(0.01f, 4)));
        CHECK("0.01000" == cat_sub(buf, fmt::real(0.01f, 5)));
        CHECK("1"       == cat_sub(buf, fmt::real(1.01f, 0)));
        CHECK("1.0"     == cat_sub(buf, fmt::real(1.01f, 1)));
        CHECK("1.01"    == cat_sub(buf, fmt::real(1.01f, 2)));
        CHECK("1.010"   == cat_sub(buf, fmt::real(1.01f, 3)));
        CHECK("1.0100"  == cat_sub(buf, fmt::real(1.01f, 4)));
        CHECK("1.01000" == cat_sub(buf, fmt::real(1.01f, 5)));
        CHECK("1"       == cat_sub(buf, fmt::real(1.234234234, 0)));
        CHECK("1.2"     == cat_sub(buf, fmt::real(1.234234234, 1)));
        CHECK("1.23"    == cat_sub(buf, fmt::real(1.234234234, 2)));
        CHECK("1.234"   == cat_sub(buf, fmt::real(1.234234234, 3)));
        CHECK("1.2342"  == cat_sub(buf, fmt::real(1.234234234, 4)));
        CHECK("1.23423" == cat_sub(buf, fmt::real(1.234234234, 5)));
        CHECK("1000000.00000" == cat_sub(buf, fmt::real(1000000.000000000, 5)));
        CHECK("1234234.23423" == cat_sub(buf, fmt::real(1234234.234234234, 5)));
        // AKA %f
        CHECK("1000000.00000" == cat_sub(buf, fmt::real(1000000.000000000, 5, FTOA_FLOAT)));  // AKA %f, same as above
        CHECK("1234234.23423" == cat_sub(buf, fmt::real(1234234.234234234, 5, FTOA_FLOAT)));  // AKA %f
        CHECK("1234234.2342"  == cat_sub(buf, fmt::real(1234234.234234234, 4, FTOA_FLOAT)));  // AKA %f
        CHECK("1234234.234"   == cat_sub(buf, fmt::real(1234234.234234234, 3, FTOA_FLOAT)));  // AKA %f
        CHECK("1234234.23"    == cat_sub(buf, fmt::real(1234234.234234234, 2, FTOA_FLOAT)));  // AKA %f
        // AKA %e
        CHECK("1.00000e+06"   == cat_sub(buf, fmt::real(1000000.000000000, 5, FTOA_SCIENT))); // AKA %e
        CHECK("1.23423e+06"   == cat_sub(buf, fmt::real(1234234.234234234, 5, FTOA_SCIENT))); // AKA %e
        CHECK("1.2342e+06"    == cat_sub(buf, fmt::real(1234234.234234234, 4, FTOA_SCIENT))); // AKA %e
        CHECK("1.234e+06"     == cat_sub(buf, fmt::real(1234234.234234234, 3, FTOA_SCIENT))); // AKA %e
        CHECK("1.23e+06"      == cat_sub(buf, fmt::real(1234234.234234234, 2, FTOA_SCIENT))); // AKA %e
        // AKA %g
        CHECK("1e+06"         == cat_sub(buf, fmt::real(1000000.000000000, 5, FTOA_FLEX)));   // AKA %g
        CHECK("1.2342e+06"    == cat_sub(buf, fmt::real(1234234.234234234, 5, FTOA_FLEX)));   // AKA %g
        CHECK("1.234e+06"     == cat_sub(buf, fmt::real(1234234.234234234, 4, FTOA_FLEX)));   // AKA %g
        CHECK("1.23e+06"      == cat_sub(buf, fmt::real(1234234.234234234, 3, FTOA_FLEX)));   // AKA %g
        CHECK("1.2e+06"       == cat_sub(buf, fmt::real(1234234.234234234, 2, FTOA_FLEX)));   // AKA %g
        // FTOA_HEXA: AKA %a (hexadecimal formatting of floats)
        CHECK("0x1.e8480p+19" == cat_sub(buf, fmt::real(1000000.000000000, 5, FTOA_HEXA)));   // AKA %a
        CHECK("0x1.2d53ap+20" == cat_sub(buf, fmt::real(1234234.234234234, 5, FTOA_HEXA)));   // AKA %a
 
        // --------------------------------------------------------------
        // fmt::raw(): dump data in raw (binary) machine format (respecting alignment)
        // --------------------------------------------------------------
        {
            C4_SUPPRESS_WARNING_GCC_CLANG_WITH_PUSH("-Wcast-align")  // we're casting the values directly, so alignment is strictly respected.
            const uint32_t payload[] = {10, 20, 30, 40, UINT32_C(0xdeadbeef)};
            // (package payload as a substr, for comparison only)
            csubstr expected = csubstr((const char *)payload, sizeof(payload));
            csubstr actual = cat_sub(buf, fmt::raw(payload));
            CHECK(!actual.overlaps(expected));
            CHECK(0 == memcmp(expected.str, actual.str, expected.len));
            // also possible with variables:
            for(const uint32_t value : payload)
            {
                // (package payload as a substr, for comparison only)
                expected = csubstr((const char *)&value, sizeof(value));
                actual = cat_sub(buf, fmt::raw(value));
                CHECK(actual.size() == sizeof(uint32_t));
                CHECK(!actual.overlaps(expected));
                CHECK(0 == memcmp(expected.str, actual.str, expected.len));
                // with non-const data, fmt::craw() may be needed for disambiguation:
                actual = cat_sub(buf, fmt::craw(value));
                CHECK(actual.size() == sizeof(uint32_t));
                CHECK(!actual.overlaps(expected));
                CHECK(0 == memcmp(expected.str, actual.str, expected.len));
                //
                // read back:
                uint32_t result = 0;
                auto reader = fmt::raw(result);
                CHECK(uncat(actual, reader));
                // and compare:
                // (vs2017/release/32bit does not reload result from cache, so force it)
                CHECK(result == value); // roundtrip completed successfully
            }
            C4_SUPPRESS_WARNING_GCC_CLANG_POP
        }
 
        // -------------------------
        // fmt::base64(): see below!
        // -------------------------
    }
}

sample_base64()

void sample_base64

(

)

demonstrates how to read and write base64-encoded blobs.

See also

Base64 encoding/decoding

Definition at line 3246 of file quickstart.cpp.

{
    // let's start by creating a tree with base64 vals and keys
    ryml::Tree tree;
    tree.rootref() |= ryml::MAP;
    struct text_and_base64 { ryml::csubstr text, base64; };
    text_and_base64 cases[] = {
        {{"Hello, World!"}, {"SGVsbG8sIFdvcmxkIQ=="}},
        {{"Brevity is the soul of wit."}, {"QnJldml0eSBpcyB0aGUgc291bCBvZiB3aXQu"}},
        {{"All that glitters is not gold."}, {"QWxsIHRoYXQgZ2xpdHRlcnMgaXMgbm90IGdvbGQu"}},
    };
    // to encode base64 and write the result to val:
    for(text_and_base64 c : cases)
        tree[c.text] << ryml::fmt::base64(c.text);
    // to encode base64 and write the result to key:
    for(text_and_base64 c : cases)
        tree.rootref().append_child() << ryml::key(ryml::fmt::base64(c.text)) << c.text;
    // check the result:
    for(text_and_base64 c : cases)
    {
        CHECK(tree[c.text].val() == c.base64);
        CHECK(tree[c.base64].val() == c.text);
    }
    // and this is how the YAML now looks:
    CHECK(ryml::emitrs_yaml<std::string>(tree) == ""
          "Hello, World!: SGVsbG8sIFdvcmxkIQ=="                                         "\n"
          "Brevity is the soul of wit.: QnJldml0eSBpcyB0aGUgc291bCBvZiB3aXQu"           "\n"
          "All that glitters is not gold.: QWxsIHRoYXQgZ2xpdHRlcnMgaXMgbm90IGdvbGQu"    "\n"
          // note that the keys below are base64-encoded
          "SGVsbG8sIFdvcmxkIQ==: Hello, World!"                                         "\n"
          "QnJldml0eSBpcyB0aGUgc291bCBvZiB3aXQu: Brevity is the soul of wit."           "\n"
          "QWxsIHRoYXQgZ2xpdHRlcnMgaXMgbm90IGdvbGQu: All that glitters is not gold."    "\n"
          "");
    char buf1_[128], buf2_[128];
    ryml::substr buf1 = buf1_;  // this is where we will write the result (using >>)
    ryml::substr buf2 = buf2_;  // this is where we will write the result (using deserialize()/deserialize_key())
    // to decode base64 and write the result to buf:
    for(const text_and_base64 c : cases)
    {
        // this decodes base64 and write the decoded result into an
        // existing buffer (buf1):
        size_t len = 0; // the decoded length
        tree[c.text] >> ryml::fmt::base64(buf1, &len);
        // The base64() tag function is used to get the
        // deserialization using c4::decode_base64(). This will
        // respect the limits of the buffer, and fail with an error if
        // the buffer is too small (or if the base64 encoding is
        // wrong). The optional second parameter is set to the decoded
        // size, ie, the length of the decoded result, which is also
        // the size required for the buffer.
        CHECK(len <= buf1.len);
        CHECK(c.text.len == len);
        CHECK(buf1.first(len) == c.text);
        // likewise for keys:
        tree[c.base64] >> ryml::key(ryml::fmt::base64(buf2, &len));
        CHECK(len <= buf2.len);
        CHECK(buf2.first(len) == c.text);
        //
        // interop with std::string:
        std::string result;
        tree[c.text] >> ryml::fmt::base64(result);
        CHECK(result == c.text);
        // likewise for keys:
        tree[c.base64] >> ryml::key(ryml::fmt::base64(result));
        CHECK(result == c.text);
    }
    // directly encode variables: integers
    {
        const uint64_t valin = UINT64_C(0xdeadbeef);
        tree["deadbeef"] << c4::fmt::base64(valin); // sometimes cbase64() is needed to avoid ambiguity
        uint64_t valout = 0;
        size_t len = 0;
        tree["deadbeef"] >> ryml::fmt::base64(valout, &len);
        CHECK(len == sizeof(valout));
        CHECK(valout == UINT64_C(0xdeadbeef)); // base64 roundtrip is bit-accurate
        // also works without length parameter:
        valout = {};
        tree["deadbeef"] >> ryml::fmt::base64(valout);
        CHECK(valout == UINT64_C(0xdeadbeef)); // base64 roundtrip is bit-accurate
    }
    // directly encode variables: floating point
    {
        const double valin = 123456.7891011;
        tree["float"] << c4::fmt::base64(valin);
        double valout = 0;
        size_t len = 0;
        tree["float"] >> ryml::fmt::base64(valout, &len);
        CHECK(len == sizeof(valout));
        CHECK(memcmp(&valout, &valin, sizeof(valout)) == 0); // base64 roundtrip is bit-accurate // NOLINT
        // also works without length parameter:
        valout = {};
        tree["float"] >> ryml::fmt::base64(valout);
        CHECK(memcmp(&valout, &valin, sizeof(valout)) == 0); // base64 roundtrip is bit-accurate // NOLINT
    }
    // directly encode memory ranges
    {
        const uint32_t data_in[11] = {0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 0xdeadbeef};
        uint32_t data_out[11] = {};
        tree["int_data"] << ryml::fmt::base64(data_in, 11);
        CHECK(memcmp(data_in, data_out, sizeof(data_in)) != 0); // before the roundtrip
        size_t len = 0;
        tree["int_data"] >> ryml::fmt::base64(data_out, 11, &len);
        CHECK(len == sizeof(data_out));
        CHECK(memcmp(data_in, data_out, sizeof(data_in)) == 0); // after the roundtrip
        // also works without length parameter:
        memset(data_out, 0, sizeof(data_out));
        tree["int_data"] >> ryml::fmt::base64(data_out);
        CHECK(memcmp(data_in, data_out, sizeof(data_in)) == 0); // after the roundtrip
    }
}

sample_user_scalar_types()

void sample_user_scalar_types

(

)

to add scalar types (ie leaf types converting to/from string), define the functions above for those types.

See Serialize/deserialize scalar types.

Definition at line 3577 of file quickstart.cpp.

{
    ryml::Tree t;
 
    auto r = t.rootref();
    r |= ryml::MAP;
 
    vec2<int> v2in{10, 11};
    vec2<int> v2out{1, 2};
    r["v2"] << v2in; // serializes to the tree's arena, and then sets the keyval
    r["v2"] >> v2out;
    CHECK(v2in.x == v2out.x);
    CHECK(v2in.y == v2out.y);
    vec3<int> v3in{100, 101, 102};
    vec3<int> v3out{1, 2, 3};
    r["v3"] << v3in; // serializes to the tree's arena, and then sets the keyval
    r["v3"] >> v3out;
    CHECK(v3in.x == v3out.x);
    CHECK(v3in.y == v3out.y);
    CHECK(v3in.z == v3out.z);
    vec4<int> v4in{1000, 1001, 1002, 1003};
    vec4<int> v4out{1, 2, 3, 4};
    r["v4"] << v4in; // serializes to the tree's arena, and then sets the keyval
    r["v4"] >> v4out;
    CHECK(v4in.x == v4out.x);
    CHECK(v4in.y == v4out.y);
    CHECK(v4in.z == v4out.z);
    CHECK(v4in.w == v4out.w);
    CHECK(ryml::emitrs_yaml<std::string>(t) == ""
          "v2: (10,11)"                 "\n"
          "v3: (100,101,102)"           "\n"
          "v4: (1000,1001,1002,1003)"   "\n"
          "");
 
    // note that only the used functions are needed:
    //   - if a type is only parsed, then only from_chars() is needed
    //   - if a type is only emitted, then only to_chars() is needed
    emit_only_vec2<int> eov2in{20, 21}; // only has to_chars()
    parse_only_vec2<int> pov2out{1, 2}; // only has from_chars()
    r["v2"] << eov2in; // serializes to the tree's arena, and then sets the keyval
    r["v2"] >> pov2out;
    CHECK(eov2in.x == pov2out.x);
    CHECK(eov2in.y == pov2out.y);
    emit_only_vec3<int> eov3in{30, 31, 32}; // only has to_chars()
    parse_only_vec3<int> pov3out{1, 2, 3}; // only has from_chars()
    r["v3"] << eov3in; // serializes to the tree's arena, and then sets the keyval
    r["v3"] >> pov3out;
    CHECK(eov3in.x == pov3out.x);
    CHECK(eov3in.y == pov3out.y);
    CHECK(eov3in.z == pov3out.z);
    emit_only_vec4<int> eov4in{40, 41, 42, 43}; // only has to_chars()
    parse_only_vec4<int> pov4out{1, 2, 3, 4}; // only has from_chars()
    r["v4"] << eov4in; // serializes to the tree's arena, and then sets the keyval
    r["v4"] >> pov4out;
    CHECK(eov4in.x == pov4out.x);
    CHECK(eov4in.y == pov4out.y);
    CHECK(eov4in.z == pov4out.z);
    CHECK(ryml::emitrs_yaml<std::string>(t) == ""
          "v2: (20,21)"        "\n"
          "v3: (30,31,32)"     "\n"
          "v4: (40,41,42,43)"  "\n"
          "");
}

sample_user_container_types()

void sample_user_container_types

(

)

shows how to serialize/deserialize container types.

See also

Serialize/deserialize container types

sample_std_types

Definition at line 3775 of file quickstart.cpp.

{
    my_type mt_in{
        {20, 21},
        {30, 31, 32},
        {40, 41, 42, 43},
        {{101, 102, 103, 104, 105, 106, 107}},
        {{{1001, 2001}, {1002, 2002}, {1003, 2003}}},
    };
    my_type mt_out;
 
    ryml::Tree t;
    t.rootref() << mt_in;  // read from this
    t.crootref() >> mt_out; // assign here
    CHECK(mt_out.v2.x == mt_in.v2.x);
    CHECK(mt_out.v2.y == mt_in.v2.y);
    CHECK(mt_out.v3.x == mt_in.v3.x);
    CHECK(mt_out.v3.y == mt_in.v3.y);
    CHECK(mt_out.v3.z == mt_in.v3.z);
    CHECK(mt_out.v4.x == mt_in.v4.x);
    CHECK(mt_out.v4.y == mt_in.v4.y);
    CHECK(mt_out.v4.z == mt_in.v4.z);
    CHECK(mt_out.v4.w == mt_in.v4.w);
    CHECK(mt_in.seq.seq_member.size() > 0);
    CHECK(mt_out.seq.seq_member.size() == mt_in.seq.seq_member.size());
    for(size_t i = 0; i < mt_in.seq.seq_member.size(); ++i)
    {
        CHECK(mt_out.seq.seq_member[i] == mt_in.seq.seq_member[i]);
    }
    CHECK(mt_in.map.map_member.size() > 0);
    CHECK(mt_out.map.map_member.size() == mt_in.map.map_member.size());
    for(auto const& kv : mt_in.map.map_member)
    {
        CHECK(mt_out.map.map_member.find(kv.first) != mt_out.map.map_member.end());
        CHECK(mt_out.map.map_member[kv.first] == kv.second);
    }
    CHECK(ryml::emitrs_yaml<std::string>(t) == ""
          "v2: (20,21)"         "\n"
          "v3: (30,31,32)"      "\n"
          "v4: (40,41,42,43)"   "\n"
          "seq:"                "\n"
          "  - 101"             "\n"
          "  - 102"             "\n"
          "  - 103"             "\n"
          "  - 104"             "\n"
          "  - 105"             "\n"
          "  - 106"             "\n"
          "  - 107"             "\n"
          "map:"                "\n"
          "  1001: 2001"        "\n"
          "  1002: 2002"        "\n"
          "  1003: 2003"        "\n"
          "");
}

sample_std_types()

void sample_std_types

(

)

demonstrates usage with the std implementations provided by ryml in the ryml_std.hpp header

See also

Serialize/deserialize container types

also the STL section in Serialization/deserialization

Definition at line 3837 of file quickstart.cpp.

{
    std::string yml_std_string = ""
        "- v2: (20,21)"              "\n"
        "  v3: (30,31,32)"           "\n"
        "  v4: (40,41,42,43)"        "\n"
        "  seq:"                     "\n"
        "    - 101"                  "\n"
        "    - 102"                  "\n"
        "    - 103"                  "\n"
        "    - 104"                  "\n"
        "    - 105"                  "\n"
        "    - 106"                  "\n"
        "    - 107"                  "\n"
        "  map:"                     "\n"
        "    1001: 2001"             "\n"
        "    1002: 2002"             "\n"
        "    1003: 2003"             "\n"
        "- v2: (120,121)"            "\n"
        "  v3: (130,131,132)"        "\n"
        "  v4: (140,141,142,143)"    "\n"
        "  seq:"                     "\n"
        "    - 1101"                 "\n"
        "    - 1102"                 "\n"
        "    - 1103"                 "\n"
        "    - 1104"                 "\n"
        "    - 1105"                 "\n"
        "    - 1106"                 "\n"
        "    - 1107"                 "\n"
        "  map:"                     "\n"
        "    11001: 12001"           "\n"
        "    11002: 12002"           "\n"
        "    11003: 12003"           "\n"
        "- v2: (220,221)"            "\n"
        "  v3: (230,231,232)"        "\n"
        "  v4: (240,241,242,243)"    "\n"
        "  seq:"                     "\n"
        "    - 2101"                 "\n"
        "    - 2102"                 "\n"
        "    - 2103"                 "\n"
        "    - 2104"                 "\n"
        "    - 2105"                 "\n"
        "    - 2106"                 "\n"
        "    - 2107"                 "\n"
        "  map:"                     "\n"
        "    21001: 22001"           "\n"
        "    21002: 22002"           "\n"
        "    21003: 22003"           "\n"
        "";
    // parse in-place using the std::string above
    ryml::Tree tree = ryml::parse_in_place(ryml::to_substr(yml_std_string));
    // my_type is a container-of-containers type. see above its
    // definition implementation for ryml.
    std::vector<my_type> vmt;
    tree.rootref() >> vmt;
    CHECK(vmt.size() == 3);
    ryml::Tree tree_out;
    tree_out.rootref() << vmt;
    CHECK(ryml::emitrs_yaml<std::string>(tree_out) == yml_std_string);
}

sample_float_precision()

void sample_float_precision

(

)

control precision of serialized floats

Definition at line 3902 of file quickstart.cpp.

{
    std::vector<double> reference{1.23234412342131234, 2.12323123143434237, 3.67847983572591234};
    // A safe precision for comparing doubles. May vary depending on
    // compiler flags. Double goes to about 15 digits, so 14 should be
    // safe enough for this test to succeed.
    const double precision_safe = 1.e-14;
    const size_t num_digits_safe = 14;
    const size_t num_digits_original = 17;
    auto get_num_digits = [](ryml::csubstr number){ return number.len - 2u; };
    //
    // no significant precision is lost when reading
    // floating point numbers:
    {
        ryml::Tree tree = ryml::parse_in_arena("[1.23234412342131234, 2.12323123143434237, 3.67847983572591234]");
        std::vector<double> output;
        tree.rootref() >> output;
        CHECK(output.size() == reference.size());
        for(size_t i = 0; i < reference.size(); ++i)
        {
            CHECK(get_num_digits(tree[(ryml::id_type)i].val()) == num_digits_original);
            CHECK(fabs(output[i] - reference[i]) < precision_safe);
        }
    }
    //
    // However, depending on the compilation settings, there may be a
    // significant precision loss when serializing with the default
    // approach, operator<<(double):
    {
        ryml::Tree serialized;
        serialized.rootref() << reference;
        std::cout << serialized;
        // Without std::to_chars() there is a loss of precision:
        #if (!C4CORE_HAVE_STD_TOCHARS) // check if std::to_chars() is available.
        CHECK((ryml::emitrs_yaml<std::string>(serialized) == ""
              "- 1.23234"   "\n"
              "- 2.12323"   "\n"
              "- 3.67848"   "\n"
              "") || (bool)"this is indicative; the exact results will vary from platform to platform.");
        C4_UNUSED(num_digits_safe);
        // ... but when using C++17 and above, the results are eminently equal:
        #else
        CHECK((ryml::emitrs_yaml<std::string>(serialized) == ""
               "- 1.2323441234213124"   "\n"
               "- 2.1232312314343424"   "\n"
               "- 3.6784798357259123"   "\n"
               "") || (bool)"this is indicative; the exact results will vary from platform to platform.");
        size_t pos = 0;
        for(ryml::ConstNodeRef child : serialized.rootref().children())
        {
            CHECK(get_num_digits(child.val()) >= num_digits_safe);
            double out = {};
            child >> out;
            CHECK(fabs(out - reference[pos++]) < precision_safe);
        }
        #endif
    }
    //
    // The difference is explained by the availability of
    // fastfloat::from_chars(), std::from_chars() and std::to_chars().
    //
    // ryml prefers the fastfloat::from_chars() version. Unfortunately
    // fastfloat does not have to_chars() (see
    // https://github.com/fastfloat/fast_float/issues/23).
    //
    // When C++17 is used, ryml uses std::to_chars(), which produces
    // good defaults.
    //
    // However, with earlier standards, or in some library
    // implementations, there's only snprintf() available. Every other
    // std library function will either disrespect the string limits,
    // or more precisely, accept no string size limits. So the
    // implementation of c4core (which ryml uses) falls back to
    // snprintf("%g"), and that picks by default a (low) number of
    // digits.
    //
    // But all is not lost for C++11/C++14 users!
    //
    // To force a particular precision when serializing, you can use
    // c4::fmt::real() (brought into the ryml:: namespace). Or you can
    // serialize the number yourself! The small downside is that you
    // have to build the container.
    //
    // First a function to check the result:
    auto check_precision = [&](ryml::Tree const& serialized){
        std::cout << serialized;
        // now it works!
        CHECK((ryml::emitrs_yaml<std::string>(serialized) == ""
               "- 1.23234412342131239"   "\n"
               "- 2.12323123143434245"   "\n"
               "- 3.67847983572591231"   "\n"
               "") || (bool)"this is indicative; the exact results will vary from platform to platform.");
        size_t pos = 0;
        for(ryml::ConstNodeRef child : serialized.rootref().children())
        {
            CHECK(get_num_digits(child.val()) == num_digits_original);
            double out = {};
            child >> out;
            CHECK(fabs(out - reference[pos++]) < precision_safe);
        }
    };
    //
    // Serialization example using fmt::real()
    {
        ryml::Tree serialized;
        ryml::NodeRef root = serialized.rootref();
        root |= ryml::SEQ;
        for(const double v : reference)
            root.append_child() << ryml::fmt::real(v, num_digits_original, ryml::FTOA_FLOAT);
        check_precision(serialized); // OK - now within bounds!
    }
    //
    // Serialization example using snprintf
    {
        ryml::Tree serialized;
        ryml::NodeRef root = serialized.rootref();
        root |= ryml::SEQ;
        char tmp[64];
        for(const double v : reference)
        {
            // reuse a buffer to serialize.
            // add 1 to the significant digits because the %g
            // specifier counts the integral digits.
            (void)snprintf(tmp, sizeof(tmp), "%.18g", v);
            // copy the serialized string to the tree (operator<<
            // copies to the arena, operator= just assigns the string
            // pointer and would be wrong in this case):
            root.append_child() << ryml::to_csubstr((const char*)tmp);
        }
        check_precision(serialized); // OK - now within bounds!
    }
}

sample_emit_to_container()

void sample_emit_to_container

(

)

demonstrates how to emit to a linear container of char

Definition at line 4039 of file quickstart.cpp.

{
 
    ryml::csubstr ymla =
        "- 1\n"
        "- 2\n"
        "";
    ryml::csubstr ymlb =
        "- a"                           "\n"
        "- b"                           "\n"
        "- x0: 1"                       "\n"
        "  x1: 2"                       "\n"
        "- champagne: Dom Perignon"     "\n"
        "  coffee: Arabica"             "\n"
        "  more:"                       "\n"
        "    vinho verde: Soalheiro"    "\n"
        "    vinho tinto: Redoma 2017"  "\n"
        "  beer:"                       "\n"
        "    - Rochefort 10"            "\n"
        "    - Busch"                   "\n"
        "    - Leffe Rituel"            "\n"
        "- foo"                         "\n"
        "- bar"                         "\n"
        "- baz"                         "\n"
        "- bat"                         "\n"
        "";
    const ryml::Tree treea = ryml::parse_in_arena(ymla);
    const ryml::Tree treeb = ryml::parse_in_arena(ymlb);
 
    // it is possible to emit to any linear container of char.
    //
    // eg, std::vector<char>
    {
        // do a blank call on an empty buffer to find the required size.
        // no overflow will occur, and returns a substr with the size
        // required to output
        ryml::csubstr output = ryml::emit_yaml(treea, treea.root_id(), ryml::substr{}, /*error_on_excess*/false);
        CHECK(output.str == nullptr);
        CHECK(output.len > 0);
        size_t num_needed_chars = output.len;
        std::vector<char> buf(num_needed_chars);
        // now try again with the proper buffer
        output = ryml::emit_yaml(treea, treea.root_id(), ryml::to_substr(buf), /*error_on_excess*/true);
        CHECK(output == ymla);
 
        // it is possible to reuse the buffer and grow it as needed.
        // first do a blank run to find the size:
        output = ryml::emit_yaml(treeb, treeb.root_id(), ryml::substr{}, /*error_on_excess*/false);
        CHECK(output.str == nullptr);
        CHECK(output.len > 0);
        CHECK(output.len == ymlb.len);
        num_needed_chars = output.len;
        buf.resize(num_needed_chars);
        // now try again with the proper buffer
        output = ryml::emit_yaml(treeb, treeb.root_id(), ryml::to_substr(buf), /*error_on_excess*/true);
        CHECK(output == ymlb);
 
        // there is a convenience wrapper performing the same as above:
        // provided to_substr() is defined for that container.
        output = ryml::emitrs_yaml(treeb, &buf);
        CHECK(output == ymlb);
 
        // or you can just output a new container:
        // provided to_substr() is defined for that container.
        std::vector<char> another = ryml::emitrs_yaml<std::vector<char>>(treeb);
        CHECK(ryml::to_csubstr(another) == ymlb);
 
        // you can also emit nested nodes:
        another = ryml::emitrs_yaml<std::vector<char>>(treeb[3][2]);
        CHECK(ryml::to_csubstr(another) == ""
              "more:"                           "\n"
              "  vinho verde: Soalheiro"        "\n"
              "  vinho tinto: Redoma 2017"      "\n"
              "");
    }
 
 
    // eg, std::string. notice this is the same code as above
    {
        // do a blank call on an empty buffer to find the required size.
        // no overflow will occur, and returns a substr with the size
        // required to output
        ryml::csubstr output = ryml::emit_yaml(treea, treea.root_id(), ryml::substr{}, /*error_on_excess*/false);
        CHECK(output.str == nullptr);
        CHECK(output.len > 0);
        size_t num_needed_chars = output.len;
        std::string buf;
        buf.resize(num_needed_chars);
        // now try again with the proper buffer
        output = ryml::emit_yaml(treea, treea.root_id(), ryml::to_substr(buf), /*error_on_excess*/true);
        CHECK(output == ymla);
 
        // it is possible to reuse the buffer and grow it as needed.
        // first do a blank run to find the size:
        output = ryml::emit_yaml(treeb, treeb.root_id(), ryml::substr{}, /*error_on_excess*/false);
        CHECK(output.str == nullptr);
        CHECK(output.len > 0);
        CHECK(output.len == ymlb.len);
        num_needed_chars = output.len;
        buf.resize(num_needed_chars);
        // now try again with the proper buffer
        output = ryml::emit_yaml(treeb, treeb.root_id(), ryml::to_substr(buf), /*error_on_excess*/true);
        CHECK(output == ymlb);
 
        // there is a convenience wrapper performing the above instructions:
        // provided to_substr() is defined for that container
        output = ryml::emitrs_yaml(treeb, &buf);
        CHECK(output == ymlb);
 
        // or you can just output a new container:
        // provided to_substr() is defined for that container.
        std::string another = ryml::emitrs_yaml<std::string>(treeb);
        CHECK(ryml::to_csubstr(another) == ymlb);
 
        // you can also emit nested nodes:
        another = ryml::emitrs_yaml<std::string>(treeb[3][2]);
        CHECK(ryml::to_csubstr(another) == ""
              "more:"                           "\n"
              "  vinho verde: Soalheiro"        "\n"
              "  vinho tinto: Redoma 2017"      "\n"
              "");
    }
}

sample_emit_to_stream()

void sample_emit_to_stream

(

)

demonstrates how to emit to a stream-like structure

Definition at line 4167 of file quickstart.cpp.

{
    ryml::csubstr ymlb =
        "- a"                           "\n"
        "- b"                           "\n"
        "- x0: 1"                       "\n"
        "  x1: 2"                       "\n"
        "- champagne: Dom Perignon"     "\n"
        "  coffee: Arabica"             "\n"
        "  more:"                       "\n"
        "    vinho verde: Soalheiro"    "\n"
        "    vinho tinto: Redoma 2017"  "\n"
        "  beer:"                       "\n"
        "    - Rochefort 10"            "\n"
        "    - Busch"                   "\n"
        "    - Leffe Rituel"            "\n"
        "- foo"                         "\n"
        "- bar"                         "\n"
        "- baz"                         "\n"
        "- bat"                         "\n"
        "";
    const ryml::Tree tree = ryml::parse_in_arena(ymlb);
 
    std::string s;
 
    // emit a full tree
    {
        std::stringstream ss;
        ss << tree; // works with any stream having .operator<<() and .write()
        s = ss.str();
        CHECK(ryml::to_csubstr(s) == ymlb);
    }
 
    // emit a full tree as json
    {
        std::stringstream ss;
        ss << ryml::as_json(tree); // works with any stream having .operator<<() and .write()
        s = ss.str();
        CHECK(s ==
              "["                                        "\n"
              "  \"a\","                                 "\n"
              "  \"b\","                                 "\n"
              "  {"                                      "\n"
              "    \"x0\": 1,"                           "\n"
              "    \"x1\": 2"                            "\n"
              "  },"                                     "\n"
              "  {"                                      "\n"
              "    \"champagne\": \"Dom Perignon\","     "\n"
              "    \"coffee\": \"Arabica\","             "\n"
              "    \"more\": {"                          "\n"
              "      \"vinho verde\": \"Soalheiro\","    "\n"
              "      \"vinho tinto\": \"Redoma 2017\""   "\n"
              "    },"                                   "\n"
              "    \"beer\": ["                          "\n"
              "      \"Rochefort 10\","                  "\n"
              "      \"Busch\","                         "\n"
              "      \"Leffe Rituel\""                   "\n"
              "    ]"                                    "\n"
              "  },"                                     "\n"
              "  \"foo\","                               "\n"
              "  \"bar\","                               "\n"
              "  \"baz\","                               "\n"
              "  \"bat\""                                "\n"
              "]"                                        "\n"
              "");
    }
 
    // emit a nested node
    {
        std::stringstream ss;
        ss << tree[3][2]; // works with any stream having .operator<<() and .write()
        s = ss.str();
        CHECK(s == ""
              "more:"                             "\n"
              "  vinho verde: Soalheiro"          "\n"
              "  vinho tinto: Redoma 2017"        "\n"
              "");
    }
 
    // emit a nested node as json
    {
        std::stringstream ss;
        ss << ryml::as_json(tree[3][2]); // works with any stream having .operator<<() and .write()
        s = ss.str();
        CHECK(ryml::to_csubstr(s) ==
              "\"more\": {"                           "\n"
              "  \"vinho verde\": \"Soalheiro\","     "\n"
              "  \"vinho tinto\": \"Redoma 2017\""    "\n"
              "}"                                     "\n"
              "");
    }
}

sample_emit_to_file()

void sample_emit_to_file

(

)

demonstrates how to emit to a FILE*

Definition at line 4264 of file quickstart.cpp.

{
    ryml::csubstr yml = ""
        "- a"                           "\n"
        "- b"                           "\n"
        "- x0: 1"                       "\n"
        "  x1: 2"                       "\n"
        "- champagne: Dom Perignon"     "\n"
        "  coffee: Arabica"             "\n"
        "  more:"                       "\n"
        "    vinho verde: Soalheiro"    "\n"
        "    vinho tinto: Redoma 2017"  "\n"
        "  beer:"                       "\n"
        "    - Rochefort 10"            "\n"
        "    - Busch"                   "\n"
        "    - Leffe Rituel"            "\n"
        "- foo"                         "\n"
        "- bar"                         "\n"
        "- baz"                         "\n"
        "- bat"                         "\n"
        "";
    const ryml::Tree tree = ryml::parse_in_arena(yml);
    // this is emitting to stdout, but of course you can pass in any
    // FILE* obtained from fopen()
    size_t len = ryml::emit_yaml(tree, tree.root_id(), stdout);
    // the return value is the number of characters that were written
    // to the file
    CHECK(len == yml.len);
}

sample_emit_nested_node()

void sample_emit_nested_node

(

)

just like parsing into a nested node, you can also emit from a nested node.

Definition at line 4298 of file quickstart.cpp.

{
    const ryml::Tree tree = ryml::parse_in_arena(""
        "- a"                            "\n"
        "- b"                            "\n"
        "- x0: 1"                        "\n"
        "  x1: 2"                        "\n"
        "- champagne: Dom Perignon"      "\n"
        "  coffee: Arabica"              "\n"
        "  more:"                        "\n"
        "    vinho verde: Soalheiro"     "\n"
        "    vinho tinto: Redoma 2017"   "\n"
        "  beer:"                        "\n"
        "    - Rochefort 10"             "\n"
        "    - Busch"                    "\n"
        "    - Leffe Rituel"             "\n"
        "    - - and so"                 "\n"
        "      - many other"             "\n"
        "      - wonderful beers"        "\n"
        "- more"                         "\n"
        "- seq"                          "\n"
        "- members"                      "\n"
        "- here"                         "\n"
        "");
    // Let's now emit the beer node. Note that its key is also
    // emitted, making the result a map with a single child:
    CHECK(ryml::emitrs_yaml<std::string>(tree[3]["beer"]) == ""
          "beer:"                    "\n"
          "  - Rochefort 10"         "\n"
          "  - Busch"                "\n"
          "  - Leffe Rituel"         "\n"
          "  - - and so"             "\n"
          "    - many other"         "\n"
          "    - wonderful beers"    "\n"
          "");
    // You can use EmitOptions to prevent the key from being
    // emitted. Note how the result is now just the contents without
    // the key, and therefore the root is now a seq:
    auto without_key = ryml::EmitOptions{}.emit_nonroot_key(false);
    CHECK(ryml::emitrs_yaml<std::string>(tree[3]["beer"], without_key) == ""
          "- Rochefort 10"         "\n"
          "- Busch"                "\n"
          "- Leffe Rituel"         "\n"
          "- - and so"             "\n"
          "  - many other"         "\n"
          "  - wonderful beers"    "\n"
          "");
    // For sequences, the behavior is opposite. Notice how the leading
    // dash defaults to omit:
    CHECK(ryml::emitrs_yaml<std::string>(tree[3]["beer"][3]) == ""
          "- and so"                 "\n"
          "- many other"             "\n"
          "- wonderful beers"        "\n"
          "");
    // And likewise, you can use EmitOptions to force the dash:
    auto with_dash = ryml::EmitOptions{}.emit_nonroot_dash(true);
    CHECK(ryml::emitrs_yaml<std::string>(tree[3]["beer"][3], with_dash) == ""
          "- - and so"                 "\n" // note the added dash and indentation
          "  - many other"             "\n"
          "  - wonderful beers"        "\n"
          "");
    // Example: emit a scalar node (seq member):
    CHECK(ryml::emitrs_yaml<std::string>(tree[3]["beer"][0]) == "Rochefort 10");
    CHECK(ryml::emitrs_yaml<std::string>(tree[3]["beer"][0], with_dash) == "- Rochefort 10\n");
    // Example: emit a scalar node (map member):
    CHECK(ryml::emitrs_yaml<std::string>(tree[3]["more"][0]) == "vinho verde: Soalheiro\n");
    CHECK(ryml::emitrs_yaml<std::string>(tree[3]["more"][0], without_key) == "Soalheiro");
}

sample_style()

void sample_style

(

)

query/set/modify node style to control formatting of emitted YAML code.

See also

See more details about formatting flow containers in sample_style_flow_formatting() (below).

see also:

• ryml::scalar_style_choose_block()
• ryml::scalar_style_choose_flow()
• ryml::scalar_style_choose_json()
• ryml::scalar_style_query_squo()
• ryml::scalar_style_query_plain()

Definition at line 4372 of file quickstart.cpp.

{
    // we will be using these helpers throughout this function
    auto tostr = [](ryml::ConstNodeRef n) {
        return ryml::emitrs_yaml<std::string>(n);
    };
    auto tostr_opts = [](ryml::ConstNodeRef n, ryml::EmitOptions opts) {
        return ryml::emitrs_yaml<std::string>(n, opts);
    };
    // let's parse this yaml:
    ryml::csubstr yaml = ""
        "block map:"                                   "\n"
        "  block key: block val"                       "\n"
        "block seq:"                                   "\n"
        "  - block val 1"                              "\n"
        "  - block val 2"                              "\n"
        "  - 'quoted'"                                 "\n"
        "flow map, singleline: {flow key: flow val}"   "\n"
        "flow seq, singleline: [flow val,flow val]"    "\n"
        "flow map, multiline: {"                       "\n"
        "    flow key: flow val"                       "\n"
        "  }"                                          "\n"
        "flow seq, multiline: ["                       "\n"
        "    flow val,"                                "\n"
        "    flow val"                                 "\n"
        "  ]"                                          "\n"
        "";
    ryml::Tree tree = ryml::parse_in_arena(yaml);
    // while parsing, ryml marks parsed nodes with their original style:
    CHECK(tree.rootref().is_block());
    CHECK(tree["block map"].is_key_plain());
    CHECK(tree["block seq"].is_key_plain());
    CHECK(tree["flow map, singleline"].is_key_plain());
    CHECK(tree["flow seq, singleline"].is_key_plain());
    CHECK(tree["flow map, multiline"].is_key_plain());
    CHECK(tree["flow seq, multiline"].is_key_plain());
    CHECK(tree["block map"].is_block());
    CHECK(tree["block seq"].is_block());
    // flow is either singleline (FLOW_SL) or multiline (FLOW_ML1)
    CHECK(tree["flow map, singleline"].is_flow_sl());
    CHECK(tree["flow seq, singleline"].is_flow_sl());
    CHECK(tree["flow map, multiline"].is_flow_ml1());
    CHECK(tree["flow seq, multiline"].is_flow_ml1());
    // is_flow() is equivalent to (is_flow_sl() || is_flow_ml1() || is_flow_mln())
    CHECK(tree["flow map, singleline"].is_flow());
    CHECK(tree["flow seq, singleline"].is_flow());
    CHECK(tree["flow map, multiline"].is_flow());
    CHECK(tree["flow seq, multiline"].is_flow());
    //
    // since the tree nodes are marked with their original parsed
    // style, emitting the parsed tree will preserve the original
    // style (minus whitespace):
    //
    CHECK(tostr(tree) == yaml); // same as before!
    //
    // you can set/modify the style programatically!
    //
    // here are more examples.
    //
    {
        ryml::NodeRef n = tree["block map"]; // Let's look at one node
        // It looks like this originally:
        CHECK(tostr(n) ==
              "block map:\n"
              "  block key: block val\n"
              "");
        // let's modify its style:
        n.set_key_style(ryml::KEY_SQUO);      // scalar style: to single-quoted scalar
        n.set_container_style(ryml::FLOW_SL); // container style: to flow singleline
        // now it looks like this:
        CHECK(tostr(n) ==
              "'block map': {block key: block val}\n"
              "");
    }
    // next example
    {
        ryml::NodeRef n = tree["block seq"];
        CHECK(tostr(n) == ""
              "block seq:\n"
              "  - block val 1\n"
              "  - block val 2\n"
              "  - 'quoted'\n"
              "");
        n.set_key_style(ryml::KEY_DQUO);       // scalar style: to double-quoted scalar
        n[2].set_val_style(ryml::VAL_PLAIN);   // scalar style: to plain
        n.set_container_style(ryml::FLOW_MLN); // container style: to flow multiline, N values per line
        CHECK(tostr(n) == ""
              "\"block seq\": [\n"
              "    block val 1,block val 2,quoted\n"
              "  ]\n");
        n.set_container_style(ryml::FLOW_MLN|ryml::FLOW_SPC); // force space after comma
        CHECK(tostr(n) == ""
              "\"block seq\": [\n"
              "    block val 1, block val 2, quoted\n"
              "  ]\n");
        auto maxcols20 = ryml::EmitOptions{}.max_cols(20); // set the max number of cols for FLOW_MLN
        CHECK(tostr_opts(n, maxcols20) == ""
              "\"block seq\": [\n"
              "    block val 1, block val 2,\n"
              "    quoted\n"
              "  ]\n");
        n.set_container_style(ryml::FLOW_MLN); // no spaces now
        CHECK(tostr_opts(n, maxcols20) == ""
              "\"block seq\": [\n"
              "    block val 1,block val 2,\n"
              "    quoted\n"
              "  ]\n");
        n.set_container_style(ryml::FLOW_SL); // to flow singleline
        CHECK(tostr(n) == ""
              "\"block seq\": [block val 1,block val 2,quoted]\n");
        n.set_container_style(ryml::FLOW_SL|ryml::FLOW_SPC); // now with space after comma
        CHECK(tostr(n) == ""
              "\"block seq\": [block val 1, block val 2, quoted]\n");
        n.set_container_style(ryml::FLOW_ML1); // to flow multiline, 1 value per line
        CHECK(tostr(n) == ""
              "\"block seq\": [\n"
              "    block val 1,\n"
              "    block val 2,\n"
              "    quoted\n"
              "  ]\n");
        /// @see See more details about formatting flow containers in
        /// @ref sample_style_flow_formatting() (below).
    }
    // next example
    {
        ryml::NodeRef n = tree["flow map, singleline"];
        CHECK(tostr(n) == "flow map, singleline: {flow key: flow val}\n");
        n.set_container_style(ryml::BLOCK);
        n["flow key"].set_val_style(ryml::VAL_LITERAL);
        CHECK(tostr(n) == ""
              "flow map, singleline:\n"
              "  flow key: |-\n"
              "    flow val\n"
              "");
    }
    // next example
    {
        ryml::NodeRef n = tree["flow map, multiline"];
        CHECK(tostr(n) == ""
              "flow map, multiline: {\n"
              "    flow key: flow val\n"
              "  }\n"
              "");
        n.set_container_style(ryml::BLOCK);
        CHECK(tostr(n) == ""
              "flow map, multiline:\n"
              "  flow key: flow val\n"
              "");
    }
    // next example
    {
        ryml::NodeRef n = tree["flow seq, singleline"];
        CHECK(tostr(n) == "flow seq, singleline: [flow val,flow val]\n");
        n.set_key_style(ryml::KEY_FOLDED);
        n.set_container_style(ryml::BLOCK);
        n[0].set_val_style(ryml::VAL_SQUO);
        n[1].set_val_style(ryml::VAL_DQUO);
        CHECK(tostr(n) == ""
              "? >-\n"
              "  flow seq, singleline\n"
              ":\n"
              "  - 'flow val'\n"
              "  - \"flow val\"\n"
              "");
    }
    // next example
    {
        ryml::NodeRef n = tree["flow seq, multiline"];
        CHECK(tostr(n) == ""
              "flow seq, multiline: [\n"
              "    flow val,\n"
              "    flow val\n"
              "  ]\n"
              "");
        n.set_container_style(ryml::FLOW_SL);
        CHECK(tostr(n) == "flow seq, multiline: [flow val,flow val]\n");
    }
    // note the full tree now:
    CHECK(tostr(tree) != yaml);
    CHECK(tostr(tree) ==
          "'block map': {block key: block val}"         "\n"
          "\"block seq\": ["                            "\n"
          "    block val 1,"                            "\n"
          "    block val 2,"                            "\n"
          "    quoted"                                  "\n"
          "  ]"                                         "\n"
          "flow map, singleline:"                       "\n"
          "  flow key: |-"                              "\n"
          "    flow val"                                "\n"
          "? >-"                                        "\n"
          "  flow seq, singleline"                      "\n"
          ":"                                           "\n"
          "  - 'flow val'"                              "\n"
          "  - \"flow val\""                            "\n"
          "flow map, multiline:"                        "\n"
          "  flow key: flow val"                        "\n"
          "flow seq, multiline: [flow val,flow val]"    "\n"
          "");
    // you can clear the style of single nodes:
    tree["block map"].clear_style();
    tree["block seq"].clear_style();
    CHECK(tostr(tree) ==
          "block map:"                                    "\n"
          "  block key: block val"                        "\n"
          "block seq:"                                    "\n"
          "  - block val 1"                               "\n"
          "  - block val 2"                               "\n"
          "  - quoted"                                    "\n"
          "flow map, singleline:"                         "\n"
          "  flow key: |-"                                "\n"
          "    flow val"                                  "\n"
          "? >-"                                          "\n"
          "  flow seq, singleline"                        "\n"
          ":"                                             "\n"
          "  - 'flow val'"                                "\n"
          "  - \"flow val\""                              "\n"
          "flow map, multiline:"                          "\n"
          "  flow key: flow val"                          "\n"
          "flow seq, multiline: [flow val,flow val]"      "\n"
          "");
    // you can clear the style recursively:
    tree.rootref().clear_style(/*recurse*/true);
    // when emitting nodes which have no style set, ryml will default
    // to block format for containers, and call
    // ryml::scalar_style_choose() to pick the style for each scalar
    // (at the cost of a scan over each scalar). Note that ryml picks
    // single-quoted for scalars containing commas:
    CHECK(tostr(tree) ==
          "block map:"              "\n"
          "  block key: block val"  "\n"
          "block seq:"              "\n"
          "  - block val 1"         "\n"
          "  - block val 2"         "\n"
          "  - quoted"              "\n"
          "flow map, singleline:"   "\n"
          "  flow key: flow val"    "\n"
          "flow seq, singleline:"   "\n"
          "  - flow val"            "\n"
          "  - flow val"            "\n"
          "flow map, multiline:"    "\n"
          "  flow key: flow val"    "\n"
          "flow seq, multiline:"    "\n"
          "  - flow val"            "\n"
          "  - flow val"            "\n"
          "");
    // you can set the style based on type conditions:
    //
    // eg, set a single key to single-quoted
    tree["block map"].set_style_conditionally(/*type_mask*/ryml::KEY,
                                              /*remflags*/ryml::KEY_STYLE,
                                              /*addflags*/ryml::KEY_SQUO,
                                              /*recurse*/false);
    CHECK(tostr(tree) ==
          "'block map':"             "\n"
          "  block key: block val"   "\n"
          "block seq:"               "\n"
          "  - block val 1"          "\n"
          "  - block val 2"          "\n"
          "  - quoted"               "\n"
          "flow map, singleline:"    "\n"
          "  flow key: flow val"     "\n"
          "flow seq, singleline:"    "\n"
          "  - flow val"             "\n"
          "  - flow val"             "\n"
          "flow map, multiline:"     "\n"
          "  flow key: flow val"     "\n"
          "flow seq, multiline:"     "\n"
          "  - flow val"             "\n"
          "  - flow val"             "\n"
          "");
    // change all keys to single-quoted:
    tree.rootref().set_style_conditionally(/*type_mask*/ryml::KEY,
                                           /*remflags*/ryml::KEY_STYLE,
                                           /*addflags*/ryml::KEY_SQUO,
                                           /*recurse*/true);
    // change all vals to double-quoted
    tree.rootref().set_style_conditionally(/*type_mask*/ryml::VAL,
                                           /*remflags*/ryml::VAL_STYLE,
                                           /*addflags*/ryml::VAL_DQUO,
                                           /*recurse*/true);
    // change all seqs to flow
    tree.rootref().set_style_conditionally(/*type_mask*/ryml::SEQ,
                                           /*remflags*/ryml::CONTAINER_STYLE,
                                           /*addflags*/ryml::FLOW_SL,
                                           /*recurse*/true);
    // change all maps to flow
    tree.rootref().set_style_conditionally(/*type_mask*/ryml::MAP,
                                           /*remflags*/ryml::CONTAINER_STYLE,
                                           /*addflags*/ryml::BLOCK,
                                           /*recurse*/true);
    // done!
    CHECK(tostr(tree) == ""
          "'block map':"                                               "\n"
          "  'block key': \"block val\""                               "\n"
          "'block seq': [\"block val 1\",\"block val 2\",\"quoted\"]"  "\n"
          "'flow map, singleline':"                                    "\n"
          "  'flow key': \"flow val\""                                 "\n"
          "'flow seq, singleline': [\"flow val\",\"flow val\"]"        "\n"
          "'flow map, multiline':"                                     "\n"
          "  'flow key': \"flow val\""                                 "\n"
          "'flow seq, multiline': [\"flow val\",\"flow val\"]"         "\n"
          "");
    // you can also set a conditional style in a single node (or its branch if recurse is true):
    tree["flow seq, singleline"].set_style_conditionally(/*type_mask*/ryml::SEQ,
                                                         /*remflags*/ryml::CONTAINER_STYLE,
                                                         /*addflags*/ryml::BLOCK,
                                                         /*recurse*/false);
    CHECK(tostr(tree) == ""
          "'block map':"                                                   "\n"
          "  'block key': \"block val\""                                   "\n"
          "'block seq': [\"block val 1\",\"block val 2\",\"quoted\"]"      "\n"
          "'flow map, singleline':"                                        "\n"
          "  'flow key': \"flow val\""                                     "\n"
          "'flow seq, singleline':"                                        "\n"
          "  - \"flow val\""                                               "\n"
          "  - \"flow val\""                                               "\n"
          "'flow map, multiline':"                                         "\n"
          "  'flow key': \"flow val\""                                     "\n"
          "'flow seq, multiline': [\"flow val\",\"flow val\"]"             "\n"
        "");
    /// see also:
    ///  - @ref ryml::scalar_style_choose_block()
    ///  - @ref ryml::scalar_style_choose_flow()
    ///  - @ref ryml::scalar_style_choose_json()
    ///  - @ref ryml::scalar_style_query_squo()
    ///  - @ref ryml::scalar_style_query_plain()
}

sample_style_flow_formatting()

void sample_style_flow_formatting

(

)

Shows how to control formatting of flow styles.

Definition at line 4704 of file quickstart.cpp.

{
    // we will be using this helper throughout this function
    auto tostr = [](ryml::ConstNodeRef n, ryml::EmitOptions opts) {
        return ryml::emitrs_yaml<std::string>(n, opts);
    };
    auto tostr_json = [](ryml::ConstNodeRef n, ryml::EmitOptions opts) {
        return ryml::emitrs_json<std::string>(n, opts);
    };
    const ryml::EmitOptions emit_defaults = ryml::EmitOptions{};
    // let's parse this, which is in FLOW_ML1 (flow multiline, 1 value per line):
    ryml::csubstr yaml = ""
        "{"              "\n"
        "  map: {"       "\n"
        "    seq: ["     "\n"
        "      0,"       "\n"
        "      1,"       "\n"
        "      2,"       "\n"
        "      3,"       "\n"
        "      [40,41]"  "\n"
        "    ]"          "\n"
        "  }"            "\n"
        "}"              "\n"
        "";
    // note that the parser defaults to detecting multiline flow
    // (FLOW_ML1) containers:
    {
        const ryml::Tree tree = ryml::parse_in_arena(yaml);
        CHECK(tree["map"].is_flow_ml1()); // etc
        CHECK(tree["map"]["seq"].is_flow_ml1()); // etc
        CHECK(tree["map"]["seq"][4].is_flow_sl()); // etc
        // emitted yaml is exactly equal to parsed yaml:
        CHECK(tostr(tree, emit_defaults) == yaml);
        // json looks like similar:
        CHECK(tostr_json(tree, emit_defaults) ==
              "{\n"
              "  \"map\": {\n"
              "    \"seq\": [\n"
              "      0,\n"
              "      1,\n"
              "      2,\n"
              "      3,\n"
              "      [40,41]\n"
              "    ]\n"
              "  }\n"
              "}\n"
              "");
    }
    // if you prefer to shorten the emitted yaml, you can set the
    // parser to disable flow multiline detection. It will then pick
    // singleline flow (FLOW_SL) for all flow containers:
    {
        const ryml::ParserOptions opts = ryml::ParserOptions{}
            .detect_flow_ml(false);
        const ryml::Tree tree = ryml::parse_in_arena(yaml, opts);
        CHECK(tree["map"].is_flow_sl()); // etc
        // notice how this is smaller now:
        CHECK(tostr(tree, emit_defaults) ==
              "{map: {seq: [0,1,2,3,[40,41]]}}");
        // and json as well
        CHECK(tostr_json(tree, emit_defaults) ==
              "{\"map\": {\"seq\": [0,1,2,3,[40,41]]}}");
        // you can also force spaces everywhere without adding
        // FLOW_SPC in individual containers:
        const ryml::EmitOptions with_spaces = ryml::EmitOptions{}
            .force_flow_spc(true);
        CHECK(tostr(tree, with_spaces) ==
              "{map: {seq: [0, 1, 2, 3, [40, 41]]}}");
        // and json as well
        CHECK(tostr_json(tree, with_spaces) ==
              "{\"map\": {\"seq\": [0, 1, 2, 3, [40, 41]]}}");
    }
    // or you can still have the default detection of flow_ml, but set
    // it to pick FLOW_MLN (multiline, n values per line), instead of
    // the default FLOW_ML1 (multiline, 1 values per line)
    {
        const ryml::ParserOptions opts = ryml::ParserOptions{}
            .flow_ml_style(ryml::FLOW_MLN);
        const ryml::Tree tree = ryml::parse_in_arena(yaml, opts);
        CHECK(tree["map"].is_flow_mln());
        CHECK(tree["map"]["seq"][4].is_flow_sl()); // [40,41] is FLOW_SL
        CHECK(tostr(tree, emit_defaults) ==
              "{"                           "\n"
              "  map: {"                    "\n"
              "    seq: ["                  "\n"
              "      0,1,2,3,[40,41]"       "\n"
              "    ]"                       "\n"
              "  }"                         "\n"
              "}"                           "\n");
        CHECK(tostr_json(tree, emit_defaults) ==
              "{"                           "\n"
              "  \"map\": {"                "\n"
              "    \"seq\": ["              "\n"
              "      0,1,2,3,[40,41]"       "\n"
              "    ]"                       "\n"
              "  }"                         "\n"
              "}"                           "\n");
        // now with spaces:
        const ryml::EmitOptions with_spaces = ryml::EmitOptions{}
            .force_flow_spc(true);
        CHECK(tostr(tree, with_spaces) ==
              "{"                           "\n"
              "  map: {"                    "\n"
              "    seq: ["                  "\n"
              "      0, 1, 2, 3, [40, 41]"  "\n"
              "    ]"                       "\n"
              "  }"                         "\n"
              "}"                           "\n");
        CHECK(tostr_json(tree, with_spaces) ==
              "{"                           "\n"
              "  \"map\": {"                "\n"
              "    \"seq\": ["              "\n"
              "      0, 1, 2, 3, [40, 41]"  "\n"
              "    ]"                       "\n"
              "  }"                         "\n"
              "}"                           "\n");
    }
    // you can also disable indentation of both FLOW_ML1 and FLOW_MLN
    // (see more details in @ref sample_style_flow_ml_indent())
    {
        const ryml::EmitOptions noindent = ryml::EmitOptions{}
            .indent_flow_ml(false);
        const ryml::Tree tree = ryml::parse_in_arena(yaml);
        CHECK(tree["map"].is_flow_ml1());
        CHECK(tree["map"]["seq"][4].is_flow_sl()); // [40,41] is FLOW_SL
        CHECK(tostr(tree, noindent) == ""
              "{"              "\n"
              "map: {"         "\n"
              "seq: ["         "\n"
              "0,"             "\n"
              "1,"             "\n"
              "2,"             "\n"
              "3,"             "\n"
              "[40,41]"        "\n"
              "]"              "\n"
              "}"              "\n"
              "}"              "\n"
              "");
        CHECK(tostr_json(tree, noindent) == ""
              "{"              "\n"
              "\"map\": {"     "\n"
              "\"seq\": ["     "\n"
              "0,"             "\n"
              "1,"             "\n"
              "2,"             "\n"
              "3,"             "\n"
              "[40,41]"        "\n"
              "]"              "\n"
              "}"              "\n"
              "}"              "\n"
              "");
    }
    // finally, you can control the number of columns in FLOW_MLN:
    {
        // let's pick a different example to make this clearer
        ryml::csubstr yaml2 = ""
            "[" "\n"
            "  0, 1, 2, 3, 4, 5, 6, 7, 8, 9,"                   "\n"
            "  10, 11, 12, 13, 14, 15, 16, 17, 18, 19, "        "\n"
            "  20, 21, 22, 23, 24, 25, 26, 27, 28, 29, "        "\n"
            "  30, 31, 32, 33, 34, 35, 36, 37, 38, 39, "        "\n"
            "  40, 41, 42, 43, 44, 45, 46, 47, 48, 49, "        "\n"
            "  50, 51, 52, 53, 54, 55, 56, 57, 58, 59, "        "\n"
            "  60, 61, 62, 63, 64, 65, 66, 67, 68, 69, "        "\n"
            "  70, 71, 72, 73, 74, 75, 76, 77, 78, 79  "        "\n"
            "]";
        // Let's force the parser to pick FLOW_MLN instead of
        // FLOW_ML1. We're doing that because wrapping is only done in
        // FLOW_MLN and -- as their names imply -- FLOW_SL is
        // single-line, and FLOW_ML1 is 1 value per line.
        const ryml::ParserOptions opts = ryml::ParserOptions{}
            .flow_ml_style(ryml::FLOW_MLN);
        const ryml::Tree tree = ryml::parse_in_arena(yaml2, opts);
        CHECK(tree.rootref().type().is_flow_mln());
        // default max columns is 80:
        CHECK(tostr(tree, emit_defaults) == ""
              "[\n"
              "  0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,\n"
              "  30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,\n"
              "  56,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71,72,73,74,75,76,77,78,79\n"
              "]\n"
              "");
        CHECK(tostr_json(tree, emit_defaults) == ""
              "[\n"
              "  0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,\n"
              "  29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,\n"
              "  55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71,72,73,74,75,76,77,78,79\n"
              "]\n"
              "");
        // let's try setting max columns to 40:
        const ryml::EmitOptions maxcols40 = ryml::EmitOptions{}
            .max_cols(40);
        CHECK(tostr(tree, maxcols40) == ""
              "[\n"
              "  0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,\n"
              "  16,17,18,19,20,21,22,23,24,25,26,27,28,\n"
              "  29,30,31,32,33,34,35,36,37,38,39,40,41,\n"
              "  42,43,44,45,46,47,48,49,50,51,52,53,54,\n"
              "  55,56,57,58,59,60,61,62,63,64,65,66,67,\n"
              "  68,69,70,71,72,73,74,75,76,77,78,79\n"
              "]\n"
              "");
        CHECK(tostr_json(tree, maxcols40) == ""
              "[\n"
              "  0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,\n"
              "  16,17,18,19,20,21,22,23,24,25,26,27,28,\n"
              "  29,30,31,32,33,34,35,36,37,38,39,40,41,\n"
              "  42,43,44,45,46,47,48,49,50,51,52,53,54,\n"
              "  55,56,57,58,59,60,61,62,63,64,65,66,67,\n"
              "  68,69,70,71,72,73,74,75,76,77,78,79\n"
              "]\n"
              "");
        // Note that you can globally force spaces everywhere through
        // the emit options:
        const ryml::EmitOptions with_spaces = ryml::EmitOptions{}
            .force_flow_spc(true);
        CHECK(tostr(tree, with_spaces) == ""
              "[\n"
              "  0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21,\n"
              "  22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41,\n"
              "  42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61,\n"
              "  62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79\n"
              "]\n"
              "");
        CHECK(tostr_json(tree, with_spaces) == ""
              "[\n"
              "  0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21,\n"
              "  22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41,\n"
              "  42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61,\n"
              "  62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79\n"
              "]\n"
              "");
        // and you can combine spaces with max columns:
        const ryml::EmitOptions maxcols40_spc = ryml::EmitOptions{}
            .max_cols(40)
            .force_flow_spc(true);
        CHECK(tostr(tree, maxcols40_spc) == ""
              "[\n"
              "  0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11,\n"
              "  12, 13, 14, 15, 16, 17, 18, 19, 20, 21,\n"
              "  22, 23, 24, 25, 26, 27, 28, 29, 30, 31,\n"
              "  32, 33, 34, 35, 36, 37, 38, 39, 40, 41,\n"
              "  42, 43, 44, 45, 46, 47, 48, 49, 50, 51,\n"
              "  52, 53, 54, 55, 56, 57, 58, 59, 60, 61,\n"
              "  62, 63, 64, 65, 66, 67, 68, 69, 70, 71,\n"
              "  72, 73, 74, 75, 76, 77, 78, 79\n"
              "]\n"
              "");
        CHECK(tostr_json(tree, maxcols40_spc) == ""
              "[\n"
              "  0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11,\n"
              "  12, 13, 14, 15, 16, 17, 18, 19, 20, 21,\n"
              "  22, 23, 24, 25, 26, 27, 28, 29, 30, 31,\n"
              "  32, 33, 34, 35, 36, 37, 38, 39, 40, 41,\n"
              "  42, 43, 44, 45, 46, 47, 48, 49, 50, 51,\n"
              "  52, 53, 54, 55, 56, 57, 58, 59, 60, 61,\n"
              "  62, 63, 64, 65, 66, 67, 68, 69, 70, 71,\n"
              "  72, 73, 74, 75, 76, 77, 78, 79\n"
              "]\n"
              "");
        // and you can combine spaces with max columns with no indentation:
        const ryml::EmitOptions maxcols40_spc_noindent = ryml::EmitOptions{}
            .max_cols(40)
            .force_flow_spc(true)
            .indent_flow_ml(false);
        CHECK(tostr(tree, maxcols40_spc_noindent) == ""
              "[\n"
              "0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12,\n"
              "13, 14, 15, 16, 17, 18, 19, 20, 21, 22,\n"
              "23, 24, 25, 26, 27, 28, 29, 30, 31, 32,\n"
              "33, 34, 35, 36, 37, 38, 39, 40, 41, 42,\n"
              "43, 44, 45, 46, 47, 48, 49, 50, 51, 52,\n"
              "53, 54, 55, 56, 57, 58, 59, 60, 61, 62,\n"
              "63, 64, 65, 66, 67, 68, 69, 70, 71, 72,\n"
              "73, 74, 75, 76, 77, 78, 79\n"
              "]\n"
              "");
        CHECK(tostr_json(tree, maxcols40_spc_noindent) == ""
              "[\n"
              "0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12,\n"
              "13, 14, 15, 16, 17, 18, 19, 20, 21, 22,\n"
              "23, 24, 25, 26, 27, 28, 29, 30, 31, 32,\n"
              "33, 34, 35, 36, 37, 38, 39, 40, 41, 42,\n"
              "43, 44, 45, 46, 47, 48, 49, 50, 51, 52,\n"
              "53, 54, 55, 56, 57, 58, 59, 60, 61, 62,\n"
              "63, 64, 65, 66, 67, 68, 69, 70, 71, 72,\n"
              "73, 74, 75, 76, 77, 78, 79\n"
              "]\n"
              "");
    }
    // Note that FLOW_SPC is /not/ detected by the parser, and that
    // depending on the parse options, either FLOW_ML1 or FLOW_MLN
    // will be used for /all/ multiline flow containers.
}

sample_style_flow_ml_indent()

void sample_style_flow_ml_indent

(

)

control the indentation of emitted flow multiline containers

Definition at line 5003 of file quickstart.cpp.

{
    // we will be using this helper throughout this function
    auto tostr = [](ryml::ConstNodeRef n, ryml::EmitOptions opts) {
        return ryml::emitrs_yaml<std::string>(n, opts);
    };
    ryml::csubstr yaml = "{map: {seq: [0, 1, 2, 3, [40, 41]]}}";
    ryml::Tree tree = ryml::parse_in_arena(yaml);
    ryml::EmitOptions defaults = {};
    ryml::EmitOptions noindent = ryml::EmitOptions{}.indent_flow_ml(false);
    CHECK(tostr(tree, defaults) == "{map: {seq: [0,1,2,3,[40,41]]}}");
    // let's now set the style to FLOW_ML1 (it was FLOW_SL)
    tree.rootref().set_container_style(ryml::FLOW_ML1);
    tree["map"].set_container_style(ryml::FLOW_ML1);
    tree["map"]["seq"].set_container_style(ryml::FLOW_ML1);
    tree["map"]["seq"][4].set_container_style(ryml::FLOW_ML1);
    // by default FLOW_ML1 prints one value per line, indented:
    CHECK(tostr(tree, defaults) ==
          "{"              "\n"
          "  map: {"       "\n"
          "    seq: ["     "\n"
          "      0,"       "\n"
          "      1,"       "\n"
          "      2,"       "\n"
          "      3,"       "\n"
          "      ["        "\n"
          "        40,"    "\n"
          "        41"     "\n"
          "      ]"        "\n"
          "    ]"          "\n"
          "  }"            "\n"
          "}"              "\n"
          "");
    // if we use the noindent options, then each value is put at the
    // beginning of the line
    CHECK(tostr(tree, noindent) ==
          "{"            "\n"
          "map: {"       "\n"
          "seq: ["       "\n"
          "0,"           "\n"
          "1,"           "\n"
          "2,"           "\n"
          "3,"           "\n"
          "["            "\n"
          "40,"          "\n"
          "41"           "\n"
          "]"            "\n"
          "]"            "\n"
          "}"            "\n"
          "}"            "\n"
          "");
    // Note that the noindent option will safely respect any prior
    // indent level from enclosing block containers! For example:
    tree.rootref().set_container_style(ryml::BLOCK);
    CHECK(tostr(tree, noindent) == ""// notice it is indented at the map level
          "map: {"       "\n"
          "  seq: ["     "\n"
          "  0,"         "\n"
          "  1,"         "\n"
          "  2,"         "\n"
          "  3,"         "\n"
          "  ["          "\n"
          "  40,"        "\n"
          "  41"         "\n"
          "  ]"          "\n"
          "  ]"          "\n"
          "  }"          "\n"
          "");
    // Let's set it one BLOCK level further:
    tree["map"].set_container_style(ryml::BLOCK);
    CHECK(tostr(tree, noindent) == // notice it is indented one more level
          "map:"         "\n"
          "  seq: ["     "\n"
          "    0,"       "\n"
          "    1,"       "\n"
          "    2,"       "\n"
          "    3,"       "\n"
          "    ["        "\n"
          "    40,"      "\n"
          "    41"       "\n"
          "    ]"        "\n"
          "    ]"        "\n"
          "");
}

sample_json()

void sample_json

(

)

shows how to parse and emit JSON.

To emit YAML parsed from JSON, see also sample_style() for info on clearing the style flags (example below).

Definition at line 5095 of file quickstart.cpp.

{
    ryml::csubstr json = ""
        "{"                                         "\n"
        "  \"doe\": \"a deer, a female deer\","     "\n"
        "  \"ray\": \"a drop of golden sun\","      "\n"
        "  \"me\": \"a name, I call myself\","      "\n"
        "  \"far\": \"a long long way to go\""      "\n"
        "}"                                         "\n"
        "";
    // Since JSON is a subset of YAML, parsing JSON is just the
    // same as YAML:
    ryml::Tree tree = ryml::parse_in_arena(json);
    // If you are sure the source is valid json, you can use the
    // appropriate parse_json overload, which is faster because json
    // has a smaller grammar:
    ryml::Tree json_tree = ryml::parse_json_in_arena(json);
    // to emit JSON:
    CHECK(ryml::emitrs_json<std::string>(tree) == json);
    CHECK(ryml::emitrs_json<std::string>(json_tree) == json);
    // to emit JSON to a stream:
    std::stringstream ss;
    ss << ryml::as_json(tree);  // <- mark it like this
    CHECK(ss.str() == json);
    // Note the following limitations:
    //
    // - YAML streams cannot be emitted as json, and are not
    //   allowed. But you can work around this by emitting the
    //   individual documents separately; see the sample_docs()
    //   below for such an example.
    //
    // - tags cannot be emitted as json, and are not allowed.
    //
    // - anchors and references cannot be emitted as json and
    //   are not allowed.
    //
 
    // Note that when parsing JSON, ryml will the style of each node
    // in the JSON. This means that if you emit as YAML it will look
    // mostly the same as the JSON:
    CHECK(ryml::emitrs_yaml<std::string>(json_tree) == json);
    // If you want to avoid this, you will need to clear the style.
    json_tree.rootref().clear_style(); // clear the style of the map (without recursing)
    // note that this is now block mode. That is because when no
    // style is set, ryml defaults to emitting in block mode.
    CHECK(ryml::emitrs_yaml<std::string>(json_tree) == ""
          "\"doe\": \"a deer, a female deer\""   "\n"
          "\"ray\": \"a drop of golden sun\""    "\n"
          "\"me\": \"a name, I call myself\""    "\n"
          "\"far\": \"a long long way to go\""   "\n"
          "");
    // if you don't want the double quotes in the scalar, you can
    // recurse:
    json_tree.rootref().clear_style(/*recurse*/true);
    // so now when emitting you will get this:
    // (the scalars with a comma are single-quote)
    CHECK(ryml::emitrs_yaml<std::string>(json_tree) == ""
          "doe: a deer, a female deer"      "\n"
          "ray: a drop of golden sun"       "\n"
          "me: a name, I call myself"       "\n"
          "far: a long long way to go"      "\n"
          "");
    // you can do custom style changes based on a type mask. this
    // will change the style of all scalar values to single-quoted
    json_tree.rootref().set_style_conditionally(ryml::VAL,
                                                /*remflags*/ryml::VAL_STYLE,
                                                /*addflags*/ryml::VAL_SQUO,
                                                /*recurse*/true);
    CHECK(ryml::emitrs_yaml<std::string>(json_tree) ==
          "doe: 'a deer, a female deer'"    "\n"
          "ray: 'a drop of golden sun'"     "\n"
          "me: 'a name, I call myself'"     "\n"
          "far: 'a long long way to go'"    "\n"
          "");
    // see in particular sample_style() for more examples
}

sample_anchors_and_aliases()

void sample_anchors_and_aliases

(

)

demonstrates usage with anchors and alias references.

Note that dereferencing is opt-in; after parsing, you have to call c4::yml::Tree::resolve() explicitly if you want resolved references in the tree. This method will resolve all references and substitute the anchored values in place of the reference.

The c4::yml::Tree::resolve() method first does a full traversal of the tree to gather all anchors and references in a separate collection, then it goes through that collection to locate the names, which it does by obeying the YAML standard diktat that

> an alias node refers to the most recent node in
> the serialization having the specified anchor

So, depending on the number of anchor/alias nodes, this is a potentially expensive operation, with a best-case linear complexity (from the initial traversal) and a worst-case quadratic complexity (if every node has an alias/anchor). This potential cost is the reason for requiring an explicit call to c4::yml::Tree::resolve()

Definition at line 5195 of file quickstart.cpp.

{
    std::string unresolved = ""
        "base: &base"                       "\n"
        "  name: Everyone has same name"    "\n"
        "foo: &foo"                         "\n"
        "  <<: *base"                       "\n"
        "  age: 10"                         "\n"
        "bar: &bar"                         "\n"
        "  <<: *base"                       "\n"
        "  age: 20"                         "\n"
        "bill_to: &id001"                   "\n"
        "  street: |-"                      "\n"
        "    123 Tornado Alley"             "\n"
        "    Suite 16"                      "\n"
        "  city: East Centerville"          "\n"
        "  state: KS"                       "\n"
        "ship_to: *id001"                   "\n"
        "&keyref key: &valref val"          "\n"
        "*valref : *keyref"                 "\n"
        "";
    std::string resolved = ""
        "base:"                             "\n"
        "  name: Everyone has same name"    "\n"
        "foo:"                              "\n"
        "  name: Everyone has same name"    "\n"
        "  age: 10"                         "\n"
        "bar:"                              "\n"
        "  name: Everyone has same name"    "\n"
        "  age: 20"                         "\n"
        "bill_to:"                          "\n"
        "  street: |-"                      "\n"
        "    123 Tornado Alley"             "\n"
        "    Suite 16"                      "\n"
        "  city: East Centerville"          "\n"
        "  state: KS"                       "\n"
        "ship_to:"                          "\n"
        "  street: |-"                      "\n"
        "    123 Tornado Alley"             "\n"
        "    Suite 16"                      "\n"
        "  city: East Centerville"          "\n"
        "  state: KS"                       "\n"
        "key: val"                          "\n"
        "val: key"                          "\n"
        "";
 
    ryml::Tree tree = ryml::parse_in_arena(ryml::to_csubstr(unresolved));
    // by default, references are not resolved when parsing:
    CHECK( ! tree["base"].has_key_anchor());
    CHECK(   tree["base"].has_val_anchor());
    CHECK(   tree["base"].val_anchor() == "base");
    CHECK(   tree["key"].key_anchor() == "keyref");
    CHECK(   tree["key"].val_anchor() == "valref");
    CHECK(   tree["*valref"].is_key_ref());
    CHECK(   tree["*valref"].is_val_ref());
    CHECK(   tree["*valref"].key_ref() == "valref");
    CHECK(   tree["*valref"].val_ref() == "keyref");
 
    // to resolve references, simply call tree.resolve(),
    // which will perform the reference instantiations:
    tree.resolve();
 
    // all the anchors and references are substistuted and then removed:
    CHECK( ! tree["base"].has_key_anchor());
    CHECK( ! tree["base"].has_val_anchor());
    CHECK( ! tree["base"].has_val_anchor());
    CHECK( ! tree["key"].has_key_anchor());
    CHECK( ! tree["key"].has_val_anchor());
    CHECK( ! tree["val"].is_key_ref()); // notice *valref is now turned to val
    CHECK( ! tree["val"].is_val_ref()); // notice *valref is now turned to val
 
    CHECK(tree["ship_to"]["city"].val() == "East Centerville");
    CHECK(tree["ship_to"]["state"].val() == "KS");
}

sample_anchors_and_aliases_create()

void sample_anchors_and_aliases_create

(

)

demonstrates how to use the API to programatically create anchors and aliases

Definition at line 5272 of file quickstart.cpp.

{
    // part 1: anchor/ref
    {
        ryml::Tree t;
        t.rootref() |= ryml::MAP|ryml::BLOCK;
        t["kanchor"] = "2";
        t["kanchor"].set_key_anchor("kanchor");
        t["vanchor"] = "3";
        t["vanchor"].set_val_anchor("vanchor");
        // to set a reference, need to call .set_val_ref()/.set_key_ref()
        t["kref"].set_val_ref("kanchor");
        t["vref"].set_val_ref("vanchor");
        t["nref"] = "*vanchor";  // NOTE: this is not set as a reference in the tree!
        CHECK(ryml::emitrs_yaml<std::string>(t) == ""
              "&kanchor kanchor: 2"  "\n"
              "vanchor: &vanchor 3"  "\n"
              "kref: *kanchor"       "\n"
              "vref: *vanchor"       "\n"
              // note that ryml emits nref with quotes to disambiguate
              // (because no style was set)
              "nref: '*vanchor'"     "\n"
              "");
        t.resolve();
        CHECK(ryml::emitrs_yaml<std::string>(t) == ""
              "kanchor: 2"           "\n"
              "vanchor: 3"           "\n"
              "kref: kanchor"        "\n"
              "vref: 3"              "\n"
              // note that nref was not resolved
              "nref: '*vanchor'"     "\n"
              "");
    }
 
    // part 2: simple inheritance (ie, adding `<<: *anchor` nodes)
    {
        ryml::Tree t = ryml::parse_in_arena(""
            "orig: &orig {foo: bar, baz: bat}"  "\n"
            "copy: {}"                          "\n"
            "notcopy: {}"                       "\n"
            "notref: {}"                        "\n"
            "");
        t["copy"]["<<"].set_val_ref("orig");
        t["notcopy"]["test"].set_val_ref("orig");
        t["notcopy"]["<<"].set_val_ref("orig");
        t["notref"]["<<"] = "*orig"; // not a reference! .set_val_ref() was not called
        CHECK(ryml::emitrs_yaml<std::string>(t) == ""
              "orig: &orig {foo: bar,baz: bat}"      "\n"
              "copy: {<<: *orig}"                    "\n"
              "notcopy: {test: *orig,<<: *orig}"     "\n"
              "notref: {<<: '*orig'}"                "\n"
              "");
        t.resolve();
        CHECK(ryml::emitrs_yaml<std::string>(t) == ""
              "orig: {foo: bar,baz: bat}"                                "\n"
              "copy: {foo: bar,baz: bat}"                                "\n"
              "notcopy: {test: {foo: bar,baz: bat},foo: bar,baz: bat}"   "\n"
              "notref: {<<: '*orig'}"                                    "\n"
              "");
    }
 
    // part 3: multiple inheritance (ie, `<<: [*ref1,*ref2,*etc]`)
    {
        ryml::Tree t = ryml::parse_in_arena(""
            "orig1: &orig1 {foo: bar}"       "\n"
            "orig2: &orig2 {baz: bat}"       "\n"
            "orig3: &orig3 {and: more}"      "\n"
            "copy: {}"                       "\n"
            "");
        ryml::NodeRef seq = t["copy"]["<<"];
        seq |= ryml::SEQ;
        seq.append_child().set_val_ref("orig1");
        seq.append_child().set_val_ref("orig2");
        seq.append_child().set_val_ref("orig3");
        CHECK(ryml::emitrs_yaml<std::string>(t) == ""
              "orig1: &orig1 {foo: bar}"                  "\n"
              "orig2: &orig2 {baz: bat}"                  "\n"
              "orig3: &orig3 {and: more}"                 "\n"
              "copy: {<<: [*orig1,*orig2,*orig3]}"        "\n"
              "");
        t.resolve();
        CHECK(ryml::emitrs_yaml<std::string>(t) == ""
              "orig1: {foo: bar}"                        "\n"
              "orig2: {baz: bat}"                        "\n"
              "orig3: {and: more}"                       "\n"
              "copy: {foo: bar,baz: bat,and: more}"      "\n");
    }
}

sample_tags()

void sample_tags

(

)

Definition at line 5364 of file quickstart.cpp.

{
    const std::string yaml = ""
        "--- !!map"           "\n"
        "a: 0"                "\n"
        "b: 1"                "\n"
        "--- !map"            "\n"
        "a: b"                "\n"
        "--- !!seq"           "\n"
        "- a"                 "\n"
        "- b"                 "\n"
        "--- !!str a b"       "\n"
        "--- !!str 'a: b'"    "\n"
        "---"                 "\n"
        "!!str a: b"          "\n"
        "--- !!set"           "\n"
        "? a"                 "\n"
        "? b"                 "\n"
        "--- !!set"           "\n"
        "a:"                  "\n"
        "--- !!seq"           "\n"
        "- !!int 0"           "\n"
        "- !!str 1"           "\n"
        "";
    const ryml::Tree tree = ryml::parse_in_arena(ryml::to_csubstr(yaml));
    const ryml::ConstNodeRef root = tree.rootref();
    CHECK(root.is_stream());
    CHECK(root.num_children() == 9);
    for(ryml::ConstNodeRef doc : root.children())
        CHECK(doc.is_doc());
    // tags are kept verbatim from the source:
    CHECK(root[0].has_val_tag());
    CHECK(root[0].val_tag() == "!!map"); // valid only if the node has a val tag
    CHECK(root[1].val_tag() == "!map");
    CHECK(root[2].val_tag() == "!!seq");
    CHECK(root[3].val_tag() == "!!str");
    CHECK(root[4].val_tag() == "!!str");
    CHECK(root[5]["a"].has_key_tag());
    CHECK(root[5]["a"].key_tag() == "!!str"); // valid only if the node has a key tag
    CHECK(root[6].val_tag() == "!!set");
    CHECK(root[7].val_tag() == "!!set");
    CHECK(root[8].val_tag() == "!!seq");
    CHECK(root[8][0].val_tag() == "!!int");
    CHECK(root[8][1].val_tag() == "!!str");
    // ryml also provides a complete toolbox to deal with tags.
    // there is an enumeration for the standard YAML tags:
    CHECK(ryml::to_tag("!map") == ryml::TAG_NONE);
    CHECK(ryml::to_tag("!!map") == ryml::TAG_MAP);
    CHECK(ryml::to_tag("!!seq") == ryml::TAG_SEQ);
    CHECK(ryml::to_tag("!!str") == ryml::TAG_STR);
    CHECK(ryml::to_tag("!!int") == ryml::TAG_INT);
    CHECK(ryml::to_tag("!!set") == ryml::TAG_SET);
    // given a tag enum, you can fetch the short tag string:
    CHECK(ryml::from_tag(ryml::TAG_NONE) == "");
    CHECK(ryml::from_tag(ryml::TAG_MAP) == "!!map");
    CHECK(ryml::from_tag(ryml::TAG_SEQ) == "!!seq");
    CHECK(ryml::from_tag(ryml::TAG_STR) == "!!str");
    CHECK(ryml::from_tag(ryml::TAG_INT) == "!!int");
    CHECK(ryml::from_tag(ryml::TAG_SET) == "!!set");
    // you can also fetch the long tag string:
    CHECK(ryml::from_tag_long(ryml::TAG_NONE) == "");
    CHECK(ryml::from_tag_long(ryml::TAG_MAP) == "<tag:yaml.org,2002:map>");
    CHECK(ryml::from_tag_long(ryml::TAG_SEQ) == "<tag:yaml.org,2002:seq>");
    CHECK(ryml::from_tag_long(ryml::TAG_STR) == "<tag:yaml.org,2002:str>");
    CHECK(ryml::from_tag_long(ryml::TAG_INT) == "<tag:yaml.org,2002:int>");
    CHECK(ryml::from_tag_long(ryml::TAG_SET) == "<tag:yaml.org,2002:set>");
    // and likewise:
    CHECK(ryml::to_tag("!map") == ryml::TAG_NONE);
    CHECK(ryml::to_tag("<tag:yaml.org,2002:map>") == ryml::TAG_MAP);
    CHECK(ryml::to_tag("<tag:yaml.org,2002:seq>") == ryml::TAG_SEQ);
    CHECK(ryml::to_tag("<tag:yaml.org,2002:str>") == ryml::TAG_STR);
    CHECK(ryml::to_tag("<tag:yaml.org,2002:int>") == ryml::TAG_INT);
    CHECK(ryml::to_tag("<tag:yaml.org,2002:set>") == ryml::TAG_SET);
    // to normalize a tag as much as possible, use normalize_tag():
    CHECK(ryml::normalize_tag("!!map") == "!!map");
    CHECK(ryml::normalize_tag("!<tag:yaml.org,2002:map>") == "!!map");
    CHECK(ryml::normalize_tag("<tag:yaml.org,2002:map>") == "!!map");
    CHECK(ryml::normalize_tag("tag:yaml.org,2002:map") == "!!map");
    CHECK(ryml::normalize_tag("!<!!map>") == "<!!map>");
    CHECK(ryml::normalize_tag("!map") == "!map");
    CHECK(ryml::normalize_tag("!my!foo") == "!my!foo");
    // and also for the long form:
    CHECK(ryml::normalize_tag_long("!!map") == "<tag:yaml.org,2002:map>");
    CHECK(ryml::normalize_tag_long("!<tag:yaml.org,2002:map>") == "<tag:yaml.org,2002:map>");
    CHECK(ryml::normalize_tag_long("<tag:yaml.org,2002:map>") == "<tag:yaml.org,2002:map>");
    CHECK(ryml::normalize_tag_long("tag:yaml.org,2002:map") == "<tag:yaml.org,2002:map>");
    CHECK(ryml::normalize_tag_long("!<!!map>") == "<!!map>");
    CHECK(ryml::normalize_tag_long("!map") == "!map");
    // The tree provides the following methods applying to every node
    // with a key and/or val tag:
    ryml::Tree normalized_tree = tree;
    normalized_tree.normalize_tags(); // normalize all tags in short form
    CHECK(ryml::emitrs_yaml<std::string>(normalized_tree) == ""
          "--- !!map"          "\n"
          "a: 0"               "\n"
          "b: 1"               "\n"
          "--- !map"           "\n"
          "a: b"               "\n"
          "--- !!seq"          "\n"
          "- a"                "\n"
          "- b"                "\n"
          "--- !!str a b"      "\n"
          "--- !!str 'a: b'"   "\n"
          "---"                "\n"
          "!!str a: b"         "\n"
          "--- !!set"          "\n"
          "a: "                "\n"
          "b: "                "\n"
          "--- !!set"          "\n"
          "a: "                "\n"
          "--- !!seq"          "\n"
          "- !!int 0"          "\n"
          "- !!str 1"          "\n"
          "");
    ryml::Tree normalized_tree_long = tree;
    normalized_tree_long.normalize_tags_long(); // normalize all tags in short form
    CHECK(ryml::emitrs_yaml<std::string>(normalized_tree_long) == ""
          "--- !<tag:yaml.org,2002:map>"          "\n"
          "a: 0"                                  "\n"
          "b: 1"                                  "\n"
          "--- !map"                              "\n"
          "a: b"                                  "\n"
          "--- !<tag:yaml.org,2002:seq>"          "\n"
          "- a"                                   "\n"
          "- b"                                   "\n"
          "--- !<tag:yaml.org,2002:str> a b"      "\n"
          "--- !<tag:yaml.org,2002:str> 'a: b'"   "\n"
          "---"                                   "\n"
          "!<tag:yaml.org,2002:str> a: b"         "\n"
          "--- !<tag:yaml.org,2002:set>"          "\n"
          "a: "                                   "\n"
          "b: "                                   "\n"
          "--- !<tag:yaml.org,2002:set>"          "\n"
          "a: "                                   "\n"
          "--- !<tag:yaml.org,2002:seq>"          "\n"
          "- !<tag:yaml.org,2002:int> 0"          "\n"
          "- !<tag:yaml.org,2002:str> 1"          "\n"
          "");
}

sample_tag_directives()

void sample_tag_directives

(

)

Definition at line 5507 of file quickstart.cpp.

{
    const std::string yaml = ""
        "%TAG !m! !my-"             "\n"
        "--- # Bulb here"           "\n"
        "!m!light fluorescent"      "\n"
        "..."                       "\n"
        "%TAG !m! !meta-"           "\n"
        "--- # Color here"          "\n"
        "!m!light green"            "\n"
        "";
    // tags are not resolved by default:
    ryml::Tree tree = ryml::parse_in_arena(ryml::to_csubstr(yaml));
    CHECK(ryml::emitrs_yaml<std::string>(tree) == ""
          "%TAG !m! !my-"               "\n"
          "--- !m!light fluorescent"    "\n"
          "..."                         "\n"
          "%TAG !m! !meta-"             "\n"
          "--- !m!light green"          "\n"
          "");
    // Use Tree::resolve_tags() to accomplish this in an existing
    // tree:
    ryml::TagCache tag_cache; // reduces memory requirements by reusing resolved tags
    tree.resolve_tags(tag_cache);
    CHECK(ryml::emitrs_yaml<std::string>(tree) == ""
          "%TAG !m! !my-"                    "\n"
          "--- !<!my-light> fluorescent"     "\n"
          "..."                              "\n"
          "%TAG !m! !meta-"                  "\n"
          "--- !<!meta-light> green"         "\n"
          "");
    // You can also Use ParserOptions to force resolution of tags
    // while parsing:
    ryml::ParserOptions opts = ryml::ParserOptions{}.resolve_tags(true);
    ryml::Tree resolved_tree = ryml::parse_in_arena(ryml::to_csubstr(yaml), opts);
    CHECK(ryml::emitrs_yaml<std::string>(resolved_tree) == ""
          "%TAG !m! !my-"                      "\n"
          "--- !<!my-light> fluorescent"       "\n"
          "..."                                "\n"
          "%TAG !m! !meta-"                    "\n"
          "--- !<!meta-light> green"           "\n"
          "");
    // see also tree.normalize_tags()
    // see also tree.normalize_tags_long()
}

sample_docs()

void sample_docs

(

)

Definition at line 5556 of file quickstart.cpp.

{
    ryml::csubstr yml = ""
        "---"        "\n"
        "a: 0"       "\n"
        "b: 1"       "\n"
        "---"        "\n"
        "c: 2"       "\n"
        "d: 3"       "\n"
        "---"        "\n"
        "- 4"        "\n"
        "- 5"        "\n"
        "- 6"        "\n"
        "- 7"        "\n"
        "";
    ryml::Tree tree = ryml::parse_in_arena(yml);
    CHECK(ryml::emitrs_yaml<std::string>(tree) == yml);
 
    // iteration through docs
    {
        // using the node API
        const ryml::ConstNodeRef stream = tree.rootref();
        CHECK(stream.is_root());
        CHECK(stream.is_stream());
        CHECK(!stream.is_doc());
        CHECK(stream.num_children() == 3);
        for(const ryml::ConstNodeRef doc : stream.children())
            CHECK(doc.is_doc());
        CHECK(tree.docref(0).id() == stream.child(0).id());
        CHECK(tree.docref(1).id() == stream.child(1).id());
        CHECK(tree.docref(2).id() == stream.child(2).id());
        // equivalent: using the lower level index API
        const ryml::id_type stream_id = tree.root_id();
        CHECK(tree.is_root(stream_id));
        CHECK(tree.is_stream(stream_id));
        CHECK(!tree.is_doc(stream_id));
        CHECK(tree.num_children(stream_id) == 3);
        for(ryml::id_type doc_id = tree.first_child(stream_id);
            doc_id != ryml::NONE;
            doc_id = tree.next_sibling(stream_id))
            CHECK(tree.is_doc(doc_id));
        CHECK(tree.doc(0) == tree.child(stream_id, 0));
        CHECK(tree.doc(1) == tree.child(stream_id, 1));
        CHECK(tree.doc(2) == tree.child(stream_id, 2));
 
        // using the node API
        CHECK(stream[0].is_doc());
        CHECK(stream[0].is_map());
        CHECK(stream[0]["a"].val() == "0");
        CHECK(stream[0]["b"].val() == "1");
        // equivalent: using the index API
        const ryml::id_type doc0_id = tree.first_child(stream_id);
        CHECK(tree.is_doc(doc0_id));
        CHECK(tree.is_map(doc0_id));
        CHECK(tree.val(tree.find_child(doc0_id, "a")) == "0");
        CHECK(tree.val(tree.find_child(doc0_id, "b")) == "1");
 
        // using the node API
        CHECK(stream[1].is_doc());
        CHECK(stream[1].is_map());
        CHECK(stream[1]["c"].val() == "2");
        CHECK(stream[1]["d"].val() == "3");
        // equivalent: using the index API
        const ryml::id_type doc1_id = tree.next_sibling(doc0_id);
        CHECK(tree.is_doc(doc1_id));
        CHECK(tree.is_map(doc1_id));
        CHECK(tree.val(tree.find_child(doc1_id, "c")) == "2");
        CHECK(tree.val(tree.find_child(doc1_id, "d")) == "3");
 
        // using the node API
        CHECK(stream[2].is_doc());
        CHECK(stream[2].is_seq());
        CHECK(stream[2][0].val() == "4");
        CHECK(stream[2][1].val() == "5");
        CHECK(stream[2][2].val() == "6");
        CHECK(stream[2][3].val() == "7");
        // equivalent: using the index API
        const ryml::id_type doc2_id = tree.next_sibling(doc1_id);
        CHECK(tree.is_doc(doc2_id));
        CHECK(tree.is_seq(doc2_id));
        CHECK(tree.val(tree.child(doc2_id, 0)) == "4");
        CHECK(tree.val(tree.child(doc2_id, 1)) == "5");
        CHECK(tree.val(tree.child(doc2_id, 2)) == "6");
        CHECK(tree.val(tree.child(doc2_id, 3)) == "7");
    }
 
    // Note: ryml emits streams as a JSON seq by default:
    CHECK(ryml::emitrs_json<std::string>(tree) ==
          "["              "\n"
          "  {"            "\n"
          "    \"a\": 0,"  "\n"
          "    \"b\": 1"   "\n"
          "  },"           "\n"
          "  {"            "\n"
          "    \"c\": 2,"  "\n"
          "    \"d\": 3"   "\n"
          "  },"           "\n"
          "  ["            "\n"
          "    4,"         "\n"
          "    5,"         "\n"
          "    6,"         "\n"
          "    7"          "\n"
          "  ]"            "\n"
          "]\n"
          "");
    // ... but you can use EmitOptions{} to set the emitter to fail if
    // it finds a stream:
    {
        ScopedErrorHandlerExample errh; // calls ryml::set_callbacks()
        ryml::Tree err_tree = ryml::parse_in_arena(yml);
        CHECK(err_tree.callbacks() == errh.callbacks());
        auto err_opts = ryml::EmitOptions{}.json_err_on_stream(true);
        CHECK(errh.check_error_occurs([&]{
            return ryml::emitrs_json<std::string>(err_tree, err_opts);
        }));
        // in which case, you can avoid the error by emitting the
        // documents one-by-one:
        const std::string expected_json[] = {
            "{"            "\n"
            "  \"a\": 0,"  "\n"
            "  \"b\": 1"   "\n"
            "}"            "\n"
            ,
            "{"            "\n"
            "  \"c\": 2,"  "\n"
            "  \"d\": 3"   "\n"
            "}"            "\n"
            ,
            "["            "\n"
            "  4,"         "\n"
            "  5,"         "\n"
            "  6,"         "\n"
            "  7"          "\n"
            "]"            "\n"
        };
        ryml::id_type count = 0;
        const ryml::ConstNodeRef stream = err_tree;
        CHECK(stream.num_children() == (ryml::id_type)C4_COUNTOF(expected_json));
        for(ryml::ConstNodeRef doc : stream.children())
            CHECK(ryml::emitrs_json<std::string>(doc, err_opts) == expected_json[count++]);
    }
}

sample_error_handler()

void sample_error_handler

(

)

demonstrates how to set a custom error handler for ryml

Definition at line 5720 of file quickstart.cpp.

{
    ErrorHandlerExample errh; // browse the implementation of this
                              // class to understand more details
    errh.check_disabled();
    // set the global error handlers. Note the error callbacks must
    // never return: they must either throw an exception, use setjmp()
    // and longjmp(), or abort. Otherwise, the parser will enter into
    // an infinite loop, or the program may crash.
    ryml::set_callbacks(errh.callbacks());
    errh.check_enabled();
    CHECK(errh.check_error_occurs([&]{
        ryml::Tree tree = ryml::parse_in_arena("errorhandler.yml", "[a: b\n}");
    }));
    ryml::set_callbacks(errh.original_callbacks); // restore defaults.
    errh.check_disabled();
}

sample_error_basic()

void sample_error_basic

(

)

Definition at line 5741 of file quickstart.cpp.

{
    auto cause_basic_error = []{
        ryml::Tree t;
        ryml::csubstr tag_handle = {}, tag_prefix = {}; // invalid, not filled
        t.add_tag_directive(tag_handle, tag_prefix, 0);
    };
    {
        ScopedErrorHandlerExample errh; // set the example callbacks (scoped)
        CHECK(errh.check_error_occurs(cause_basic_error));
    }
#ifdef _RYML_WITH_EXCEPTIONS
    bool gotit = false;
    try
    {
        cause_basic_error();
    }
    catch(ryml::ExceptionBasic const& exc)
    {
        gotit = true;
        ryml::csubstr msg = ryml::to_csubstr(exc.what());
        CHECK(!exc.errdata_basic.location.name.empty());
        CHECK(!msg.empty());
    }
    CHECK(gotit);
#endif
}

sample_error_parse()

void sample_error_parse

(

)

Definition at line 5770 of file quickstart.cpp.

{
    ryml::csubstr ymlsrc = ""
        "{"       "\n"
        "  a: b"  "\n"
        "   ["    "\n"
        "";
    ryml::csubstr ymlfile = "file.yml";
    auto cause_parse_error = [&]{
        return ryml::parse_in_arena(ymlfile, ymlsrc);
    };
    // the YAML in ymlsrc must cause a parse error while it is being
    // parsed. We use our error handler to catch that error, and save
    // the error info:
    ErrorHandlerExample errh;
    {
        ryml::set_callbacks(errh.callbacks());
        CHECK(errh.check_error_occurs(cause_parse_error));
        // the handler in errh saves the error info in itself. Let's
        // use that to see the messages we get.
        //
        // this message is the short message passed into the parse
        // error handler:
        CHECK(errh.saved_msg_short == "invalid character: '['");
        // this message was created inside the handler, by calling
        // ryml::err_parse_format():
        CHECK(ryml::to_csubstr(errh.saved_msg_full).begins_with("file.yml:3: col=4 (12B): ERROR: [parse] invalid character: '['"));
        // If you keep the YAML source buffer around, you can also use
        // it to create/print a larger error message showing the
        // YAML source code context which causes the error:
        std::string msg_ctx = errh.saved_msg_full + "\n";
        ryml::location_format_with_context([&msg_ctx](ryml::csubstr s){
            msg_ctx.append(s.str, s.len);
        }, errh.saved_parse_loc, ymlsrc, "err");
        CHECK(ryml::to_csubstr(msg_ctx).begins_with("file.yml:3: col=4 (12B): ERROR: [parse] invalid character: '['"));
        CHECK(ryml::to_csubstr(msg_ctx).ends_with(
                  "file.yml:3: col=4 (12B): err:"     "\n"
                  "err:"                              "\n"
                  "err:        ["                     "\n"
                  "err:         |"                    "\n"
                  "err:         (here)"               "\n"
                  "err:"                              "\n"
                  "err: see region:"                  "\n"
                  "err:"                              "\n"
                  "err:     {"                        "\n"
                  "err:       a: b"                   "\n"
                  "err:        ["                     "\n"
                  "err:         |"                    "\n"
                  "err:         (here)"               "\n"
                  ""));
        //
        // Let's now check the location (see the message above):
        CHECK(errh.saved_parse_loc.name == ymlfile);
        CHECK(errh.saved_parse_loc.line == 3);
        CHECK(errh.saved_parse_loc.col == 4);
        CHECK(errh.saved_parse_loc.offset == 12);
        CHECK(errh.saved_parse_loc.offset <= ymlsrc.len);
        // ... and this is the location in the ryml source code file where
        // this error was found:
        CHECK(!errh.saved_basic_loc.name.empty());
        CHECK(errh.saved_basic_loc.line > 0);
        CHECK(errh.saved_basic_loc.col > 0);
        CHECK(errh.saved_basic_loc.offset > 0);
        ryml::set_callbacks(errh.original_callbacks);
    }
    // A parse error is also a basic error. If no parse error handler
    // is set, then ryml falls back to a basic error:
    {
        ryml::Callbacks cb = errh.callbacks();
        cb.m_error_parse = nullptr;
        ryml::set_callbacks(cb);
        CHECK(ryml::get_callbacks().m_error_parse == nullptr);
        CHECK(errh.check_error_occurs(cause_parse_error));
        // we got a basic error instead of a parse error:
        CHECK(errh.saved_msg_short == "invalid character: '['");
        // notice that the full message now displays this as a basic
        // error:
        CHECK(errh.saved_msg_full == "file.yml:3: col=4 (12B): ERROR: [basic] invalid character: '['");
        // the yml location is now in the location saved from the basic error
        CHECK(errh.saved_basic_loc.name == ymlfile);
        CHECK(errh.saved_basic_loc.line == 3);
        CHECK(errh.saved_basic_loc.col == 4);
        CHECK(errh.saved_basic_loc.offset == 12);
        CHECK(errh.saved_basic_loc.offset <= ymlsrc.len);
        CHECK(errh.saved_parse_loc.name == ryml::csubstr{});
        CHECK(errh.saved_parse_loc.line == ryml::csubstr::npos);
        CHECK(errh.saved_parse_loc.col == ryml::csubstr::npos);
        CHECK(errh.saved_parse_loc.offset == ryml::csubstr::npos);
        ryml::set_callbacks(errh.original_callbacks);
    }
#ifdef _RYML_WITH_EXCEPTIONS
    bool gotit = false;
    try
    {
        cause_parse_error();
    }
    catch(ryml::ExceptionParse const& exc)
    {
        gotit = true;
        ryml::csubstr msg = ryml::to_csubstr(exc.what());
        CHECK(exc.errdata_parse.ymlloc.name == ymlfile);
        CHECK(exc.errdata_parse.ymlloc.line == 3);
        CHECK(exc.errdata_parse.ymlloc.col == 4);
        CHECK(exc.errdata_parse.ymlloc.offset == 12);
        CHECK(exc.errdata_parse.ymlloc.offset <= ymlsrc.len);
        // the message saved in the exception is just the concrete error description:
        CHECK(msg == "invalid character: '['");
        // to print richer error messages, ryml provides helpers to
        // format that description into a complete error message,
        // containing location and source context indication:
        std::string full;
        auto dumpfn = [&full](ryml::csubstr s) { full.append(s.str, s.len); };
        ryml::err_parse_format(dumpfn, msg, exc.errdata_parse);
        full += '\n';
        ryml::location_format_with_context(dumpfn, exc.errdata_parse.ymlloc, ymlsrc, "err", 3);
        CHECK(ryml::to_csubstr(full).begins_with("file.yml:3: col=4 (12B): ERROR: [parse] invalid character: '['"));
        CHECK(ryml::to_csubstr(full).ends_with(
                  "file.yml:3: col=4 (12B): err:"      "\n"
                  "err:"                               "\n"
                  "err:        ["                      "\n"
                  "err:         |"                     "\n"
                  "err:         (here)"                "\n"
                  "err:"                               "\n"
                  "err: see region:"                   "\n"
                  "err:"                               "\n"
                  "err:     {"                         "\n"
                  "err:       a: b"                    "\n"
                  "err:        ["                      "\n"
                  "err:         |"                     "\n"
                  "err:         (here)"                "\n"
                  ""));
    }
    CHECK(gotit);
    gotit = false;
    try
    {
        cause_parse_error();
    }
    catch(ryml::ExceptionBasic const& exc) // use references! don't slice the exception
    {
        gotit = true;
        ryml::csubstr msg = ryml::to_csubstr(exc.what());
        CHECK(!exc.errdata_basic.location.name.empty());
        CHECK(!msg.empty());
    }
    CHECK(gotit);
#endif
}

sample_error_visit()

void sample_error_visit

(

)

Visit errors happen when an error is triggered while reading from a node.

Definition at line 5922 of file quickstart.cpp.

{
    ryml::csubstr ymlfile = "file.yml";
    ryml::csubstr ymlsrc = "float: 123.456";
    ErrorHandlerExample errh;
    {
        ryml::set_callbacks(errh.callbacks());
        ryml::Tree tree = ryml::parse_in_arena(ymlfile, ymlsrc);
        CHECK(errh.check_error_occurs([&]{
            int intval = 0;
            tree["float"] >> intval; // cannot deserialize 123.456 to int
        }));
        // the handler in errh saves the error info in itself. Let's
        // use that to see the messages we get.
        //
        // this message is the short message passed into the visit error
        CHECK(errh.saved_msg_short == "could not deserialize value");
        // this message was created inside the handler, by calling
        // ryml::err_visit_format():
        CHECK(ryml::csubstr::npos != ryml::to_csubstr(errh.saved_msg_full).find("ERROR: [visit] could not deserialize value"));
        // The location of the visit error is of the C++ source file where
        // the error was detected -- NOT of the YAML source file:
        CHECK(errh.saved_basic_loc.name != ymlfile);
        // However, note that the tree and node id are available:
        CHECK(errh.saved_visit_tree == &tree);
        CHECK(errh.saved_visit_id == tree["float"].id());
        // see sample_error_visit_location() for an example on how
        // to extract the location.
    }
    // visit errors also fall back to basic errors when the visit
    // handler is not set (similar to the behavior of ExceptionVisit):
    {
        ryml::Callbacks cb = errh.callbacks();
        cb.m_error_visit = nullptr;
        ryml::set_callbacks(cb);
        CHECK(ryml::get_callbacks().m_error_visit == nullptr);
        ryml::Tree tree = ryml::parse_in_arena(ymlfile, ymlsrc);
        CHECK(errh.check_error_occurs([&]{
            int intval = 0;
            tree["float"] >> intval; // cannot deserialize 123.456 to int
        }));
        // we got a basic error instead of a visit error:
        CHECK(errh.saved_msg_short == "could not deserialize value");
        // notice that the full message now displays this as a basic
        // error:
        CHECK(ryml::csubstr::npos != ryml::to_csubstr(errh.saved_msg_full).find("ERROR: [basic] could not deserialize value"));
        // the tree and id are not set, because this was called as a basic error
        CHECK(errh.saved_visit_tree == nullptr);
        CHECK(errh.saved_visit_id == ryml::NONE);
        ryml::set_callbacks(errh.original_callbacks);
    }
#ifdef _RYML_WITH_EXCEPTIONS
    // when using the default ryml callbacks (see
    // RYML_NO_DEFAULT_CALLBACKS), and
    // RYML_DEFAULT_CALLBACK_USES_EXCEPTIONS is defined, the ryml
    // parse handler throws an exception of type ryml::ExceptionVisit,
    // which is derived from ryml::ExceptionBasic.
    {
        const ryml::Tree tree = ryml::parse_in_arena(ymlfile, ymlsrc);
        bool gotit = false;
        try
        {
            int intval = 0;
            tree["float"] >> intval; // cannot deserialize 123.456 to int
        }
        catch(ryml::ExceptionVisit const& exc)
        {
            gotit = true;
            ryml::csubstr msg = ryml::to_csubstr(exc.what());
            CHECK(!exc.errdata_visit.cpploc.name.empty());
            CHECK(exc.errdata_visit.tree == &tree);
            CHECK(exc.errdata_visit.node == tree["float"].id());
            CHECK(!msg.empty());
        }
        CHECK(gotit);
    }
    // you can also catch the exception as its base,
    // ryml::ExceptionBasic:
    {
        const ryml::Tree tree = ryml::parse_in_arena(ymlfile, ymlsrc);
        bool gotit = false;
        try
        {
            int intval = 0;
            tree["float"] >> intval; // cannot deserialize 123.456 to int
        }
        catch(ryml::ExceptionBasic const& exc) // use references! don't slice the exception
        {
            gotit = true;
            ryml::csubstr msg = ryml::to_csubstr(exc.what());
            CHECK(!exc.errdata_basic.location.name.empty());
            CHECK(!msg.empty());
        }
        CHECK(gotit);
    }
#endif
}

sample_error_visit_location()

void sample_error_visit_location

(

)

It is possible to obtain the YAML location from a visit error: when the tree is obtained from parsing YAML, the messages may be enriched by using a parser set to track the locations.

See sample_location_tracking() for more details on how to use locations.

Definition at line 6026 of file quickstart.cpp.

{
    ScopedErrorHandlerExample errh;
    // we will use locations to show the YAML source context of the
    // node where the visit error was triggered. This is a very
    // convenient feature to show detailed messages when deserializing
    // data read from a file (but do note this is opt-in, and it is
    // not mandatory). See sample_location_tracking() for more details
    // on location tracking.
    ryml::ParserOptions opts = ryml::ParserOptions{}.locations(true);
    ryml::EventHandlerTree evt_handler{};
    ryml::Parser parser(&evt_handler, opts);
    ryml::csubstr ymlfile = "file.yml";
    ryml::csubstr ymlsrc = ""
        "foo: bar"           "\n"
        "char: a"            "\n"
        "int: a"             "\n"
        "float: 123.456"     "\n"
        "";
    const ryml::Tree tree = ryml::parse_in_arena(&parser, ymlfile, ymlsrc);
    // This function will cause a visit error when being called:
    auto cause_visit_error = [&]{
        int intval = 0;
        tree["float"] >> intval; // cannot deserialize 123.456 to int
    };
    // Like with the parse error, we will use our error handler to
    // catch that visit error, and save the error info:
    CHECK(evt_handler.callbacks() == errh.callbacks());
    CHECK(parser.callbacks() == errh.callbacks());
    CHECK(tree.callbacks() == errh.callbacks());
    {
        CHECK(errh.check_error_occurs(cause_visit_error));
        // the handler in errh saves the error info in itself. Let's
        // use that to see the messages we get.
        //
        // this message is the short message passed into the visit error
        CHECK(errh.saved_msg_short == "could not deserialize value");
        // this message was created inside the handler, by calling
        // ryml::err_visit_format():
        CHECK(ryml::csubstr::npos != ryml::to_csubstr(errh.saved_msg_full).find("ERROR: [visit] could not deserialize value"));
        // The location of the visit error is of the C++ source file where
        // the error was detected -- NOT of the YAML source file:
        CHECK(errh.saved_basic_loc.name != ymlfile);
        // However, note that the tree and node id are available:
        CHECK(errh.saved_visit_tree == &tree);
        CHECK(errh.saved_visit_id == tree["float"].id());
        // ... which we can use to get the location in the YAML source
        // from the parser (but see @ref sample_location_tracking()):
        ryml::Location ymlloc = errh.saved_visit_tree->location(parser, errh.saved_visit_id);
        CHECK(ymlloc.name == ymlfile);
        // In turn, we can use format_location_context() to
        // print/create an error message pointing at the YAML source
        // code:
        std::string msg = errh.saved_msg_full;
        ryml::location_format_with_context([&msg](ryml::csubstr s){
            msg.append(s.str, s.len);
        }, ymlloc, ymlsrc, "err", /*number of lines to show before the error*/3);
        CHECK(ryml::to_csubstr(msg).ends_with(
                  "file.yml:3: col=0 (24B): err:"     "\n"
                  "err:"                              "\n"
                  "err:     float: 123.456"           "\n"
                  "err:     |"                        "\n"
                  "err:     (here)"                   "\n"
                  "err:"                              "\n"
                  "err: see region:"                  "\n"
                  "err:"                              "\n"
                  "err:     foo: bar"                 "\n"
                  "err:     char: a"                  "\n"
                  "err:     int: a"                   "\n"
                  "err:     float: 123.456"           "\n"
                  "err:     |"                        "\n"
                  "err:     (here)"                   "\n"
                  ""));
    }
}

sample_global_allocator()

void sample_global_allocator

(

)

demonstrates how to set the global allocator for ryml

Definition at line 6200 of file quickstart.cpp.

{
    GlobalAllocatorExample mem;
 
    // save the existing callbacks for restoring
    ryml::Callbacks defaults = ryml::get_callbacks();
 
    // set to our callbacks
    ryml::set_callbacks(mem.callbacks());
 
    // verify that the allocator is in effect
    ryml::Callbacks const& current = ryml::get_callbacks();
    CHECK(current.m_allocate == &mem.s_allocate);
    CHECK(current.m_free == &mem.s_free);
 
    // so far nothing was allocated
    CHECK(mem.alloc_size == 0);
 
    // parse one tree and check
    (void)ryml::parse_in_arena(R"({foo: bar})");
    mem.check_and_reset();
 
    // parse another tree and check
    (void)ryml::parse_in_arena(R"([a, b, c, d, {foo: bar, money: pennys}])");
    mem.check_and_reset();
 
    // verify that by reserving we save allocations
    {
        ryml::EventHandlerTree evt_handler;
        ryml::Parser parser(&evt_handler); // reuse a parser
        ryml::Tree tree;          // reuse a tree
 
        tree.reserve(10);         // reserve the number of nodes
        tree.reserve_arena(100);  // reserve the arena size
        parser.reserve_stack(10); // reserve the parser depth.
 
        // since the parser stack uses Small Storage Optimization,
        // allocations will only happen with capacities higher than 16.
        CHECK(mem.num_allocs == 2); // tree, tree_arena and NOT the parser
 
        parser.reserve_stack(20); // reserve the parser depth.
        CHECK(mem.num_allocs == 3); // tree, tree_arena and now the parser as well
 
        // verify that no other allocations occur when parsing
        size_t size_before = mem.alloc_size;
        parse_in_arena(&parser, "", R"([a, b, c, d, {foo: bar, money: pennys}])", &tree);
        CHECK(mem.alloc_size == size_before);
        CHECK(mem.num_allocs == 3);
    }
    mem.check_and_reset();
 
    // restore defaults.
    ryml::set_callbacks(defaults);
}

sample_per_tree_allocator()

void sample_per_tree_allocator

(

)

Definition at line 6323 of file quickstart.cpp.

{
    PerTreeMemoryExample mrp;
    PerTreeMemoryExample mr1;
    PerTreeMemoryExample mr2;
 
    // the trees will use the memory in the resources above,
    // with each tree using a separate resource
    {
        // Watchout: ensure that the lifetime of the callbacks target
        // exceeds the lifetime of the tree.
        ryml::EventHandlerTree evt_handler(mrp.callbacks());
        ryml::Parser parser(&evt_handler);
        ryml::Tree tree1(mr1.callbacks());
        ryml::Tree tree2(mr2.callbacks());
 
        ryml::csubstr yml1 = "{a: b}";
        ryml::csubstr yml2 = "{c: d, e: f, g: [h, i, 0, 1, 2, 3]}";
 
        parse_in_arena(&parser, "file1.yml", yml1, &tree1);
        parse_in_arena(&parser, "file2.yml", yml2, &tree2);
    }
 
    CHECK(mrp.num_allocs == 0); // YAML depth not large enough to warrant a parser allocation
    CHECK(mr1.alloc_size <= mr2.alloc_size); // because yml2 has more nodes
}

sample_static_trees()

void sample_static_trees

(

)

shows how to work around the static initialization order fiasco when using a static-duration ryml tree

See also

https://en.cppreference.com/w/cpp/language/siof

Definition at line 6357 of file quickstart.cpp.

{
    // Static trees may incur a static initialization order
    // problem. This happens because a default-constructed tree will
    // obtain the callbacks from the current global setting, which may
    // not have been initialized due to undefined static
    // initialization order:
    //
    // ERROR! depends on ryml::get_callbacks() which may not have been initialized.
    //static ryml::Tree tree;
    //
    // To work around the issue, declare static callbacks
    // to explicitly initialize the static tree:
    static ryml::Callbacks callbacks = default_callbacks(); // use default callback members
    static ryml::Tree tree(callbacks); // OK
    // now you can use the tree as normal:
    ryml::parse_in_arena(R"(doe: "a deer, a female deer")", &tree);
    CHECK(tree["doe"].val() == "a deer, a female deer");
}

sample_location_tracking()

void sample_location_tracking

(

)

demonstrates how to obtain the (zero-based) location of a node from a recently parsed tree

Definition at line 6383 of file quickstart.cpp.

{
    // NOTE: locations are zero-based. If you intend to show the
    // location to a human user, you may want to pre-increment the line
    // and column by 1.
    ryml::csubstr yaml = ""
        "{"                          "\n"
        "aa: contents,"              "\n"
        "foo: [one, [two, three]]"   "\n"
        "}"                          "\n"
        "";
    // A parser is needed to track locations, and it has to be
    // explicitly set to do it. Location tracking is disabled by
    // default.
    ryml::ParserOptions opts = {};
    opts.locations(true); // enable locations, default is false
    ryml::EventHandlerTree evt_handler = {};
    ryml::Parser parser(&evt_handler, opts);
    CHECK(parser.options().locations());
    // When locations are enabled, the first task while parsing will
    // consist of building and caching (in the parser) a
    // source-to-node lookup structure to accelerate location lookups.
    //
    // The cost of building the location accelerator is linear in the
    // size of the source buffer. This increased cost is the reason
    // for the opt-in requirement. When locations are disabled there
    // is no cost.
    //
    // Building the location accelerator may trigger an allocation,
    // but this can and should be avoided by reserving prior to
    // parsing:
    parser.reserve_locations(50u); // reserve for 50 lines
    // Now the structure will be built during parsing:
    ryml::Tree tree = parse_in_arena(&parser, "source.yml", yaml);
    // After this, we are ready to query the location from the parser:
    ryml::Location loc = tree.rootref().location(parser);
    // As for the complexity of the query: for large buffers it is
    // O(log(numlines)). For short source buffers (30 lines and less),
    // it is O(numlines), as a plain linear search is faster in this
    // case.
    CHECK(parser.location_contents(loc).begins_with("{"));
    CHECK(loc.offset == 0u);
    CHECK(loc.line == 0u);
    CHECK(loc.col == 0u);
    // on the next call, we only pay O(log(numlines)) because the
    // rebuild is already available:
    loc = tree["aa"].location(parser);
    CHECK(parser.location_contents(loc).begins_with("aa"));
    CHECK(loc.offset == 2u);
    CHECK(loc.line == 1u);
    CHECK(loc.col == 0u);
    // KEYSEQ in flow style: points at the key
    loc = tree["foo"].location(parser);
    CHECK(parser.location_contents(loc).begins_with("foo"));
    CHECK(loc.offset == 16u);
    CHECK(loc.line == 2u);
    CHECK(loc.col == 0u);
    loc = tree["foo"][0].location(parser);
    CHECK(parser.location_contents(loc).begins_with("one"));
    CHECK(loc.line == 2u);
    CHECK(loc.col == 6u);
    // SEQ in flow style: location points at the opening '[' (there's no key)
    loc = tree["foo"][1].location(parser);
    CHECK(parser.location_contents(loc).begins_with("["));
    CHECK(loc.line == 2u);
    CHECK(loc.col == 11u);
    loc = tree["foo"][1][0].location(parser);
    CHECK(parser.location_contents(loc).begins_with("two"));
    CHECK(loc.line == 2u);
    CHECK(loc.col == 12u);
    loc = tree["foo"][1][1].location(parser);
    CHECK(parser.location_contents(loc).begins_with("three"));
    CHECK(loc.line == 2u);
    CHECK(loc.col == 17u);
    // NOTE. The parser locations always point at the latest buffer to
    // be parsed with the parser object, so they must be queried using
    // the corresponding latest tree to be parsed. This means that if
    // the parser is reused, earlier trees will loose the possibility
    // of querying for location. It is undefined behavior to query the
    // parser for the location of a node from an earlier tree:
    ryml::Tree docval = parse_in_arena(&parser, "docval.yaml", "this is a docval");
    // From now on, none of the locations from the previous tree can
    // be queried:
    //loc = tree.rootref().location(parser); // ERROR, undefined behavior
    loc = docval.rootref().location(parser); // OK. this is the latest tree from this parser
    CHECK(parser.location_contents(loc).begins_with("this is a docval"));
    CHECK(loc.line == 0u);
    CHECK(loc.col == 0u);
 
    // NOTES ABOUT CONTAINER LOCATIONS
    ryml::Tree tree2 = parse_in_arena(&parser, "containers.yaml",
        ""                            "\n"
        "a new: buffer"               "\n"
        "to: be parsed"               "\n"
        "map with key:"               "\n"
        "  first: value"              "\n"
        "  second: value"             "\n"
        "seq with key:"               "\n"
        "  - first value"             "\n"
        "  - second value"            "\n"
        "  -"                         "\n"
        "    - nested first value"    "\n"
        "    - nested second value"   "\n"
        "  -"                         "\n"
        "    nested first: value"     "\n"
        "    nested second: value"    "\n"
        "");
    // (Likewise, the docval tree can no longer be used to query.)
    //
    // For key-less block-style maps, the location of the container
    // points at the first child's key. For example, in this case
    // the root does not have a key, so its location is taken
    // to be at the first child:
    loc = tree2.rootref().location(parser);
    CHECK(parser.location_contents(loc).begins_with("a new"));
    CHECK(loc.offset == 1u);
    CHECK(loc.line == 1u);
    CHECK(loc.col == 0u);
    // note the first child points exactly at the same place:
    loc = tree2["a new"].location(parser);
    CHECK(parser.location_contents(loc).begins_with("a new"));
    CHECK(loc.offset == 1u);
    CHECK(loc.line == 1u);
    CHECK(loc.col == 0u);
    loc = tree2["to"].location(parser);
    CHECK(parser.location_contents(loc).begins_with("to"));
    CHECK(loc.line == 2u);
    CHECK(loc.col == 0u);
    // but of course, if the block-style map is a KEYMAP, then the
    // location is the map's key, and not the first child's key:
    loc = tree2["map with key"].location(parser);
    CHECK(parser.location_contents(loc).begins_with("map with key"));
    CHECK(loc.line == 3u);
    CHECK(loc.col == 0u);
    loc = tree2["map with key"]["first"].location(parser);
    CHECK(parser.location_contents(loc).begins_with("first"));
    CHECK(loc.line == 4u);
    CHECK(loc.col == 2u);
    loc = tree2["map with key"]["second"].location(parser);
    CHECK(parser.location_contents(loc).begins_with("second"));
    CHECK(loc.line == 5u);
    CHECK(loc.col == 2u);
    // same thing for KEYSEQ:
    loc = tree2["seq with key"].location(parser);
    CHECK(parser.location_contents(loc).begins_with("seq with key"));
    CHECK(loc.line == 6u);
    CHECK(loc.col == 0u);
    loc = tree2["seq with key"][0].location(parser);
    CHECK(parser.location_contents(loc).begins_with("first value"));
    CHECK(loc.line == 7u);
    CHECK(loc.col == 4u);
    loc = tree2["seq with key"][1].location(parser);
    CHECK(parser.location_contents(loc).begins_with("second value"));
    CHECK(loc.line == 8u);
    CHECK(loc.col == 4u);
    // SEQ nested in SEQ: container location points at the first child's "- " dash
    loc = tree2["seq with key"][2].location(parser);
    CHECK(parser.location_contents(loc).begins_with("- nested first value"));
    CHECK(loc.line == 10u);
    CHECK(loc.col == 4u);
    loc = tree2["seq with key"][2][0].location(parser);
    CHECK(parser.location_contents(loc).begins_with("nested first value"));
    CHECK(loc.line == 10u);
    CHECK(loc.col == 6u);
    // MAP nested in SEQ: same as above: point to key
    loc = tree2["seq with key"][3].location(parser);
    CHECK(parser.location_contents(loc).begins_with("nested first: "));
    CHECK(loc.line == 13u);
    CHECK(loc.col == 4u);
    loc = tree2["seq with key"][3][0].location(parser);
    CHECK(parser.location_contents(loc).begins_with("nested first: "));
    CHECK(loc.line == 13u);
    CHECK(loc.col == 4u);
}