From cb9a212cc7a10231dd36bc164394721afe26969f Mon Sep 17 00:00:00 2001
From: radishzzz <shuyinsen@gmail.com>
Date: Sun, 23 Mar 2025 18:50:14 +0000
Subject: [PATCH] fix: prevent HTML tags from being escaped in article
 descriptions

---
 .vscode/settings.json    |  1 +
 src/utils/description.ts | 22 ++++++++++++----------
 2 files changed, 13 insertions(+), 10 deletions(-)

diff --git a/.vscode/settings.json b/.vscode/settings.json
index 5826c03..90b090b 100644
--- a/.vscode/settings.json
+++ b/.vscode/settings.json
@@ -62,6 +62,7 @@
     "abbrlink",
     "antfu",
     "apiflash",
+    "apos",
     "Artículos",
     "astrodotbuild",
     "astrojs",
diff --git a/src/utils/description.ts b/src/utils/description.ts
index 04018e8..82e740e 100644
--- a/src/utils/description.ts
+++ b/src/utils/description.ts
@@ -1,12 +1,12 @@
 import type { CollectionEntry } from 'astro:content'
 import { defaultLocale } from '@/config'
 import MarkdownIt from 'markdown-it'
-import sanitizeHtml from 'sanitize-html'
-
-const parser = new MarkdownIt()
 
 type ExcerptScene = 'list' | 'meta' | 'og' | 'rss'
 
+const parser = new MarkdownIt()
+const isCJKLang = (lang: string) => ['zh', 'zh-tw', 'ja'].includes(lang)
+
 // Excerpt length in different scenarios
 const EXCERPT_LENGTHS: Record<ExcerptScene, {
   cjk: number
@@ -30,8 +30,6 @@ const EXCERPT_LENGTHS: Record<ExcerptScene, {
   },
 }
 
-const isCJKLang = (lang: string) => ['zh', 'zh-tw', 'ja'].includes(lang)
-
 // Generate an excerpt from Markdown content
 export function generateExcerpt(
   content: string,
@@ -45,11 +43,15 @@ export function generateExcerpt(
     ? EXCERPT_LENGTHS[scene].cjk
     : EXCERPT_LENGTHS[scene].other
 
-  // Convert Markdown to plain text
-  const plainText = sanitizeHtml(parser.render(content), {
-    allowedTags: [],
-    allowedAttributes: {},
-  })
+  // Remove all HTML tags and decode HTML entities
+  const plainText = parser.render(content)
+    .replace(/<[^>]*>/g, '')
+    .replace(/&lt;/g, '<')
+    .replace(/&gt;/g, '>')
+    .replace(/&amp;/g, '&')
+    .replace(/&quot;/g, '"')
+    .replace(/&apos;/g, '\'')
+    .replace(/&nbsp;/g, ' ')
 
   // Replace line breaks with spaces
   const normalizedText = plainText.replace(/\s+/g, ' ')