Add tables of contents (using the VS Code Markdown All-In-One extension) and reformat lists to consistently use 4-space indents. No content changes.
4.9 KiB
Working on UI development
The UI is presented from a single HTML page (index.html) and any number of Javascript files, css files, images, etc. These are "packed" together using webpack.
For ongoing development it is possible to have the UI running in a web browser using "hot loading". This means that as you make changes to source files, they will be detected, the webpack will be recompiled and generated and then the browser will be informed to reload things. In combination with the debugger built into modern browsers this makes for a reasonable process.
For a production build, the same process is followed, except with different settings. In particular, no hot loading development server will be started and more effort is expended on packing and minimizing the components of the application as represented in the various "bundles". Read more about this in the webpack documentation.
Getting started
Checkout the branch you want to work on and type
$ cd ui
$ npm run start
This will pack and prepare a development setup. By default the development server that serves up the web page(s) will listen on http://localhost:3000/ so you can direct your browser there. It assumes the Moonfire NVR server is running at http://localhost:8080/ and will proxy API requests there.
Make any changes to the source code as you desire (look at existing code for examples and typical style), and the browser will hot-load your changes. Often times you will make mistakes. Anything from a coding error (for which you can use the browser's debugger), or compilation breaking Javascript errors. The latter will often be reported as errors during the webpack assembly process, but some will show up in the browser console, or both.
Overriding defaults
The UI is setup with Create React App.
npm run start
will honor any of the environment variables described in their
Advanced Configuration,
as well as Moonfire NVR's custom PROXY_TARGET
variable. Quick reference:
variable | description | default |
---|---|---|
PROXY_TARGET |
base URL of the backing Moonfire NVR server (see ui/src/setupProxy.js ) |
http://localhost:8080/ |
PORT |
port to listen on | 3000 |
HOST |
host/IP to listen on (or 0.0.0.0 for all) |
0.0.0.0 |
Thus one could connect to a remote Moonfire NVR by specifying its URL as follows:
$ PROXY_TARGET=https://nvr.example.com/ npm run start
This allows you to test a new UI against your stable, production Moonfire NVR installation with real data.
You can also set environment variables in .env
files, as described in
Adding Custom Environment Variables.
A note on https
Commonly production setups require credentials and run over https
, as
described in secure.md. Furthermore, Moonfire NVR will set the
secure
attribute on cookies it receives over https
, so that the browser
will only send them over a https
connection.
This is great for security and somewhat inconvenient for proxying. Fundamentally, there are three ways to make it work:
- Configure the proxy server with valid credentials to supply on every request, without requiring the browser to authenticate.
- Configure the proxy server to strip out the
secure
attribute from cookie response headers, so the browser will send them to the proxy server. - Configure the proxy server with a TLS certificate. a. using a self-signed certificate manually added to the browser's store. b. using a certificate from a "real" Certificate Authority (such as letsencrypt).
Currently the configuration only implements method 2. It's easy to configure but has a couple caveats:
- if you alternate between proxying to a test Moonfire NVR installation and a real one, your browser won't know the difference. It will supply whichever credentials were sent to it last.
- if you connect via a host other than localhost, your browser will have a
production cookie that it's willing to send to a remote host over a
non-
https
connection. If you ever load this website using an untrustworthy DNS server, your credentials can be compromised.
We might add support for method 3 in the future. It's less convenient to configure but can avoid these problems.