@davidwells/extract-deps 0.1.2 → 0.1.3

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@davidwells/extract-deps",
-  "version": "0.1.2",
+  "version": "0.1.3",
   "description": "Get imports and requires from a codebase without AST",
   "main": "src/index.js",
   "exports": {

package/src/detect-syntax.js

@@ -1,4 +1,5 @@
 // via https://github.com/unjs/mlly/blob/6bf5ad4d784e314319f872a8df41d64edd4e5505/src/syntax.ts#L7
+const { stripComments } = require('./utils')
 
 const ESM_RE =
   /([\s;]|^)(import\b[\s\w*,{}]*from|import\b\s*["'*{]|export\b\s*(?:[*{]|default|class|type|interface|enum|function|const|var|let|async function)|import\.meta\b)/m;
@@ -6,8 +7,6 @@ const ESM_RE =
 const CJS_RE =
   /([\s;]|^)(module\.exports\b|exports\.\w|require\s*\(|global\.\w)/m;
 
-const COMMENT_RE = /\/\*.+?\*\/|\/\/.*(?:\r?\n|$)/g;
-
 const BUILTIN_EXTENSIONS = new Set([".mjs", ".cjs", ".node", ".wasm"]);
 
 /**
@@ -25,8 +24,8 @@ const BUILTIN_EXTENSIONS = new Set([".mjs", ".cjs", ".node", ".wasm"]);
  */
 function hasESMSyntax(code, opts = {}) {
   // Always strip comments to avoid false positives from commented syntax
-  code = code.replace(COMMENT_RE, "");
-  return ESM_RE.test(code);
+  code = stripComments(code)
+  return ESM_RE.test(code)
 }
 
 /**
@@ -38,8 +37,8 @@ function hasESMSyntax(code, opts = {}) {
  */
 function hasCJSSyntax(code, opts = {}) {
   // Always strip comments to avoid false positives from commented syntax
-  code = code.replace(COMMENT_RE, "");
-  return CJS_RE.test(code);
+  code = stripComments(code)
+  return CJS_RE.test(code)
 }
 
 /**
@@ -51,16 +50,16 @@ function hasCJSSyntax(code, opts = {}) {
  */
 function detectSyntax(code, opts = {}) {
   // Always strip comments to avoid false positives from commented syntax
-  code = code.replace(COMMENT_RE, "");
+  code = stripComments(code)
   // We strip comments once hence not passing opts down to hasESMSyntax and hasCJSSyntax
-  const hasESM = hasESMSyntax(code, {});
-  const hasCJS = hasCJSSyntax(code, {});
+  const hasESM = hasESMSyntax(code, {})
+  const hasCJS = hasCJSSyntax(code, {})
 
   return {
     hasESM,
     hasCJS,
     isMixed: hasESM && hasCJS,
-  };
+  }
 }
 
 module.exports = {

package/src/get-exports.js

@@ -1,4 +1,4 @@
-const { getNames, extractExportName } = require('./utils')
+const { getNames, extractExportName, stripComments } = require('./utils')
 
 // https://regex101.com/r/example/1
 const EXPORT_DEFAULT_REGEX = /export\s+default\s+(?:(?:async\s+)?function\s+([\w$]+)|class\s+([\w$]+)|([\w$]+|\([^)]*\)\s*=>|{[\s\S]*?}|\[[\s\S]*?\]|'(?:[^'\\]|\\.)*'|"(?:[^"\\]|\\.)*"|`(?:[^`\\]|\\.)*`|[\w$]+(?:\s*[+\-*/%]\s*[\w$]+)*))/gm
@@ -15,7 +15,9 @@ const EXPORT_TYPE_REGEX = /export\s+(?:type|interface)\s+([\w$]+)/gm
 // https://regex101.com/r/example/7
 const EXPORT_ASSERTION_REGEX = /export\s+(?:["'\s]*([\w*{}\n\r\t, ]+)from\s*)?["'\s](.*[@\w_-]+)["'\s]+(assert|with)\s*\{([^}]+)\};?$/gm
 
-function getExports(code, options = {}) {
+function getExports(code = '', options = {}) {
+  // Strip comments to avoid false positives
+  code = stripComments(code)
   const defaultExports = getDefaultExports(code)
   const namedExports = getNamedExports(code)
   const namedDeclarationExports = getNamedDeclarationExports(code)

package/src/get-imports.js

@@ -1,4 +1,4 @@
-const { isBuiltIn, isRelativeImport, isRelativeImportWithQuotes, getNames } = require('./utils')
+const { isBuiltIn, isRelativeImport, isRelativeImportWithQuotes, getNames, stripComments } = require('./utils')
 
 // https://regex101.com/r/cfEmlm/1
 const IMPORT_REGEX = /import(?:["'\s]*([\w*{}\n\r\t, ]+)from\s*)?["'\s](.*[@\w_-]+)["'\s]+;?$/gm
@@ -9,7 +9,9 @@ const DYNAMIC_IMPORT_REGEX = /(?:(const|var|let)\s+([\w*{}\n\r\t,:= ]+)\s*=\s*)?
 // https://regex101.com/r/example/1
 const ASSERTION_IMPORT_REGEX = /(?:^|\n)\s*import\s+(?:["'\s]*([\w*{}\n\r\t, ]+)from\s*)?["'\s](.*[@\w_-]+)["'\s]+(assert|with)\s*\{([^}]+)\};?$/gm
 
-function getImports(code, options = {}) {
+function getImports(code = '', options = {}) {
+  // Strip comments to avoid false positives
+  code = stripComments(code)
   const staticImports = getStaticImports(code)
   const dynamicImports = getDynamicImports(code)
   if (options.includeAssertionImports) {

package/src/utils.js

@@ -1,6 +1,26 @@
 const { inspectParameters, getParametersNames } = require('./vendor/inspect-parameters-declaration')
 const builtinModules = require('module').builtinModules
 
+const COMMENT_RE = /\/\*[\s\S]*?\*\/|\/\/.*(?:\r?\n|$)/g
+const PROTOCOL_PLACEHOLDER = '__PROTO_SLASH__'
+
+/**
+ * Strip comments from code, preserving :// in URLs
+ * @param {string} code
+ * @returns {string}
+ */
+function stripComments(code) {
+  const hasProtocol = code.indexOf('://') > -1
+  if (hasProtocol) {
+    code = code.replace(/:\/\//g, PROTOCOL_PLACEHOLDER)
+  }
+  code = code.replace(COMMENT_RE, '')
+  if (hasProtocol) {
+    code = code.replace(/__PROTO_SLASH__/g, '://')
+  }
+  return code
+}
+
 /**
  * Check if a module specifier is a Node.js built-in
  * @param {string} moduleName - The module specifier (e.g., 'fs', 'node:fs/promises')
@@ -211,6 +231,7 @@ function extractExportName(exportValue) {
 }
 
 module.exports = {
+  stripComments,
   isBuiltIn,
   isRelativeImport,
   isRelativeImportWithQuotes,

package/tests/get-exports.test.js

@@ -532,4 +532,35 @@ export {
   assert.equal(foundExports[0].exportAlias, 'myHandler', 'exportAlias should be "myHandler" without trailing comma')
 })
 
+test('should ignore commented-out exports', () => {
+  const code = `
+// export const foo = 'foo'
+/* export default bar */
+/*
+export const multiLine = 'test'
+*/
+export const realExport = 'real'
+  `
+
+  const foundExports = getExports(code)
+
+  assert.equal(foundExports.length, 1, 'Should only find 1 real export')
+  assert.equal(foundExports[0].exportName, 'realExport', 'Should find realExport')
+})
+
+test('should preserve URLs with :// in strings', () => {
+  const code = `
+export const url = 'http://example.com/path'
+export const https = 'https://secure.com'
+export const real = 'value' // this comment should be stripped
+  `
+
+  const foundExports = getExports(code)
+
+  assert.equal(foundExports.length, 3, 'Should find 3 exports')
+  assert.equal(foundExports[0].exportName, 'url', 'Should find url export')
+  assert.equal(foundExports[1].exportName, 'https', 'Should find https export')
+  assert.equal(foundExports[2].exportName, 'real', 'Should find real export')
+})
+
 test.run() 

package/tests/get-imports.test.js

@@ -412,4 +412,31 @@ import styles from './styles.css' with { type: 'css' }
   assert.is(hasWithImport, true, 'Should have with import')
 })
 
+test('should ignore commented-out imports', () => {
+  const code = `
+// import foo from 'foo'
+/* import bar from 'bar' */
+/*
+import multiLine from 'multi'
+*/
+import realImport from 'real'
+  `
+
+  const importsFound = getImports(code)
+
+  assert.equal(importsFound.length, 1, 'Should only find 1 real import')
+  assert.equal(importsFound[0].from, 'real', 'Should find real import')
+})
+
+test('should preserve URLs with :// in strings', () => {
+  const code = `
+import foo from 'foo' // http://example.com
+  `
+
+  const importsFound = getImports(code)
+
+  assert.equal(importsFound.length, 1, 'Should find 1 import')
+  assert.equal(importsFound[0].from, 'foo', 'Should find foo import')
+})
+
 test.run()