micro-app | 微前端的使用

For anyone who comes to this question - I found a great article that I managed to follow to get what I needed from this.

https://medium.com/@timon.grassl/how-to-use-routing-in-angular-web-components-c6a76449cdb

The gist of the article is that you can use a named router outlet inside the element app. You then choose either to use the RouterTestingModule to handle routes and not have the overall application URL change on nevigate, or do some poking around to get the named router outlet to affect the URL. Well worth a read.

You check my sample project here https://github.com/xmlking/angular-elements-in-angular-shell I am still planing improve developer experience

When you are working you split your application down into services based on certain guidelines( be it biz driven or logic entities being together) and this will lead to similar problem as you have mentioned. My take on that is when you working on your service you want to take control of what your view of that service is. User service is free to define task as they seem fit and you may use User as is or you can change user model as suited for Task service. Important thing is you get to decide (ofcourse aligned to the User service). They are source of truth for user and you are for Task.

Your view of User is built from what ever User service says but you can structure it way you want to. So you do have to define User model but it not tightly coupled. Important thing for you is to make sure that you never share your model of user out to the world. Always the reference.

User service is free to add anything to the user model ( and remove provided others are not using and if they are, User service is breaking a contract) And you are free to pick and choose from the User Response you get from user service.

Thats how two are decoupled. If you talk about code redundancy, where in different services have to create a user model and make api calls and handle failures, well thats a different issue, and you can mitigate it by asking services to expose packages with models and basics calls.

I don't know for sure. But looking at your dependencies, when I google for AppsFly it seems like a platform which delivers dynamic code to apps. And Google Play policy says (as you quote)

An app distributed on Google Play may not modify, replace, or update itself using any method other than Google Play's update mechanism.

So using AppsFly might be against the Policy. To clarify, I know nothing about AppsFly as their website doesn't have much information, but it sure sounds like it breaks the Google Play policy.

I received an answer from @dna2github, the creator of NodeBase (thanks a lot!) that I'll include here (with permission):

Hi,

Thx for your question. I will do a brief answer in my view.

1. Running V8 javascript engine on android which includes NodeJS

pros:

• integrated with Java world; can get full control of code.

cons:

• a little hard to integrate with 3rd packages (need time to learn how).
• need to learn about NodeJS and V8 things and also J2V8 docs (it consume long time).

2. Compile NodeJS as a native library (using node-on-android)

pros:

• focus on js dev and no need to consider android side.
• less learning time; similar to Cordova phonegap ....

cons:

• js app => apk is a black box.

3. Running NodeJS on Android using Termux

pros:

• flexible

cons:

• no gui

4. Other interesting approaches

Not familar with LiquidCore; build micro service especially from url, I think, is to resolve no direct available storage on iOS. react-native-node the Android part is based on NodeBase method and use the prebuilt binary.

For NodeBase:

pros:

• similar to 3; difference is that it has its own gui to start/stop app.
• it can be a template for everything; for example, if would like to run django, you just need to replace node to python; rails, ruby...

cons:

• native process access problem; the process cannot inherit access from Android app.
• happy toy happy open source not like a commercial app; need more design if want to distribute to customers

At first, I run node in terminal; I find only dev can easily to use it to start js app. My friends and families also wanna some tools for example make water mark on picture in batch. NodeBase is created for them to easy to start/stop app. Then they just need to open browser to use it. My another idea to create NodeBase is that we can build sharable applications that can be shared in the same Wi-Fi. When host starts an app, it can be visited by near people. Then they can work and play together. For example, we play werewolf and when there is no judge, we will start the werewolf app to have a judge for the first round. We can also share files between devices via download/upload.

For me, I can build what I want flexibly for example, I would like to make my Android as a machine learning runner; it can help me run machine learning programs at anytime (with node and python, thus in my another repo: dna2oslab is focus on building binaries) to make use of phone running time.

For you, if wanna port your app in a short time, I recommend 2; if you have time and other resources, 1 is better. 3 if you just make a toy/demo. 4 other is always possible and just do your imagination to create works.

Best wishes, Seven

One way to think about contexts is to view them as a layer of abstraction when you are designing your app. As it is put in the docs,

Phoenix projects are structured like Elixir and any other Elixir project – we split our code into contexts. A context will group related functionality, such as posts and comments, often encapsulating patterns such as data access and data validation. By using contexts, we decouple and isolate our systems into manageable, independent parts.

Models themselves with their underlying schema are therefore the units which you group into contexts. Models can have more fine-grained api which is necessary for their inner workings. Contexts in turn only expose methods which are useful to other application components or contexts, thus hiding the underlying implementation detail of the models.

It is not necessary to extend the context boundary down through the models level for absolute isolation. Instead you create your schema as usual, with all the many-to-many relations you need on the models. Then implement whatever low-level methods on the models directly. Put in the context only those methods which are either public api of the context or private methods touching more than one model.

UPDATE

In your scenario you might have e.g. Chats context where you put the rooms method. Its signature might be e.g. Chats.rooms_of(user). Repo.get() also goes to the context as, e.g. def get(id), do: Repo.get(User, id). The models end up containing changesets, validation methods, private methods with no dependencies to other models. The context in turn gets most of the group's publicly available business logic methods.

So the issue ended up being a problem we could fix by updating our resolvers.

In the location block we specified the resolvers for amazons DNS service, while we left the other resolvers the same to work for some internal proxy stuff.

This is now giving us the expected speed.

You have to add the store to each Vue instance:

This can be accomplished by manually bootstrapping your root level component(s) in the NgModule ngDoBootstrap method.

(Note that in Angular 5+ this method may no longer be required, see this Angular PR)

We first find all root elements we want to bootstrap and give them a unique ID. Then for each instance, hack the component factory selector with the new ID and trigger the bootstrap.