fix(gatsby): show meaningful error message when engines try to bundle…

… ts-node (#35762)

Co-authored-by: Lennart <lekoarts@gmail.com>

packages/gatsby-cli/src/structured-errors/error-map.ts

@@ -64,6 +64,17 @@ const errors = {
     type: Type.WEBPACK,
     level: Level.ERROR,
   },
+  "98011": {
+    text: (context): string =>
+      `Rendering Engines attempted to use unsupported "${
+        context.package
+      }" package${
+        context.importedBy ? ` (imported by "${context.importedBy}")` : ``
+      }${context.advisory ? `\n\n${context.advisory}` : ``}`,
+    type: Type.WEBPACK,
+    level: Level.ERROR,
+    category: ErrorCategory.USER,
+  },
   "98123": {
     text: (context): string =>
       `${context.stageLabel} failed\n\n${

packages/gatsby/src/schema/graphql-engine/bundle-webpack.ts

@@ -2,7 +2,8 @@
 
 import * as path from "path"
 import * as fs from "fs-extra"
-import webpack from "webpack"
+import webpack, { Module, NormalModule, Compilation } from "webpack"
+import ConcatenatedModule from "webpack/lib/optimize/ConcatenatedModule"
 import { printQueryEnginePlugins } from "./print-plugins"
 import mod from "module"
 import { WebpackLoggingPlugin } from "../../utils/webpack/plugins/webpack-logging"
@@ -77,6 +78,16 @@ export async function createGraphqlEngineBundle(
     ],
     module: {
       rules: [
+        {
+          test: /\.ts$/,
+          exclude: /node_modules/,
+          use: {
+            loader: `babel-loader`,
+            options: {
+              presets: [`@babel/preset-typescript`],
+            },
+          },
+        },
         {
           oneOf: [
             {
@@ -93,7 +104,7 @@ export async function createGraphqlEngineBundle(
             },
             {
               // specific set of loaders for gatsby-node files - our babel transform that removes lifecycles not needed for engine -> relocator
-              test: /gatsby-node\.([cm]?js)$/,
+              test: /gatsby-node\.(cjs|mjs|js|ts)$/,
               // it is recommended for Node builds to turn off AMD support
               parser: { amd: false },
               use: [
@@ -106,7 +117,7 @@ export async function createGraphqlEngineBundle(
             {
               // generic loader for all other cases than lmdb or gatsby-node - we don't do anything special other than using relocator on it
               // For node binary relocations, include ".node" files as well here
-              test: /\.([cm]?js|node)$/,
+              test: /\.(cjs|mjs|js|ts|node)$/,
               // it is recommended for Node builds to turn off AMD support
               parser: { amd: false },
               use: assetRelocatorUseEntry,
@@ -124,16 +135,6 @@ export async function createGraphqlEngineBundle(
             },
           },
         },
-        {
-          test: /\.ts$/,
-          exclude: /node_modules/,
-          use: {
-            loader: `babel-loader`,
-            options: {
-              presets: [`@babel/preset-typescript`],
-            },
-          },
-        },
         {
           test: /\.txt/,
           type: `asset/resource`,
@@ -150,6 +151,7 @@ export async function createGraphqlEngineBundle(
         inquirer: false,
         // only load one version of lmdb
         lmdb: require.resolve(`lmdb`),
+        "ts-node": require.resolve(`./shims/ts-node`),
       },
     },
     plugins: [
@@ -166,16 +168,71 @@ export async function createGraphqlEngineBundle(
   })
 
   return new Promise((resolve, reject) => {
-    compiler.run((err, stats) => {
-      compiler.close(closeErr => {
-        if (err) {
-          return reject(err)
+    compiler.run((err, stats): void => {
+      function getResourcePath(
+        webpackModule?: Module | NormalModule | ConcatenatedModule | null
+      ): string | undefined {
+        if (webpackModule && !(webpackModule instanceof ConcatenatedModule)) {
+          return (webpackModule as NormalModule).resource
+        }
+
+        if (webpackModule?.modules) {
+          // ConcatenatedModule is a collection of modules so we have to go deeper to actually get a path,
+          // at this point we won't know which one so we just grab first module here
+          const [firstSubModule] = webpackModule.modules
+          return getResourcePath(firstSubModule)
         }
-        if (closeErr) {
-          return reject(closeErr)
+
+        return undefined
+      }
+
+      function iterateModules(
+        webpackModules: Set<Module>,
+        compilation: Compilation
+      ): void {
+        for (const webpackModule of webpackModules) {
+          if (webpackModule instanceof ConcatenatedModule) {
+            iterateModules(
+              (webpackModule as ConcatenatedModule).modules,
+              compilation
+            )
+          } else {
+            const resourcePath = getResourcePath(webpackModule)
+            if (resourcePath?.includes(`ts-node`)) {
+              const importedBy = getResourcePath(
+                compilation.moduleGraph.getIssuer(webpackModule)
+              )
+              const structuredError = {
+                id: `98011`,
+                context: {
+                  package: `ts-node`,
+                  importedBy,
+                  advisory: `Gatsby is supporting TypeScript natively (see https://gatsby.dev/typescript). "ts-node" might not be needed anymore at all, consider removing it.`,
+                },
+              }
+              throw structuredError
+            }
+          }
         }
-        return resolve(stats?.compilation)
-      })
+      }
+
+      try {
+        if (stats?.compilation.modules) {
+          iterateModules(stats.compilation.modules, stats.compilation)
+        }
+
+        compiler.close(closeErr => {
+          if (err) {
+            return reject(err)
+          }
+          if (closeErr) {
+            return reject(closeErr)
+          }
+          return resolve(stats?.compilation)
+        })
+      } catch (e) {
+        reject(e)
+      }
     })
   })
 }

packages/gatsby/src/schema/graphql-engine/shims/ts-node.js

@@ -0,0 +1,12 @@
+export function register() {
+  // no-op
+  //
+  // We are actually failing the build when `ts-node` exists in webpack's dependency graph
+  // because it's known to not work and cause failures.
+  //
+  // This shim for `ts-node` just skips trying to bundle the actual `ts-node`
+  // so webpack has less work to do during bundling.
+  //
+  // Using or not this shim, functionally doesn't make a difference - we will still
+  // fail the build with same actionable error anyway.
+}

packages/gatsby/src/schema/graphql-engine/standalone-regenerate.ts

@@ -0,0 +1,70 @@
+#!/usr/bin/env node
+
+/*
+this is used for development purposes only
+to be able to run `gatsby build` once to source data
+and print schema and then just rebundle graphql-engine
+with source file changes and test re-built engine quickly
+
+Usage:
+There need to be at least one successful `gatsby build`
+before starting to use this script (warm up datastore,
+generate "page-ssr" bundle). Once that's done you can
+run following command in test site directory:
+
+```shell
+node node_modules/gatsby/dist/schema/graphql-engine/standalone-regenerate.js
+```
+*/
+
+import { createGraphqlEngineBundle } from "./bundle-webpack"
+import reporter from "gatsby-cli/lib/reporter"
+import { loadConfigAndPlugins } from "../../utils/worker/child/load-config-and-plugins"
+import * as fs from "fs-extra"
+import { validateEngines } from "../../utils/validate-engines"
+
+async function run(): Promise<void> {
+  // load config
+  console.log(`loading config and plugins`)
+  await loadConfigAndPlugins({
+    siteDirectory: process.cwd(),
+  })
+
+  try {
+    console.log(`clearing webpack cache\n\n`)
+    // get rid of cache if it exist
+    await fs.remove(process.cwd() + `/.cache/webpack/query-engine`)
+  } catch (e) {
+    // eslint-disable no-empty
+  }
+
+  // recompile
+  const buildActivityTimer = reporter.activityTimer(
+    `Building Rendering Engines`
+  )
+  try {
+    buildActivityTimer.start()
+    await createGraphqlEngineBundle(process.cwd(), reporter, true)
+  } catch (err) {
+    buildActivityTimer.panic(err)
+  } finally {
+    buildActivityTimer.end()
+  }
+
+  // validate
+  const validateEnginesActivity = reporter.activityTimer(
+    `Validating Rendering Engines`
+  )
+  validateEnginesActivity.start()
+  try {
+    await validateEngines(process.cwd())
+  } catch (error) {
+    validateEnginesActivity.panic({ id: `98001`, context: {}, error })
+  } finally {
+    validateEnginesActivity.end()
+  }
+
+  console.log(`DONE`)
+}
+
+run()