aboutsummaryrefslogtreecommitdiffstats

iro-sass

iro-sass is a multi-purpose Sass library for designing websites. Its main feature is a BEM system, but it also includes some other useful features:

  • Easing background gradients
  • Context stacks: A temporary data storage
  • Property trees: A persistent data storage
  • Responsive properties: A generalization of responsive typography
  • Modular scales

Contents

Getting started

Include iro-sass in your Sass stylesheet:

@import 'iro-sass/src/main';

iro-sass uses function, mixin and variable names that are prefixed with 'iro-' to avoid clashes with other libraries. There are, however, shorter versions of many mixins and functions available (referred to as "shortcodes"). Just import one of these files to use a specific set of shortcodes:

  • iro-sass/src/bem-shortcodes: BEM
  • iro-sass/src/responsive-shortcodes: Responsive properties
  • iro-sass/src/harmony-shortcodes: Modular scales

Development

Clone the repository, then run npm install or yarn to install all dependencies.

The following npm scripts are available:

  • lint: Lint the source code.
  • livelint: Lint automatically whenever the code changes.
  • doc: Generate the SassDoc documentation.
  • test: Run unit tests.

Features

BEM system

iro-sass' main feature is its BEM system which was developed over the course of two years. Features include:

  • Full BEM and BEMIT support: Namespaced blocks, suffixes, states, and so on.
  • Robustness: Most selector-related operations use Sass' native selector functions instead of manual parsing and assembling.
  • Quality: All mixins generate optimal selectors with a minimal degree of specificity.
  • Safety: All mixins perform checks if they are used correctly.
  • Flexibility: Mix BEM selectors and other selectors however you like — the BEM system will adapt.
  • Strictness: The BEM system allows you to define rules that control how the BEM mixins may or may not be used.

Below is a basic example showing how the BEM system can be used:

@include iro-bem-object('media') {
    display:         flex;
    align-items:     flex-start;
    justify-content: flex-start;

    @include iro-bem-element('image') {
        display:  block;
        flex:     0 0 auto;
        order:    1;
        overflow: hidden;
    }

    @include iro-bem-element('body') {
        order: 2;
    }

    @include iro-bem-modifier('rtl') {
        justify-content: flex-end;

        @include iro-bem-element('image') {
            order: 2;
        }

        @include iro-bem-element('body') {
            order: 1;
        }
    }
}

The result is this CSS:

.o-media {
    display:         flex;
    align-items:     flex-start;
    justify-content: flex-start;
}

.o-media__image {
    display:  block;
    flex:     0 0 auto;
    order:    1;
    overflow: hidden;
}

.o-media__body {
    order: 2;
}

.o-media--rtl {
    justify-content: flex-end;
}

.o-media--rtl .o-media__image {
    order: 2;
}

.o-media--rtl .o-media__body {
    order: 1;
}

Easing background gradients

The background gradients generated by browsers usually have a pretty hard transition from one color to another. In some situations, this results in a clearly visible edge where the transition ends.

Andreas Larsen wrote an article on CSS-Tricks where this whole problem is explained in detail. He also developed a solution, a PostCSS plugin which automatically converts linear gradients using an easing function into regular linear-gradients.

This solution works for simple use cases, but unfortunately it wasn't suitable for me. First, I wanted to use easing radial-gradients as well, which aren't supported by the PostCSS plugin. And second, I wanted to freely position the color stops and not be locked to 0% for the start and 100% for the end.

The easing gradients provided by iro-sass address the above problems. Moreover, they have one more major feature: You can use multiple color stops with varying easing functions.

The syntax is kept as close to the new CSSWG proposal as possible to make the transition to the native easing gradients easier later on.

Example usage:

.test {
    background-image: iro-easing-linear-gradient(
        to right,
        #000 2em,
        #f00,
        ease-in-out-sine,
        transparent 10em
    );
}

This will generate a linear-gradient where black normally fades into red, from 2em to 6em. Then, red smoothly fades into transparent, from 6em to 10em. After that, the gradient remains transparent.

Context stacks

Context stacks are a temporary data storage and, as the name suggests, are used like a conventional stack data structure. This means: Whenever you want to store a context -- which is an identifier and any kind of data, such as a map, a list, a string, etc... -- you push it to the stack. From then on, this context is publicly accessible. In order to remove it, you pop the stack.

This feature becomes extremely useful when paired with mixins and their @content directive: Pushing a context to the stack before @content and popping the stack afterwards gives it the role of a call stack. Thats how the BEM system, for example, attaches metadata to the selectors it generates. These information are used to generate optimal selectors without much parsing.

Below is an example of how context stacks can be used:

$context-id: 'some-context-stack';

@mixin anything($p) {
    @include iro-context-push($context-id, 'anything', (
        --this: 1,
        --is:   true,
        --the:  'test',
        --data: $p
    ));

    @content;

    @include iro-context-pop($context-id);
}

// Usage:

.test {
    @include anything('hello') {
        $context-data: nth(iro-context-get($context-id, 'anything'), 2);
        $this: map-get($context-data, --this); // 1
        $is:   map-get($context-data, --is);   // true
        $the:  map-get($context-data, --the);  // 'test'
        $data: map-get($context-data, --data); // 'hello'
    }
}

Property trees

Property trees are basically global maps that are immutable as long as you just use the intended functions.

It's a very simple feature, but it makes managing large sets of structured data much easier.

Example usage:

@include iro-props-store((
    --accent:      #f00,
    --accent-text: #fff,

    --background:     #fff,
    --text:           #222,

    --link: (
        --idle: (
            --text:      #000,
            --underline: #f00
        ),
        --hover: (
            --text:      #f00,
            --underline: #f00
        )
    )
), 'light');

// Usage:

p {
    color:            iro-props-get-static(--text, 'light');       // #222
    background-color: iro-props-get-static(--background, 'light'); // #fff
}

a {
    color:            iro-props-get-static(--link --idle --text, 'light');                // #000
    border-bottom:    1px solid iro-props-get-static(--link --idle --underline, 'light'); // #f00
    text-decoration:  none;

    &:hover {
        color:               iro-props-get-static(--link --hover --text, 'light');      // #f00
        border-bottom-color: iro-props-get-static(--link --hover --underline, 'light'); // #f00
    }
}

Responsive properties

Responsive properties allow you to assign values to properties depending on the current viewport width. iro-sass provides a large number of mixins for this task to cover many use cases. The most simple one is the following:

.title {
    @include iro-responsive-property(padding, ( 20rem: 2.1rem, 40rem: 2.6rem, 60rem: 3.5rem ));
}

The padding will be 2.1rem if the viewport is 20rem wide, 2.6rem if it's 40rem wide, and 3.5rem if it's 60rem wide. If the viewport is narrower than 20rem, the padding will stick with 2.1rem. If the viewport is wider than 60rem, the padding will stick with 3.5rem.

By default, iro-sass will dynamically scale the property value between viewport widths, which is technique known from fluid typography. This behavior can be switched off if it's undesired.

If you use include-media, all responsive mixins also support named viewports. The example above could then be written like this:

.title {
    @include iro-responsive-property(padding, ( phone: 2.1rem, tablet: 2.6rem, desktop: 3.5rem ));
}

Modular scales

From the description of modularscale-sass:

A modular scale is a list of values that share the same relationship. These values are often used to size type and create a sense of harmony in a design. Proportions within modular scales are all around us from the spacing of the joints on our fingers to branches on trees. These natural proportions have been used since the time of the ancient Greeks in architecture and design and can be a tremendously helpful tool to leverage for web designers.

iro-sass provides a mixin to create basic and multi-stranded modular scales. It's a lightweight alternative to modularscale-sass.

Example with a multi-stranded modular scale:

$mod-scale: 1em 2em, 1.1;

h1 {
   font-size: iro-harmony-modular-scale(3, $mod-scale...); // Will be: 1.128em
}
h2 {
   font-size: iro-harmony-modular-scale(2, $mod-scale...); // Will be: 1.1em
}
h3 {
   font-size: iro-harmony-modular-scale(1, $mod-scale...); // Will be: 1.026em
}

Combined with iro-sass' responsive properties:

$responsive-mod-scale: (
    320px: (1rem 2rem, 1.1),
    640px: (1rem 2rem, 1.2)
);

h1 {
    @include iro-responsive-modular-scale(font-size, 3, $responsive-mod-scale);
}
h2 {
    @include iro-responsive-modular-scale(font-size, 2, $responsive-mod-scale);
}
h3 {
    @include iro-responsive-modular-scale(font-size, 1, $responsive-mod-scale);
}