Go Concepts - Lexical Structure

From NovaOrdis Knowledge Base
Jump to navigation Jump to search

TO DEPLETE

Source Files

Go code goes into source files. They have the .go extension.

Each source file is mandatory declared to belong to a package.

They can then optionally import code from other packages, using the import keyword.

Each source file has its own file block.

Thera are no restrictions on the file name, it is not required to have any relationship with the name of the package it contains code for.

Semicolons

Go programs may omit most of the semicolons at the end of line if the following conditions apply:

  1. When the input is broken into tokens, a semicolon is automatically inserted into the token stream immediately after a line's final token if that token is:
    1. an identifier
    2. an integer, floating-point, imaginary, rune, or string literal
    3. one of the keywords break, continue, fall through, or return
    4. one of the operators and delimiters ++, --, ), ], or }
  2. To allow complex statements to occupy a single line, a semicolon may be omitted before a closing ")" or "}".

For more details, see the specification:

https://golang.org/ref/spec#Semicolons

Blocks

A block is a sequence of declarations and statements within matching curly braces. The block may be empty. The most common blocks are those defined in the source code - they define block lexical scopes (see variable scopes). Aside from those, Go recognizes the following implicit blocks:

  • The universe block which contains all Go source text.
  • The package block.
  • The file block that contains all the source text in the file.
  • Each if, for and switch statements is considered to be in its own implicit block.
  • Each clause in a switch and select statements acts as an implicit block.

Identifiers

Identifiers name program entities, such as constants, variables, types, functions, packages, etc. An identifier is a sequence of one more letter and digits. The first character must be a letter. The letters are unicode letters or underscore "_". It must not be a keyword, keywords are reserved by the language.

Identifiers can be exported, and the way the first letter is capitalized is relevant to whether the identifier is exported or not. See exported identifiers.

It is idiomatic to use camel case style for Go identifiers.

Exported Identifiers

Exported Identifiers

Keywords

break default func select
case defer go map struct
chan else goto package switch
const fallthrough if type
continue for import return var

Literals

A literal is a notation for representing a fixed value in source code. Go provides literals for booleans, integers, floating-point values, strings, built-in types such as arrays, slices and maps, user-defined types (structs), functions, etc. The compiler is always able to infer the literal's type.

Declarations

A declaration binds an identifier to a constant, type, variable, function, label or package. Every identifier in the program must be declared. No identifier may be declared twice in the same block, and no identifier may be declared in both the file and package block.

Variables

A variable is a storage location in memory. Go being a statically typed language, a variable always has a specific type. A variable also has a name, which is referred to as variable identifier.

The Go compiler will generate an error if a variable is declared and not used.

Variable Declarations

Long Variable Declaration

This is the idiomatic way to specify we want a variable initialized with its zero value. If the variable is initialized to something else than the zero value, the short variable declaration should be used.

var <variable_identifier> <variable_type_literal> = <initial_value_literal>

Example:

var a string = "blue"
var b = "red"
var c string
var f = func (i int) int {
        return i
}

Declaring the type is optional if an initial value is specified, the compiler is able to infer the type of the literal, and the variable will be built to have that type.

Assigning a value is optional, if no value is explicitly assigned, the variable is initialized to the type's zero value. However, the type must be specified in this case.

Multiple variables are declared as such:

var (
     a = 1
     b = 2
     c = "something"
)

Short Variable Declaration

The short variable declaration uses the short variable declaration operator (:=). It implies a variable declaration (var) so it cannot be used for variable that have been already declared. It always assigns an initial value and also infers the variable type based on the literal type, so there is no need to declare the variable's type. The short variable declaration is the idiomatic way to declare variables that must be initialized to values other than the zero value. To initialize variables to their zero values, use the long variable declaration.

<variable_identifier> := <initial_value_literal>

The compiler uses the type of the right hand expression to infer the type of the variable being declared.

Example:

a := "blah"

This syntax can be used in a for loop.

Note that this form cannot be used for the declaration of package-level variable. These are other example where the short variable declaration does not work:

Short Variable Declaration Invalid Cases

When to use the Long and when to use the Short Declaration

Generally, the shorter value should be used whenever possible. Some sources claim that the long value is not "idiomatic".

Other sources suggest to use the long form when declaring variables that will be initialized to their zero value and to use the short form when providing extra initialization or making a function call.

For more details on variable names, see identifiers.

Variable Scopes

The range of places where a variable is visible from it is called the scope of the variable.

Go is lexically scoped using blocks.

A variable is visible within the block delimited by the nearest curly braces {}, but not outside them. The variable is visible inside any sub-blocks.

If a variable with the same name is defined in a sub-block, the that variable takes precedence, in the sub-block and its sub-blocks, over the variable declared in the upper block.

Package-Level Variable

A package-level variable is a variable declared outside the scope of any function.

An unexported variable declared at the package level (outside any function) is only visible throughout the package.

An exported variable declared at the package level is visible throughout the package and in all packages that import the exporting package.

Functions and Variable Scopes

Functions and Variable Scopes

Pointers

A pointer variable or a pointer is a pointer type variable. It contains an address to a memory location. Pointers are not commonly used with built-in types, but are very useful when paired with structs.

Pointers are useful when functions need to access and change state of a variable that was declared within the scope of a different function and possibly a different goroutine. It is convenient, but also dangerous if the variable is modified concurrently without mutual exclusion protection.

In absence of explicit initialization, pointer are initialized to nil.

On a 64-bit architecture, a pointer variable requires 8 bytes of memory.