I'm looking to use https://github.com/flatiron/nconf in the browser. I have tried to use it with browserify but because nconf calls fs.readdirSync when it is required to scan directories for configuration, it fails in the browser.
// config.js
'use strict';
var nconf = require('nconf'); // this triggers the fs.readdirSync
var globalConfig = { my: { global: 1 }};
nconf.use('global', { type: 'literal', store: globalConfig });
var envConfig = { my: { env: 2 }};
nconf.use('env', { type: 'literal', store: envConfig });
module.exports = nconf;
Is it possible to use some sort of browserify transform (I didn't see a way to make it force the use of BRFS in nconf) or a way to use nconf (or some other similar library) to manage client side configuration?
If not nconf itself, then just something that would allow me to do something similar to this:
config.load('user-settings', { my : { user: 1 } });
config.load('global-settings', { my: { global: 2 } } );
config.get('my.user'); // 1
config.get('my.global'); // 2
config.unload('global-settings');
config.get('my.global'); // undefined
I recently had this issue myself. I decided to just put together my own config file, rather than pull in another library. here's what I ended up with:
/*
This file determines which environment we're in, and provides the app with the appropriate config
*/
export const defaults = {
searchURI: 'http://www.google.com/',
anotherURI: 'https://stackoverflow.com/',
};
export const dev = {
searchURI: 'https://www.bing.com/',
};
export const test = {
searchURI: 'https://www.yahoo.com/',
};
export const prod = {
searchURI: 'https://duckduckgo.com/',
};
const config = () => {
switch (process.env.YOUR_ENV_NAME) {
case 'dev':
return dev;
case 'test':
return test;
default:
return prod;
}
};
export default {
...defaults,
...config(),
};
using this pattern, you can import config just like you would nconf:
import config from './path/to/above/file.js';
const { searchURI, anotherURI } = config;
// OR
const searchURI = config.searchURI;
const anotherURI = config.anotherURI;
Related
How can I pass a variable between plugins in Rollup?
What I've tried:
// plugin-a.js
const pluginA = () => {
return {
name: 'pluginA',
async options(options) {
options.define = options.define || {};
options.define['foo'] = 'bar';
}
}
}
// plugin-b.js
const pluginB = (options = {}) => {
return {
name: 'pluginB',
buildStart: async (options) => {
console.log(options)
}
}
}
I'm getting a warning:
(!) You have passed an unrecognized option
Unknown input options: define. Allowed options: acorn, acornInjectPlugins, cache, context, experimentalCacheExpiry, external, inlineDynamicImports, input, makeAbsoluteExternalsRelative, manualChunks, maxParallelFileOps, maxParallelFileReads, moduleContext, onwarn, perf, plugins, preserveEntrySignatures, preserveModules, preserveSymlinks, shimMissingExports, strictDeprecations, treeshake, watch
It seems passing data should be done by what Rollup refers to as Direct plugin communication. This is working for me. I feel this is very hard coupled though.
function parentPlugin() {
return {
name: 'parent',
api: {
//...methods and properties exposed for other plugins
doSomething(...args) {
// do something interesting
}
}
// ...plugin hooks
};
}
function dependentPlugin() {
let parentApi;
return {
name: 'dependent',
buildStart({ plugins }) {
const parentName = 'parent';
const parentPlugin = plugins.find(plugin => plugin.name === parentName);
if (!parentPlugin) {
// or handle this silently if it is optional
throw new Error(`This plugin depends on the "${parentName}" plugin.`);
}
// now you can access the API methods in subsequent hooks
parentApi = parentPlugin.api;
},
transform(code, id) {
if (thereIsAReasonToDoSomething(id)) {
parentApi.doSomething(id);
}
}
};
}
There's also Custom module meta-data, however when I read the meta I always get null.
I'm trying to leverage nuxtjs SSG capabilities by creating a static web site where the pages content and navigation are fetched from an API.
I already found my way around on how to dynamically generate the routes by defining a module where I use the generate:before hook to fetch the pages content and routes. When creating the routes I store the page content as the route payload. The following code does just that and works as intended.
modules/dynamicRoutesGenerator.js
const generator = function () {
//Before hook to generate our custom routes
this.nuxt.hook('generate:before', async (generator, generatorOptions) => {
generator.generateRoutes(await generateDynamicRoutes())
})
}
let generateDynamicRoutes = async function() {
//...
return routes
}
export default generator
Now the problem I'm facing is that I have some navigation components that need the generated routes and I was thinking to store them into the vuex store.
I tried the generate:done hook but I don't know how to get the vuex store context from there. What I ended up using was the nuxtServerInit() action because as stated in the docs:
If nuxt generate is ran, nuxtServerInit will be executed for every dynamic route generated.
This is exactly what I need so I'm trying to use it with the following code:
store/index.js
export const actions = {
nuxtServerInit (context, nuxtContext) {
context.commit("dynamicRoutes/addRoute", nuxtContext)
}
}
store/dynamicRoutes.js
export const state = () => ({
navMenuNivel0: {}
})
export const mutations = {
addRoute (state, { ssrContext }) {
//Ignore static generated routes
if (!ssrContext.payload || !ssrContext.payload.entrada) return
//If we match this condition then it's a nivel0 route
if (!ssrContext.payload.navMenuNivel0) {
console.log(JSON.stringify(state.navMenuNivel0, null, 2));
//Store nivel0 route, we could use url only but only _id is guaranteed to be unique
state.navMenuNivel0[ssrContext.payload._id] = {
url: ssrContext.url,
entrada: ssrContext.payload.entrada,
navMenuNivel1: []
}
console.log(JSON.stringify(state.navMenuNivel0, null, 2));
//Nivel1 route
} else {
//...
}
}
}
export const getters = {
navMenuNivel0: state => state.navMenuNivel0
}
The action is indeed called and I get all the expected values, however it seems like that with each call of nuxtServerInit() the store state gets reset. I printed the values in the console (because I'm not sure even if it's possible to debug this) and this is what they look like:
{}
{
"5fc2f4f15a691a0fe8d6d7e5": {
"url": "/A",
"entrada": "A",
"navMenuNivel1": []
}
}
{}
{
"5fc2f5115a691a0fe8d6d7e6": {
"url": "/B",
"entrada": "B",
"navMenuNivel1": []
}
}
I have searched all that I could on this subject and altough I didn't find an example similar to mine, I put all the pieces I could together and this was what I came up with.
My idea was to make only one request to the API (during build time), store everything in vuex then use that data in the components and pages.
Either there is a way of doing it better or I don't fully grasp the nuxtServerInit() action. I'm stuck and don't know how to solve this problem and can't see another solution.
If you made it this far thanks for your time!
I came up a with solution but I don't find it very elegant.
The idea is to store the the API requests data in a static file. Then create a plugin to have a $staticAPI object that expose the API data and some functions.
I used the build:before hook because it runs before generate:before and builder:extendPlugins which means that by the time the route generation or plugin creation happen, we already have the API data stored.
dynamicRoutesGenerator.js
const generator = function () {
//Add hook before build to create our static API files
this.nuxt.hook('build:before', async (plugins) => {
//Fetch the routes and pages from API
let navMenuRoutes = await APIService.fetchQuery(QueryService.navMenuRoutesQuery())
let pages = await APIService.fetchQuery(QueryService.paginasQuery())
//Cache the queries results into staticAPI file
APIService.saveStaticAPIData("navMenuRoutes", navMenuRoutes)
APIService.saveStaticAPIData("pages", pages)
})
//Before hook to generate our custom routes
this.nuxt.hook('generate:before', async (generator, generatorOptions) => {
console.log('generate:before')
generator.generateRoutes(await generateDynamicRoutes())
})
}
//Here I can't find a way to access via $staticAPI
let generateDynamicRoutes = async function() {
let navMenuRoutes = APIService.getStaticAPIData("navMenuRoutes")
//...
}
The plugin staticAPI.js:
import APIService from '../services/APIService'
let fetchPage = function(fetchUrl) {
return this.pages.find(p => { return p.url === fetchUrl})
}
export default async (context, inject) => {
//Get routes and files from the files
let navMenuRoutes = APIService.getStaticAPIData("navMenuRoutes")
let pages = APIService.getStaticAPIData("pages")
//Put the objects and functions in the $staticAPI property
inject ('staticAPI', { navMenuRoutes, pages, fetchPage })
}
The APIService helper to save/load data to the file:
//...
let fs = require('fs');
let saveStaticAPIData = function (fileName = 'test', fileContent = '{}') {
fs.writeFileSync("./static-api-data/" + fileName + ".json", JSON.stringify(fileContent, null, 2));
}
let getStaticAPIData = function (fileName = '{}') {
let staticData = {};
try {
staticData = require("../static-api-data/" + fileName + ".json");
} catch (ex) {}
return staticData;
}
module.exports = { fetchQuery, apiUrl, saveStaticAPIData, getStaticAPIData }
nuxt.config.js
build: {
//Enable 'fs' module
extend (config, { isDev, isClient }) {
config.node = { fs: 'empty' }
}
},
plugins: [
{ src: '~/plugins/staticAPI.js', mode: 'server' }
],
buildModules: [
'#nuxtjs/style-resources',
'#/modules/staticAPIGenerator',
'#/modules/dynamicRoutesGenerator'
]
This seems dumb, but I have it setup like this:
in config/index.js:
module.exports = {
API_LOCATION: 'http://localhost:8080/api/'
}
then in src/app.js I have:
import Vue from 'vue'
import VueRouter from 'vue-router'
import VueResource from 'vue-resource';
Vue.use(VueRouter);
Vue.use(VueResource);
const App = require("./app.vue");
const home = require("./components/home.vue");
const config = require('../config');
window.config = config;
Then in src/components/home.vue, I have a script block that uses it like so:
<script>
module.exports = {
data: function() {
return {
obj: null
}
},
created: function() {
this.$http.get(config.API_LOCAITON + '/call').then(res => {
// Do some business
}, res => {
// Handle some error
});
}
}
</script>
This works but it strikes me as a bad idea to use window to handle an application configuration. What's the more canonical approach here?
Import it.
<script>
import config from "../config"
module.exports = {
data: function() {
return {
obj: null
}
},
created: function() {
this.$http.get(config.API_LOCATION + '/call').then(res => {
// Do some business
}, res => {
// Handle some error
});
}
}
</script>
Or just the location.
<script>
import { API_LOCATION } from "../config"
module.exports = {
data: function() {
return {
obj: null
}
},
created: function() {
this.$http.get(API_LOCATION + '/call').then(res => {
// Do some business
}, res => {
// Handle some error
});
}
}
</script>
PROD: config/prod.env.js append your VAR='"value"'
'use strict'
module.exports = {
NODE_ENV: '"production"',
API_LOCATION: '"https://production URL"'
}
DEV: config/dev.env.js append your VAR='"value"'
'use strict'
const merge = require('webpack-merge')
const prodEnv = require('./prod.env')
module.exports = merge(prodEnv, {
NODE_ENV: '"development"',
API_LOCATION: '"http://localhost"'
})
Your variable will available in process.env.API_LOCATION or
process.env.VAR_NAME
TL;DR; Global configuration in Vue is done using the .env and .env.production files that you create in the route folder of the app (next to package.json)
I can confirm that in Vue using the Vue CLI .env file gets loaded automatically when you run npm run serve
But keep in mind the following that I checked:
Variables in the .env files have to start with VUE_APP prefix to be automatically picked up and available as process.env.VUE_APP_MYVAR in the code
If you're defining JS objects such as 'VUE_APP_MY_OBJECT={a:100}then you'll have to parse it to JSON before using it in the codeJSON.parse(process.env.VUE_APP_MY_OBJECT)`
If you're not using Webpack you might have to fiddle a bit with it to pick these config files up. According to this answer it looks like webpack should do it automatically. Dunno
Simply set ip path(or localhost) in local storage when login successfull and get value from local storage where you need through out the project.
here how you set value in localstrage.
// Set value in IpAdress
localstorage.setItem('IpAddress','192.168.100.100:8080');
// Get value from IpAddress
localstorage.getItem('IpAddress');
in my case whole path looks like:
localstorage.getItem('IpAddress')+/api/controller/method|
Stuck with this one.
I am using laravel elxir with tsify to generate my js. I run the typescript through factor-bundle to split common js modules into a seperate files. I don't think though that will be a problem in this case because everything is in a spec.js
spec.ts
/// <reference path="../../../typings/index.d.ts" />
import "jasmine-jquery";
// #start widgets
import "./widgets/common/widget-factory/test";
factory-widget/index.ts
export class WidgetFactory {
.... this contains a require call to browser.service which i need to mock
}
factory-widget/test.ts
...
import {WidgetFactory} from "./index";
const proxyRequire = require("proxyquire");
it("should output the factory items", ()=> {
proxyRequire('./widgets/browser.service/index',{
"#global": true,
});
}
browser-service.ts
...
export class BrowserService implements IBrowserService{
//details
}
Getting an error Uncaught TypeError: require.resolve is not a function on line 262.
Here is the code ( yeah it's over 20,000 lines ) how else are you supposed to debug this stuff . ¯_(ツ)_/¯
I've looked at Stubbing with proxyquire. I am not holding my breath getting an answer on this one.
Edit: 06-09-2016
Proxquire is needed to overide the require call in the boot method of the WidgetFactory class
In factory-widget/index.ts:
boot(output = true):any {
let required = {};
if (this._sorted.length) {
this._sorted.forEach((key)=> {
if (output) {
console.log(`${this._path}${key}/index`);
// this is where is need to overide the call to require.
required[key] = require(`${this._path}${key}/index`);
}
});
this._sorted.forEach((key)=> {
let dependencies = {},
module = this._factory[key];
if (module.hasOwnProperty(this.dependencyKey)) {
module[this.dependencyKey].map((key)=> {
dependencies[_.camelCase(key)] = this.isService(module) ? new required[key] : key;
});
}
if (this.isTag(module)) {
if (output) {
document.addEventListener("DOMContentLoaded", ()=> {
riot.mount(key, dependencies);
});
}
//console.log(key,dependencies);
}
else {
}
})
}
}
I've added a proxyquireify example to the tsify GitHub repo. It's based on the simple example in the proxyquireify README.md.
The significant parts are the re-definition of require to call proxyquire in foo-spec.ts:
const proxyquire = require('proxyquireify')(require);
require = function (name) {
const stubs = {
'./bar': {
kinder: function () { return 'schokolade'; },
wunder: function () { return 'wirklich wunderbar'; }
}
};
return proxyquire(name, stubs);
} as NodeRequire;
and the configuration of the proxyquire plugin in build.js:
browserify()
.plugin(tsify)
.plugin(proxyquire.plugin)
.require(require.resolve('./src/foo-spec.ts'), { entry: true })
.bundle()
.pipe(process.stdout);
If you build the bundle.js and run it under Node.js, you should see that the message written to the console includes strings returned by the functions in the stubbed ./bar module.
I have file foo.js:
export function bar (m) {
console.log(m);
}
And another file that uses foo.js, cap.js:
import { bar } from 'foo';
export default m => {
// Some logic that I need to test
bar(m);
}
I have test.js:
import cap from 'cap'
describe('cap', () => {
it('should bar', () => {
cap('some');
});
});
Somehow I need override implementation of bar(m) in test. Is there any way to do this?
P.S. I use babel, webpack and mocha.
Ouch.. I found solution, so I use sinon to stub and import * as foo from 'foo' to get object with all exported functions so I can stub them.
import sinon from 'sinon';
import cap from 'cap';
import * as foo from 'foo';
sinon.stub(foo, 'bar', m => {
console.log('confirm', m);
});
describe('cap', () => {
it('should bar', () => {
cap('some');
});
});
You can replace/rewrite/stub exports only from within the module itself. (Here's an explanation)
If you rewrite 'foo.js' like this:
var bar = function bar (m) {
console.log(m);
};
export {bar}
export function stub($stub) {
bar = $stub;
}
You can then override it in your test like this:
import cap from 'cap'
import {stub} from 'foo'
describe('cap', () => {
it('should bar', () => {
stub(() => console.log('stubbed'));
cap('some'); // will output 'stubbed' in the console instead of 'some'
});
});
I've created a Babel plugin that transforms all the exports automatically so that they can be stubbed: https://github.com/asapach/babel-plugin-rewire-exports
While #Mike solution would work in old versions of sinon, it has been removed since sinon 3.0.0.
Now instead of:
sinon.stub(obj, "meth", fn);
you should do:
stub(obj, 'meth').callsFake(fn)
Example of mocking google oauth api:
import google from 'googleapis';
const oauth2Stub = sinon.stub();
sinon.stub(google, 'oauth2').callsFake(oauth2Stub);
oauth2Stub.withArgs('v2').returns({
tokeninfo: (accessToken, params, callback) => {
callback(null, { email: 'poo#bar.com' }); // callback with expected result
}
});
You can use babel-plugin-rewire (npm install --save-dev babel-plugin-rewire)
And then in test.js use the __Rewire__ function on the imported module to replace the function in that module:
// test.js
import sinon from 'sinon'
import cap from 'cap'
describe('cap', () => {
it('should bar', () => {
const barStub = sinon.stub().returns(42);
cap.__Rewire__('bar', barStub); // <-- Magic happens here
cap('some');
expect(barStub.calledOnce).to.be.true;
});
});
Be sure to add rewire to your babel plugins in .babelrc:
// .babelrc
{
"presets": [
"es2015"
],
"plugins": [],
"env": {
"test": {
"plugins": [
"rewire"
]
}
}
}
Lastly, as you can see the babel-plugin-rewire plugin is only enabled in the test environment, so you should call you test runner with the BABEL_ENV environment variable set to test (which you're probably doing already):
env BABEL_ENV=test mocha --compilers js:babel-core/register test-example.js
Note: I couldn't get babel-plugin-rewire-exports to work.
This was definitely a gotcha for me too...
I created a little util to workaround this limitation of sinon. (Available in js too).
// mockable.ts 👇
import sinon from 'sinon'
export function mockable<T extends unknown[], Ret>(fn: (...fnArgs: T) => Ret) {
let mock: sinon.SinonStub<T, Ret> | undefined
const wrapper = (...args: T) => {
if (mock) return mock(...args)
return fn(...args)
}
const restore = () => {
mock = undefined
}
wrapper.mock = (customMock?: sinon.SinonStub<T, Ret>) => {
mock = customMock || sinon.stub()
return Object.assign(mock, { restore })
}
wrapper.restore = restore
return wrapper
}
If you paste the above snippet into your project you can use it like so
foo.js
import { mockable } from './mockable'
// we now need to wrap the function we wish to mock
export const foo = mockable((x) => {
console.log(x)
})
main.js
import { foo } from './foo'
export const main = () => {
foo('asdf') // use as normal
}
test.js
import { foo } from './foo'
import { main } from './main'
// mock the function - optionally pass in your own mock
const mock = foo.mock()
// test the function
main()
console.assert(mock.calledOnceWith('asdf'), 'not called')
// restore the function
stub.restore()
The benefit of this approach is that you don't have to remember to always import the function in a certain way. import { foo } from './foo' works just as well as import * as foo from './foo'. Automatic imports will likely just work in your IDE.