A simple parser for whitespace-significant html, intended for use with reshape.
Note: This project is in early development, and versioning is a little different. Read this for more details.
If you are interested in using reshape, but enjoy whitespace-significant html syntax, like jade, slim, etc. You have come to the right place. Alternately, if you are simply looking for a strong alternative to jade for any other reason, read on.
This parser's syntax is heavily inspired by jade/pug. It is a much smaller and simpler version, containing only the bare minimum necessary to write clean whitespace-significant html. It is intended for use as a reshape plugin, and returns a Reshape AST, but could in theory be used for any other purpose as well.
If you are looking for the ability to add expressions, variables, loops, layouts, etc. you can implement them through other reshape plugins in addition to this parser. This type of functionality is outside the scope of this parser as it is intended to be light, simple, modular, and accessible to contributors.
The source is not very long or complicated and is heavily commented for clarity. Take a look and feel free to contribute!
npm install sugarml --save
Note: This project is compatible with node v6+ only
const {readFileSync} = require('fs')
const reshape = require('reshape')
const sugarml = require('sugarml')
reshape({ parser: sugarml })
.process(readFileSync('./index.sml', 'utf8'))
.then((result) => console.log(result.output()))
This parser is very loose with its rules and standards. It is not responsible for enforcing good style or conventions, it's simply responsible for compiling your code. This means that you can use all sorts of invalid characters in attribute and tag names, and indentation rules are extremely loose.
If you're looking for beautiful syntax highlighting, look no further. Available plugins listed below:
If you made a plugin for another editor, please send in a pull request to have it featured here!
This parser determines how tags are nested based on indentation. For example:
.first-level
.second-level
.third-level hi!
.second-level-2
This would be compiled into the following html output:
<div class="first-level">
<div class="second-level">
<div class="third-level">hi!</div>
</div>
<div class="second-level-2"></div>
</div>
As long as one line is indented with more characters than the last, it will be nested. It doesn't matter if the number of characters that you use is consistent, or if they are spaces or tabs, or anything else. It's just the number of space characters used to indent, that's it. So you can get away with very messy code, if you want, but that's what linters are for.
A tag is written simply as the name of the tag. Tag names must start with a letter, then after that can contain any character other than #
, .
, (
, :
, or a space/tab. These character limitation are in place solely because of the language's syntax requirements.
So, for example, these tag names are valid and will compile correctly (although I would not advise using a tag with characters other than letters and hyphens personally):
div
custom-name
wow@mytagiscool!
hello___12world
However, these tag names will not compile into the results you expect:
foo(bar
hello.min.js
iam:atag
my cool tag
Fortunately, it is not advisable to have custom html tags that look anything like these anyway, so you should be in the clear as long as you are writing reasonable html.
There is a shorthand for adding classes and IDs to your tags, which is exactly the same as it is in just about every other whitespace-significant html parser, and the same as CSS. For example:
p#main >>> <p id="main"></p>
p.staff >>> <p class="staff"></p>
p#main.staff.active >>> <p id="main" class="staff active"></p>
You can chain as many classes and IDs in this manner as you want. If you do not use an element name, it will assume you want a div. For example:
#main >>> <div id="main"></div>
Attributes fall between parentheses directly after a tag's name. They are space-separated and can be either boolean (key, no value), or key/value pairs. For example:
input(type='text') >>> <input type="text">
input(checked) >>> <input checked>
input(type='checkbox' checked) >>> <input type="checkbox" checked>
You can quote your attribute values or not, your choice (although we would recommend quoting). If you quote your attribute, it can contain any value other than an end-quote that matches the type of your starting quote (obviously). If you do not quote your attribute, it can contain any character other than a single or double quote, a close paren, or a space. As such, you are less likely to run into issues when quoting your attribute, which is why we recommend it. For example:
div(class=foo) >>> <div class="foo"></div>
div(class='foo bar') >>> <div class="foo bar"></div>
div(class=foo bar) >>> <div class="foo" bar></div>
div(class="foo('wow')") >>> <div class="foo('wow')"></div>
div(class=foo("wow")) >>> Syntax error
Attributes can contain any character other than =
or a space. If you value is quoted, it can contain any value other than a quote (that will end the attribute), and if it's not quoted, it can contain any value other than a quote or space. So even attributes with special characters (found sometimes in certain front-end frameworks like vue and angular) work fine. For example:
div(:bind='focus') >>> <div :bind="click"></div>
div(*ngFor='foo in bar') >>> <div *ngFor="foo in bar"></div>
div(@click='doSomething') >>> <div @click="doSomething"></div>
Sometimes you don't really want to indent nested tags when they are short enough to be placed on one line. When this happens, you can use a colon instead of a newline and indent to nest. For example:
ul
li: a(href='#') link 1
li: a(href='#') link 2
You can nest as many wrapper tags on a single line as you want, as long as they are all separated by a colon immediately following the tag. However If the tag has content, it cannot be nested with a colon. For example:
.wrap: .wrap2: .wrap3 hello! // this works fine
.wrap hello!: .wrap2 // this does not
If you need to mix up text content alongside inline tags and such, you can use the pipe character for this as such:
p
| Here's some text
strong And some bold text
| ...and some more text
This would render as:
<p>Here's some text <strong>and some bold text</strong> ...and some more text</p>
For any type of content transforms that are more complex than this, we recommend checking out reshape-content.
If you have a block of text you don't really want to put pipes before, you can use a text block for this. Put a .
after the end of the tag name, then indent the content in to the next line as such:
p.
Here's some content
Split into multiple lines, wow!
span will render as text not a span
ok that's the end!
This would render as:
<p>Here's some content
Split into multiple lines, wow!
span will render as text not a span
ok that's the end!</p>
Note that you will not be able to render any elements within a text block. The block continues until there's a line that is not indented at a greater level than the parent element.
The doctype is a special tag, and is handled specially. If the first word on the first line of your template is doctype
, it will be parsed as a doctype. Anything after this word will be added to the tag. So for example:
doctype html >>> <!DOCTYPE html>
doctype HTML PUBLIC "-//W3C//DTD HTML 4.01//EN" "http://www.w3.org/TR/html4/strict.dtd" >>> <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN" "http://www.w3.org/TR/html4/strict.dtd">
Input:
doctype html
html
head
title Testing
body#index
h1 Hello world!
p.intro Wow what a great little language! Some features:
ul(data-list='yep' @sortable)
li: a(href='#') whitespace significant!
li: a(href='#') simple classes and ids!
footer
| Thanks for visiting
span see you next time!
Setup:
'use strict'
const {readFileSync} = require('fs')
const reshape = require('reshape')
const sugarml = require('sugarml')
reshape({ parser: sugarml })
.process(readFileSync('./index.sml', 'utf8'))
.then((result) => console.log(result.output()))
Output:
<!DOCTYPE html>
<html>
<head>
<title>Testing</title>
</head>
<body id="index">
<h1>Hello world!</h1>
<p class="intro">Wow what a great little language! Some features:</p>
<ul data-list="yep" @sortable>
<li><a href="#">whitespace significant!</a></li>
<li><a href="#">simple classes and ids!</a></li>
</ul>
<footer>
Thanks for visiting
<span>see you next time!</span>
</footer>
</body>
</html>
- Details on the license can be found here
- Details on running tests and contributing can be found here