Introduction

Expressions in hyperscript are a mix of familiar and new. Many of the expected operations from javascript are there:

  • Common literals: 1 true "this is a string" null
  • List and object literals: [1, 2, 3], {foo:"bar", [computeKey()]: computeValue()}

Most of the common comparison and mathematical operators are there as well:

  x > 10
  y == null
  lst.length < 5
  a + b < 42

There are some minor improvements to them:

  • Property names in object literals can contain hyphens. This is convenient for representing CSS:

    add { font-weight: bold } to the #element in me

However, once you get past the basics, hyperscript starts to get a little wild.

Let's start with CSS literals.

CSS Literals

Hyperscript gives you the ability to embed CSS literals directly in your code to select elements. There are four main expression types:

ID Literals

You can refer to an element by ID directly in hyperscript as follows:

<div _="on click put 'Clicked!' into #example.innerHTML">Click Me</div>
<div id="example"></div>

The #example is an ID literal and will evaluate to the element with the given id. Here we put some text into its innerHTML when the top div is clicked.

Class Literals

You can refer to a group of elements by class directly in hyperscript as follows:

<div _="on click put 'Clicked!' into .example.innerHTML">Click Me</div>
<div class="example"></div>
<div class="example"></div>

The #example is an ID literal and will evaluate all the elements with the class example on them. Here we put some text into their innerHTML when the top div is clicked. Note that the put command can work with collections as well as single values, so it can put the given value into all the returned elements.

Query Literals

You can refer to a group of elements by an arbitrary CSS selector by enclosing the selector in a < and />:

<div _="on click put 'Clicked!' into <div/>.innerHTML">Click Me</div>
<div class="example"></div>
<div class="example"></div>

This example will put "Clicked" into every div on the page!

<div _="on click put 'Clicked!' into <div:not(.example)/>.innerHTML">Click Me</div>
<div class="example"></div>
<div class="example"></div>

This example will put "Clicked" into every div that does not have the example class on it.

Attribute Literals

Finally, you can refer to an attribute with two syntaxes:

<div foo="bar" _="on click put @foo into me">Click Me</div>
<div _="on click toggle [@foo='bar'] into me">Click Me</div>

The short syntax, @<attribute name> can be used to get or set attribute values, and may be chained with possessives:

for anchor in <a/>
  log the anchor's @href
end

or with non-possessive property chains:

for anchor in <a/>
  log anchor@href
end

The longer syntax, surrounding the @<attribute-name> with square brackets, may be used for queries that require a value, or for commands like toggle or add that require a value

for anchor in [@href]
  log anchor@href
end

In Expressions

The in expression isn't a literal, but can be used in conjunction with them for common patterns:

<div _="on click put 'Clicked!' into (<p/> in me).innerHTML">Click Me
  <p></p>
  <p></p>
</div>

The in expression in this case evaluates the query in the given context on the right hand side. Here we are looking up paragraph tags inside the clicked element (me) and setting their innerHTML to "Clicked!"

Comparisons

In addition to the typical comparison operators, such as == and !=, hyperscript supports the following natural language aliases

  • is - equivalent to ==: 1 is 1
  • is not - equivalent to !=: 1 is not 1
  • am - equivalent to ==: I am 1
  • am not - equivalent to !=: I am not 1
  • no - equivalent to != null: no .example in me

Note that I is an alias for me, the current element.

Furthermore, hyperscript supports two more comparison operators: matches and contains which can be used in the following forms:

  I match <:hover/>
  it matches <:hover/>
  I do not match <:hover/>
  it does not match <:hover/>
  I contain <:focus/>
  it contains <:focus/>
  I do not contain <:focus/>
  it does not contain <:focus/>

Strings

Strings are similar to javascript, and can start with " or '.

<div _="on click set world to 'hyperscript' put 'Hello $world' into my.innerHTML">
  Click Me
</div>

In a few places, hyperscript allows "naked" strings, strings without a leading quote or double quote. An example is the fetch command, which can take a URL as a naked string:

<button _="on click fetch /example then put it into my.innerHTML">
    Fetch It!
</button>

Here the /example element is an example of a naked string. Naked strings are ended by whitespace.

Javascript string templates are supported by using the same syntax, enclosing backtics:

<button _="on click fetch /example then put `result: ${the result}` into my.innerHTML">
    Fetch It!
</button>

Possessive Expressions

The possessive expression is an expression that starts with my or its or a symbol followed by a `'s' and that is roughly the equivalent of a property acesss

<div _="on click put the window's location into me">
  Click Me
</div>

This is equivalent to:

<div _="on click put window.location into me">
  Click Me
</div>

You may also access and set DOM attributes using the possessive with attribute literals:

Click on Links To See The URL

Conversions

Hyperscript centralizes conversions into a single construct, the as expression:

  10 as String
  "10" as Int
  "10.3" as Float

Out of the box hyperscript provides the following conversions:

  • String - converts to string
  • Int - converts to an integer
  • Float - converts to a float
  • Number - converts to a number
  • Date - converts to a date

Null Safety

The . operator in hyperscript is null safe, so elt.parent will evaluate to null if elt is null

Array Expansion

Properties on arrays, except for length, are expanded via a flat map

  set divs to </div>                      -- get all divs in the document
  set divParents to divs.parentElement    -- get all parents of those divs
  set divChildren to divs.children        -- get all children of those divs

Blocks

Hyperscript does not have anonymous functions or complex arrow functions. Because hyperscript is async transparent complicated callbacks are generally not necessary.

However it does support a simple, expression-only version of arrow functions called "blocks", with a slightly different syntax:

    set strs to ["a", "ab", "abc"]
    set lens to strs.map( \ s -> s.length )

Blocks start with a backslash, followed by args, then an -> and then an expression value.

Async

By default, hyperscript synchronizes on any Promises that go through its runtime. Consider the following code:

<script type="text/hyperscript">
def waitThenReturn10()
    wait 10ms
    return 10
end
</script>
<button _="on click put 'The answer was $waitThenReturn10()' into my.innerHTML">
  Click Me...
</button>

Here we have an asynchronous function with a wait in it that will cause the function to return a Promise rather than the value. Hyperscript will "pause" evaluation of the event handler until that promise resolves and can provide a value to the string expression, and then continue. This is very cool and is usually what you want.

However, there may be a case where you don't want hyperscript to pause and, instead, want to pass the raw promise on somewhere else. To do this, you can use the async prefix for the expression

<button _="on click call handleAPromise(async waitThenReturn10())">
  Click Me...
</button>

Here we pass the promise returned by waitThenReturn10() out to handleAPromise(), as a Promise, rather than resolving it.

Time Expressions

In a few places in the hyperscript grammar you can use "time expressions", which are just natural ways to specify a time interval. The wait command is one such place:

<button _="on click wait 2s then put 'I waited!' into my.innerHTML">
  Click to Wait...
</button>

You can use the following formats for time expressions:

  • 10 ms - milliseconds
  • 10 milliseconds - milliseconds
  • 10 s - seconds
  • 10 seconds - seconds

Note that a space between the number and modifier are not necessary, and that the ms modifier is just for clarity since milliseconds is the normal interpretation for things like setTimeout(), which the wait command is based on.

Of Expressions

The of expression allows you to reverse the normal object oriented syntax and write logic in more natural english

  <button _="on click call window.location.refresh()">
    Refresh the Location
  </button>

Can be rewritten like this:

  <button _="on click refresh() the location of the window">
    Refresh the Location
  </button>

Closest Expressions

The closest expression allows you to find the closest match of a CSS selector

<%!-- logs the closest section to the div -->
<div _="on click log the closest <section/>">
  ...
</div>

If you pass an attribute literal to the closest expression, it will evaluate to the value of that attribute on the closest element that has it:

<%!-- logs the data-example attribute's value from the body tag -->
<body data-example="An example attribute">
    <div _="on click log the closest @data-example">
      ...
    </div>
</body>

You can use the parent modifier in the closest expression to begin the search from the parent element of the current element:

<%!-- logs the closest parent div to this div -->
<div _="on click log the closest parent <div/>">
  ...
</div>

Positional Expressions

The positional exprssions, first, last and random allows you to get the first, last or a random element from an array-like object

<%!-- logs the first section in the document -->
<div _="on click log the first <section/>">
  ...
</div>

Relative Positional Expressions

The relative positional exprssions, next and previous allows you refer to the next or previous element of a given type within a linear, depth first forward or backward scan of the elements in the DOM tree

<%!-- add the focused class to the next div in the dom with the header class on it -->
<div _="on click add .focused to the next <div.header/>">
  ...
</div>

</div.header>