Skip to content

A fully open source & end-to-end encrypted note taking alternative to Evernote.

core: fix escaped html turning into markup in search results (#10519)

Abdullah Atta87a058e3c0fc
diff --git a/packages/core/__tests__/lookup.test.js b/packages/core/__tests__/lookup.test.js
index 30d593c56..1ba083674 100644
--- a/packages/core/__tests__/lookup.test.js
+++ b/packages/core/__tests__/lookup.test.js
@@ -184,6 +184,29 @@ describe("notesWithHighlighting", () => {
expect(item.item).toBeDefined();
}));
+ test("search should keep escaped html escaped in highlighted content", () =>
+ databaseTest().then(async (db) => {
+ await db.notes.add({
+ title: "escaped",
+ content: {
+ type: "tiptap",
+ data: `<p>&lt;img src=x onerror=alert(1)&gt; hello</p><p><a href="https://example.com/?a=1&amp;b=&quot;2&quot;">hello link</a></p>`
+ }
+ });
+
+ const filtered = await db.lookup.notesWithHighlighting(
+ "hello",
+ db.notes.all
+ );
+ const { item } = await filtered.item(0);
+ expect(item.rawContent).not.toContain("<img");
+ expect(item.rawContent).toContain("&lt;img src=x onerror=alert(1)&gt; ");
+ expect(item.rawContent).toContain(
+ `href="https://example.com/?a=1&amp;b=&quot;2&quot;"`
+ );
+ expect(item.content[0][0].prefix).toBe("<img src=x onerror=alert(1)> ");
+ }));
+
test("search should be diacritic agnostic", () =>
databaseTest().then(async (db) => {
await db.notes.add({ title: "outdoor café" });
diff --git a/packages/core/src/api/lookup.ts b/packages/core/src/api/lookup.ts
index 71130041e..c502d46ac 100644
--- a/packages/core/src/api/lookup.ts
+++ b/packages/core/src/api/lookup.ts
@@ -18,7 +18,7 @@ along with this program. If not, see <http://www.gnu.org/licenses/>.
*/
import { match } from "fuzzyjs";
-import { decodeHTML } from "entities";
+import { decodeHTML, escapeAttribute, escapeText } from "entities";
import Database from "./index.js";
import {
HighlightedResult,
@@ -982,11 +982,14 @@ function highlightHtmlContent(html: string, queries: string[]): string {
{
ontext(text) {
const normalizedText = removeDiacritics(text);
+ // text is decoded by the parser so it must be escaped again
+ // otherwise escaped text (e.g. &lt;) would turn into markup
const highlighted = highlightRegexMatches(
text,
normalizedText,
searchRegex,
- matchIdCounter
+ matchIdCounter,
+ escapeText
);
matchIdCounter = highlighted.nextId;
@@ -1039,7 +1042,7 @@ function highlightHtmlContent(html: string, queries: string[]): string {
continue;
}
- html += ` ${key}="${value}"`;
+ html += ` ${key}="${escapeAttribute(value)}"`;
}
html += `>${element.buffer}</${element.name}>`;
@@ -1193,7 +1196,8 @@ function highlightRegexMatches(
text: string,
normalizedText: string,
regex: RegExp,
- startId: number
+ startId: number,
+ encode: (text: string) => string = (text) => text
): { result: string; hasMatches: boolean; nextId: number } {
let matchIdCounter = startId;
let hasMatches = false;
@@ -1203,15 +1207,15 @@ function highlightRegexMatches(
while ((m = regex.exec(normalizedText)) !== null) {
hasMatches = true;
- result += text.slice(lastIndex, m.index);
+ result += encode(text.slice(lastIndex, m.index));
result += createSearchResultTag(
- text.slice(m.index, m.index + m[0].length),
+ encode(text.slice(m.index, m.index + m[0].length)),
`match-${++matchIdCounter}`
);
lastIndex = m.index + m[0].length;
}
- result += text.slice(lastIndex);
+ result += encode(text.slice(lastIndex));
return { result, hasMatches, nextId: matchIdCounter };
}