From 21ed6ee42e298b7eb640ed35912e9c0355c1270d Mon Sep 17 00:00:00 2001
From: Michael Jackson <michael@jackson.us>
Date: Tue, 4 Sep 2018 15:58:52 -0700
Subject: [PATCH] Add ?module support for HTML files

Closes #113
---
 modules/actions/serveFile.js                  | 14 ++-
 modules/actions/serveHTMLModule.js            | 50 ++++++++++
 modules/actions/serveJavaScriptModule.js      | 28 +-----
 modules/middleware/findFile.js                |  6 +-
 modules/utils/rewriteBareModuleIdentifiers.js | 23 +++++
 package.json                                  |  1 +
 yarn.lock                                     | 96 ++++++++++++++++++-
 7 files changed, 187 insertions(+), 31 deletions(-)
 create mode 100644 modules/actions/serveHTMLModule.js
 create mode 100644 modules/utils/rewriteBareModuleIdentifiers.js

diff --git a/modules/actions/serveFile.js b/modules/actions/serveFile.js
index 366ee91..1d1a556 100644
--- a/modules/actions/serveFile.js
+++ b/modules/actions/serveFile.js
@@ -1,4 +1,5 @@
 const serveAutoIndexPage = require("./serveAutoIndexPage");
+const serveHTMLModule = require("./serveHTMLModule");
 const serveJavaScriptModule = require("./serveJavaScriptModule");
 const serveStaticFile = require("./serveStaticFile");
 const serveMetadata = require("./serveMetadata");
@@ -16,7 +17,18 @@ function serveFile(req, res) {
   }
 
   if (req.query.module != null) {
-    return serveJavaScriptModule(req, res);
+    if (req.entry.contentType === "application/javascript") {
+      return serveJavaScriptModule(req, res);
+    }
+
+    if (req.entry.contentType === "text/html") {
+      return serveHTMLModule(req, res);
+    }
+
+    return res
+      .status(403)
+      .type("text")
+      .send("?module mode is available only for JavaScript and HTML files");
   }
 
   serveStaticFile(req, res);
diff --git a/modules/actions/serveHTMLModule.js b/modules/actions/serveHTMLModule.js
new file mode 100644
index 0000000..1b2cdf2
--- /dev/null
+++ b/modules/actions/serveHTMLModule.js
@@ -0,0 +1,50 @@
+const etag = require("etag");
+const cheerio = require("cheerio");
+
+const getContentTypeHeader = require("../utils/getContentTypeHeader");
+const rewriteBareModuleIdentifiers = require("../utils/rewriteBareModuleIdentifiers");
+
+function serveHTMLModule(req, res) {
+  try {
+    const $ = cheerio.load(req.entry.content.toString("utf8"));
+
+    $("script[type=module]").each((index, element) => {
+      $(element).html(
+        rewriteBareModuleIdentifiers($(element).html(), req.packageConfig)
+      );
+    });
+
+    const code = $.html();
+
+    res
+      .set({
+        "Content-Length": Buffer.byteLength(code),
+        "Content-Type": getContentTypeHeader(req.entry.contentType),
+        "Cache-Control": "public, max-age=31536000, immutable", // 1 year
+        ETag: etag(code),
+        "Cache-Tag": "file, html-file, html-module"
+      })
+      .send(code);
+  } catch (error) {
+    console.error(error);
+
+    const errorName = error.constructor.name;
+    const errorMessage = error.message.replace(
+      /^.*?\/unpkg-.+?\//,
+      `/${req.packageSpec}/`
+    );
+    const codeFrame = error.codeFrame;
+    const debugInfo = `${errorName}: ${errorMessage}\n\n${codeFrame}`;
+
+    res
+      .status(500)
+      .type("text")
+      .send(
+        `Cannot generate module for ${req.packageSpec}${
+          req.filename
+        }\n\n${debugInfo}`
+      );
+  }
+}
+
+module.exports = serveHTMLModule;
diff --git a/modules/actions/serveJavaScriptModule.js b/modules/actions/serveJavaScriptModule.js
index ad7ec60..5772c95 100644
--- a/modules/actions/serveJavaScriptModule.js
+++ b/modules/actions/serveJavaScriptModule.js
@@ -1,35 +1,9 @@
 const etag = require("etag");
-const babel = require("babel-core");
 
 const getContentTypeHeader = require("../utils/getContentTypeHeader");
-const unpkgRewrite = require("../plugins/unpkgRewrite");
-
-function rewriteBareModuleIdentifiers(code, packageConfig) {
-  const dependencies = Object.assign(
-    {},
-    packageConfig.peerDependencies,
-    packageConfig.dependencies
-  );
-
-  const options = {
-    // Ignore .babelrc and package.json babel config
-    // because we haven't installed dependencies so
-    // we can't load plugins; see #84
-    babelrc: false,
-    plugins: [unpkgRewrite(dependencies)]
-  };
-
-  return babel.transform(code, options).code;
-}
+const rewriteBareModuleIdentifiers = require("../utils/rewriteBareModuleIdentifiers");
 
 function serveJavaScriptModule(req, res) {
-  if (req.entry.contentType !== "application/javascript") {
-    return res
-      .status(403)
-      .type("text")
-      .send("?module mode is available only for JavaScript files");
-  }
-
   try {
     const code = rewriteBareModuleIdentifiers(
       req.entry.content.toString("utf8"),
diff --git a/modules/middleware/findFile.js b/modules/middleware/findFile.js
index 92d3344..f55a494 100644
--- a/modules/middleware/findFile.js
+++ b/modules/middleware/findFile.js
@@ -84,8 +84,10 @@ function searchEntries(tarballStream, entryName, wantsHTML) {
         // and the client wants HTML.
         if (
           entry.name === entryName ||
-          // Allow accessing e.g. `/lib/index.html` using `/lib/`
-          (wantsHTML && entry.name === `${entryName}/index.html`) ||
+          // Allow accessing e.g. `/index.html` using `/`
+          (wantsHTML &&
+            entry.name ===
+              (entryName === "" ? "index.html" : `${entryName}/index.html`)) ||
           // Allow accessing e.g. `/index.js` or `/index.json` using
           // `/index` for compatibility with CommonJS
           (!wantsHTML && entry.name === `${entryName}.js`) ||
diff --git a/modules/utils/rewriteBareModuleIdentifiers.js b/modules/utils/rewriteBareModuleIdentifiers.js
new file mode 100644
index 0000000..603d7bd
--- /dev/null
+++ b/modules/utils/rewriteBareModuleIdentifiers.js
@@ -0,0 +1,23 @@
+const babel = require("babel-core");
+
+const unpkgRewrite = require("../plugins/unpkgRewrite");
+
+function rewriteBareModuleIdentifiers(code, packageConfig) {
+  const dependencies = Object.assign(
+    {},
+    packageConfig.peerDependencies,
+    packageConfig.dependencies
+  );
+
+  const options = {
+    // Ignore .babelrc and package.json babel config
+    // because we haven't installed dependencies so
+    // we can't load plugins; see #84
+    babelrc: false,
+    plugins: [unpkgRewrite(dependencies)]
+  };
+
+  return babel.transform(code, options).code;
+}
+
+module.exports = rewriteBareModuleIdentifiers;
diff --git a/package.json b/package.json
index 7f679a7..5eec884 100644
--- a/package.json
+++ b/package.json
@@ -15,6 +15,7 @@
     "babel-preset-stage-2": "^6.24.1",
     "babel-register": "^6.26.0",
     "body-parser": "^1.18.2",
+    "cheerio": "^1.0.0-rc.2",
     "cors": "^2.8.1",
     "countries-list": "^1.3.2",
     "date-fns": "^1.28.1",
diff --git a/yarn.lock b/yarn.lock
index 22ca323..3f0b4b3 100644
--- a/yarn.lock
+++ b/yarn.lock
@@ -68,6 +68,10 @@
     lodash "^4.2.0"
     to-fast-properties "^2.0.0"
 
+"@types/node@*":
+  version "10.9.4"
+  resolved "https://registry.yarnpkg.com/@types/node/-/node-10.9.4.tgz#0f4cb2dc7c1de6096055357f70179043c33e9897"
+
 abab@^1.0.4:
   version "1.0.4"
   resolved "https://registry.yarnpkg.com/abab/-/abab-1.0.4.tgz#5faad9c2c07f60dd76770f71cf025b62a63cfd4e"
@@ -1210,6 +1214,10 @@ bonjour@^3.5.0:
     multicast-dns "^6.0.1"
     multicast-dns-service-types "^1.1.0"
 
+boolbase@~1.0.0:
+  version "1.0.0"
+  resolved "https://registry.yarnpkg.com/boolbase/-/boolbase-1.0.0.tgz#68dff5fbe60c51eb37725ea9e3ed310dcc1e776e"
+
 boom@2.x.x:
   version "2.10.1"
   resolved "https://registry.yarnpkg.com/boom/-/boom-2.10.1.tgz#39c8918ceff5799f83f9492a848f625add0c766f"
@@ -1563,6 +1571,17 @@ charenc@~0.0.1:
   version "0.0.2"
   resolved "https://registry.yarnpkg.com/charenc/-/charenc-0.0.2.tgz#c0a1d2f3a7092e03774bfa83f14c0fc5790a8667"
 
+cheerio@^1.0.0-rc.2:
+  version "1.0.0-rc.2"
+  resolved "https://registry.yarnpkg.com/cheerio/-/cheerio-1.0.0-rc.2.tgz#4b9f53a81b27e4d5dac31c0ffd0cfa03cc6830db"
+  dependencies:
+    css-select "~1.2.0"
+    dom-serializer "~0.1.0"
+    entities "~1.1.1"
+    htmlparser2 "^3.9.1"
+    lodash "^4.15.0"
+    parse5 "^3.0.1"
+
 chokidar@^1.7.0:
   version "1.7.0"
   resolved "https://registry.yarnpkg.com/chokidar/-/chokidar-1.7.0.tgz#798e689778151c8076b4b360e5edd28cda2bb468"
@@ -1996,6 +2015,15 @@ css-loader@0.26.1:
     postcss-modules-values "^1.1.0"
     source-list-map "^0.1.4"
 
+css-select@~1.2.0:
+  version "1.2.0"
+  resolved "https://registry.yarnpkg.com/css-select/-/css-select-1.2.0.tgz#2b3a110539c5355f1cd8d314623e870b121ec858"
+  dependencies:
+    boolbase "~1.0.0"
+    css-what "2.1"
+    domutils "1.5.1"
+    nth-check "~1.0.1"
+
 css-selector-tokenizer@^0.7.0:
   version "0.7.0"
   resolved "https://registry.yarnpkg.com/css-selector-tokenizer/-/css-selector-tokenizer-0.7.0.tgz#e6988474ae8c953477bf5e7efecfceccd9cf4c86"
@@ -2004,6 +2032,10 @@ css-selector-tokenizer@^0.7.0:
     fastparse "^1.1.1"
     regexpu-core "^1.0.0"
 
+css-what@2.1:
+  version "2.1.0"
+  resolved "https://registry.yarnpkg.com/css-what/-/css-what-2.1.0.tgz#9467d032c38cfaefb9f2d79501253062f87fa1bd"
+
 cssesc@^0.1.0:
   version "0.1.0"
   resolved "https://registry.yarnpkg.com/cssesc/-/cssesc-0.1.0.tgz#c814903e45623371a0477b40109aaafbeeaddbb4"
@@ -2278,16 +2310,51 @@ doctrine@^2.0.2, doctrine@^2.1.0:
   dependencies:
     esutils "^2.0.2"
 
+dom-serializer@0, dom-serializer@~0.1.0:
+  version "0.1.0"
+  resolved "https://registry.yarnpkg.com/dom-serializer/-/dom-serializer-0.1.0.tgz#073c697546ce0780ce23be4a28e293e40bc30c82"
+  dependencies:
+    domelementtype "~1.1.1"
+    entities "~1.1.1"
+
 domain-browser@^1.1.1:
   version "1.2.0"
   resolved "https://registry.yarnpkg.com/domain-browser/-/domain-browser-1.2.0.tgz#3d31f50191a6749dd1375a7f522e823d42e54eda"
 
+domelementtype@1, domelementtype@^1.3.0:
+  version "1.3.0"
+  resolved "https://registry.yarnpkg.com/domelementtype/-/domelementtype-1.3.0.tgz#b17aed82e8ab59e52dd9c19b1756e0fc187204c2"
+
+domelementtype@~1.1.1:
+  version "1.1.3"
+  resolved "https://registry.yarnpkg.com/domelementtype/-/domelementtype-1.1.3.tgz#bd28773e2642881aec51544924299c5cd822185b"
+
 domexception@^1.0.0:
   version "1.0.1"
   resolved "https://registry.yarnpkg.com/domexception/-/domexception-1.0.1.tgz#937442644ca6a31261ef36e3ec677fe805582c90"
   dependencies:
     webidl-conversions "^4.0.2"
 
+domhandler@^2.3.0:
+  version "2.4.2"
+  resolved "https://registry.yarnpkg.com/domhandler/-/domhandler-2.4.2.tgz#8805097e933d65e85546f726d60f5eb88b44f803"
+  dependencies:
+    domelementtype "1"
+
+domutils@1.5.1:
+  version "1.5.1"
+  resolved "https://registry.yarnpkg.com/domutils/-/domutils-1.5.1.tgz#dcd8488a26f563d61079e48c9f7b7e32373682cf"
+  dependencies:
+    dom-serializer "0"
+    domelementtype "1"
+
+domutils@^1.5.1:
+  version "1.7.0"
+  resolved "https://registry.yarnpkg.com/domutils/-/domutils-1.7.0.tgz#56ea341e834e06e6748af7a1cb25da67ea9f8c2a"
+  dependencies:
+    dom-serializer "0"
+    domelementtype "1"
+
 dot-prop@^4.1.0:
   version "4.2.0"
   resolved "https://registry.yarnpkg.com/dot-prop/-/dot-prop-4.2.0.tgz#1f19e0c2e1aa0e32797c49799f2837ac6af69c57"
@@ -2381,6 +2448,10 @@ enhanced-resolve@^3.4.0:
     object-assign "^4.0.1"
     tapable "^0.2.7"
 
+entities@^1.1.1, entities@~1.1.1:
+  version "1.1.1"
+  resolved "https://registry.yarnpkg.com/entities/-/entities-1.1.1.tgz#6e5c2d0a5621b5dadaecef80b90edfb5cd7772f0"
+
 errno@^0.1.3:
   version "0.1.7"
   resolved "https://registry.yarnpkg.com/errno/-/errno-0.1.7.tgz#4684d71779ad39af177e3f007996f7c67c852618"
@@ -3449,6 +3520,17 @@ html-minifier@^3.5.8:
     relateurl "0.2.x"
     uglify-js "3.3.x"
 
+htmlparser2@^3.9.1:
+  version "3.9.2"
+  resolved "https://registry.yarnpkg.com/htmlparser2/-/htmlparser2-3.9.2.tgz#1bdf87acca0f3f9e53fa4fcceb0f4b4cbb00b338"
+  dependencies:
+    domelementtype "^1.3.0"
+    domhandler "^2.3.0"
+    domutils "^1.5.1"
+    entities "^1.1.1"
+    inherits "^2.0.1"
+    readable-stream "^2.0.2"
+
 http-deceiver@^1.2.7:
   version "1.2.7"
   resolved "https://registry.yarnpkg.com/http-deceiver/-/http-deceiver-1.2.7.tgz#fa7168944ab9a519d337cb0bec7284dc3e723d87"
@@ -4606,7 +4688,7 @@ lodash.uniq@^4.5.0:
   version "4.5.0"
   resolved "https://registry.yarnpkg.com/lodash.uniq/-/lodash.uniq-4.5.0.tgz#d0225373aeb652adc1bc82e4945339a842754773"
 
-lodash@^4.13.1, lodash@^4.17.10:
+lodash@^4.13.1, lodash@^4.15.0, lodash@^4.17.10:
   version "4.17.10"
   resolved "https://registry.yarnpkg.com/lodash/-/lodash-4.17.10.tgz#1b7793cf7259ea38fb3661d4d38b3260af8ae4e7"
 
@@ -5179,6 +5261,12 @@ npmlog@^4.0.2:
     gauge "~2.7.3"
     set-blocking "~2.0.0"
 
+nth-check@~1.0.1:
+  version "1.0.1"
+  resolved "https://registry.yarnpkg.com/nth-check/-/nth-check-1.0.1.tgz#9929acdf628fc2c41098deab82ac580cf149aae4"
+  dependencies:
+    boolbase "~1.0.0"
+
 num2fraction@^1.2.2:
   version "1.2.2"
   resolved "https://registry.yarnpkg.com/num2fraction/-/num2fraction-1.2.2.tgz#6f682b6a027a4e9ddfa4564cd2589d1d4e669ede"
@@ -5406,6 +5494,12 @@ parse5@4.0.0:
   version "4.0.0"
   resolved "https://registry.yarnpkg.com/parse5/-/parse5-4.0.0.tgz#6d78656e3da8d78b4ec0b906f7c08ef1dfe3f608"
 
+parse5@^3.0.1:
+  version "3.0.3"
+  resolved "https://registry.yarnpkg.com/parse5/-/parse5-3.0.3.tgz#042f792ffdd36851551cf4e9e066b3874ab45b5c"
+  dependencies:
+    "@types/node" "*"
+
 parseurl@~1.3.2:
   version "1.3.2"
   resolved "https://registry.yarnpkg.com/parseurl/-/parseurl-1.3.2.tgz#fc289d4ed8993119460c156253262cdc8de65bf3"