From b68151a179fe9a9e02388b2aad9352a34cffec75 Mon Sep 17 00:00:00 2001
From: seaHi <me@seahi.me>
Date: Wed, 16 Sep 2026 21:32:47 +0800
Subject: [PATCH] =?UTF-8?q?refactor:=20=E5=B0=86=E6=96=87=E7=AB=A0?=
 =?UTF-8?q?=E6=91=98=E8=A6=81=E7=94=9F=E6=88=90=E7=AE=80=E5=8C=96=E4=B8=BA?=
 =?UTF-8?q?=E7=BA=AF=E6=96=87=E6=9C=AC=E5=A4=84=E7=90=86=E5=B9=B6=E7=A7=BB?=
 =?UTF-8?q?=E9=99=A4=20HTML=20=E6=B8=B2=E6=9F=93?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

---
 src/lib/markdown/excerpt.test.ts             |  62 +++++------
 src/lib/markdown/excerpt.ts                  | 109 ++-----------------
 src/routes/+page.server.ts                   |   6 +-
 src/routes/+page.svelte                      |   3 +-
 src/routes/categories/[slug]/+page.server.ts |   4 +-
 src/routes/categories/[slug]/+page.svelte    |   3 +-
 src/routes/tags/[tag]/+page.server.ts        |   4 +-
 src/routes/tags/[tag]/+page.svelte           |   3 +-
 8 files changed, 47 insertions(+), 147 deletions(-)

diff --git a/src/lib/markdown/excerpt.test.ts b/src/lib/markdown/excerpt.test.ts
index 3edfedf..0dbfd49 100644
--- a/src/lib/markdown/excerpt.test.ts
+++ b/src/lib/markdown/excerpt.test.ts
@@ -1,63 +1,57 @@
 import { describe, expect, it } from 'vitest';
-import { createExcerptHtml } from './excerpt';
+import { createExcerpt } from './excerpt';
 
-describe('createExcerptHtml', () => {
-	it('renders inline styles instead of stripping them', () => {
-		expect(createExcerptHtml('这是**粗体**与*斜体*，还有~~删除线~~。')).toBe(
-			'这是<strong>粗体</strong>与<em>斜体</em>，还有<del>删除线</del>。...'
+describe('createExcerpt', () => {
+	it('strips inline markers instead of rendering them', () => {
+		expect(createExcerpt('这是**粗体**与*斜体*，还有~~删除线~~。')).toBe(
+			'这是粗体与斜体，还有删除线。'
 		);
 	});
 
-	it('keeps inline code and drops the backticks', () => {
-		expect(createExcerptHtml('先执行 `pnpm run build` 再发布')).toBe(
-			'先执行 <code>pnpm run build</code> 再发布...'
-		);
+	it('drops the backticks of inline code', () => {
+		expect(createExcerpt('先执行 `pnpm run build` 再发布')).toBe('先执行 pnpm run build 再发布');
 	});
 
 	it('reduces links and images to readable text', () => {
-		expect(createExcerptHtml('参考[官方文档](https://example.com)与![配图](cover.png)说明')).toBe(
-			'参考官方文档与配图说明...'
+		expect(createExcerpt('参考[官方文档](https://example.com)与![配图](cover.png)说明')).toBe(
+			'参考官方文档与配图说明'
 		);
 	});
 
 	it('removes block level markers so the excerpt stays a single line', () => {
-		expect(createExcerptHtml('# 标题\n\n- 第一项\n- 第二项\n\n> 引用内容')).toBe(
-			'标题 第一项 第二项 引用内容...'
+		expect(createExcerpt('# 标题\n\n- 第一项\n- 第二项\n\n> 引用内容')).toBe(
+			'标题 第一项 第二项 引用内容'
 		);
 	});
 
 	it('drops Hugo shortcodes and raw HTML', () => {
 		expect(
-			createExcerptHtml('{{< admonition note "提示" >}}正文{{< /admonition >}}<span>片段</span>')
-		).toBe('正文片段...');
+			createExcerpt('{{< admonition note "提示" >}}正文{{< /admonition >}}<span>片段</span>')
+		).toBe('正文片段');
 	});
 
-	it('closes inline tags that are cut off by the length limit', () => {
-		const html = createExcerptHtml(`**${'a'.repeat(200)}**`, 10);
+	it('truncates to the length limit without appending an ellipsis', () => {
+		const excerpt = createExcerpt('a'.repeat(200), 10);
 
-		expect(html).toBe(`<strong>${'a'.repeat(10)}</strong>...`);
+		expect(excerpt).toBe('a'.repeat(10));
+		expect(excerpt).not.toContain('...');
 	});
 
-	it('counts HTML entities as a single visible character', () => {
-		expect(createExcerptHtml('&'.repeat(30), 5)).toBe(`${'&amp;'.repeat(5)}...`);
+	it('never returns markdown markers or markup', () => {
+		const excerpt = createExcerpt('**粗体** 和 `code` 以及 <b>html</b>');
+
+		expect(excerpt).toBe('粗体 和 code 以及 html');
+		expect(excerpt).not.toMatch(/[*`~<>]/);
 	});
 
-	it('never keeps a dangling inline tag when the cut lands on it', () => {
-		const html = createExcerptHtml(`${'a'.repeat(9)}**bc**`, 10);
-
-		expect(html).toBe(`aaaaaaaaa<strong>b</strong>...`);
-	});
-
-	it('escapes characters that would otherwise break the markup', () => {
-		const html = createExcerptHtml('对比 1 < 2 的结果 <script>alert(1)</script>');
-
-		expect(html).toContain('1 &lt; 2');
-		expect(html).not.toContain('<script');
+	it('keeps plain text that only looks like markup', () => {
+		expect(createExcerpt('对比 1 < 2 的结果是安全的')).toBe('对比 1 < 2 的结果是安全的');
 	});
 
 	it('returns an empty string when there is nothing to show', () => {
-		expect(createExcerptHtml('')).toBe('');
-		expect(createExcerptHtml('   \n\n  ')).toBe('');
-		expect(createExcerptHtml('{{< admonition note >}}{{< /admonition >}}')).toBe('');
+		expect(createExcerpt('')).toBe('');
+		expect(createExcerpt('   \n\n  ')).toBe('');
+		expect(createExcerpt('{{< admonition note >}}{{< /admonition >}}')).toBe('');
+		expect(createExcerpt('***')).toBe('');
 	});
 });
diff --git a/src/lib/markdown/excerpt.ts b/src/lib/markdown/excerpt.ts
index 3e98f1d..f88f332 100644
--- a/src/lib/markdown/excerpt.ts
+++ b/src/lib/markdown/excerpt.ts
@@ -1,32 +1,12 @@
-import { Marked } from 'marked';
-import sanitizeHtml from 'sanitize-html';
-
-/** 摘要默认保留的可见字符数 */
+/** 摘要默认保留的字符数 */
 export const EXCERPT_LENGTH = 150;
 
-/** 摘要中允许保留的行内样式标签（粗体、斜体、删除线、行内代码） */
-export const EXCERPT_INLINE_TAGS = ['strong', 'em', 'del', 'code'];
-
-const excerptMarked = new Marked();
-
-/** 仅匹配摘要白名单内的行内标签，用于截断时配对 */
-const INLINE_TAG_PATTERN = new RegExp(
-	`^<\\s*(/?)\\s*(${EXCERPT_INLINE_TAGS.join('|')})\\s*>$`,
-	'i'
-);
-
-/** 匹配没有任何内容的行内标签，例如 `<strong></strong>` */
-const EMPTY_TAG_PATTERN = new RegExp(`<(${EXCERPT_INLINE_TAGS.join('|')})>\\s*</\\1>`, 'gi');
-
 /** 匹配 HTML 标签，要求以字母开头，避免误伤 `1 < 2` 这类正文 */
 const HTML_TAG_PATTERN = /<\/?[a-zA-Z][^>]*>/g;
 
-/** 匹配 HTML 实体，截断时按一个可见字符计算 */
-const HTML_ENTITY_PATTERN = /^&(?:#\d+|#x[0-9a-fA-F]+|[a-zA-Z]+);/;
-
 /**
- * 把文章内容整理成适合做摘要的单行文本：
- * 去掉 shortcode / HTML / 块级 Markdown 标记，保留粗体、斜体等行内标记。
+ * 把文章内容整理成适合做摘要的单行纯文本：
+ * 去掉 shortcode、HTML 与 Markdown 标记，只保留正文文字。
  */
 function normalizeExcerptSource(source: string): string {
 	return source
@@ -42,88 +22,17 @@ function normalizeExcerptSource(source: string): string {
 		.replace(/^[ \t]{0,3}\|?[ \t]*:?-{3,}:?[ \t]*(?:\|[ \t]*:?-{3,}:?[ \t]*)*\|?[ \t]*$/gm, '') // 表格分隔行
 		.replace(/^[ \t]{0,3}(?:[*_-][ \t]*){3,}$/gm, '') // 分隔线
 		.replace(/\|/g, ' ') // 表格竖线
+		.replace(/[*`~]/g, '') // 行内标记：粗体、斜体、删除线、行内代码
 		.replace(/\s+/g, ' ')
 		.trim();
 }
 
 /**
- * 按可见字符数截断行内 HTML，并补齐被截断的标签。
- * 只接受白名单内的行内标签，遇到其它标签立即停止，保证输出安全。
+ * 由文章内容生成纯文本摘要：不保留任何 Markdown 标记，也不追加省略号。
+ * 无内容时返回空字符串。
  */
-function truncateInlineHtml(html: string, maxLength: number): string {
-	let result = '';
-	let visible = 0;
-	const openTags: string[] = [];
+export function createExcerpt(source: string, maxLength: number = EXCERPT_LENGTH): string {
+	const plainText = normalizeExcerptSource(source || '');
 
-	for (let index = 0; index < html.length && visible < maxLength;) {
-		const char = html[index];
-
-		if (char === '<') {
-			const end = html.indexOf('>', index);
-			if (end === -1) break;
-
-			const tag = html.slice(index, end + 1);
-			const match = INLINE_TAG_PATTERN.exec(tag);
-			if (!match) break;
-
-			const name = match[2].toLowerCase();
-			if (match[1]) {
-				const position = openTags.lastIndexOf(name);
-				if (position !== -1) openTags.splice(position, 1);
-			} else {
-				openTags.push(name);
-			}
-
-			result += tag;
-			index = end + 1;
-			continue;
-		}
-
-		if (char === '&') {
-			const entity = HTML_ENTITY_PATTERN.exec(html.slice(index));
-			if (entity) {
-				result += entity[0];
-				visible += 1;
-				index += entity[0].length;
-				continue;
-			}
-		}
-
-		result += char;
-		visible += 1;
-		index += 1;
-	}
-
-	for (const name of openTags.reverse()) {
-		result += `</${name}>`;
-	}
-
-	let cleaned = result;
-	for (;;) {
-		const next = cleaned.replace(EMPTY_TAG_PATTERN, '').replace(/\s+$/, '');
-		if (next === cleaned) break;
-		cleaned = next;
-	}
-
-	return cleaned;
-}
-
-/**
- * 由文章内容生成带行内样式的摘要 HTML（粗体、斜体、删除线、行内代码）。
- * 返回值已做白名单净化，可安全地用 `{@html}` 渲染；无内容时返回空字符串。
- */
-export function createExcerptHtml(source: string, maxLength: number = EXCERPT_LENGTH): string {
-	const normalized = normalizeExcerptSource(source || '');
-	if (!normalized) return '';
-
-	// 先解析再截断：源文本必须先完整解析，否则成对的行内标记会被切断而失效
-	const html = excerptMarked.parseInline(normalized) as string;
-	const sanitized = sanitizeHtml(html, {
-		allowedTags: EXCERPT_INLINE_TAGS,
-		allowedAttributes: {},
-		disallowedTagsMode: 'discard'
-	});
-
-	const truncated = truncateInlineHtml(sanitized, maxLength).trim();
-	return truncated ? `${truncated}...` : '';
+	return plainText.slice(0, maxLength).trim();
 }
diff --git a/src/routes/+page.server.ts b/src/routes/+page.server.ts
index 092034c..cbedebf 100644
--- a/src/routes/+page.server.ts
+++ b/src/routes/+page.server.ts
@@ -1,6 +1,6 @@
 /* eslint-disable @typescript-eslint/no-explicit-any */
 
-import { createExcerptHtml } from '$lib/markdown/excerpt';
+import { createExcerpt } from '$lib/markdown/excerpt';
 import type { PageServerLoad } from './$types';
 
 export const load: PageServerLoad = async ({ platform }) => {
@@ -51,8 +51,8 @@ export const load: PageServerLoad = async ({ platform }) => {
 			// Prefer the manually marked summary, otherwise derive it from the beginning
 			const contentStr = ((post.content as string) || '').split('<!--more-->')[0];
 
-			// Inline styles (bold, italic, ...) are kept as safe HTML for the cards
-			const excerpt = createExcerptHtml(contentStr);
+			// The excerpt is plain text: no Markdown markers and no trailing ellipsis
+			const excerpt = createExcerpt(contentStr);
 
 			// Do not send full content to the list view to save bandwidth
 
diff --git a/src/routes/+page.svelte b/src/routes/+page.svelte
index 5049c9f..4bfc3f3 100644
--- a/src/routes/+page.svelte
+++ b/src/routes/+page.svelte
@@ -24,8 +24,7 @@
 					{/if}
 					<div class="content-wrap">
 						<h1 class="title">{post.title}</h1>
-						<!-- eslint-disable-next-line svelte/no-at-html-tags -->
-						<p class="excerpt">{@html post.excerpt}</p>
+						<p class="excerpt">{post.excerpt}</p>
 					</div>
 				</article>
 			</a>
diff --git a/src/routes/categories/[slug]/+page.server.ts b/src/routes/categories/[slug]/+page.server.ts
index a45a068..f377a53 100644
--- a/src/routes/categories/[slug]/+page.server.ts
+++ b/src/routes/categories/[slug]/+page.server.ts
@@ -1,5 +1,5 @@
 import { error, isHttpError } from '@sveltejs/kit';
-import { createExcerptHtml } from '$lib/markdown/excerpt';
+import { createExcerpt } from '$lib/markdown/excerpt';
 import type { PageServerLoad } from './$types';
 
 const PAGE_SIZE = 20;
@@ -72,7 +72,7 @@ export const load: PageServerLoad = async ({ platform, params, url, parent }) =>
 							year: 'numeric',
 							timeZone: 'UTC'
 						}),
-				excerpt: createExcerptHtml(post.excerpt_source || '')
+				excerpt: createExcerpt(post.excerpt_source || '')
 			};
 		});
 
diff --git a/src/routes/categories/[slug]/+page.svelte b/src/routes/categories/[slug]/+page.svelte
index 01427c7..85f7be4 100644
--- a/src/routes/categories/[slug]/+page.svelte
+++ b/src/routes/categories/[slug]/+page.svelte
@@ -38,8 +38,7 @@
 						{/if}
 						<div class="content-wrap">
 							<h1 class="title">{post.title}</h1>
-							<!-- eslint-disable-next-line svelte/no-at-html-tags -->
-							<p class="excerpt">{@html post.excerpt}</p>
+							<p class="excerpt">{post.excerpt}</p>
 						</div>
 					</article>
 				</a>
diff --git a/src/routes/tags/[tag]/+page.server.ts b/src/routes/tags/[tag]/+page.server.ts
index 38649a8..057ff5c 100644
--- a/src/routes/tags/[tag]/+page.server.ts
+++ b/src/routes/tags/[tag]/+page.server.ts
@@ -1,5 +1,5 @@
 import { error, isHttpError } from '@sveltejs/kit';
-import { createExcerptHtml } from '$lib/markdown/excerpt';
+import { createExcerpt } from '$lib/markdown/excerpt';
 import type { PageServerLoad } from './$types';
 
 const PAGE_SIZE = 20;
@@ -73,7 +73,7 @@ export const load: PageServerLoad = async ({ platform, params, url, parent }) =>
 							year: 'numeric',
 							timeZone: 'UTC'
 						}),
-				excerpt: createExcerptHtml(post.excerpt_source || '')
+				excerpt: createExcerpt(post.excerpt_source || '')
 			};
 		});
 
diff --git a/src/routes/tags/[tag]/+page.svelte b/src/routes/tags/[tag]/+page.svelte
index 4d41671..f3686fc 100644
--- a/src/routes/tags/[tag]/+page.svelte
+++ b/src/routes/tags/[tag]/+page.svelte
@@ -37,8 +37,7 @@
 						{/if}
 						<div class="content-wrap">
 							<h1 class="title">{post.title}</h1>
-							<!-- eslint-disable-next-line svelte/no-at-html-tags -->
-							<p class="excerpt">{@html post.excerpt}</p>
+							<p class="excerpt">{post.excerpt}</p>
 						</div>
 					</article>
 				</a>