feat(regexp): add escape function (denoland#3334)

mxdvl · May 16, 2023 · b0b5185 · b0b5185
1 parent d4f79d8
commit b0b5185
Show file tree

Hide file tree

Showing 3 changed files with 194 additions and 0 deletions.
diff --git a/regexp/escape.ts b/regexp/escape.ts
@@ -0,0 +1,83 @@
+// Copyright 2018-2023 the Deno authors. All rights reserved. MIT license.
+
+// // For future forward-compatibility with regexp `v` flag, reservedCharMap is
+// // autogenerated from the ClassSetReservedDoublePunctuator,
+// // ClassSetSyntaxCharacter, and ClassSetReservedPunctuator categories in the
+// // draft spec.
+// // See https://github.com/tc39/proposal-regexp-v-flag#how-is-the-v-flag-different-from-the-u-flag
+// // and https://arai-a.github.io/ecma262-compare/snapshot.html?pr=2418#prod-ClassSetReservedDoublePunctuator
+// const reservedChars = [...new Set(['ClassSetReservedDoublePunctuator', 'ClassSetSyntaxCharacter', 'ClassSetReservedPunctuator'].map(n =>
+//   document.querySelector(`[name=${n}] emu-rhs`).textContent.replaceAll(/\s/g, '')
+// ).join(''))]
+// const reservedCharMap = Object.fromEntries(reservedChars
+//   .map(x => {
+//     try {
+//       for (const flag of 'gimsuy') {
+//         new RegExp(`\\${x}`, flag)
+//         new RegExp(`[\\${x}]`, flag)
+//       }
+//       return [x, `\\${x}`]
+//     } catch (e) {
+//       return [x, `\\x${x.codePointAt(0).toString(16).padStart(2, '0')}`]
+//     }
+//   }))
+
+const reservedCharMap = {
+  "&": "\\x26",
+  "!": "\\x21",
+  "#": "\\x23",
+  "$": "\\$",
+  "%": "\\x25",
+  "*": "\\*",
+  "+": "\\+",
+  ",": "\\x2c",
+  ".": "\\.",
+  ":": "\\x3a",
+  ";": "\\x3b",
+  "<": "\\x3c",
+  "=": "\\x3d",
+  ">": "\\x3e",
+  "?": "\\?",
+  "@": "\\x40",
+  "^": "\\^",
+  "`": "\\x60",
+  "~": "\\x7e",
+  "(": "\\(",
+  ")": "\\)",
+  "[": "\\[",
+  "]": "\\]",
+  "{": "\\{",
+  "}": "\\}",
+  "/": "\\/",
+  "-": "\\x2d",
+  "\\": "\\\\",
+  "|": "\\|",
+};
+
+const RX_REGEXP_ESCAPE = new RegExp(
+  `[${Object.values(reservedCharMap).join("")}]`,
+  "gu",
+);
+
+/**
+ * Escapes arbitrary text for interpolation into a regexp, such that it will
+ * match exactly that text and nothing else.
+ *
+ * @example
+ * ```ts
+ * import { escape } from "https://deno.land/std@$STD_VERSION/regexp/mod.ts";
+ * import { assertEquals, assertMatch, assertNotMatch } from "https://deno.land/std@$STD_VERSION/testing/asserts.ts";
+ *
+ * const re = new RegExp(`^${escape(".")}$`, "u");
+ *
+ * assertEquals("^\\.$", re.source);
+ * assertMatch(".", re);
+ * assertNotMatch("a", re);
+ * ```
+ */
+export function escape(str: string) {
+  return str.replaceAll(
+    RX_REGEXP_ESCAPE,
+    (m) => reservedCharMap[m as keyof typeof reservedCharMap],
+  );
+}
diff --git a/regexp/escape_test.ts b/regexp/escape_test.ts
@@ -0,0 +1,100 @@
+// Copyright 2018-2023 the Deno authors. All rights reserved. MIT license.
+
+import { escape } from "./escape.ts";
+import {
+  assertEquals,
+  assertMatch,
+  assertNotMatch,
+} from "../testing/asserts.ts";
+
+const ALL_ASCII =
+  "\x00\x01\x02\x03\x04\x05\x06\x07\b\t\n\v\f\r\x0E\x0F\x10\x11\x12\x13\x14\x15\x16\x17\x18\x19\x1A\x1B\x1C\x1D\x1E\x1F !\"#$%&'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\\]^_`abcdefghijklmnopqrstuvwxyz{|}~\x7F";
+const ALL_REGEXP_FLAGS = "gimsuy";
+
+Deno.test("regexp", async (t) => {
+  await t.step("escape", async (t) => {
+    await t.step("examples", async (t) => {
+      await t.step("`.` matches literal `.`", () => {
+        const re = new RegExp(`^${escape(".")}$`, "u");
+
+        assertEquals("^\\.$", re.source);
+        assertMatch(".", re);
+        assertNotMatch("a", re);
+      });
+      await t.step("`$` matches literal `$`", () => {
+        const re = new RegExp(`^${escape("$")}$`);
+
+        assertMatch("$", re);
+        assertNotMatch("", re);
+      });
+      await t.step("`*` matches literal `*`", () => {
+        const re = new RegExp(`^${escape("a*")}$`);
+
+        assertMatch("a*", re);
+        assertNotMatch("", re);
+        assertNotMatch("aaa", re);
+      });
+      await t.step("escapes work correctly within character class", () => {
+        const re = new RegExp(`^[${escape(".$*+[](){}|\\<>")}]$`);
+
+        assertMatch(".", re);
+        assertMatch("$", re);
+        assertMatch("*", re);
+        assertMatch("+", re);
+        assertMatch("[", re);
+        assertMatch("]", re);
+        assertMatch("(", re);
+        assertMatch(")", re);
+        assertMatch("{", re);
+        assertMatch("}", re);
+        assertMatch("|", re);
+        assertMatch("\\", re);
+        assertMatch("<", re);
+        assertMatch(">", re);
+
+        assertNotMatch("a", re);
+      });
+    });
+    await t.step("all ASCII", async (t) => {
+      await t.step("interpolates without erroring", async (t) => {
+        await t.step("outside character class", () => {
+          for (const char of ALL_ASCII) {
+            for (const flag of ALL_REGEXP_FLAGS) {
+              new RegExp(escape(char), flag);
+            }
+          }
+        });
+        await t.step("within character class", () => {
+          for (const char of ALL_ASCII) {
+            for (const flag of ALL_REGEXP_FLAGS) {
+              new RegExp(`[${escape(char)}]`, flag);
+            }
+          }
+        });
+        await t.step("matches self", () => {
+          for (const char of ALL_ASCII) {
+            for (const flag of ALL_REGEXP_FLAGS) {
+              assertMatch(char, new RegExp(`^${escape(char)}$`, flag));
+            }
+          }
+        });
+        await t.step("doesn't match any other chars", () => {
+          for (const char of ALL_ASCII) {
+            for (const flag of ALL_REGEXP_FLAGS) {
+              if (flag === "i") continue;
+
+              for (const char2 of ALL_ASCII) {
+                if (char2 === char) continue;
+
+                assertNotMatch(
+                  char2,
+                  new RegExp(`^${escape(char)}$`, flag),
+                );
+              }
+            }
+          }
+        });
+      });
+    });
+  });
+});
diff --git a/regexp/mod.ts b/regexp/mod.ts
@@ -0,0 +1,11 @@
+// Copyright 2018-2023 the Deno authors. All rights reserved. MIT license.
+// This module is browser compatible.
+
+/**
+ * Functions for tasks related to regular expression (regexps), such as
+ * escaping text for interpolation into a regexp
+ *
+ * @module
+ */
+
+export * from "./escape.ts";