Partial content
\n```\n\n## CoffeeScript support\n\nHaml and CoffeeScript are a winning team, both use indention for blocks and are a perfect match for this reason. You can\nuse CoffeeScript instead of Ruby in your Haml tags and the attributes.\n\n**It's not recommended to put too much logic into the template.**\n\n### Attributes\n\nWhen you define an attribute value without putting it into quotes (single or double quotes), it's considered to be\nCoffeeScript code to be run at render time. By default, attributes values from CoffeeScript code are escaped before\ninserting into the document. You can change this behaviour by setting the appropriate compiler option.\n\nHTML style attributes are the most limited and can only assign a simple variable:\n\n```haml\n%img(src='/images/demo.png' width=@width height=@height alt=alt)\n```\n\nBoth the `@width` and `@height` values must be passed as locals when rendering the template and `alt` must be defined\nbefore the `%img` tag.\n\nRuby style tags can be more complex and can call functions:\n\n```haml\n%header\n %user{ :class => App.currentUser.get('status') }= App.currentUser.getDisplayName()\n```\n\nAttribute definitions are also supported in the Ruby 1.9 style:\n\n```haml\n%header\n %user{ class: App.currentUser.get('status') }= App.currentUser.getDisplayName()\n```\n\nMore fancy stuff can be done when use interpolation within a double quoted attribute value:\n\n```haml\n%header\n %user{ class: \"#{ if @user.get('roles').indexOf('admin') is -1 then 'normal' else 'admin' }\" }= @user.getDisplayName()\n```\n\n_But think twice about it before putting such fancy stuff into your template, there are better places like models,\nviews or helpers to put heavy logic into._\n\nYou can define your attributes over multiple lines and the next line must not be correctly indented, so you can align\nthem properly:\n\n```haml\n%input#password.hint{ type: 'password', name: 'registration[password]',\n data: { hint: 'Something very important', align: 'left' } }\n```\n\nIn the above example you also see the usage for generating HTML5 data attributes.\n\n### Running Code\n\nYou can run any CoffeeScript code in your template:\n\n```haml\n- for project in @projects\n - if project.visible\n .project\n %h1= project.name\n %p&= project.description\n```\n\nThere are several supported types to run your code:\n\n* Run code without insert anything into the document: `-`\n* Run code and insert the result into the document: `=`\n\nAll inserted content from running code is escaped by default. You can change this behaviour by setting the appropriate\ncompiler option.\n\nThere are three variations to run code and insert its result into the document, two of them to change the escaping style\nchosen in the compile option:\n\n* Run code and do not escape the result: `!=`\n* Run code and escape the result: `&=`\n* Preserve whitespace when insert the result: `~`\n\nAgain, please consult the official [Haml reference](http://haml-lang.com/docs/yardoc/file.HAML_REFERENCE.html) for more\ndetails. Haml Coffee implements the same functionality like Ruby Haml, only for CoffeeScript.\n\n#### Interpolation\n\nIf you use the CoffeeScript interpolation without explicitly run code with `=` and `-`, the interpolation output is not\nbeing escaped. You can manually force escaping by using the HTML escape reference `$e`:\n\n```Haml\n%p\n foo #{ $e '\n foo <bar>\n
\n```\n\n#### Multiline code blocks\n\nRunning code must be placed on a single line and unlike Ruby Haml, you cannot stretch a it over multiple lines by\nputting a comma at the end.\n\nHowever, you can use multiline endings `|` to stretch your code over multiple lines to some extend:\n\n```Haml\n- links = { |\n home: '/', |\n docs: '/docs', |\n about: '/about' |\n } |\n\n%ul\n - for name, link of links\n %li\n %a{ href: link }= name\n```\n\nPlease note, that since the line is concatenated before the compilation, you cannot omit the curly braces and the\ncommas in the above example, like you'd do in normal CoffeeScript code. Therefore it's recommended to use the\nCoffeeScript filter to have real multiline code blocks:\n\n```Haml\n:coffeescript\n links =\n home: '/'\n docs: '/docs'\n about: '/about'\n\n%ul\n - for name, link of links\n %li\n %a{ href: link }= name\n```\n\n#### Functions\n\nYou can also create functions that generate Haml:\n\n```haml\n- sum = (a, b) ->\n %div\n %span= a\n %span= b\n %span= a+b\n= sum(1,2)\n= sum(3,4)\n```\n\nor pass generated HTML output through a function for post-processing.\n\n```haml\n= postProcess ->\n %a{ href: '/' }\n```\n\nThe content of the `:coffeescript` filter is run when the template is rendered and doesn't output anything into the\nresulting document. This comes in handy when you have code to run over multiple lines and don't want to prefix each line\nwith `-`:\n\n```haml\n%body\n :coffeescript\n tags = ['CoffeeScript', 'Haml']\n project = 'Haml Coffee'\n %h2= project\n %ul\n - for tag in tags\n %li= tag\n```\n\n## Compiler options\n\nThe following section describes all the available compiler options that you can use through the JavaScript API,\nas Express view option or as argument to the command line utility.\n\nThe command line arguments may be slightly different. For example instead of passing `--escape-html=false` you have to\nuse the `--disable-html-escaping` argument. You can see a list of all the command line arguments by executing\n`haml-coffee --help`.\n\n### HTML generation options\n\nThe HTML options change the way how the generated HTML will look like.\n\n#### Output format\n\n* Name: 'format'\n* Type: `String`\n* Default: `html5`\n\nThe Haml parser knows different HTML formats to which a given template can be rendered and it must be one of:\n\n* xhtml\n* html4\n* html5\n\nDoctype, self-closing tags and attributes handling depends on this setting. Please consult the official\n[Haml reference](http://haml-lang.com/docs/yardoc/file.HAML_REFERENCE.html) for more details.\n\n#### Uglify output\n\n* Name: `uglify`\n* Type: `Boolean`\n* Default: `false`\n\nAll generated HTML tags are properly indented by default, so the output looks nice. This can be helpful when debugging.\nYou can skip the indention by setting the `uglify` option to false. This save you some bytes and you'll have increased\nrendering speed.\n\n#### HTML escape\n\n* Name: `escapeHtml`\n* Type: `Boolean`\n* Default: `true`\n\nThe reserved HTML characters `\"`, `'`, `&`, `<` and `>` are converted to their HTML entities by default when they are\ninserted into the HTML document from evaluated CoffeeScript.\n\nYou can always change the escaping mode within the template to either force escaping with `&=` or force unescaping with\n`!=`.\n\n#### Attributes escape\n\n* Name: `escapeAttributes`\n* Type: `Boolean`\n* Default: `true`\n\nAll HTML attributes that are generated by evaluating CoffeeScript are also escaped by default. You can turn of HTML\nescaping of the attributes only by setting `escapeAttributes` to false. You can't change this behaviour in the template\nsince there is no Haml markup for this to instruct the compiler to change the escaping mode.\n\n#### Clean CoffeeScript values\n\n* Name: `cleanValue`\n* Type: `Boolean`\n* Default: `true`\n\nEvery output that is generated from evaluating CoffeeScript code is cleaned before inserting into the document. The\ndefault implementation converts `null` or `undefined` values into an empty string and marks real boolean values with a\nhidden marker character. The hidden marker character is necessary to distinguish between String values like `'true'`,\n`'false'` and real boolean values `true`, `false` in the markup, so that a boolean attribute conversion can quickly\nconvert these values to the correct HTML5/XHTML/HTML4 representation.\n\n#### Preserve whitespace tags\n\n* Name: `preserve`\n* Type: `String`\n* Default: `textarea,pre`\n\nThe `preserve` option defines a list of comma separated HTML tags that are whitespace sensitive. Content from these tags\nmust be preserved, so that the indention has no influence on the displayed content. This is simply done by converting\nthe newline characters to their equivalent HTML entity.\n\n#### Autoclose tags\n\n* Name: `autoclose`\n* Type: `String`\n* Default: `meta,img,link,br,hr,input,area,param,col,base`\n\nThe autoclose option defines a list of tag names that should be automatically closed if they have no content.\n\n#### Module loader support\n\n* Name: `placement`\n* Type: `String`\n* Default: `global`\n\nThe `placement` option defines where the template function is inserted\nupon compilation.\n\nPossible values are:\n\n* `global`