This is based off work by @mrkishi, and is an attempt to do the following:
set things that have actually changedGiven a situation like this...
| <InputAdd></InputAdd> | |
| <List></List> | |
| <script> | |
| import InputAdd from './InputAdd.html' | |
| import List from './List.html' | |
| import { Store } from 'svelte/store.js'; | |
| javascript:!function()%7B%22use%20strict%22%3Bfunction%20t(t)%7Breturn%20t*%3D-.002%2C%5B50*Math.sin(t)%2C50*Math.cos(t)%5D%7Dfunction%20e(t)%7Breturn%22hsl(%22%2BMath.max(0%2C90-.5*t)%2B%22%2C%20100%25%2C%2050%25)%22%7Dfunction%20n(e)%7Bfor(var%20n%3D%5B%5B0%2C0%5D%5D%2Ci%3De.lastTime%3Bi%3Ce.time%3Bi%2B%3D4)n.push(t(i))%3Breturn%20n.push(t(e.time))%2C%22M%22%2Bn.join(%22L%22)%2B%22Z%22%7Dfunction%20i()%7Bvar%20t%3Dfunction(t)%7Breturn%20document.createElement(t)%7D(%22style%22)%3Bt.id%3D%22svelte-731919397-style%22%2Ct.textContent%3D%22svg%5Bsvelte-731919397%5D%2C%5Bsvelte-731919397%5D%20svg%7Bposition%3Afixed%3Bright%3A1em%3Bbottom%3A1em%3Bwidth%3A100px%3Bheight%3A100px%7D%22%2Cf(t%2Cdocument.head)%7Dfunction%20r(t%2Ce)%7Bfor(var%20n%2Ci%2Cr%2Co%3De.frames%2C_%3D%5B%5D%2Cm%3D0%3Bm%3Co.length%3Bm%2B%3D1)_%5Bm%5D%3Ds(t%2Ch(%7B%7D%2Ce%2C%7Bframe%3Ao%5Bm%5D%2Ci%3Am%7D))%3Breturn%7Bc%3Afunction()%7Bn%3Dc(%22svg%22)%2Ci%3Dc(%22g%22)%2Cr%3Dc(%22circle%22)%3Bfor(var%20t%3D0%3Bt%3C_.length%3Bt%2B%3D1)_%5Bt%5D.c()%3 |
| #include <stdio.h> | |
| #include <dirent.h> | |
| int main (int argc, char *argv[ ]) | |
| { | |
| DIR *dp; | |
| struct dirent *dirp; | |
| if (argc != 2) { | |
| fprintf(stderr, "usage: ls directory_name\n"); |
The fundamental problem with top-level await is that it prevents two unrelated modules doing asynchronous work in parallel, because module evaluation order is strictly deterministic.
What if we provided an escape valve that relaxed that constraint?
// main.js
import foo from async './foo.js';
import bar from async './bar.js';
import baz from './baz.js';
Am planning to open source this when I get some breathing room, but a gist will do for now.
Usage:
<Scroller top=0.2 bottom=0.8 threshold=0.5 bind:index bind:offset bind:progress parallax>
<div slot='background'>
<!-- fixed, or parallaxing (if the parallax option is set) background -->
</div>
| <body> | |
| <header> | |
| <h1>Hello world!</h1> | |
| </header> | |
| <main> | |
| <noscript id='__await_block_0_start'></noscript> | |
| <p>loading...</p> | |
| <noscript id='__await_block_0_end'></noscript> | |
| </main> |
Bear with me while I think aloud about this (please comment there, not here!). Goals:
Let's start with a single store that is external to the component tree. Our top-level <App> component connects to it:
From this Twitter convo. It would be cool to have a plugin that allowed you to use bare module specifiers (import * as React from 'react') in your app, and converted them to unpkg.com URLs.
Say you have these two source files:
// main.js
import * as React from 'react';
import * as ReactDOM from 'react-dom';
import message from './message.js';
Rollup 0.48 introduces a few changes to the options object, because the current options are confusingly different between the CLI and the options exported by your config file.
entry is now inputsourceMap and sourceMapFile are now sourcemap and sourcemapFile (note casing)moduleName is now nameuseStrict is now strictThe dest and format options are now grouped together as a single output: { file, format, ... } object. output can also be an array of { file, format, ... } objects, in which case it behaves similarly to the current targets. Other output options — exports, paths and so on — can be added to the output object (though they will fall back to their top-level namesakes, if unspecified).