feat(repeater): add --proxy-domains-bypass flag (#616)

this flag allows specific requests with URLs matching any of the hosts provided in the flag’s input to bypass the proxy when it is enabled.
NeuraLegion · Nov 25, 2024 · 4a8084c · 4a8084c
1 parent 557bd5d
commit 4a8084c
Show file tree

Hide file tree

Showing 3 changed files with 40 additions and 8 deletions.
diff --git a/src/Commands/RunRepeater.ts b/src/Commands/RunRepeater.ts
@@ -135,17 +135,19 @@ export class RunRepeater implements CommandModule {
         requiresArg: true,
         array: true,
         describe:
-          'Comma-separated list of domains that should be routed through the proxy. This option is only applicable when using the --proxy option'
+          'Space-separated list of domains that should be routed through the proxy. This option is only applicable when using the --proxy option'
+      })
+      .option('proxy-domains-bypass', {
+        requiresArg: true,
+        array: true,
+        describe:
+          'Space-separated list of domains that should not be routed through the proxy. This option is only applicable when using the --proxy option'
       })
       .conflicts({
         daemon: 'remove-daemon',
-        ntlm: [
-          'proxy',
-          'proxy-bright',
-          'proxy-target',
-          'experimental-connection-reuse'
-        ]
+        ntlm: ['proxy', 'experimental-connection-reuse']
       })
+      .conflicts('proxy-domains', 'proxy-domains-bypass')
       .env('REPEATER')
       .middleware((args: Arguments) => {
         if (Object.hasOwnProperty.call(args, '')) {
@@ -196,7 +198,8 @@ export class RunRepeater implements CommandModule {
                 { type: 'application/ld+json', allowTruncation: false },
                 { type: 'application/graphql', allowTruncation: false }
               ],
-              proxyDomains: args.proxyDomains as string[]
+              proxyDomains: args.proxyDomains as string[],
+              proxyDomainsBypass: args.proxyDomainsBypass as string[]
             }
           })
           .register<DefaultRepeaterServerOptions>(

diff --git a/src/RequestExecutor/HttpRequestExecutor.ts b/src/RequestExecutor/HttpRequestExecutor.ts
@@ -40,6 +40,7 @@ export class HttpRequestExecutor implements RequestExecutor {
   private readonly httpAgent?: http.Agent;
   private readonly httpsAgent?: https.Agent;
   private readonly proxyDomains?: RegExp[];
+  private readonly proxyDomainsBypass?: RegExp[];
 
   get protocol(): Protocol {
     return Protocol.HTTP;
@@ -67,11 +68,23 @@ export class HttpRequestExecutor implements RequestExecutor {
       this.httpAgent = new http.Agent(agentOptions);
     }
 
+    if (this.options.proxyDomains && this.options.proxyDomainsBypass) {
+      throw new Error(
+        'cannot use both proxyDomains and proxyDomainsBypass at the same time'
+      );
+    }
+
     if (this.options.proxyDomains) {
       this.proxyDomains = this.options.proxyDomains.map((domain) =>
         Helpers.wildcardToRegExp(domain)
       );
     }
+
+    if (this.options.proxyDomainsBypass) {
+      this.proxyDomainsBypass = this.options.proxyDomainsBypass.map((domain) =>
+        Helpers.wildcardToRegExp(domain)
+      );
+    }
   }
 
   public async execute(options: Request): Promise<Response> {
@@ -197,12 +210,27 @@ export class HttpRequestExecutor implements RequestExecutor {
   }
 
   private getRequestAgent(options: Request) {
+    // do not use proxy for domains that are not in the list
     if (
       this.proxyDomains &&
       !this.proxyDomains.some((domain) =>
         domain.test(parseUrl(options.url).hostname)
       )
     ) {
+      logger.debug("Not using proxy for URL '%s'", options.url);
+
+      return options.secureEndpoint ? this.httpsAgent : this.httpAgent;
+    }
+
+    // do not use proxy for domains that are in the bypass list
+    if (
+      this.proxyDomainsBypass &&
+      this.proxyDomainsBypass.some((domain) =>
+        domain.test(parseUrl(options.url).hostname)
+      )
+    ) {
+      logger.debug("Bypassing proxy for URL '%s'", options.url);
+
       return options.secureEndpoint ? this.httpsAgent : this.httpAgent;
     }
 

diff --git a/src/RequestExecutor/RequestExecutorOptions.ts b/src/RequestExecutor/RequestExecutorOptions.ts
@@ -15,6 +15,7 @@ export interface RequestExecutorOptions {
   maxContentLength?: number;
   reuseConnection?: boolean;
   proxyDomains?: string[];
+  proxyDomainsBypass?: string[];
 }
 
 export const RequestExecutorOptions: unique symbol = Symbol(