Since the system is deployed in a distributed manner and has multiple domain names, the problem of obtaining the URL is often involved. This is a capability that needs to be provided at the system framework level. Otherwise, each module needs to find its own way to obtain the IP, which will be very confusing, and bugs will easily occur when going online
There are several main problems that need to be solved:
1. Able to automatically distinguish between development environment and production environment. For example, when deployed online, the URL may be http://www.xxx.com/svc/hello, but when developing locally it should be http://127.0.0.1/svc/hello. It cannot be written to death, otherwise development and deployment will have to be switched back and forth, which is very troublesome
2. Be able to distinguish URLs according to different services. For example, to obtain verification code services, you should call http://www.xxx.com/svc/getCode, and for WeChat-related services, you should call http://wx.xxx.com/svc/xxx
This article summarizes and shares some ideas:
1. The application has a corresponding configuration file, which explains whether to start in development mode or production mode. And separate the URLs, such as authentication-related URLs, WeChat-related URLs, common service-related URLs, etc.
2. At the same time, there are multiple configuration files, such as topo-dev.json, topo-production .json, topo-image.json, etc. This isolates different environments. If it is started in development mode, topo-dev.json is loaded, and the configured URLs are all like 127.0.0.1
3. When starting, load this configuration file, and put the key information under the global._g_env global variable, you can easily obtain the environment and URL information during runtime
The server side code is also It runs in the node environment, so it is very simple to obtain the URL. Through _g_env.url, you can get the path in the configuration file
The front-end page often also You need to send an ajax request, so you also need to know the url. However, static js has no way to obtain the server's environment information and URL, so it needs to obtain this information from the server. A feasible approach is:
First, the server has a service that specifically downloads this information. Send:
function clientSettingScript(req, res, next){ var script = "window.global = {_g_server:{}}; \n"+ ";global[\"_g_server\"].staticurl=\"" +global["_g_topo"].clientAccess.staticurl + "\"\n"+ ";global[\"_g_server\"].uploadurl=\"" +global["_g_topo"].clientAccess.uploadurl + "\"\n"+ ";global[\"_g_server\"].authurl=\"" +global["_g_topo"].clientAccess.authurl + "\"\n"+ ";global[\"_g_server\"].serviceurl=\"" +global["_g_topo"].clientAccess.serviceurl + "\"\n"+ ";global[\"_g_server\"].wxserviceurl=\"" +global["_g_topo"].clientAccess.wxserviceurl + "\"\n"+ ";global[\"_g_server\"].nail_pc_url=\"" +global["_g_topo"].connector.nail_pc_url + "\"\n"+ ";global[\"_g_env\"] =\"" +global["_g_topo"].env+ "\";\n"; res.end(script);}
This is an ordinary express service, but it is actually a js script. On the front-end page, use the script tag to load it
<script src="/svc/portal/setting"></script>
In this way, when the browser gets the response, it will use it as a piece of js To execute the script, a global variable global is placed on the window, which contains environment information and URL information
At the same time, the URL only contains the domain name, and the page assembles the complete URL according to the actual situation, such as:
security_code_url: global["_g_server"].serviceurl + "/getCode/"
The key to this approach is:
1. Combine the URL and environment information Put it in a separate configuration file instead of hard-coding it in the code. At the same time, isolate different configuration files according to the development environment, production environment, and mirror environment
2. Write a dedicated service on the server side to give these configuration information to the client page, and the client page does not need to be written to death