fix(eslint): add support for custom pageExtensions in no-html-link-fo…

[Srajal1311]

…r-pages rule

[ijjk]

Allow CI Workflow Run

• approve CI run for commit: bfff7d1

Note: this should only be enabled once the PR is ready to go and can only be enabled by a maintainer

[ijjk]

Allow CI Workflow Run

• approve CI run for commit: bfff7d1

Note: this should only be enabled once the PR is ready to go and can only be enabled by a maintainer

[vercel]

The new file extension check will never match because it tests the normalized URL (which has a trailing slash) against a regex that expects the string to end with a file extension.

View Details

📝 Patch Details

diff --git a/packages/eslint-plugin-next/src/rules/no-html-link-for-pages.ts b/packages/eslint-plugin-next/src/rules/no-html-link-for-pages.ts
index f38216066d..4b849d660c 100644
--- a/packages/eslint-plugin-next/src/rules/no-html-link-for-pages.ts
+++ b/packages/eslint-plugin-next/src/rules/no-html-link-for-pages.ts
@@ -162,14 +162,15 @@ export default defineRule({
           return
         }
 
-        const hrefPath = normalizeURL(href.value.value)
-        // Outgoing links are ignored
-        if (/^(https?:\/\/|\/\/)/.test(hrefPath)) {
+        // --- NEW CODE: skip internal links with allowed extensions ---
+        // Check before normalization to avoid the trailing slash added by normalizeURL
+        if (allowedExtRegex.test(href.value.value)) {
           return
         }
 
-        // --- NEW CODE: skip internal links with allowed extensions ---
-        if (allowedExtRegex.test(hrefPath)) {
+        const hrefPath = normalizeURL(href.value.value)
+        // Outgoing links are ignored
+        if (/^(https?:\/\/|\/\/)/.test(hrefPath)) {
           return
         }
 

Analysis

File extension check never matches due to trailing slash from URL normalization

What fails: In packages/eslint-plugin-next/src/rules/no-html-link-for-pages.ts, the allowedExtRegex check at line 172 never matches links with file extensions (.js, .jsx, .ts, .tsx) because it tests against the normalized URL which has a trailing slash appended.

How to reproduce:

// In the ESLint rule, with href="/public/script.js"
const hrefPath = normalizeURL("/public/script.js");  // Returns "/public/script.js/"
const allowedExtRegex = /\.(js|jsx|ts|tsx)$/;
allowedExtRegex.test(hrefPath);  // Returns false (expects to end with extension, but ends with "/")

Result: Links to static files like <a href="/public/script.js"> are incorrectly flagged by the ESLint rule, even though they should be skipped.

Expected: The regex should match and skip these links, as they point to static resources rather than Next.js pages. Testing against the original href.value.value before normalization correctly identifies file extensions.

Fix: Check for file extensions before URL normalization, since normalizeURL() appends a trailing slash that breaks the end-of-string anchor ($) in the regex pattern.