Remove punycode package

We can get the same functionality with the URL constructor

Author: jhnns

package-lock.json

@@ -64,8 +64,7 @@
   },
   "dependencies": {
     "is-ip": "^3.1.0",
-    "node-fetch": "^2.6.1",
-    "punycode": "^2.1.1"
+    "node-fetch": "^2.6.1"
   },
   "devDependencies": {
     "@types/jest": "^27.4.0",

package.json

@@ -1,11 +1,11 @@
-import { toASCII } from "punycode";
 import { buildTries } from "./build-tries";
 import {
   PUBLIC_SUFFIX_MARKER_ICANN_START,
   PUBLIC_SUFFIX_MARKER_ICANN_END,
   PUBLIC_SUFFIX_MARKER_PRIVATE_START,
   PUBLIC_SUFFIX_MARKER_PRIVATE_END,
 } from "../config";
+import { toASCII } from "../punycode";
 
 describe("buildTries()", () => {
   test("normalizes rules (to punycode, to lower case)", () => {

serialized-tries/info.json

@@ -1,10 +1,10 @@
-import { toASCII } from "punycode";
 import {
   PUBLIC_SUFFIX_MARKER_ICANN_START,
   PUBLIC_SUFFIX_MARKER_ICANN_END,
   PUBLIC_SUFFIX_MARKER_PRIVATE_START,
   PUBLIC_SUFFIX_MARKER_PRIVATE_END,
 } from "../config";
+import { toASCII } from "../punycode";
 import { createTrieFromList } from "../trie/create-trie";
 
 const matchNewLine = /\r?\n/u;
@@ -34,22 +34,6 @@ const extractByMarkers = (
 const containsRule = (line: string) =>
   matchComment.test(line) === false && matchWhitespace.test(line) === false;
 
-const normalizeRule = (rule: string) =>
-  /*
-	Users of parse-domain should only pass parsed hostnames (e.g. via new URL("...").hostname)
-	to parseDomain(). This frees us from the burden of URL parsing.
-	See also https://github.com/peerigon/parse-domain/issues/14
-
-	The problem is that new URL("...").hostname will also translate all non-ASCII characters
-	into punycode (e.g. 大分.jp becomes xn--kbrq7o.jp). This means for us that our serialized
-	trie needs to use punyencoded hostnames.
-
-	The downside of this is that punycode hostnames are bigger in terms of file size.
-	However, that's still better than including a punycode library into the runtime code of parse-domain.
-	Gzip is also very effective in removing duplicate characters.
-	*/
-  toASCII(rule).toLowerCase();
-
 const parsePsl = (listContent: string) => {
   return {
     icann: extractByMarkers(
@@ -59,15 +43,15 @@ const parsePsl = (listContent: string) => {
     )
       .split(matchNewLine)
       .filter(containsRule)
-      .map(normalizeRule),
+      .map(toASCII),
     private: extractByMarkers(
       listContent,
       PUBLIC_SUFFIX_MARKER_PRIVATE_START,
       PUBLIC_SUFFIX_MARKER_PRIVATE_END
     )
       .split(matchNewLine)
       .filter(containsRule)
-      .map(normalizeRule),
+      .map(toASCII),
   };
 };
 

src/psl/build-tries.test.ts

@@ -1,5 +1,5 @@
-import { toASCII } from "punycode";
 import { parseDomain, ParseResult, ParseResultType } from "./main";
+import { toASCII } from "./punycode";
 
 describe("Official test suite from https://raw.githubusercontent.com/publicsuffix/list/main/tests/test_psl.txt", () => {
   const canonical = (hostname: string) =>

src/psl/build-tries.ts

@@ -0,0 +1,3 @@
+export const toASCII = (hostname: string) => {
+  return new URL(`http://${hostname}`).hostname;
+};