fix: better detect for IP addresses, add test

i18n/en.json

@@ -13,8 +13,9 @@
         },
         "search-help-text": "Search {engine}"
     },
-    "404": {
+    "notfound": {
-        "title": "Page Not Found"
+        "title": "page not found",
+        "desc": "Please check if there is a typo in the URL. <br/>If SparkHome brought you to this page,<br/> please <a style=\"text-decoration:underline;\" href=\"mailto:contact@alikia2x.com\">contact us.</a>"
     },
     "about": {
         "title": "SparkHome"

lib/url/validLink.ts

@@ -1,22 +1,26 @@
-import punycode from "punycode/";
+import punycode from "punycode";
 import { tldList } from "./tldList";
 
 export default function validLink(link: string) {
-    let finalURL = '';
+    let finalURL;
     try {
         const url = new URL(link);
-        finalURL = url.origin;
+        finalURL = url;
         return true;
     } catch (error) {
         // if the URL is invalid, try to add the protocol
         try {
             const urlWithHTTP = new URL("http://" + link);
-            finalURL = urlWithHTTP.origin;
+            finalURL = urlWithHTTP;
         } catch (error) {
             return false;
         }
     }
-    if (validTLD(finalURL)) {
+    if (
+        validTLD(finalURL.host) ||
+        isValidIPv6(finalURL.host.slice(1, finalURL.host.length - 1)) ||
+        isValidIPv4(finalURL.host)
+    ) {
         return true;
     } else {
         return false;
@@ -31,3 +35,54 @@ export function validTLD(domain: string): boolean {
         return false;
     }
 }
+
+export function isValidIPv6(ip: string): boolean {
+    const length = ip.length;
+    let groups = 1;
+    let groupDigits = 0;
+    let doubleColonCount = 0;
+    for (let i = 0; i < length; i++) {
+        const char = ip[i];
+        if ("0" <= char && char <= "9") {
+            groupDigits++;
+        } else if ("a" <= char && char <= "f") {
+            groupDigits++;
+        } else if ("A" <= char && char <= "F") {
+            groupDigits++;
+        } else if (char === ":" && i + 1 < length && ip[i + 1] !== ":") {
+            groups++;
+            groupDigits = 0;
+        } else if (char === ":" && i + 1 < length && ip[i + 1] === ":") {
+            doubleColonCount++;
+            i++;
+            groupDigits = 0;
+        } else {
+            return false;
+        }
+        if (groups > 8) {
+            return false;
+        } else if (groupDigits > 4) {
+            return false;
+        } else if (doubleColonCount > 1) {
+            return false;
+        }
+    }
+    if (doubleColonCount === 0 && groups !== 8) {
+        return false;
+    }
+    return true;
+}
+
+export function isValidIPv4(ip: string): boolean {
+    const parts = ip.split(".");
+    if (parts.length !== 4) {
+        return false;
+    }
+    for (const part of parts) {
+        const num = Number(part);
+        if (isNaN(num) || num < 0 || num > 255 || !part.match(/^\d+$/)) {
+            return false;
+        }
+    }
+    return true;
+}

package.json

@@ -1,7 +1,7 @@
 {
   "name": "sparkhome",
   "private": false,
-  "version": "5.2.1",
+  "version": "5.2.2",
   "type": "module",
   "scripts": {
     "dev": "bun server.ts",
@@ -21,6 +21,7 @@
     "i18next": "^23.11.5",
     "i18next-browser-languagedetector": "^8.0.0",
     "i18next-icu": "^2.3.0",
+    "jest": "^29.7.0",
     "jotai": "^2.8.3",
     "node-nlp": "^4.27.0",
     "react": "^18.3.1",

pages/[...].tsx

@@ -0,0 +1,17 @@
+import { useTranslation } from "react-i18next";
+
+export default function NotFound() {
+    const { t } = useTranslation();
+    return (
+        <div className="relative w-screen h-screen flex justify-center items-center">
+            <div className="flex items-center">
+                <h1 className="text-7xl font-thin">404</h1>
+                <div className="relative h-20 mx-4 w-[0.15rem] bg-black dark:bg-white"></div>
+                <div className="flex flex-col">
+                    <div className="uppercase text-3xl font-light">{t("notfound.title")}</div>
+                    <div className="text-sm" dangerouslySetInnerHTML={{__html:t("notfound.desc")}}></div>
+                </div>
+            </div>
+        </div>
+    );
+}

test/validLink.test.ts

@@ -0,0 +1,70 @@
+import { describe, expect, test } from "@jest/globals";
+import validLink, { validTLD } from "../lib/url/validLink";
+
+describe("Check if a string is an accessible domain/URL/IP", () => {
+    test("Plain, full URL", () => {
+        // Plain form
+        expect(validLink("http://example.com")).toBe(true);
+        // With https and path
+        expect(validLink("https://jestjs.io/docs/getting-started/")).toBe(true);
+        // With anchor
+        expect(validLink("https://difftastic.wilfred.me.uk/zh-CN/git.html#git-difftool")).toBe(true);
+        // With params
+        expect(validLink("https://www.bilibili.com/list/ml2252204359?oid=990610203&bvid=BV1sx4y1g7Hh")).toBe(true);
+    });
+    test("Punycode URL", () => {
+        expect(validLink("https://原神大学.com/")).toBe(true);
+        expect(validLink("中国原神大学.com")).toBe(true);
+    });
+    test("Invalid TLD with protocol", () => {
+        expect(validLink("https://www.example.notexist")).toBe(true);
+    });
+    test("Invalid TLD with no protocol", () => {
+        expect(validLink("www.example.notexist")).toBe(false);
+    });
+    test("IPv4 without protocol", () => {
+        expect(validLink("127.0.0.1")).toBe(true);
+    });
+    test("IPv6 without protocol", () => {
+        expect(validLink("[::]")).toBe(true);
+    });
+});
+
+// Reference: https://www.iana.org/domains/root/db
+describe("Check if the given TLD exist and assigned.", () => {
+    test("Valid normal TLD", () => {
+        expect(validTLD("com")).toBe(true);
+        expect(validTLD("top")).toBe(true);
+        expect(validTLD("net")).toBe(true);
+        expect(validTLD("org")).toBe(true);
+    });
+    test("Valid new TLDs", () => {
+        // they really exist!
+        expect(validTLD("foo")).toBe(true);
+        expect(validTLD("bar")).toBe(true);
+    });
+    test("Exist but not assigned TLD", () => {
+        expect(validTLD("active")).toBe(false);
+        expect(validTLD("off")).toBe(false);
+    });
+    test("with dot", () => {
+        expect(validTLD(".com")).toBe(true);
+        expect(validTLD(".us")).toBe(true);
+        expect(validTLD(".cn")).toBe(true);
+        expect(validTLD(".io")).toBe(true);
+    });
+    test("Punycode TLDs", () => {
+        expect(validTLD(".中国")).toBe(true);
+        expect(validTLD(".РФ")).toBe(true);
+        expect(validTLD(".कॉम")).toBe(true);
+        expect(validTLD("ایران")).toBe(true);
+        expect(validTLD("இலங்கை")).toBe(true);
+        expect(validTLD("გე")).toBe(true);
+        expect(validTLD("ポイント")).toBe(true);
+    });
+    test("Punycode TLDs but not assigned", () => {
+        expect(validTLD("テスト")).toBe(false);
+        expect(validTLD("परीक्षा")).toBe(false);
+        expect(validTLD("测试")).toBe(false);
+    });
+});