Slim is a template language whose goal is to reduce the view syntax to the essential parts without becoming cryptic. It started as an exercise to see how much could be removed from a standard html template (<, >, closing tags, etc...). As more people took an interest in Slim, the functionality grew and so did the flexibility of the syntax.
A short list of the features...
- Elegant syntax
- Short syntax without closing tags (Using indentation instead)
- HTML style mode with closing tags
- Configurable shortcut tags (
#for<div id="...">and.for<div class="...">in the default configuration)
- Safety
- Automatic HTML escaping by default
- Support for Rails'
html_safe?
- Highly configurable and extendable via plugins
- Logic less mode similar to Mustache, realized as plugin
- Translator/I18n, realized as plugin
- High performance
- Comparable speed to ERB/Erubis
- Streaming support in Rails
- Supported by all major frameworks (Rails, Sinatra, ...)
- Full Unicode support for tags and attributes on Ruby 1.9
- Embedded engines like Markdown and Textile
NOTE: Slim 2.0 is not yet released, but you can try the preview versions.
If you are already using Slim 1.3 or 1.2 and want to upgrade to the newest version 2.0, you should at first upgrade to Slim 1.3.7 which will emit warnings for deprecated features. This way you can easily see if your application is already Slim 2.0 compliant!
Slim 2.0 removes deprecated features from the 1.3 series and cleans up some minor inconsistencies in the syntax (Attribute wrapper and quoted attribute escaping). Of course it also includes new features and bugfixes. See the CHANGES for details.
In general don't be afraid of upgrading! We try to be backward compatible and follow semantic versions as good as possible.
Slim is a fast, lightweight templating engine with support for Rails 3 and 4. It has been heavily tested on all major ruby implementations. We use continous integration (travis-ci).
Slim's core syntax is guided by one thought: "What's the minimum required to make this work".
As more people have contributed to Slim, there have been syntax additions influenced from their use of Haml and Jade. The Slim team is open to these additions because we know beauty is in the eye of the beholder.
Slim uses Temple for parsing/compilation and is also integrated into Tilt, so it can be used together with Sinatra or plain Rack.
The architecture of Temple is very flexible and allows the extension of the parsing and compilation process without monkey-patching. This is used by the logic less plugin and the translator plugin which provides I18n.
Within the Rails community, Erb and Haml are without doubt the two most popular templating engines. However, Erb's syntax is cumbersome and Haml's syntax can be quite cryptic to the uninitiated.
There is also some development in logic-less engines (e.g. Mustache which requires you to write standard HTML). You can also use Slim in logic-less mode if you like the Slim syntax to build your HTML but don't want to write Ruby in your templates.
Slim was born to bring a minimalist syntax approach with speed. If people chose not to use Slim, it would not be because of speed.
Yes, Slim is speedy! Benchmarks are done for every commit at http://travis-ci.org/#!/slim-template/slim. Don't trust the numbers? That's as it should be. Please try the benchmark rake task yourself!
Install Slim as a gem:
gem install slim
Include Slim in your Gemfile with gem 'slim' or require it with require 'slim'. That's it! Now, just use the .slim extension and you're good to go.
Here's a quick example to demonstrate what a Slim template looks like:
doctype html
html
head
title Slim Examples
meta name="keywords" content="template language"
meta name="author" content=author
link rel="icon" type="image/png" href=file_path("favicon.png")
javascript:
alert('Slim supports embedded javascript!')
body
h1 Markup examples
#content
p This example shows you how a basic Slim file looks like.
= yield
- if items.any?
table#items
- for item in items do
tr
td.name = item.name
td.price = item.price
- else
p No items found Please add some inventory.
Thank you!
div id="footer"
= render 'footer'
| Copyright © #{@year} #{@author}
Indentation matters, but the indentation depth can be chosen as you like. If you want to first indent 2 spaces, then 5 spaces, it's your choice. To nest markup you only need to indent by one space, the rest is gravy.
The pipe tells Slim to just copy the line. It essentially escapes any processing. Each following line that is indented greater than the pipe is copied over.
body
p
|
This is a test of the text block.
The parsed result of the above:
<body><p>This is a test of the text block.</p></body>
The left margin is set at the indent of the pipe + one space. Any additional spaces will be copied over.
body
p
| This line is on the left margin.
This line will have one space in front of it.
This line will have two spaces in front of it.
And so on...
You can also embed html in the text line
- articles.each do |a|
| <tr><td>#{a.name}</td><td>#{a.description}</td></tr>
The single quote tells Slim to copy the line (similar to |), but makes sure that a single trailing white space is appended.
You can write html tags directly in Slim which allows you to write your templates in a more html like style with closing tags or mix html and Slim style.
<html>
head
title Example
<body>
- if articles.empty?
- else
table
- articles.each do |a|
<tr><td>#{a.name}</td><td>#{a.description}</td></tr>
</body>
</html>
The dash denotes control code. Examples of control code are loops and conditionals. end is forbidden behind -. Blocks are defined only by indentation.
If your ruby code needs to use multiple lines, append a backslash \ at the end of the lines. If your line ends with comma , (e.g because of a method call) you don't need the additional backslash before the linebreak.
body
- if articles.empty?
| No inventory
The equal sign tells Slim it's a Ruby call that produces output to add to the buffer. If your ruby code needs to use multiple lines, append a backslash \ at the end of the lines, for example:
= javascript_include_tag \
"jquery", \
"application"
If your line ends with comma , (e.g because of a method call) you don't need the additional backslash before the linebreak.
Same as the single equal sign (=), except that it adds a trailing white space.
Same as the single equal sign (=), but does not go through the escape_html method.
Same as the double equal sign (==), except that it adds a trailing white space.
Use the forward slash for code comments - anything after it won't get displayed in the final render. Use / for code comments and /! for html comments
body
p
/ This line won't get displayed.
Neither does this line.
/! This will get displayed as html comments.
The parsed result of the above:
<body><p><!--This will get displayed as html comments.--></p></body>
Use the forward slash immediately followed by an exclamation mark for html comments (<!-- ... -->).
/[if IE]
p Get a better browser.
renders as
<!--[if IE]><p>Get a better browser.</p><![endif]-->
The doctype tag is a special tag which can be used to generate the complex doctypes in a very simple way.
XML VERSION
doctype xml
<?xml version="1.0" encoding="utf-8" ?>
doctype xml ISO-8859-1
<?xml version="1.0" encoding="iso-8859-1" ?>
XHTML DOCTYPES
doctype html
<!DOCTYPE html>
doctype 5
<!DOCTYPE html>
doctype 1.1
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.1//EN"
"http://www.w3.org/TR/xhtml11/DTD/xhtml11.dtd">
doctype strict
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN"
"http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd">
doctype frameset
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Frameset//EN"
"http://www.w3.org/TR/xhtml1/DTD/xhtml1-frameset.dtd">
doctype mobile
<!DOCTYPE html PUBLIC "-//WAPFORUM//DTD XHTML Mobile 1.2//EN"
"http://www.openmobilealliance.org/tech/DTD/xhtml-mobile12.dtd">
doctype basic
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML Basic 1.1//EN"
"http://www.w3.org/TR/xhtml-basic/xhtml-basic11.dtd">
doctype transitional
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"
"http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
HTML 4 DOCTYPES
doctype strict
<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01//EN"
"http://www.w3.org/TR/html4/strict.dtd">
doctype frameset
<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Frameset//EN"
"http://www.w3.org/TR/html4/frameset.dtd">
doctype transitional
<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN"
"http://www.w3.org/TR/html4/loose.dtd">
You can close tags explicitly by appending a trailing /.
img src="image.png"/
Note, that this is usually not necessary since the standard html tags (img, br, ...) are closed automatically.
Sometimes you may want to be a little more compact and inline the tags.
ul
li.first: a href="/a" A link
li: a href="/b" B link
For readability, don't forget you can wrap the attributes.
ul
li.first: a[href="/a"] A link
li: a[href="/b"] B link
Either start on the same line as the tag
body
h1 id="headline" Welcome to my site.
Or nest it. You must use a pipe or a backtick to escape processing
body
h1 id="headline"
| Welcome to my site.
Can make the call on the same line
body
h1 id="headline" = page_headline
Or nest it.
body
h1 id="headline"
= page_headline
You write attributes directly after the tag. For normal text attributes you must use double " or single quotes ' (Quoted attributes).
a href="http://slim-lang.com" title='Slim Homepage' Goto the Slim homepage
You can use text interpolation in the quoted attributes.
If a delimiter makes the syntax more readable for you,
you can use the characters {...}, (...), [...] to wrap the attributes.
body
h1(id="logo") = page_logo
h2[id="tagline" class="small tagline"] = page_tagline
If you wrap the attributes, you can spread them across multiple lines:
h2[id="tagline"
class="small tagline"] = page_tagline
You may use spaces around the wrappers and assignments:
h1 id = "logo" = page_logo
h2 [ id = "tagline" ] = page_tagline
Example:
a href="http://slim-lang.com" title='Slim Homepage' Goto the Slim homepage
You can use text interpolation in the quoted attributes:
a href="http://#{url}" Goto the #{url}
The attribute value will be escaped by default. Use == if you want to disable escaping in the attribute.
a href=="&"
You can break quoted attributes with backslash \
a data-title="help" data-content="extremely long help text that goes on\
and one and one and then starts over...."
Write the ruby code directly after the =. If the code contains spaces you have to wrap
the code into parentheses (...). You can also directly write hashes {...} and arrays [...].
body
table
- for user in users do
td id="user_#{user.id}" class=user.role
a href=user_action(user, :edit) Edit #{user.name}
a href=(path_to_user user) = user.name
The attribute value will be escaped by default. Use == if you want to disable escaping in the attribute.
a href==action_path(:start)
You can also break ruby attributes with backslash \ or trailing , as describe for control sections.
The attribute values true, false and nil are interpreted
as booleans. If you use the attribut wrapper you can omit the attribute assigment
input type="text" disabled="disabled"
input type="text" disabled=true
input(type="text" disabled)
input type="text"
input type="text" disabled=false
input type="text" disabled=nil
You can configure attributes to be merged if multiple are given (See option :merge_attrs). In the default configuration
this is done for class attributes with the white space as delimiter.
a.menu class="highlight" href="http://slim-lang.com/" Slim-lang.com
This renders as
<a class="menu highlight" href="http://slim-lang.com/">Slim-lang.com</a>
You can also use an Array as attribute value and the array elements will be merged using the delimiter.
a class=["menu","highlight"]
a class=:menu,:highlight
The splat shortcut allows you turn a hash in to attribute/value pairs
.card*{'data-url'=>place_path(place), 'data-id'=>place.id} = place.name
renders as
<div class="card" data-id="1234" data-url="/place/1234">Slim's house</div>
You can also use methods or instance variables which return a hash as shown here:
.card *method_which_returns_hash = place.name
.card *@hash_instance_variable = place.name
The hash attributes which support attribute merging (see Slim option :merge_attrs) can be given as an Array
.first *{:class => [:second, :third]} Text
renders as
div class="first second third"
You can create completely dynamic tags using the splat attributes. Just create a method which returns a hash with the :tag key.
ruby:
def a_unless_current
@page_current ? {:tag => 'span'} : {:tag => 'a', :href => 'http://slim-lang.com/'}
end
- @page_current = true
*a_unless_current Link
- @page_current = false
*a_unless_current Link
renders as
<span>Link</span><a href="http://slim-lang.com/">Link</a>
You can define custom tag shortcuts by setting the option :shortcut.
Slim::Engine.set_default_options :shortcut => {'c' => {:tag => 'container'}, '#' => {:attr => 'id'}, '.' => {:attr => 'class'} }
We can use it in Slim code like this
c.content Text
which renders to
<container class="content">Text</container>
You can define custom shortcuts (Similar to # for id and . for class).
In this example we add & to create a shortcut for the input elements with type attribute.
Slim::Engine.set_default_options :shortcut => {'&' => {:tag => 'input', :attr => 'type'}, '#' => {:attr => 'id'}, '.' => {:attr => 'class'}}
We can use it in Slim code like this
&text name="user"
&password name="pw"
&submit
which renders to
<input type="text" name="user" />
<input type="password" name="pw" />
<input type="submit" />
In another example we add @ to create a shortcut for the role attribute.
Slim::Engine.set_default_options :shortcut => {'@' => {:attr => 'role'}, '#' => {:attr => 'id'}, '.' => {:attr => 'class'}}
We can use it in Slim code like this
.person@admin = person.name
which renders to
<div class="person" role="admin">Daniel</div>
Similarly to Haml, you can specify the id and class attributes in the following shortcut form
body
h1#headline
= page_headline
h2#tagline.small.tagline
= page_tagline
.content
= show_content
This is the same as
body
h1 id="headline"
= page_headline
h2 id="tagline" class="small tagline"
= page_tagline
div class="content"
= show_content
Use standard Ruby interpolation. The text will be html escaped by default.
body
h1 Welcome #{current_user.name} to the show.
| Unescaped #{{content}} is also possible.
To escape the interpolation (i.e. render as is)
body
h1 Welcome \#{current_user.name} to the show.
Thanks to Tilt, Slim has impressive support for embedding other template engines.
Examples:
coffee:
square = (x) -> x * x
markdown:
#Header
Hello from #{"Markdown!"}
Second Line!
Supported engines:
| Filter | Required gems | Type | Description |
| ruby: | none | Shortcut | Shortcut to embed ruby code |
| javascript: | none | Shortcut | Shortcut to embed javascript code and wrap in script tag |
| css: | none | Shortcut | Shortcut to embed css code and wrap in style tag |
| sass: | sass | Compile time | Embed sass code and wrap in style tag |
| scss: | sass | Compile time | Embedd scss code and wrap in style tag |
| less: | less | Compile time | Embed less css code and wrap in style tag |
| styl: | styl | Compile time | Embed stylus css code and wrap in style tag |
| coffee: | coffee-script | Compile time | Compile coffee script code and wrap in script tag |
| asciidoc: | asciidoctor | Compile time + Interpolation | Compile AsciiDoc code and interpolate #\{variables} in text |
| markdown: | redcarpet/rdiscount/kramdown | Compile time + Interpolation | Compile markdown code and interpolate #\{variables} in text |
| textile: | redcloth | Compile time + Interpolation | Compile textile code and interpolate #\{variables} in text |
| creole: | creole | Compile time + Interpolation | Compile creole code and interpolate #\{variables} in text |
| wiki:, mediawiki: | wikicloth | Compile time + Interpolation | Compile wiki code and interpolate #\{variables} in text |
| rdoc: | rdoc | Compile time + Interpolation | Compile rdoc code and interpolate #\{variables} in text |
| builder: | builder | Precompiled | Embed builder code |
| nokogiri: | nokogiri | Precompiled | Embed nokogiri builder code |
| erb: | none | Precompiled | Embed erb code |
The embedded engines can be configured in Slim by setting the options directly on the Slim::Embedded filter. Example:
Slim::Embedded.default_options[:markdown] = {:auto_ids => false}
Slim and the underlying Temple framework are highly configurable.
The way how you configure Slim depends a bit on the compilation mechanism (Rails or Tilt). It is always possible to set default options per Slim::Engine class. This can be done in Rails' environment files. For instance, in config/environments/development.rb you probably want:
# Indent html for pretty debugging and do not sort attributes (Ruby 1.8)
Slim::Engine.set_default_options :pretty => true, :sort_attrs => false
# Indent html for pretty debugging and do not sort attributes (Ruby 1.9)
Slim::Engine.set_default_options pretty: true, sort_attrs: false
You can also access the option hash directly:
Slim::Engine.default_options[:pretty] = true
There are two ways to set options at runtime. For Tilt templates (Slim::Template) you can set
the options when you instatiate the template:
Slim::Template.new('template.slim', optional_option_hash).render(scope)
The other possibility is to set the options per thread which is interesting mostly for Rails:
Slim::Engine.with_options(option_hash) do
# Any Slim engines which are created here use the option_hash
# For example in Rails:
render :page, :layout => true
end
You have to be aware that the compiled engine code and the options are cached per template in Rails and you cannot change the option afterwards.
# First render call
Slim::Engine.with_options(:pretty => true) do
render :page, :layout => true
end
# Second render call
Slim::Engine.with_options(:pretty => false) do
render :page, :layout => true # :pretty is still true because it is cached
end
The following options are exposed by the Slim::Engine and can be set with Slim::Engine.set_default_options.
There are a lot of them but the good thing is, that Slim checks the configuration keys and reports an error if you try to use an invalid configuration key.
| Type | Name | Default | Purpose |
| String | :file | nil | Name of parsed file, set automatically by Slim::Template |
| Integer | :tabsize | 4 | Number of white spaces per tab (used by the parser) |
| String | :encoding | "utf-8" | Set encoding of template |
| String | :default_tag | "div" | Default tag to be used if tag name is omitted |
| Hash | :shortcut | \{'.' => {:attr => 'class'}, '#' => {:attr => 'id'}} | Attribute shortcuts |
| Array<Symbol,String> | :enable_engines | nil (All enabled) | List of enabled embedded engines (whitelist) |
| Array<Symbol,String> | :disable_engines | nil (None disabled) | List of disabled embedded engines (blacklist) |
| Boolean | :disable_capture | false (true in Rails) | Disable capturing in blocks (blocks write to the default buffer |
| Boolean | :disable_escape | false | Disable automatic escaping of strings |
| Boolean | :use_html_safe | false (true in Rails) | Use String#html_safe? from ActiveSupport (Works together with :disable_escape) |
| Symbol | :format | :xhtml | HTML output format (Possible formats :xhtml, :html4, :html5, :html) |
| String | :attr_quote | '"' | Character to wrap attributes in html (can be ' or ") |
| Hash | :merge_attrs | \{'class' => ' '} | Joining character used if multiple html attributes are supplied (e.g. class="class1 class2") |
| Array<String> | :hyphen_attrs | %w(data) | Attributes which will be hyphenated if a Hash is given (e.g. data={a:1,b:2} will render as data-a="1" data-b="2") |
| Boolean | :sort_attrs | true | Sort attributes by name |
| Symbol | :js_wrapper | nil | Wrap javascript by :comment, :cdata or :both. You can also :guess the wrapper based on :format. |
| Boolean | :pretty | false | Pretty html indenting (This is slower!) |
| String | :indent | ' ' | Indentation string |
| Boolean | :streaming | false (true in Rails > 3.1) | Enable output streaming |
| Class | :generator | Temple::Generators::ArrayBuffer/RailsOutputBuffer | Temple code generator (default generator generates array buffer) |
| String | :buffer | '_buf' ('@output_buffer' in Rails) | Variable used for buffer |
There are more options which are supported by the Temple filters but which are not exposed and are not officially supported. You have to take a look at the Slim and Temple code for that.
For developers who know more about Slim and Temple architecture it is possible to override default options at different positions. Temple uses an inheritance mechanism to allow subclasses to override options of the superclass. The option priorities are as follows:
Slim::Templateoptions passed at engine instatinationSlim::Template.default_optionsSlim::Engine.thread_options,Slim::Engine.default_options- Parser/Filter/Generator
thread_options,default_options(e.gSlim::Parser,Slim::Compiler)
It is also possible to set options for superclasses like Temple::Engine. But this will affect all temple template engines then.
Slim::Engine < Temple::Engine
Slim::Compiler < Temple::Filter
Slim currently provides plugins for logic less mode and I18n. See the plugin documentation.
Slim uses Tilt to compile the generated code. If you want to use the Slim template directly, you can use the Tilt interface.
Tilt.new['template.slim'].render(scope)
Slim::Template.new('template.slim', optional_option_hash).render(scope)
Slim::Template.new(optional_option_hash) { source }.render(scope)
The optional option hash can have to options which were documented in the section above.
require 'sinatra'
require 'slim'
get('/') { slim :index }
__END__
@@ index
doctype html
html
head
title Sinatra With Slim
body
h1 Slim Is Fun!
Rails generators are provided by slim-rails. slim-rails
is not necessary to use Slim in Rails though. Just install Slim and add it to your Gemfile with gem 'slim'.
Then just use the .slim extension and you're good to go.
HTTP streaming is enabled by default if you use a Rails version which supports it.
The gem 'slim' comes with the small tool 'slimrb' to test Slim from the command line.
$ slimrb --help
Usage: slimrb [options]
-s, --stdin Read input from standard input instead of an input file
--trace Show a full traceback on error
-c, --compile Compile only but do not run
-r, --rails Generate rails compatible code (Implies --compile)
-t, --translator Enable translator plugin
-l, --logic-less Enable logic less plugin
-p, --pretty Produce pretty html
-o, --option [NAME=CODE] Set slim option
-h, --help Show this message
-v, --version Print version
Start 'slimrb', type your code and press Ctrl-d to send EOF. Example usage:
$ slimrb markdown: First paragraph. Second paragraph. * one * two * three //Enter Ctrl-d <p>First paragraph </p> <p>Second paragraph </p> <ul> <li>one</li> <li>two</li> <li>three</li> </ul>
There are plugins for various text editors (including the most important ones - Vim, Emacs and Textmate):
Yes, Slim is one of the fastest Ruby template engines out there! In production mode Slim is nearly as fast as Erubis (which is the fastest template engine). But we would be happy if you chose Slim also for any other reason, we assure you performance will not be an obstacle.
Run the benchmarks with rake bench. You can add the option slow to
run the slow parsing benchmark which needs more time. You can also increase the number of iterations.
rake bench slow=1 iterations=1000
We run the benchmarks for every commit on Travis-CI. Take a look at the newest benchmarking results: http://travis-ci.org/#!/slim-template/slim
Slim provides an extensive test-suite based on minitest. You can run the tests with 'rake test' and the rails integration tests with 'rake test:rails'.
We are currently experimenting with human-readable literate tests which are written as markdown files: TESTS.md
Travis-CI is used for continous integration testing: http://travis-ci.org/#!/slim-template/slim
Slim is working well on all major Ruby implementations:
- Ruby 1.8.7
- Ruby 1.9.2
- Ruby 1.9.3
- Ruby EE
- JRuby
- Rubinius 2.0
If you'd like to help improve Slim, clone the project with Git by running:
$ git clone git://github.com/slim-template/slim
Work your magic and then submit a pull request. We love pull requests!
Please remember to test against Ruby versions 1.9.2 and 1.8.7.
If you find the documentation lacking (and you probably will), help us out and update this README.md. If you don't have the time to work on Slim, but found something we should know about, please submit an issue.
Slim is released under the MIT license.
- Daniel Mendler (Lead developer)
- Andrew Stone
- Fred Wu
Template compilation framework:
Framework support:
Syntax highlighting:
Template Converters (HAML, ERB, ...):
Language ports/Similar languages: