From 7025c312864417d9ef8591ab239c5c675769fa80 Mon Sep 17 00:00:00 2001
From: Chris Kirk <chris@parabola.io>
Date: Tue, 6 Feb 2018 17:01:44 -0800
Subject: [PATCH 1/3] Implementing proxy.

---
 index.js               |  1 +
 src/puppeteer_utils.js | 32 ++++++++++++++++++++++----------
 2 files changed, 23 insertions(+), 10 deletions(-)

diff --git a/index.js b/index.js
index fabc2513..1b1dfc69 100644
--- a/index.js
+++ b/index.js
@@ -29,6 +29,7 @@ const defaultOptions = {
   puppeteerExecutablePath: undefined,
   puppeteerIgnoreHTTPSErrors: false,
   publicPath: "/",
+  proxy: {},
   minifyCss: {},
   minifyHtml: {
     collapseBooleanAttributes: true,
diff --git a/src/puppeteer_utils.js b/src/puppeteer_utils.js
index 753e3081..100ae8a7 100644
--- a/src/puppeteer_utils.js
+++ b/src/puppeteer_utils.js
@@ -7,19 +7,31 @@ const path = require("path");
 const fs = require("fs");
 
 /**
- * @param {{page: Page, options: {skipThirdPartyRequests: true}, basePath: string }} opt
+ * @param {{page: Page, options: {skipThirdPartyRequests: true, proxy: {}}, basePath: string }} opt
  * @return {Promise<void>}
  */
-const skipThirdPartyRequests = async opt => {
+const handleThirdPartyRequests = async opt => {
   const { page, options, basePath } = opt;
-  if (!options.skipThirdPartyRequests) return;
+  if (!options.skipThirdPartyRequests || !options.proxy) return;
   await page.setRequestInterception(true);
   page.on("request", request => {
-    if (request.url().startsWith(basePath)) {
-      request.continue();
-    } else {
-      request.abort();
+    if (options.proxy) {
+      for (proxyUrl in options.proxy) {
+        if (request.url().startsWith(proxyUrl)) {
+          const requestChanges = {};
+          if (typeof options.proxy[proxyUrl] === 'string') {
+            requestChanges.url = request.url().replace(proxyUrl, options.proxy[proxyUrl]);
+          }
+          request.continue(requestChanges);
+          return;
+        }
+      }
     }
+
+    if (options.skipThirdPartyRequests && !request.url().startsWith(basePath))
+      request.abort();
+
+    request.continue();
   });
 };
 
@@ -164,8 +176,8 @@ const crawl = async opt => {
       try {
         const page = await browser.newPage();
         if (options.viewport) await page.setViewport(options.viewport);
-        if (options.skipThirdPartyRequests)
-          await skipThirdPartyRequests({ page, options, basePath });
+        if (options.skipThirdPartyRequests || options.proxy)
+          await handleThirdPartyRequests({ page, options, basePath });
         enableLogging({
           page,
           options,
@@ -217,7 +229,7 @@ const crawl = async opt => {
     });
 };
 
-exports.skipThirdPartyRequests = skipThirdPartyRequests;
+exports.handleThirdPartyRequests = handleThirdPartyRequests;
 exports.enableLogging = enableLogging;
 exports.getLinks = getLinks;
 exports.crawl = crawl;

From 100a6d7141f55e6efe40f5f13ca5fb711c573fb4 Mon Sep 17 00:00:00 2001
From: Chris Kirk <chris@parabola.io>
Date: Tue, 6 Feb 2018 17:15:31 -0800
Subject: [PATCH 2/3] adding logging and fixing proxy errors.

---
 src/puppeteer_utils.js | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)

diff --git a/src/puppeteer_utils.js b/src/puppeteer_utils.js
index 100ae8a7..0f4bd686 100644
--- a/src/puppeteer_utils.js
+++ b/src/puppeteer_utils.js
@@ -17,7 +17,9 @@ const handleThirdPartyRequests = async opt => {
   page.on("request", request => {
     if (options.proxy) {
       for (proxyUrl in options.proxy) {
+        console.log('proxyUrl', proxyUrl);
         if (request.url().startsWith(proxyUrl)) {
+          console.log('proxy match!', request.url());
           const requestChanges = {};
           if (typeof options.proxy[proxyUrl] === 'string') {
             requestChanges.url = request.url().replace(proxyUrl, options.proxy[proxyUrl]);
@@ -28,8 +30,10 @@ const handleThirdPartyRequests = async opt => {
       }
     }
 
-    if (options.skipThirdPartyRequests && !request.url().startsWith(basePath))
+    if (options.skipThirdPartyRequests && !request.url().startsWith(basePath)) {
       request.abort();
+      return;
+    }
 
     request.continue();
   });

From 5821dbf8f395c6e42a846ee4ff7570c5f247184c Mon Sep 17 00:00:00 2001
From: Chris Kirk <chris@parabola.io>
Date: Tue, 6 Feb 2018 17:20:45 -0800
Subject: [PATCH 3/3] removing console logs.

---
 src/puppeteer_utils.js | 2 --
 1 file changed, 2 deletions(-)

diff --git a/src/puppeteer_utils.js b/src/puppeteer_utils.js
index 0f4bd686..84dc3ced 100644
--- a/src/puppeteer_utils.js
+++ b/src/puppeteer_utils.js
@@ -17,9 +17,7 @@ const handleThirdPartyRequests = async opt => {
   page.on("request", request => {
     if (options.proxy) {
       for (proxyUrl in options.proxy) {
-        console.log('proxyUrl', proxyUrl);
         if (request.url().startsWith(proxyUrl)) {
-          console.log('proxy match!', request.url());
           const requestChanges = {};
           if (typeof options.proxy[proxyUrl] === 'string') {
             requestChanges.url = request.url().replace(proxyUrl, options.proxy[proxyUrl]);