self-hosted/joplin
1
0
Fork 0
mirror of https://github.com/laurent22/joplin.git synced 2024-12-30 10:36:35 +02:00
Code Issues Releases Activity

Desktop: Fixes #9868: Fix markdown (frontmatter) notes that start with a UTF-8 BOM aren't imporetd correctly (#9875)

Browse Source
This commit is contained in:
Henry Heino 2024-02-08 08:55:34 -08:00 committed by GitHub
parent 628877147e
commit d614d40cfd
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
4 changed files with 28 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

8
packages/app-cli/tests/support/test_notes/yaml/note_with_byte_order_mark.md Normal file
View File

@ -0,0 +1,8 @@
---
title: Frontmatter test
tags:
- tag1
- tag2
---
This note begins with an invisible byte order mark, just before its frontmatter.

4
packages/lib/services/interop/InteropService_Importer_Md.ts
View File

@ -13,6 +13,7 @@ import { unique } from '../../ArrayUtils';
const { pregQuote } = require('../../string-utils-common');
import { MarkupToHtml } from '@joplin/renderer';
import { isDataUrl } from '@joplin/utils/url';
import { stripBom } from '../../string-utils';
export default class InteropService_Importer_Md extends InteropService_Importer_Base {
protected importedNotes: Record<string, NoteEntity> = {};
@ -167,7 +168,8 @@ export default class InteropService_Importer_Md extends InteropService_Importer_
if (!stat) throw new Error(`Cannot read ${resolvedPath}`);
const ext = fileExtension(resolvedPath);
const title = filename(resolvedPath);
const body = await shim.fsDriver().readFile(resolvedPath);
const body = stripBom(await shim.fsDriver().readFile(resolvedPath));
const note = {
parent_id: parentFolderId,
title: title,

9
packages/lib/services/interop/InteropService_Importer_Md_frontmatter.test.ts
View File

@ -173,4 +173,13 @@ describe('InteropService_Importer_Md_frontmatter: importMetadata', () => {
const note = await importTestFile('note_with_dataurl_image.md');
expect(note.body).toBe('<img src="data:image/svg+xml,%3Csvg%20xmlns%3D%22http%3A//www.w3.org/2000/svg%22%20viewBox%3D%220%200%2023%2038%22%3E%3Cpath%20d%3D%22M16.6%2038.1h-5.5l-.2-2.9-.2%202.9h-5.5L5%2025.3l-.8%202a1.53%201.53%200%2001-1.9.9l-1.2-.4a1.58%201.58%200%2001-1-1.9v-.1c.3-.9%203.1-11.2%203.1-11.2a2.66%202.66%200%20012.3-2l.6-.5a6.93%206.93%200%20014.7-12%206.8%206.8%200%20014.9%202%207%207%200%20012%204.9%206.65%206.65%200%2001-2.2%205l.7.5a2.78%202.78%200%20012.4%202s2.9%2011.2%202.9%2011.3a1.53%201.53%200%2001-.9%201.9l-1.3.4a1.63%201.63%200%2001-1.9-.9l-.7-1.8-.1%2012.7zm-3.6-2h1.7L14.9%2020.3l1.9-.3%202.4%206.3.3-.1c-.2-.8-.8-3.2-2.8-10.9a.63.63%200%2000-.6-.5h-.6l-1.1-.9h-1.9l-.3-2a4.83%204.83%200%20003.5-4.7A4.78%204.78%200%200011%202.3H10.8a4.9%204.9%200%2000-1.4%209.6l-.3%202h-1.9l-1%20.9h-.6a.74.74%200%2000-.6.5c-2%207.5-2.7%2010-3%2010.9l.3.1L4.8%2020l1.9.3.2%2015.8h1.6l.6-8.4a1.52%201.52%200%20011.5-1.4%201.5%201.5%200%20011.5%201.4l.9%208.4zm-10.9-9.6zm17.5-.1z%22%20style%3D%22isolation%3Aisolate%22%20fill%3D%22%23333%22%20opacity%3D%22.7%22/%3E%3Cpath%20d%3D%22M5.9%2013.6l1.1-.9h7.8l1.2.9%22%20fill%3D%22%23ce592c%22/%3E%3Cellipse%20cx%3D%2210.9%22%20cy%3D%2213.1%22%20rx%3D%222.7%22%20ry%3D%22.3%22%20style%3D%22isolation%3Aisolate%22%20fill%3D%22%23ce592c%22%20opacity%3D%22.5%22/%3E%3Cpath%20d%3D%22M20.6%2026.1l-2.9-11.3a1.71%201.71%200%2000-1.6-1.2H5.699999999999999a1.69%201.69%200%2000-1.5%201.3l-3.1%2011.3a.61.61%200%2000.3.7l1.1.4a.61.61%200%2000.7-.3l2.7-6.7.2%2016.8h3.6l.6-9.3a.47.47%200%2001.44-.5h.06c.4%200%20.4.2.5.5l.6%209.3h3.6L15.7%2020.3l2.5%206.6a.52.52%200%2000.66.31l1.2-.4a.57.57%200%2000.5-.7z%22%20fill%3D%22%23fdbf2d%22/%3E%3Cpath%20d%3D%22M7%2013.6l3.9%206.7%203.9-6.7%22%20style%3D%22isolation%3Aisolate%22%20fill%3D%22%23cf572e%22%20opacity%3D%22.6%22/%3E%3Ccircle%20cx%3D%2210.9%22%20cy%3D%227%22%20r%3D%225.9%22%20fill%3D%22%23fdbf2d%22/%3E%3C/svg%3E" alt="Street View Pegman Control" style="height:30px;width:30px;position:absolute;transform:translate(-50%,-50%);pointer-events:none">');
});
it('should recognize frontmatter in a file that starts with a UTF8 byte order mark', async () => {
const note = await importTestFile('note_with_byte_order_mark.md');
expect(note.title).toBe('Frontmatter test');
expect(note.body).toBe('This note begins with an invisible byte order mark, just before its frontmatter.\n');
const tags = (await Tag.tagsByNoteId(note.id)).map(tag => tag.title).sort();
expect(tags).toMatchObject(['tag1', 'tag2']);
});
});

8
packages/lib/string-utils.ts
View File

@ -303,3 +303,11 @@ export function scriptType(s: string) {
return 'en';
}
// A UTF-8/UTF-16 byte order mark can appear at the start of a file and
// can break logic that relies on a file starting with specific text.
// See https://github.com/laurent22/joplin/issues/9868
export const stripBom = (text: string) => {
// Remove the UTF-16 BOM --- NodeJS seems to convert UTF-8 BOMs to UTF-16 BOMs
// when reading files.
return text.replace(/^\ufeff/u, '');
};