Important: This documentation covers Yarn 1 (Classic).
For Yarn 2+ docs and migration guide, see yarnpkg.com.

Package detail

siph

lastonoga24ISC0.0.6TypeScript support: included

Yet another reactive framework based on Sinuous

readme

Sinuous

Yet another reactive framework based on Sinuous

Why?

There is no way to use reactive frameworks for websites because of bad performance (FID, TTI).

Thats why even github use jquery-like libs.

What?

This framework is in pre-alpha and was created to test partial hydration combined with jquery-like reactivity (thanks Sinuous).

Sinuous is fast, but doesnt have modern features like loops, if statements, components (single file components) and... Partial hydration.

Partial hydration is hydration that works only with dynamic and statefull parts of application. It helps to use component and reactive paradigm for website development (not applications and SPA).

How?

  • Partial hydration. hydrate only dynamic parts of components
  • Single file components. webpack loader and compiler (code transpiler) + component system
  • Automagic reactivity. define a reactive variable and all methods/props that use that variable become reactive

Benefits

  • Small. hello world at 7.1kB gzip.
  • Fast. Thanks Sinuous and partial hydration.
  • Truly reactive. automatically derived from the app state.
  • Tips & Tricks. Slots, Loops, Statements and Props as we get used to

Performance

It was dirty check, but +- in similar conditions.

Index page with 10k static components:

<template>
  <div>
    <sbutton v-for="(item, key) in items" :key="key">
      Button {{ item }}
    </sbutton>
  </div>
</template>
`

sButton:

<template>
  <div class="button">
    <slot />
  </div>
</template>
`
Metric SinuousCC SinuousCC (Fn) Sinuous NuxtJS NuxtJS (Fn)
Render 1018ms 865ms 297ms (single obj) 3743ms 1825ms
Hydration 85ms 68ms 5000ms+ 455ms 238ms
Hydration (with onclick) 165ms 117ms - 707ms 390ms
Hydration (x1000 dynamic) 75ms - - 185ms -
Full page size 43kb 39kb 60kb 169kb 169kb

That framework is faster then Vue + NuxtJS in:

  1. Full static: x3.5
  2. Static with event: x3
  3. Dynamic: x2.5
  4. Page size: x4

Features

  • <input checked="" disabled="" type="checkbox"> Single file components
  • <input checked="" disabled="" type="checkbox"> Statements (v-if, v-else-if, v-else)
  • <input checked="" disabled="" type="checkbox"> Slots
  • <input checked="" disabled="" type="checkbox"> Loops
  • <input checked="" disabled="" type="checkbox"> Functional components (automatic)
  • <input checked="" disabled="" type="checkbox"> Attributes inheritence to component
  • <input checked="" disabled="" type="checkbox"> Props (Can be used as Init data for component)
  • <input disabled="" type="checkbox"> Component lazy load
  • <input disabled="" type="checkbox"> Dynamic component
  • <input disabled="" type="checkbox"> Server side render
  • <input disabled="" type="checkbox"> Server side data fetching (once)

When Statements, slots, loops and props will be ready i gonna make full page perfromance to understand how Partial hydration works.

Bugs

  • Slots need to be fixed for render function in statement and loops (on refactor step)
  • Component children register is broken (because of loops, need to be fixed)
  • Disallow to use multiple nodes in loops, statements (so everywehere for now)

Problem of multiple nodes Its hard to understand which nodes should be hydrated because there is no tree before hydration/render. On render step its not necessary because of H function. But on hydration step there is no way to understand how many nodes should be replaced with new value.

Component example

<template>
  <div v-if="testProp" v-for="(item) in [1, 2, 3]" @click="reactive_click">
    test {{ visible }} - {{ item }}
  </div>
</template>

<script>
import { d } from 'import-some-thing'

/**
 * Reactive (stateful) data defines with $
 */
let $visible = d;
let $clicks2 = {
  a: 2
};
/**
 * Normal data defines as usual variable
 */
let clicks = 1;
let testProp = false;

/**
 * Will become computed prop
 */
let computed = () {
  let k = [];

  for(let d in [1,2,3]) {
    k.push(visible);
  }

  return visible * 2 * 5;
}

/**
 * Usual methods
 */
function click(event) {
  clicks++;
  alert(this._data.clicks + `test ${ computed }`)
}

function reactive_click(event2) {
  // Work with reactive as with usual variable
  visible = visible + 1;
}
</script>

Render

import Sinuous from '@sinuous/i';
import render from '@sinuous/render';
import TestComponent from 'components/test';
import TestPage from 'page/test';

Sinuous.registerComponent(TestComponent);

render(TestPage, document.getElementById('layout'));

Hydration

import Sinuous from '@sinuous/i';
import hydration from '@sinuous/hydration';
import TestComponent from 'components/test';
import TestPage from 'page/test';

Sinuous.registerComponent(TestComponent);

hydration(TestPage, document.getElementById('layout'));

Partial Hydration

All components have 2 functions that create on component compilation step (webpack-loader):

  1. Render
  2. Hydration

Thats why on Server side (like NuxtJS) components should be imported with ?ssrOnly and on client side with ?runtimeOnly You can use ?ssrOnly and ?runtimeOnlyparams to minify build code.

Server side

import TestComponent from 'components/test?ssrOnly';
import TestPage from 'page/test?ssrOnly';

Client side

import TestComponent from 'components/test?runtimeOnly';
import TestPage from 'page/test?runtimeOnly';

Webpack

Use @sinuous/loader to compile SFC. Use parseName to make component autonaming

{
  test: /\.sin/,
  use: [
      {
    loader: '@sinuous/loader',
    options: {
      parseName(file) {
        file = camelize(file);
        let rootPath = camelize(path.resolve(__dirname, '../'));

          let componentPath = file
            .split(rootPath)
            .join('')
            .replace(/\.sin/i, '')
            .replace(/Components/, '')
            .replace(/(\s|\/)/g, '');

          return componentPath;
        }
      }
      }
  ]
},

changelog

v2.1.3 [code, diff]

  • Fix: An out of memory bug when parsing numbers has been fixed. ([#228], [#229])

v2.1.2 [code, diff]

  • Fix: Bump minimist to v1.2.5. ([#222])

v2.1.1 [code, diff]

  • New: package.json and package.json5 include a module property so bundlers like webpack, rollup and parcel can take advantage of the ES Module build. (#208)
  • Fix: stringify outputs \0 as \\x00 when followed by a digit. (#210)
  • Fix: Spelling mistakes have been fixed. (#196)

v2.1.0 [code, diff]

  • New: The index.mjs and index.min.mjs browser builds in the dist directory support ES6 modules. (#187)

v2.0.1 [code, diff]

  • Fix: The browser builds in the dist directory support ES5. (#182)

v2.0.0 [code, diff]

  • Major: JSON5 officially supports Node.js v6 and later. Support for Node.js v4 has been dropped. Since Node.js v6 supports ES5 features, the code has been rewritten in native ES5, and the dependence on Babel has been eliminated.

  • New: Support for Unicode 10 has been added.

  • New: The test framework has been migrated from Mocha to Tap.

  • New: The browser build at dist/index.js is no longer minified by default. A minified version is available at dist/index.min.js. (#181)

  • Fix: The warning has been made clearer when line and paragraph separators are used in strings.

  • Fix: package.json5 has been restored, and it is automatically generated and committed when the version is bumped. A new build-package NPM script has been added to facilitate this.

v1.0.1 [code, diff]

This release includes a bug fix and minor change.

  • Fix: parse throws on unclosed objects and arrays.

  • New: package.json5 has been removed until an easier way to keep it in sync with package.json is found.

v1.0.0 [code, diff]

This release includes major internal changes and public API enhancements.

  • Major: JSON5 officially supports Node.js v4 and later. Support for Node.js v0.10 and v0.12 have been dropped.

  • New: Unicode property names and Unicode escapes in property names are supported. (#1)

  • New: stringify outputs trailing commas in objects and arrays when a space option is provided. (#66)

  • New: JSON5 allows line and paragraph separator characters (U+2028 and U+2029) in strings in order to be compatible with JSON. However, ES5 does not allow these characters in strings, so JSON5 gives a warning when they are parsed and escapes them when they are stringified. (#70)

  • New: stringify accepts an options object as its second argument. The supported options are replacer, space, and a new quote option that specifies the quote character used in strings. (#71)

  • New: The CLI supports STDIN and STDOUT and adds --out-file, --space, and --validate options. See json5 --help for more information. (#72, #84, and #108)

  • New: In addition to the white space characters space \t, \v, \f, \n, \r, and \xA0, the additional white space characters \u2028, \u2029, and all other characters in the Space Separator Unicode category are allowed.

  • New: In addition to the character escapes \', \", \\, \b, \f, \n, \r, and \t, the additional character escapes \v and \0, hexadecimal escapes like \x0F, and unnecessary escapes like \a are allowed in string values and string property names.

  • New: stringify outputs strings with single quotes by default but intelligently uses double quotes if there are more single quotes than double quotes inside the string. (i.e. stringify('Stay here.') outputs 'Stay here.' while stringify('Let\'s go.') outputs "Let's go.")

  • New: When a character is not allowed in a string, stringify outputs a character escape like \t when available, a hexadecimal escape like \x0F when the Unicode code point is less than 256, or a Unicode character escape like \u01FF, in that order.

  • New: stringify checks for a toJSON5 method on objects and, if it exists, stringifies its return value instead of the object. toJSON5 overrides toJSON if they both exist.

  • New: To require or import JSON5 files, use require('json5/lib/register') or import 'json5/lib/register'. Previous versions used json5/lib/require, which still exists for backward compatibility but is deprecated and will give a warning.

  • New: To use JSON5 in browsers, use the file at dist/index.js or https://unpkg.com/json5@^1.0.0.

  • Fix: stringify properly outputs Infinity and NaN. (#67)

  • Fix: isWord no longer becomes a property of JSON5 after calling stringify. (#68 and #89)

  • Fix: stringify no longer throws when an object does not have a prototype. (#154)

  • Fix: stringify properly handles the key argument of toJSON(key) methods. toJSON5(key) follows this pattern.

  • Fix: stringify accepts Number and String objects as its space argument.

  • Fix: In addition to a function, stringify also accepts an array of keys to include in the output as its replacer argument. Numbers, Number objects, and String objects will be converted to a string if they are given as array values.

v0.5.1 [code, diff]

This release includes a minor fix for indentations when stringifying empty arrays.

  • Fix: Indents no longer appear in empty arrays when stringified. (#134)

v0.5.0 [code, diff]

This release includes major internal changes and public API enhancements.

  • Major: JSON5 officially supports Node.js v4 LTS and v5. Support for Node.js v0.6 and v0.8 have been dropped, while support for v0.10 and v0.12 remain.

  • Fix: YUI Compressor no longer fails when compressing json5.js. (#97)

  • New: parse and the CLI provide line and column numbers when displaying error messages. (#101; awesome work by @amb26.)

v0.4.0 [code, diff]

Note that v0.3.0 was tagged, but never published to npm, so this v0.4.0 changelog entry includes v0.3.0 features.

This is a massive release that adds stringify support, among other things.

  • Major: JSON5.stringify() now exists! This method is analogous to the native JSON.stringify(); it just avoids quoting keys where possible. See the usage documentation for more. (#32; huge thanks and props @aeisenberg!)

  • New: NaN and -NaN are now allowed number literals. (#30; thanks @rowanhill.)

  • New: Duplicate object keys are now allowed; the last value is used. This is the same behavior as JSON. (#57; thanks @jordanbtucker.)

  • Fix: Properly handle various whitespace and newline cases now. E.g. JSON5 now properly supports escaped CR and CRLF newlines in strings, and JSON5 now accepts the same whitespace as JSON (stricter than ES5). (#58, #60, and #63; thanks @jordanbtucker.)

  • New: Negative hexadecimal numbers (e.g. -0xC8) are allowed again. (They were disallowed in v0.2.0; see below.) It turns out they are valid in ES5, so JSON5 supports them now too. (#36; thanks @jordanbtucker!)

v0.2.0 [code, diff]

This release fixes some bugs and adds some more utility features to help you express data more easily:

  • Breaking: Negative hexadecimal numbers (e.g. -0xC8) are rejected now. While V8 (e.g. Chrome and Node) supported them, it turns out they're invalid in ES5. This has been fixed in V8 (and by extension, Chrome and Node), so JSON5 officially rejects them now, too. (#36)

  • New: Trailing decimal points in decimal numbers are allowed again. (They were disallowed in v0.1.0; see below.) They're allowed by ES5, and differentiating between integers and floats may make sense on some platforms. (#16; thanks @Midar.)

  • New: Infinity and -Infinity are now allowed number literals. (#30; thanks @pepkin88.)

  • New: Plus signs (+) in front of numbers are now allowed, since it can be helpful in some contexts to explicitly mark numbers as positive. (E.g. when a property represents changes or deltas.)

  • Fix: unescaped newlines in strings are rejected now. (#24; thanks @Midar.)

v0.1.0 [code, diff]

This release tightens JSON5 support and adds helpful utility features:

  • New: Support hexadecimal numbers. (Thanks @MaxNanasy.)

  • Fix: Reject octal numbers properly now. Previously, they were accepted but improperly parsed as base-10 numbers. (Thanks @MaxNanasy.)

  • Breaking: Reject "noctal" numbers now (base-10 numbers that begin with a leading zero). These are disallowed by both JSON5 and JSON, as well as by ES5's strict mode. (Thanks @MaxNanasy.)

  • New: Support leading decimal points in decimal numbers. (Thanks @MaxNanasy.)

  • Breaking: Reject trailing decimal points in decimal numbers now. These are disallowed by both JSON5 and JSON. (Thanks @MaxNanasy.)

  • Breaking: Reject omitted elements in arrays now. These are disallowed by both JSON5 and JSON.

  • Fix: Throw proper SyntaxError instances on errors now.

  • New: Add Node.js require() hook. Register via json5/lib/require.

  • New: Add Node.js json5 executable to compile JSON5 files to JSON.

v0.0.1 [code, diff]

This was the first implementation of this JSON5 parser.

  • Support unquoted object keys, including reserved words. Unicode characters and escape sequences sequences aren't yet supported.

  • Support single-quoted strings.

  • Support multi-line strings.

  • Support trailing commas in arrays and objects.

  • Support comments, both inline and block.

v0.0.0 [code]

Let's consider this to be Douglas Crockford's original json_parse.js — a parser for the regular JSON format.