node package manager


vacuum is another node.js module for templating.


This is not necessarily what the code already does, e.g. I have no idea whether it's fast.

  • be fast
  • be secure
  • be streaming (e.g. send out a static head while the DB is still looking up some data)
  • be easy to understand (no big pile of special cases)

Basic usage

Look into the "example" folder for a working example.

Setup code:

var vacuum = require('vacuum')
// Load all .html files from that folder and register them by name. 
var renderTemplate = vacuum.loadSync(__dirname+'/templates')

Rendering a template to a HTTP response (article.html is the file name of the template):

renderTemplate('article', {articleID: articleID, title: articleTitle}, httpResponse)

The template files are normal HTML with some special-syntax tags inside. Example:

<!DOCTYPE html>
      {var name="title"}
    {childblock of="document"}

This could be a HTML document template. It only contains bodyless special tags, the syntax for them is {tagName key1="value1" key2="value2" ...}. The tag name determines which template should be inserted. There are two kinds of templates:

  • template files (like this one)
  • template functions (like var and childblock)

Template functions are JS functions that can be used inside of templates. Because of them, there's something called "context". In the renderTemplate example above, the initial context is {articleID: articleID, title: articleTitle}, but context can also be changed by template functions - however, these changes only affect descendants of that template function. Attributes also change the context - in the HTML template above, the {var name="title"} inclusion calls the var template with the context {articleID: articleID, title: articleTitle, name: 'title'}. The var template then does (this is somewhat simplified) chunk(context[]); done().

Here's an example that uses the HTML template defined above:

{#document title="Test"}
  Hello You!

This example contains an inclusion with body - it has an opening tag with # and a closing tag with /. The body of the inclusion becomes a template which is given to document's context as $block. document can then do things with it - it could e.g. call it multiple times with different contexts. However, here it's only used with the {childblock} default template function.

Making your own template functions

You can make your own template functions by attaching them to renderTemplate:

renderTemplate.connection = function CONNECTION(template, functions, context, chunk, done) {
  var connection = vacuum.getFromContext(context, 'name')
  var address = connection.remoteAddress

As you can see, there are five parameters.

template is a code representation of the tag used to reference this template. If the tag has a body, that body is stored in the parts property of template.

functions is the same as your renderTemplate.

context is the context (a modified copy of the context of the template inclusions parent).

chunk is the function used to write rendered data as a string. It takes one argument.

done is a normal callback - call it without arguments for success (after you've finished all calls to chunk), call it with an error if an error occurs.

Default functions

You can make your own template functions, but there are also some defaults:


This calls the body that was given to it for each element in the given array.

Important context variables:

  • list - name of the context variable which contains the array
  • element - name of the context variable inside of which the element from the array should be stored


This prints the value from the variable whose name is stored in name.


This takes the template stored in $block_<of> on the context and renders it. It needs a of context variable that specifies the template from whose inclusion the child block should come.