Remove length restriction from JSX entities, and ignore `Object.proto…

…type` (#14327) * Allow any length in unicode CP jsx entities * Add failing test * Don't access entities from Object.prototype * Add `&#0_0;` test * Fix `` parsing
babel · Mar 6, 2022 · 4bb9b89 · 4bb9b89
1 parent 5cf3118
commit 4bb9b89
Show file tree

Hide file tree

Showing 11 changed files with 272 additions and 35 deletions.
diff --git a/packages/babel-parser/src/plugins/jsx/index.js b/packages/babel-parser/src/plugins/jsx/index.js
@@ -22,9 +22,6 @@ import type { Position } from "../../util/location";
 import { isNewLine } from "../../util/whitespace";
 import { Errors, makeErrorTemplates, ErrorCodes } from "../../parser/error";
 
-const HEX_NUMBER = /^[\da-fA-F]+$/;
-const DECIMAL_NUMBER = /^\d+$/;
-
 /* eslint sort-keys: "error" */
 const JsxErrors = makeErrorTemplates(
   {
@@ -190,39 +187,54 @@ export default (superClass: Class<Parser>): Class<Parser> =>
     }
 
     jsxReadEntity(): string {
-      let str = "";
-      let count = 0;
-      let entity;
-      let ch = this.input[this.state.pos];
-
       const startPos = ++this.state.pos;
-      while (this.state.pos < this.length && count++ < 10) {
-        ch = this.input[this.state.pos++];
-        if (ch === ";") {
-          if (str[0] === "#") {
-            if (str[1] === "x") {
-              str = str.substr(2);
-              if (HEX_NUMBER.test(str)) {
-                entity = String.fromCodePoint(parseInt(str, 16));
-              }
-            } else {
-              str = str.substr(1);
-              if (DECIMAL_NUMBER.test(str)) {
-                entity = String.fromCodePoint(parseInt(str, 10));
-              }
-            }
-          } else {
-            entity = XHTMLEntities[str];
+      if (this.codePointAtPos(this.state.pos) === charCodes.numberSign) {
+        ++this.state.pos;
+
+        let radix = 10;
+        if (this.codePointAtPos(this.state.pos) === charCodes.lowercaseX) {
+          radix = 16;
+          ++this.state.pos;
+        }
+
+        const codePoint = this.readInt(
+          radix,
+          /* len */ undefined,
+          /* forceLen */ false,
+          /* allowNumSeparator */ "bail",
+        );
+        if (
+          codePoint !== null &&
+          this.codePointAtPos(this.state.pos) === charCodes.semicolon
+        ) {
+          ++this.state.pos;
+          return String.fromCodePoint(codePoint);
+        }
+      } else {
+        let count = 0;
+        let semi = false;
+        while (
+          count++ < 10 &&
+          this.state.pos < this.length &&
+          !(semi = this.codePointAtPos(this.state.pos) == charCodes.semicolon)
+        ) {
+          ++this.state.pos;
+        }
+
+        if (semi) {
+          const desc = this.input.slice(startPos, this.state.pos);
+          const entity = XHTMLEntities[desc];
+          ++this.state.pos;
+
+          if (entity) {
+            return entity;
           }
-          break;
         }
-        str += ch;
-      }
-      if (!entity) {
-        this.state.pos = startPos;
-        return "&";
       }
-      return entity;
+
+      // Not a valid entity
+      this.state.pos = startPos;
+      return "&";
     }
 
     // Read a JSX identifier (valid tag or attribute name).

diff --git a/packages/babel-parser/src/plugins/jsx/xhtml.js b/packages/babel-parser/src/plugins/jsx/xhtml.js
@@ -1,6 +1,7 @@
 // @flow
 
-const entities: { [name: string]: string } = {
+const entities: { [name: string]: string, __proto__: null } = {
+  __proto__: null,
   quot: "\u0022",
   amp: "&",
   apos: "\u0027",

diff --git a/packages/babel-parser/src/tokenizer/index.js b/packages/babel-parser/src/tokenizer/index.js
@@ -1174,7 +1174,7 @@ export default class Tokenizer extends ParserErrors {
     radix: number,
     len?: number,
     forceLen?: boolean,
-    allowNumSeparator: boolean = true,
+    allowNumSeparator: boolean | "bail" = true,
   ): number | null {
     const start = this.state.pos;
     const forbiddenSiblings =
@@ -1197,7 +1197,7 @@ export default class Tokenizer extends ParserErrors {
       const code = this.input.charCodeAt(this.state.pos);
       let val;
 
-      if (code === charCodes.underscore) {
+      if (code === charCodes.underscore && allowNumSeparator !== "bail") {
         const prev = this.input.charCodeAt(this.state.pos - 1);
         const next = this.input.charCodeAt(this.state.pos + 1);
         if (allowedSiblings.indexOf(next) === -1) {

diff --git a/packages/babel-parser/test/fixtures/jsx/html-entities/code-point/input.js b/packages/babel-parser/test/fixtures/jsx/html-entities/code-point/input.js
@@ -0,0 +1,7 @@
+<a>
+    &#1234;
+    &#xABC;
+    &#00000000000000000000001234;
+    &#x0000000000000000000000ABC;
+    &#0000;
+</a>
diff --git a/packages/babel-parser/test/fixtures/jsx/html-entities/code-point/output.json b/packages/babel-parser/test/fixtures/jsx/html-entities/code-point/output.json
@@ -0,0 +1,52 @@
+{
+  "type": "File",
+  "start":0,"end":112,"loc":{"start":{"line":1,"column":0,"index":0},"end":{"line":7,"column":4,"index":112}},
+  "program": {
+    "type": "Program",
+    "start":0,"end":112,"loc":{"start":{"line":1,"column":0,"index":0},"end":{"line":7,"column":4,"index":112}},
+    "sourceType": "script",
+    "interpreter": null,
+    "body": [
+      {
+        "type": "ExpressionStatement",
+        "start":0,"end":112,"loc":{"start":{"line":1,"column":0,"index":0},"end":{"line":7,"column":4,"index":112}},
+        "expression": {
+          "type": "JSXElement",
+          "start":0,"end":112,"loc":{"start":{"line":1,"column":0,"index":0},"end":{"line":7,"column":4,"index":112}},
+          "openingElement": {
+            "type": "JSXOpeningElement",
+            "start":0,"end":3,"loc":{"start":{"line":1,"column":0,"index":0},"end":{"line":1,"column":3,"index":3}},
+            "name": {
+              "type": "JSXIdentifier",
+              "start":1,"end":2,"loc":{"start":{"line":1,"column":1,"index":1},"end":{"line":1,"column":2,"index":2}},
+              "name": "a"
+            },
+            "attributes": [],
+            "selfClosing": false
+          },
+          "closingElement": {
+            "type": "JSXClosingElement",
+            "start":108,"end":112,"loc":{"start":{"line":7,"column":0,"index":108},"end":{"line":7,"column":4,"index":112}},
+            "name": {
+              "type": "JSXIdentifier",
+              "start":110,"end":111,"loc":{"start":{"line":7,"column":2,"index":110},"end":{"line":7,"column":3,"index":111}},
+              "name": "a"
+            }
+          },
+          "children": [
+            {
+              "type": "JSXText",
+              "start":3,"end":108,"loc":{"start":{"line":1,"column":3,"index":3},"end":{"line":7,"column":0,"index":108}},
+              "extra": {
+                "rawValue": "\n    Ӓ\n    ઼\n    Ӓ\n    ઼\n    \u0000\n",
+                "raw": "\n    &#1234;\n    &#xABC;\n    &#00000000000000000000001234;\n    &#x0000000000000000000000ABC;\n    &#0000;\n"
+              },
+              "value": "\n    Ӓ\n    ઼\n    Ӓ\n    ઼\n    \u0000\n"
+            }
+          ]
+        }
+      }
+    ],
+    "directives": []
+  }
+}
diff --git a/packages/babel-parser/test/fixtures/jsx/html-entities/html4/input.js b/packages/babel-parser/test/fixtures/jsx/html-entities/html4/input.js
@@ -0,0 +1 @@
+<a>&amp;</a>
diff --git a/packages/babel-parser/test/fixtures/jsx/html-entities/html4/output.json b/packages/babel-parser/test/fixtures/jsx/html-entities/html4/output.json
@@ -0,0 +1,52 @@
+{
+  "type": "File",
+  "start":0,"end":12,"loc":{"start":{"line":1,"column":0,"index":0},"end":{"line":1,"column":12,"index":12}},
+  "program": {
+    "type": "Program",
+    "start":0,"end":12,"loc":{"start":{"line":1,"column":0,"index":0},"end":{"line":1,"column":12,"index":12}},
+    "sourceType": "script",
+    "interpreter": null,
+    "body": [
+      {
+        "type": "ExpressionStatement",
+        "start":0,"end":12,"loc":{"start":{"line":1,"column":0,"index":0},"end":{"line":1,"column":12,"index":12}},
+        "expression": {
+          "type": "JSXElement",
+          "start":0,"end":12,"loc":{"start":{"line":1,"column":0,"index":0},"end":{"line":1,"column":12,"index":12}},
+          "openingElement": {
+            "type": "JSXOpeningElement",
+            "start":0,"end":3,"loc":{"start":{"line":1,"column":0,"index":0},"end":{"line":1,"column":3,"index":3}},
+            "name": {
+              "type": "JSXIdentifier",
+              "start":1,"end":2,"loc":{"start":{"line":1,"column":1,"index":1},"end":{"line":1,"column":2,"index":2}},
+              "name": "a"
+            },
+            "attributes": [],
+            "selfClosing": false
+          },
+          "closingElement": {
+            "type": "JSXClosingElement",
+            "start":8,"end":12,"loc":{"start":{"line":1,"column":8,"index":8},"end":{"line":1,"column":12,"index":12}},
+            "name": {
+              "type": "JSXIdentifier",
+              "start":10,"end":11,"loc":{"start":{"line":1,"column":10,"index":10},"end":{"line":1,"column":11,"index":11}},
+              "name": "a"
+            }
+          },
+          "children": [
+            {
+              "type": "JSXText",
+              "start":3,"end":8,"loc":{"start":{"line":1,"column":3,"index":3},"end":{"line":1,"column":8,"index":8}},
+              "extra": {
+                "rawValue": "&",
+                "raw": "&amp;"
+              },
+              "value": "&"
+            }
+          ]
+        }
+      }
+    ],
+    "directives": []
+  }
+}
diff --git a/packages/babel-parser/test/fixtures/jsx/html-entities/invalid/input.js b/packages/babel-parser/test/fixtures/jsx/html-entities/invalid/input.js
@@ -0,0 +1,7 @@
+<a>
+    &amp
+    &ampa;
+    &amp ;
+    &xamp;
+    &#0_0;
+</a>
diff --git a/packages/babel-parser/test/fixtures/jsx/html-entities/invalid/output.json b/packages/babel-parser/test/fixtures/jsx/html-entities/invalid/output.json
@@ -0,0 +1,52 @@
+{
+  "type": "File",
+  "start":0,"end":61,"loc":{"start":{"line":1,"column":0,"index":0},"end":{"line":7,"column":4,"index":61}},
+  "program": {
+    "type": "Program",
+    "start":0,"end":61,"loc":{"start":{"line":1,"column":0,"index":0},"end":{"line":7,"column":4,"index":61}},
+    "sourceType": "script",
+    "interpreter": null,
+    "body": [
+      {
+        "type": "ExpressionStatement",
+        "start":0,"end":61,"loc":{"start":{"line":1,"column":0,"index":0},"end":{"line":7,"column":4,"index":61}},
+        "expression": {
+          "type": "JSXElement",
+          "start":0,"end":61,"loc":{"start":{"line":1,"column":0,"index":0},"end":{"line":7,"column":4,"index":61}},
+          "openingElement": {
+            "type": "JSXOpeningElement",
+            "start":0,"end":3,"loc":{"start":{"line":1,"column":0,"index":0},"end":{"line":1,"column":3,"index":3}},
+            "name": {
+              "type": "JSXIdentifier",
+              "start":1,"end":2,"loc":{"start":{"line":1,"column":1,"index":1},"end":{"line":1,"column":2,"index":2}},
+              "name": "a"
+            },
+            "attributes": [],
+            "selfClosing": false
+          },
+          "closingElement": {
+            "type": "JSXClosingElement",
+            "start":57,"end":61,"loc":{"start":{"line":7,"column":0,"index":57},"end":{"line":7,"column":4,"index":61}},
+            "name": {
+              "type": "JSXIdentifier",
+              "start":59,"end":60,"loc":{"start":{"line":7,"column":2,"index":59},"end":{"line":7,"column":3,"index":60}},
+              "name": "a"
+            }
+          },
+          "children": [
+            {
+              "type": "JSXText",
+              "start":3,"end":57,"loc":{"start":{"line":1,"column":3,"index":3},"end":{"line":7,"column":0,"index":57}},
+              "extra": {
+                "rawValue": "\n    &amp\n    &ampa;\n    &amp ;\n    &xamp;\n    &#0_0;\n",
+                "raw": "\n    &amp\n    &ampa;\n    &amp ;\n    &xamp;\n    &#0_0;\n"
+              },
+              "value": "\n    &amp\n    &ampa;\n    &amp ;\n    &xamp;\n    &#0_0;\n"
+            }
+          ]
+        }
+      }
+    ],
+    "directives": []
+  }
+}
diff --git a/packages/babel-parser/test/fixtures/jsx/html-entities/no-object-prototype/input.js b/packages/babel-parser/test/fixtures/jsx/html-entities/no-object-prototype/input.js
@@ -0,0 +1 @@
+<a>&valueOf;</a>
diff --git a/packages/babel-parser/test/fixtures/jsx/html-entities/no-object-prototype/output.json b/packages/babel-parser/test/fixtures/jsx/html-entities/no-object-prototype/output.json
@@ -0,0 +1,52 @@
+{
+  "type": "File",
+  "start":0,"end":16,"loc":{"start":{"line":1,"column":0,"index":0},"end":{"line":1,"column":16,"index":16}},
+  "program": {
+    "type": "Program",
+    "start":0,"end":16,"loc":{"start":{"line":1,"column":0,"index":0},"end":{"line":1,"column":16,"index":16}},
+    "sourceType": "script",
+    "interpreter": null,
+    "body": [
+      {
+        "type": "ExpressionStatement",
+        "start":0,"end":16,"loc":{"start":{"line":1,"column":0,"index":0},"end":{"line":1,"column":16,"index":16}},
+        "expression": {
+          "type": "JSXElement",
+          "start":0,"end":16,"loc":{"start":{"line":1,"column":0,"index":0},"end":{"line":1,"column":16,"index":16}},
+          "openingElement": {
+            "type": "JSXOpeningElement",
+            "start":0,"end":3,"loc":{"start":{"line":1,"column":0,"index":0},"end":{"line":1,"column":3,"index":3}},
+            "name": {
+              "type": "JSXIdentifier",
+              "start":1,"end":2,"loc":{"start":{"line":1,"column":1,"index":1},"end":{"line":1,"column":2,"index":2}},
+              "name": "a"
+            },
+            "attributes": [],
+            "selfClosing": false
+          },
+          "closingElement": {
+            "type": "JSXClosingElement",
+            "start":12,"end":16,"loc":{"start":{"line":1,"column":12,"index":12},"end":{"line":1,"column":16,"index":16}},
+            "name": {
+              "type": "JSXIdentifier",
+              "start":14,"end":15,"loc":{"start":{"line":1,"column":14,"index":14},"end":{"line":1,"column":15,"index":15}},
+              "name": "a"
+            }
+          },
+          "children": [
+            {
+              "type": "JSXText",
+              "start":3,"end":12,"loc":{"start":{"line":1,"column":3,"index":3},"end":{"line":1,"column":12,"index":12}},
+              "extra": {
+                "rawValue": "&valueOf;",
+                "raw": "&valueOf;"
+              },
+              "value": "&valueOf;"
+            }
+          ]
+        }
+      }
+    ],
+    "directives": []
+  }
+}