Main Repo   ·    Mirror   ·    Dub Package Registry   ·    Documentation

toml-foolery

Toml-foolery is a library for the D programming language which simplifies encoding and decoding of TOML-formatted data. There are no intermediate types that couple your code to this library — Toml-foolery parses TOML data directly into any struct, and structs can be converted into TOML.

Toml-foolery v2.1.0 is compatible with TOML v1.0.0.

Usage

The parseToml template function decodes TOML code into a D data structure. It may optionally receive a pre-made instance of that struct as its second argument.

The tomlify function encodes D data structure as TOML code. It accepts any struct and returns a string containing TOML data.

Note that toml-foolery doesn't do any file I/O.

Each field of the given struct becomes a TOML key-value pair, where the key is the name of the field in D. This can be customized by applying the @TomlName attribute to a field. The string passed to this attribute is the key name to look for when decoding TOML data, and the key name to write when encoding the struct into TOML.

Example

import std.datetime;
import std.file;
import std.stdio;

import toml_foolery;


void main()
{
    string tomlSource = `
        name = "Rincewind"
        dateOfBirth = 2032-10-23
        profession = "Wizard"

        [alma-mater]
        name = "Unseen University"
        city = "Ankh-Morpork"

        [[pets]]
        name = "Luggage"
        species = "Arca vulgaris"
        profession = "Thief"
    `;

    Person person = parseToml!Person(tomlSource);

    writeln(person);
    writeln();
    writeln("And back into TOML:");
    writeln();
    writeln(tomlify(person));
}

struct Person
{
    enum Profession
    {
        Professor,
        Witch,
        Wizard,
        Thief,
        Patrician,
        Death
    }

    struct Animal
    {
        string name;
        string species;
        Profession profession;
    }

    struct Establishment
    {
        string name;
        string city;
    }

    string name;
    Date dateOfBirth;
    Profession profession;

    @TomlName("alma-mater")
    Establishment almaMater;

    Animal[] pets;
}

Prints:

Person("Rincewind", 2032-Oct-23, Wizard, Establishment("Unseen University", "Ankh-Morpork"), [Animal("Luggage", "Arca vulgaris", Thief)])

And back into TOML:

name = "Rincewind"
dateOfBirth = 2032-10-23
profession = "Wizard"

[alma-mater]
name = "Unseen University"
city = "Ankh-Morpork"

[[pets]]
name = "Luggage"
species = "Arca vulgaris"
profession = "Thief"

TOML–D type correspondence

Type of field	Resulting TOML type
string, wstring, dstring, char, wchar, dchar	String
byte, short, int, long, ubyte, ushort, uint, ulong¹	Integer
float, double, real	Floating point
bool	Boolean
enum	String²
std.datetime.systime.SysTime	Offset Date-Time
std.datetime.date.DateTime	Local Date-Time³
std.datetime.date.Date	Local Date
std.datetime.date.TimeOfDay	Local Time³
Array of the above types	Array of the corresponding TOML type
struct	Table
Array of structs	Array of tables
Associative array	Table

¹ The TOML specification requires Integers to be in the range [long.min, long.max], so toml-foolery will throw a TomlTypeException if the input contains an integer outside that range.

² Parsing is case-sensitive.

³ The TOML specification expects at least millisecond precision for local date-times and local times, but the D standard library's corresponding data structures are precise only to the second. Any fractional-second precision will be lost upon parsing.

Notes

• If the parsed TOML data contains keys that don't match any fields in the given struct, toml-foolery will ignore those keys.
• If a destination field is a static array, but the corresponding array in the TOML code is too big to fit, additional entries will be ignored. Dynamic arrays will be resized as needed.
• The TOML specification allows mixed-type arrays, but toml-foolery does not yet support them.
• Toml-foolery does not support classes or pointers.
• Line-separators are preserved when decoding multiline strings.
• Toml-foolery can parse all features of TOML, but when encoding data into TOML, some formats will not be created:
  • Inline tables (Regular []-syntax is always used)
  • Dotted keys (same as above)
  • Literal strings (regular "strings" used instead)
  • Multiline strings (same as above)