Does it still make sense to use Express.js with io.js or does io.js have some sort of built-in functionality now that Express.js gave or do we still rely on Express.js. I mean Express.js doesn't extend io.js, it extends node.js.
I'm trying to figure out how you'd do the Web Server part if using io.js.
Think of io.js just as a new release of node.js. They have essentially identical feature sets and are essentially drop-in compatible with each other. You use any library from npm in the same matter on either io.js or node.js. Express.js and most other libraries in npm work on both just fine. There are still many npm modules that contain binary add-ons that don't yet work with io.js, but that should sort itself out quickly enough.
Related
I have a flutter application running on the Dart programming language. I have published an npm package that I want to use in my flutter application. Is there any way I can use the existing npm package in dart or somehow automatically convert npm package to dart package?
Dart and Javascript are different programming languages. While it might be possible to transpile you lib from js to dart, it is not easily doable, nor maintainable, and most likely not generalisable.
Unless you are comfortable with both languages and their limitations, you should not try to convert an NPM package to some Dart code unless you don't have any other choice.
I would advise you to rewrite the NPM package in Dart instead.
REMARK : this answer is based on the hypothesis that you don't want to integrate JS code in a webview but actually use the code inside the Dart VM.
It really depends on what you are going for,
Are you compiling your flutter app for web ?
Dart's JS library might be what you're looking for...
Are you going to compile for Android?
Then there's Android JS, Flutter_liquidcore, Interactive_WebView.
It Really depends on what you are expecting the application to compile to.
But don't expect native app experience...
We converted a fairly large js library to dart (manually) it was surprisingly easy as the two languages are very similar.
Tl:Dr: Use a WebView or run a Node/Deno/SpiderNode subprocess.
Let me answer the title of your question; you're going to have to wrap it in a webview or a node.js subprocess that can communicate over IPC using something like DBUS. This is cutting edge hard experimental tech as of 2021.
If it's a small we'll tested package, you could babel transpile it to Dart and automatically run the tests.
if it's a large package like Nuxt, forget about it, or put it in a WebView.
If it's your own package, consider translating it to Dart and then transpiling to JavaScript. It's Trivial to go from Dart to readable Node/ES2021.
But do you want to really use Dart for everything? Will the community for Dart be as strong as the Trois, Vue, WebXR typescript community in 5 to 80 years?
My advice would be to use Dart for what it's good for, client side JavaScript. And use well tested packages that can transpile from TypeScript to Dart, or a webview.
Could you use glue subprocesses to run node/deno/spidernode? see the Termux F-Droid debacle, where if you want to run a Node server on the latest version of Termux, you can't use the Google Play Store, but F-Droid. Basically, JIT will make your App Store approval process slower, but they will only be able to ban it from a policy standpoint, not from a technical one. You should not use npm install, but instead bundle the installed modules with the released app build.
How does the deprecation of JIT play in a world of Fuku native like PWA's and Spectre? From a security perspective, deploying to Kotlin/Swift is a narrow minded way of outsourcing the responsibility of making compiler patches happen, you should be building at least one a month, or leave the business. From a privacy perspective, the handshakes and huristic privacy protections will come to PWAs as hard as they will for binary apps.
I understand that there are basically two ways of deploying a vue app. One is to include the library directly with:
<script src="https://unpkg.com/vue"></script>
and the other method is to use vue cli, then build with npm
I saw many tutorials, but I could not get a quite understanding of what exactly is different, or what are the advantages of each side?
Vue can be used through -
CLI - es NPM / yarn to install Vue. Sets up default config pretty nicely
NPM - (or yarn) install Vue directly and configure yourself [recommended for large projects]
CDN - refer directly through in your HTML /script
If you want to just include Vue in your script that has a larger objective, or accomplish something 'comparatively' simple, just use CDN. This is a good way to use the power of Vue without worrying about the intermediate build steps specific to Vue. Your completed product will continue to refer to Vue from CDN.
Also, each CDN lookup will mean an additional server request - this is likely to serve content faster than your own server, but an additional lookup nevertheless. Having tens of CDNs is not ideal.
For larger projects, NPM is ideal. You would want the power of Single File Components with (files with a .vue extension) that play well with editors, and provide a more structured way to develop the application (incl. things like scoped CSS).
Also, the build step through local Vue is a must if you use functions not supported by browsers today, or want to support older browsers.
https://v2.vuejs.org/v2/guide/installation.html
I've been developing in Aurelia-CLI for about 3 months and like it so far. I think it's a solid framework and obviously escalating in support and usage. That's a good thing!
Before I develop much more of my large app, I'm wondering if I'm using the best build system. I've only tried Aurelia-CLI and am not really familiar with Webpack or JSPM, and therefore I don't know what I'm missing. Are there any clear advantages or disadvantages in using either of the other two build systems, or is using the CLI the most clean and supported approach? Since I'm developing independently, I don't have any external constraints.
Thanks for your help.
UPDATE
This answer is almost two years old. Feel free to research updates and provide another more complete answer and I can replace this answer or point to that answer. Thanks!
Aurelia CLI
Aurelia CLI is great for getting started. It's important to understand that under the covers the CLI is using require.js but proxies the configuration through aurelia.json in your application. This means that you need to understand how to configure aurelia to work with require.js at the moment. Once you need to start configuring to match your workflow or change build steps up it gets a bit cumbersome at the moment. We are working to improve this. There are many features planned for the Aurelia CLI but given at the time of writing this that it is in an alpha / beta state it should generally be used on proof of concept or other smaller apps, not production-ready large scale apps yet.
WebPack
WebPack is arguably the most popular kid on the block at the moment. WebPack is not a module loader, but a bundler. It's important to understand this because while we strive to make Aurelia work great with all module loaders WebPack by default is not in charge of loading modules so a dynamically loaded application requires the developer to expand on this. WebPack is strong in creating optimized bundles and can be easy to use as long as you are comfortable with configure WebPack. WebPack has considerably more GitHub stars due to the popularity from React using WebPack it's hard to say whether the choice is better when using Aurelia simply because of the number of GitHub stars.
JSPM / System.js
Some of the skeletons use JSPM and System.js. The reason is that these are the closest to 'spec compliant' solutions. JSPM tries to help as much as possible when loading from the JSPM registry. If not yet available in the registry you can load from NPM or GitHub directly. From a module loading perspective you use a config.js file that is (usually) automatically maintained when installing dependencies to improve the developer workflow.
Side biased note
On most larger apps at the moment I typically prefer using JSPM / System.js simply because I have a great understanding of the tooling and prefer the control that I am provided. I work on a great number of Aurelia apps that are in production and typically reserve CLI for smaller proof of concept apps and WebPack is a great alternative but I prefer the flexibility and understanding I have with JSPM / System.js at the moment.
The CLI isn't currently feature complete, but it is a much simpler setup. Webpack can basically do anything you want to do, but you'll be maintaining your webpack configuration just as much as you maintain your Aurelia code.
Ok, maybe not just as much, but you'll have to learn Webpack to use webpack. The Aurelia CLI is simple to get started, but has some definite limitations. For example, CSS files that reference external resources won't work w/the Aurelia CLI, but they should work fine with Webpack.
First, I can understand if this post gets shutdown due to its subjective nature.
I believe it's time to re-visit the answers about Aurelia CLI being a second-class tool. I respect both PW Kad and Ashley Grant immensely, but I am just not convinced that a statement like this is true anymore:
There are many features planned for the Aurelia CLI but given at the
time of writing this that it is in an alpha / beta state it should
generally be used on proof of concept or other smaller apps, not
production-ready large scale apps yet.
Notably, I have a production application that way back in the day I started with Aurelia CLI, and changed it to JSPM precisely for the reasons noted. But recently, I rebuilt the same app from scratch using the CLI and I realize that it is much easier to use, particularly managing modules and publishing! And this is an app with Google Maps, Google Analytics, Auth0, DevExpress, Bootstrap, etc.
Just think it is time to give Aurelia CLI a little love. It's ready.
Aurelia CLI is the most preferred option with this announcement.
http://aurelia.io/blog/2017/08/18/aurelia-cli-webpack-update/
Now It has more flexibility for your choice of preferences.
I'm writing an aurelia application with aurelia-fetch-client library.
When I try to post an object to my api service I get the error:
TypeMismatchError
In console only from Edge. Other browsers (Chrome, Firefox and IE11) have no problems. There are no description or any other details about it.
In all but the newest versions of Edge you need to include the fetch polyfill in order for aurelia-fetch-client to work. I think it's supported since version 14 but I wouldn't necessarily rely on it. Edge is known to be quirky with some of these things (the Promise implementation is also horribly slow, which is why I personally always use bluebird)
You can install it with npm i whatwg-fetch --save and make sure to import it + include it in your bundle config (the instructions for this depend on which build system you're using)
I have an online site and Im going to overhaul it. While I cannot find any information about if I can use npm and some kind of package manager (e.g Webpack) on live site which is on shared host (which has latest Node.js, npm support etc), Im going to develope it locally and worry about "publishing" it later on.
Is there a way to somehow covert my current downloaded project to npm project or Im better off just starting a new project? This is rather confusing, I've never used npm before.
Im using WordPress, everything is run with PHP atm but Im going to overhaul it and use Node.js.
NPM can be used to collect and manage Javascript dependencies for the browser so as to create a stand alone front-end JS app. BUT, bower is probably a better choice.
NPM is designed to manage Javascript dependencies for projects that use NodeJS or use the CommonJS module format for requiring modules.
Bower is specifically a package manager (like npm or composer) but it is meant to manage browser based javascript dependencies.
Currently, javascript doesn't have a formally defined module/import system, so a number of competing require() functions are been produced. NodeJS comes with a require() function that searches the npm/ folder for modules. Browserify is a pre-processor that can scan the npm folder for a dependency and all of its dependencies and bundle them into 1 file for a browser to download (because browser's don't have a require() function because the JS standard doesn't have define one)
I might be rambling here, but you should probably look at bower, and then - at some point in the future - look at either browserify or requirejs to combine and optimize all the JS plugins for your front-end app.
Edit for clarity:
Basically there are 2 engines to run your JS application: the browser or NodeJS. NodeJS needs npm, the browser has no idea what npm is. If you want to write a clean, single page app, all JS front-end for a PHP backend, you don't need NodeJS, and therefore don't need npm. Using npm will prematurely complicate the development of a front-end browser app because it will force you to decide on a require() implementation (Browserify or Requirejs) right from the start.