Collections on TrustLang

Type system

Mon, 01 Jan 0001 00:00:00 +0000

Principles

TrustLang has a closed type system: any new data type necessarily inherits (extends) one of the already existing ones. This resembles the Object model in Java, but without physical inheritance (without boxing simple types into objects). This approach is the basis for generic programming, when a derived type inherits the set of operations of the “parent” type, and continues to work with all its descendants.

The language is dynamically typed: the type of a variable can be specified explicitly or left to automatic inference. The specified type is used for compatibility checks on assignment, for checking the type of a function’s return value and for controlling the placement in memory of native types. Automatic casting is possible only between compatible types; otherwise an explicit type conversion is required.

Dictionaries and sets (enumerations)

Mon, 01 Jan 0001 00:00:00 +0000

Working examples with the actual behavior of the current version can be run in the playground: dictionary, dictionary with mixed types, enumeration and variant.

Dictionary (:Dictionary)

A dictionary is a set of data of an arbitrary type with access to elements by an integer index or by the name of the element (if present) (it resembles both a tuple and a structure). A dictionary is always one-dimensional, but each element can contain an arbitrary number of elements of any type, including other dictionaries.

Classes

Mon, 01 Jan 0001 00:00:00 +0000

The data type :Class is similar to a dictionary, but all its fields must have names (access to properties by index is also possible). When creating an instance of a class, a new type is created that copies the properties and methods of all parents. Unlike a dictionary, the set of fields of a class is determined at compile time and can only be extended relative to the parent types.

Native types

Mon, 01 Jan 0001 00:00:00 +0000

Native data types

Native (machine) data can be both variables and functions, and the name of native objects starts with the percent sign “%”.

In the case of variables, this is binary data that is located in one continuous region of machine memory at a certain address and has a strictly defined format.

To use existing libraries, you must either import a native object, thereby creating a TrustLang object but with an implementation in another language, for example C/C++, or make a forward declaration of the native object.

Ranges, arrays and tensors

Mon, 01 Jan 0001 00:00:00 +0000

Ranges, arrays and (in perspective) tensors are container (collection) types for working with sequences and multidimensional data.

Ranges (:Range)

A range is a type, an approximate analog of a “generator” in Python: it alternately yields elements in a given interval with a given step. In the text, a range is specified by two or three elements separated by two dots: 1..5 — from one to five with the default step 1; the step is specified by the third element: 0..10..2. As the boundaries and the step, not only literals but also variables and rational numbers can be used: 0.1..$stop..0.1 — from 0.1 to $stop with step 0.1; 0..100:Rational — a range of rational numbers.