Whitespace

Unlike most languages, whitespace matters in Z#. Similar to Python and F#.

---

Spaces and Tabs

Spaces are used to separate code elements in order to make them distinct.

iftrue does not mean anything, while if true represents an if statement with a condition (true). The space separates the parts and makes them meaningful.

---

Indentation

Z# uses whitespace -indents- to signal scope. Where a lot of languages use {} to declare a scope, Z# attempts to minimize noise and do away with those characters.

An indent is represented by a sequence of spaces. Typically 4 spaces are used. The number of spaces must be consistent throughout a single file but can differ between files. The first indent that is encountered is taken as the template for all others to follow.

The start of the line represents the root scope. An indent is always placed below a statement or instruction that belongs to a higher scope (one up).

A simple example of a function and its implementation:

MyFuncion: ()
    implementation_here
    more_code_here

The function is at root scope and all of its implementation is one indent removed from that origin.

These scopes can of course be nested.

Here is an example of a function with nested scopes.

MyFuncion: ()
    if true
        conditional_code_here
    implementation_here

The ‘conditional code’ is only executed if the if branch is taken. The ‘implementation’ is always executed.

---

Scope Names

Names for the indentation levels that create scopes:

global-scope        // exports
<File>
    top-level (file-scope)

    <Function>
        function-scope

    <Type>
        type-scope

---

Breaking up lines

As a general rule: broken up lines continue on the next line one indent further than a new (child) scope would be: a double indent.

---

Literal Strings

No extra quotes are required. The new-line (EOL) character(s) and the indent white-space(s) will NOT become part of the string.

s := "this is a very long string to demonstrate
        how to break up long literal strings"

So in this example there is a trailing space after ‘demonstrate’ in order to separate that word from the next word ‘how’.

---

Functions

The function name cannot be split. Parameters can be spread out over multiple lines. The opening parenthesis is always next to the function name. The closing parenthesis can either be directly after the last parameter or at the beginning (same indent/scope as the function name) of a new line. Commas that separate the parameters go directly after its previous parameter and cannot be on the start of a new line (after a double indent).

my_very_long_function_name_with_lots_of_parameters: (
        p1: U8, p2: U16, p3: U32, p4: U64,
        p5: Str
)
    ...

Use new-lines instead of commas.

function_with_params_but_no_commas: (
        p1: U8
        p2: U16
        p3: U32
        p4: U64
        p5: Str
)
    ...

The () are becoming the list delimiters. Does not using commas also apply to general lists?

lst := (
        1
        2
        3
        4
        5
        6
)

// would be the same as
lst := (1, 2, 3, 4, 5, 6)

---

Merging lines

Use the -> operator to continue a line of code.

fn: ()
    // you cannot write this
    if true doStuff() -> else doNothing()
    // use this instead
    true ? doStuff() : doNothing()

    a := 42
    // useful for match case continuation
    x := match a
        Error -> "This is an error"
        U8 -> "This is an unsigned 8-bit.
        _ -> "I don't know what this is"

    // also used for lambda's
    spaces := x.Where((c) -> c = ' ')

---

Comments

Comments can really throw of the flow of a block. Therefor comments must start at the same indentation as the block.

Example of hard to read code:

// indentation block
fn: ()
    a := 42
    x := 101
// comment
    y := 2112    // unclear if this belongs to the block above the comment

Corrected example:

// indentation block
fn: ()
    a := 42
    x := 101
    // comment
    y := 2112

---

TBD

Allow {} as an escape of indent based parsing. When after certain keywords an opening { is detected, the indentation rules are dropped and the closing } ends the code block. This could be a nice balance between both types of block detection mechanisms.

{} could also be used as an inline block.