Compare commits

..

8 Commits

Author SHA1 Message Date
Philipinho 206961e842 feat: properly preserve table width 2026-04-30 02:21:39 +01:00
Philipinho 980521f957 v0.80.1 2026-04-27 16:06:32 +01:00
Philipinho fe44dc92a9 sync 2026-04-27 15:51:23 +01:00
Philip Okugbe fad410ef23 chore: add undici for oidc proxy support (#2132) 2026-04-27 15:50:42 +01:00
Philipinho 15b8908b1a update postcss 2026-04-27 15:23:47 +01:00
Philipinho 8e15b22d8c package updates 2026-04-27 15:22:02 +01:00
Philipinho ec83fc82d5 fix: refactor sanitize 2026-04-27 15:16:26 +01:00
Philipinho a573acedd0 fix: local storage, and package overrides 2026-04-22 14:13:25 +01:00
13 changed files with 1201 additions and 951 deletions
+6 -6
View File
@@ -1,7 +1,7 @@
{
"name": "client",
"private": true,
"version": "0.80.0",
"version": "0.80.1",
"scripts": {
"dev": "vite",
"build": "tsc && vite build",
@@ -31,8 +31,8 @@
"emoji-mart": "^5.6.0",
"file-saver": "^2.0.5",
"highlightjs-sap-abap": "^0.3.0",
"i18next": "^25.10.1",
"i18next-http-backend": "^3.0.2",
"i18next": "25.10.1",
"i18next-http-backend": "3.0.6",
"jotai": "^2.18.1",
"jotai-optics": "^0.4.0",
"js-cookie": "^3.0.5",
@@ -42,7 +42,7 @@
"mantine-form-zod-resolver": "^1.3.0",
"mermaid": "^11.13.0",
"mitt": "^3.0.1",
"posthog-js": "1.363.1",
"posthog-js": "1.372.2",
"react": "^18.3.1",
"react-arborist": "3.4.0",
"react-clear-modal": "^2.0.18",
@@ -50,7 +50,7 @@
"react-drawio": "^1.0.7",
"react-error-boundary": "^6.1.1",
"react-helmet-async": "^3.0.0",
"react-i18next": "^16.5.8",
"react-i18next": "16.5.8",
"react-router-dom": "^7.13.1",
"semver": "^7.7.4",
"socket.io-client": "^4.8.3",
@@ -74,7 +74,7 @@
"eslint-plugin-react-refresh": "^0.5.2",
"globals": "^15.13.0",
"optics-ts": "^2.4.1",
"postcss": "^8.5.8",
"postcss": "^8.5.12",
"postcss-preset-mantine": "^1.18.0",
"postcss-simple-vars": "^7.0.1",
"prettier": "^3.8.1",
@@ -80,10 +80,12 @@ export const MarkdownClipboard = Extension.create({
const { from, to } = view.state.selection;
const parsed = markdownToHtml(text.replace(/\n+$/, ""));
const body = elementFromString(parsed);
normalizeTableColumnWidths(body);
const contentNodes = DOMParser.fromSchema(
this.editor.schema,
).parseSlice(elementFromString(parsed), {
).parseSlice(body, {
preserveWhitespace: true,
});
@@ -137,3 +139,92 @@ function elementFromString(value) {
return new window.DOMParser().parseFromString(wrappedValue, "text/html").body;
}
const DEFAULT_PASTE_COL_WIDTH_PX = 150;
function parsePixelWidth(el: Element): number | null {
const attr = el.getAttribute("width");
if (attr) {
const n = parseInt(attr, 10);
if (Number.isFinite(n) && n > 0) return n;
}
const style = el.getAttribute("style") || "";
const m = style.match(/(?:^|;)\s*width\s*:\s*([\d.]+)\s*px/i);
if (m) {
const n = parseInt(m[1], 10);
if (Number.isFinite(n) && n > 0) return n;
}
return null;
}
function getFirstRow(table: Element): Element | null {
const tbodyRow = table.querySelector(":scope > tbody > tr");
if (tbodyRow) return tbodyRow;
const theadRow = table.querySelector(":scope > thead > tr");
if (theadRow) return theadRow;
return table.querySelector(":scope > tr");
}
function deriveColumnWidths(table: Element): (number | null)[] | null {
const cols = table.querySelectorAll(":scope > colgroup > col");
if (cols.length > 0) {
const widths: (number | null)[] = [];
cols.forEach((col) => widths.push(parsePixelWidth(col)));
if (widths.some((w) => w !== null)) return widths;
}
const firstRow = getFirstRow(table);
if (!firstRow) return null;
const widths: (number | null)[] = [];
Array.from(firstRow.children)
.filter((c) => c.tagName === "TD" || c.tagName === "TH")
.forEach((cell) => {
const colspan = parseInt(cell.getAttribute("colspan") || "1", 10) || 1;
const w = parsePixelWidth(cell);
for (let i = 0; i < colspan; i++) {
widths.push(w !== null ? Math.round(w / colspan) : null);
}
});
if (widths.length === 0 || widths.every((w) => w === null)) return null;
return widths;
}
// Mirror of server normalizeTableColumnWidths (see import/utils/table-utils.ts):
// markdown source has no widths, so without this every pasted table renders
// at table-layout:fixed/100% and squashes columns to fit the editor instead of
// letting .tableWrapper's overflow-x: auto scroll.
export function normalizeTableColumnWidths(root: Element): void {
root.querySelectorAll("table").forEach((table) => {
const firstRow = getFirstRow(table);
if (!firstRow) return;
let colWidths = deriveColumnWidths(table);
if (!colWidths) {
let count = 0;
Array.from(firstRow.children)
.filter((c) => c.tagName === "TD" || c.tagName === "TH")
.forEach((cell) => {
count += parseInt(cell.getAttribute("colspan") || "1", 10) || 1;
});
if (count === 0) return;
colWidths = new Array(count).fill(DEFAULT_PASTE_COL_WIDTH_PX);
}
let col = 0;
Array.from(firstRow.children)
.filter((c) => c.tagName === "TD" || c.tagName === "TH")
.forEach((cell) => {
if (cell.getAttribute("colwidth")) {
col += parseInt(cell.getAttribute("colspan") || "1", 10) || 1;
return;
}
const colspan = parseInt(cell.getAttribute("colspan") || "1", 10) || 1;
const slice = colWidths!.slice(col, col + colspan);
col += colspan;
if (slice.length === 0 || slice.every((w) => w === null)) return;
const values = slice.map((w) => (w == null ? 100 : w));
cell.setAttribute("colwidth", values.join(","));
});
});
}
+21 -17
View File
@@ -1,6 +1,6 @@
{
"name": "server",
"version": "0.80.0",
"version": "0.80.1",
"description": "",
"author": "",
"private": true,
@@ -33,13 +33,13 @@
"@ai-sdk/google": "^3.0.52",
"@ai-sdk/openai": "^3.0.47",
"@ai-sdk/openai-compatible": "^2.0.37",
"@aws-sdk/client-s3": "3.1014.0",
"@aws-sdk/lib-storage": "3.1014.0",
"@aws-sdk/s3-request-presigner": "3.1014.0",
"@aws-sdk/client-s3": "3.1037.0",
"@aws-sdk/lib-storage": "3.1037.0",
"@aws-sdk/s3-request-presigner": "3.1037.0",
"@clickhouse/client": "^1.18.2",
"@fastify/cookie": "^11.0.2",
"@fastify/multipart": "^9.4.0",
"@fastify/static": "^9.0.0",
"@fastify/multipart": "^10.0.0",
"@fastify/static": "^9.1.3",
"@keyv/redis": "^5.1.6",
"@langchain/core": "1.1.39",
"@langchain/textsplitters": "1.0.1",
@@ -48,19 +48,19 @@
"@nestjs-labs/nestjs-ioredis": "^11.0.4",
"@nestjs/bullmq": "^11.0.4",
"@nestjs/cache-manager": "^3.1.0",
"@nestjs/common": "^11.1.18",
"@nestjs/config": "^4.0.3",
"@nestjs/core": "^11.1.18",
"@nestjs/common": "^11.1.19",
"@nestjs/config": "^4.0.4",
"@nestjs/core": "^11.1.19",
"@nestjs/event-emitter": "^3.0.1",
"@nestjs/jwt": "11.0.2",
"@nestjs/mapped-types": "^2.1.1",
"@nestjs/passport": "^11.0.5",
"@nestjs/platform-fastify": "^11.1.18",
"@nestjs/platform-socket.io": "^11.1.18",
"@nestjs/schedule": "^6.1.1",
"@nestjs/platform-fastify": "^11.1.19",
"@nestjs/platform-socket.io": "^11.1.19",
"@nestjs/schedule": "^6.1.3",
"@nestjs/terminus": "^11.1.1",
"@nestjs/throttler": "^6.5.0",
"@nestjs/websockets": "^11.1.18",
"@nestjs/websockets": "^11.1.19",
"@node-saml/passport-saml": "^5.1.0",
"@react-email/components": "1.0.10",
"@react-email/render": "2.0.4",
@@ -69,7 +69,7 @@
"ai-sdk-ollama": "^3.8.1",
"bcrypt": "^6.0.0",
"bowser": "^2.14.1",
"bullmq": "^5.71.0",
"bullmq": "^5.76.0",
"cache-manager": "^7.2.8",
"cheerio": "^1.2.0",
"class-transformer": "^0.5.1",
@@ -110,22 +110,23 @@
"react": "^18.3.1",
"reflect-metadata": "^0.2.2",
"rxjs": "^7.8.2",
"sanitize-filename-ts": "1.0.2",
"sanitize-filename": "1.6.3",
"socket.io": "^4.8.3",
"stripe": "^17.7.0",
"tlds": "^1.261.0",
"tmp-promise": "^3.0.3",
"tseep": "^1.3.1",
"typesense": "^3.0.5",
"undici": "7.24.0",
"ws": "^8.20.0",
"yauzl": "^3.2.1",
"zod": "^4.3.6"
},
"devDependencies": {
"@eslint/js": "^9.28.0",
"@nestjs/cli": "^11.0.18",
"@nestjs/cli": "^11.0.21",
"@nestjs/schematics": "^11.0.10",
"@nestjs/testing": "^11.1.18",
"@nestjs/testing": "^11.1.19",
"@types/bcrypt": "^6.0.0",
"@types/debounce": "^1.2.4",
"@types/fs-extra": "^11.0.4",
@@ -165,6 +166,9 @@
"transform": {
"^.+\\.(t|j)s$": "ts-jest"
},
"transformIgnorePatterns": [
"/node_modules/(?!(\\.pnpm/)?(nanoid|uuid|image-dimensions|marked)(@|/))"
],
"collectCoverageFrom": [
"**/*.(t|j)s"
],
Binary file not shown.
+28 -6
View File
@@ -1,6 +1,6 @@
import * as path from 'path';
import * as bcrypt from 'bcrypt';
import { sanitize } from 'sanitize-filename-ts';
import sanitize = require('sanitize-filename');
import { FastifyRequest } from 'fastify';
import { Readable, Transform } from 'stream';
@@ -72,11 +72,33 @@ export function extractDateFromUuid7(uuid7: string) {
return new Date(timestamp);
}
export function sanitizeFileName(fileName: string): string {
const sanitizedFilename = sanitize(fileName)
.replace(/ /g, '_')
.replace(/#/g, '_');
return sanitizedFilename.slice(0, 255);
export type SanitizeFileNameOptions = {
/** Keep spaces and `#` instead of replacing them with `_`. Useful for
* download filenames where readability matters. Defaults to false. */
preserveSpaces?: boolean;
};
export function sanitizeFileName(
fileName: string,
options: SanitizeFileNameOptions = {},
): string {
// Decode percent-encoded sequences so that bypasses like "..%2F" reach
// sanitize() as literal "../" and get stripped. sanitize-filename only
// strips literal characters and won't catch encoded path separators
// on its own.
const decoded = fileName.replace(/%[0-9a-fA-F]{2}/g, (m) => {
try {
return decodeURIComponent(m);
} catch {
return m;
}
});
const sanitized = sanitize(decoded);
if (options.preserveSpaces) {
return sanitized;
}
return sanitized.replace(/ /g, '_').replace(/#/g, '_');
}
export function removeAccent(str: string): string {
@@ -53,7 +53,6 @@ import { EnvironmentService } from '../../integrations/environment/environment.s
import { TokenService } from '../auth/services/token.service';
import { JwtAttachmentPayload, JwtType } from '../auth/dto/jwt-payload';
import * as path from 'path';
import { sanitize } from 'sanitize-filename-ts';
import { AttachmentInfoDto, RemoveIconDto } from './dto/attachment.dto';
import { PageAccessService } from '../page/page-access/page-access.service';
import { AuditEvent, AuditResource } from '../../common/events/audit-events';
@@ -357,13 +356,19 @@ export class AttachmentController {
throw new BadRequestException('Invalid image attachment type');
}
if (!fileName || sanitize(fileName) !== fileName) {
if (!fileName) {
throw new BadRequestException('Invalid file name');
}
const filenameWithoutExt = path.basename(fileName, path.extname(fileName));
if (!isValidUUID(filenameWithoutExt)) {
throw new BadRequestException('Invalid file id');
const ext = path.extname(fileName);
const filenameWithoutExt = path.basename(fileName, ext);
if (
!ext ||
!isValidUUID(filenameWithoutExt) ||
`${filenameWithoutExt}${ext}` !== fileName
) {
throw new BadRequestException('Invalid file name');
}
const filePath = `${getAttachmentFolderPath(attachmentType, workspace.id)}/${fileName}`;
@@ -23,9 +23,12 @@ import {
SpaceCaslSubject,
} from '../../core/casl/interfaces/space-ability.type';
import { FastifyReply } from 'fastify';
import { sanitize } from 'sanitize-filename-ts';
import { getExportExtension } from './utils';
import { getMimeType, getPageTitle } from '../../common/helpers';
import {
getMimeType,
getPageTitle,
sanitizeFileName,
} from '../../common/helpers';
import * as path from 'path';
import { AuditEvent, AuditResource } from '../../common/events/audit-events';
import {
@@ -85,7 +88,9 @@ export class ExportController {
if (result.type === 'file') {
const ext = getExportExtension(dto.format);
const fileName = sanitize(page.title || 'untitled') + ext;
const fileName =
sanitizeFileName(page.title || 'untitled', { preserveSpaces: true }) +
ext;
const contentType = getMimeType(path.extname(fileName));
res.headers({
@@ -96,7 +101,9 @@ export class ExportController {
res.send(result.content);
} else {
const fileName = sanitize(page.title || 'untitled') + '.zip';
const fileName =
sanitizeFileName(page.title || 'untitled', { preserveSpaces: true }) +
'.zip';
res.headers({
'Content-Type': 'application/zip',
@@ -144,7 +151,9 @@ export class ExportController {
'Content-Type': 'application/zip',
'Content-Disposition':
'attachment; filename="' +
encodeURIComponent(sanitize(exportFile.fileName)) +
encodeURIComponent(
sanitizeFileName(exportFile.fileName, { preserveSpaces: true }),
) +
'"',
});
@@ -1,7 +1,6 @@
import { BadRequestException, Injectable, Logger } from '@nestjs/common';
import { PageRepo } from '@docmost/db/repos/page/page.repo';
import { MultipartFile } from '@fastify/multipart';
import { sanitize } from 'sanitize-filename-ts';
import * as path from 'path';
import {
htmlToJson,
@@ -30,6 +29,8 @@ import { InjectQueue } from '@nestjs/bullmq';
import { Queue } from 'bullmq';
import { QueueJob, QueueName } from '../../queue/constants';
import { ModuleRef } from '@nestjs/core';
import { load } from 'cheerio';
import { normalizeImportHtml } from '../utils/import-formatter';
@Injectable()
export class ImportService {
@@ -53,8 +54,8 @@ export class ImportService {
const file = await filePromise;
const fileBuffer = await file.toBuffer();
const fileExtension = path.extname(file.filename).toLowerCase();
const fileName = sanitize(
path.basename(file.filename, fileExtension).slice(0, 255),
const fileName = sanitizeFileName(
path.basename(file.filename, fileExtension),
);
const fileContent = fileBuffer.toString();
@@ -137,7 +138,9 @@ export class ImportService {
async processHTML(htmlInput: string): Promise<any> {
try {
return htmlToJson(htmlInput);
const $ = load(htmlInput);
normalizeImportHtml($, $.root());
return htmlToJson($.html() || '');
} catch (err) {
throw err;
}
@@ -5,6 +5,7 @@ import { v7 } from 'uuid';
import { InsertableBacklink } from '@docmost/db/types/entity.types';
import { Cheerio, CheerioAPI, load } from 'cheerio';
import slugify from '@sindresorhus/slugify';
import { normalizeTableColumnWidths } from './table-utils';
// Check if text contains Unicode characters (for emojis/icons)
function isUnicodeCharacter(text: string): boolean {
@@ -51,9 +52,7 @@ export async function formatImportHtml(opts: {
}
}
notionFormatter($, $root);
xwikiFormatter($, $root);
defaultHtmlFormatter($, $root);
normalizeImportHtml($, $root);
const backlinks = await rewriteInternalLinksToMentionHtml(
$,
@@ -73,6 +72,23 @@ export async function formatImportHtml(opts: {
};
}
/**
* Contextless HTML cleanup shared by every import path.
* - notionFormatter: no-op on non-Notion HTML (class-selector-based).
* - xwikiFormatter: no-op on non-XWiki HTML (looks for #xwikicontent).
* - defaultHtmlFormatter: table column widths + provider auto-embeds.
*
* Does NOT run rewriteInternalLinksToMentionHtml — that requires zip context.
*/
export function normalizeImportHtml(
$: CheerioAPI,
$root: Cheerio<any>,
): void {
notionFormatter($, $root);
xwikiFormatter($, $root);
defaultHtmlFormatter($, $root);
}
export function xwikiFormatter($: CheerioAPI, $root: Cheerio<any>) {
const $content = $root.find('#xwikicontent');
if ($content.length) {
@@ -82,6 +98,8 @@ export function xwikiFormatter($: CheerioAPI, $root: Cheerio<any>) {
}
export function defaultHtmlFormatter($: CheerioAPI, $root: Cheerio<any>) {
normalizeTableColumnWidths($, $root);
$root.find('a[href]').each((_, el) => {
const $el = $(el);
const url = $el.attr('href')!;
@@ -0,0 +1,107 @@
import { CheerioAPI, Cheerio } from 'cheerio';
const DEFAULT_IMPORT_COL_WIDTH_PX = 150;
/**
* Extracts a pixel-integer width from either the `width` attribute or
* `style="width: Npx"` on a <col>/<td>/<th>. Returns null when absent,
* non-numeric, or a non-px unit (em, %).
*/
function parsePixelWidth(el: Cheerio<any>): number | null {
const attr = el.attr('width');
if (attr) {
const n = parseInt(attr, 10);
if (Number.isFinite(n) && n > 0) return n;
}
const style = el.attr('style') || '';
const m = style.match(/(?:^|;)\s*width\s*:\s*([\d.]+)\s*px/i);
if (m) {
const n = parseInt(m[1], 10);
if (Number.isFinite(n) && n > 0) return n;
}
return null;
}
/**
* Derives per-column widths for a table, in visual column order.
* Priority: <colgroup><col> → first-row cells' own width style.
* Returns an array of length = number of columns, with null entries
* for columns whose width couldn't be determined.
*/
function deriveColumnWidths(
$: CheerioAPI,
table: Cheerio<any>,
): (number | null)[] | null {
const cols = table.find('> colgroup > col');
if (cols.length > 0) {
const widths: (number | null)[] = [];
cols.each(function () {
widths.push(parsePixelWidth($(this)));
});
if (widths.some((w) => w !== null)) return widths;
}
// Fallback: first row's cells.
const firstRow = table.find('> tbody > tr, > thead > tr, > tr').first();
if (!firstRow.length) return null;
const widths: (number | null)[] = [];
firstRow.children('td, th').each(function () {
const cell = $(this);
const colspan = parseInt(cell.attr('colspan') || '1', 10) || 1;
const w = parsePixelWidth(cell);
for (let i = 0; i < colspan; i++) {
widths.push(w !== null ? Math.round(w / colspan) : null);
}
});
if (widths.every((w) => w === null)) return null;
return widths;
}
/**
* Apply colwidth attributes to the first row of each table based on
* derived column widths. Accounts for colspan. Idempotent — re-running
* on already-normalized markup is a no-op.
*
* This lives upstream of tiptap's generateJSON: tiptap reads
* `colwidth="N[,N...]"` on <td>/<th> to build the runtime <colgroup>.
*/
export function normalizeTableColumnWidths(
$: CheerioAPI,
$root: Cheerio<any>,
): void {
$root.find('table').each(function () {
const table = $(this);
const firstRow = table.find('> tbody > tr, > thead > tr, > tr').first();
if (!firstRow.length) return;
let colWidths = deriveColumnWidths($, table);
if (!colWidths) {
// No widths anywhere (e.g. markdown-sourced tables). Apply a default
// per-column width so the table's intrinsic width can exceed the
// editor container, letting .tableWrapper's overflow-x: auto scroll
// instead of cramming columns into the available width.
let count = 0;
firstRow.children('td, th').each(function () {
count += parseInt($(this).attr('colspan') || '1', 10) || 1;
});
if (count === 0) return;
colWidths = new Array(count).fill(DEFAULT_IMPORT_COL_WIDTH_PX);
}
let col = 0;
firstRow.children('td, th').each(function () {
const cell = $(this);
if (cell.attr('colwidth')) {
col += parseInt(cell.attr('colspan') || '1', 10) || 1;
return;
}
const colspan = parseInt(cell.attr('colspan') || '1', 10) || 1;
const slice = colWidths.slice(col, col + colspan);
col += colspan;
if (slice.length === 0 || slice.every((w) => w === null)) return;
const values = slice.map((w) => (w == null ? 100 : w));
cell.attr('colwidth', values.join(','));
});
});
}
+10 -9
View File
@@ -1,7 +1,7 @@
{
"name": "docmost",
"homepage": "https://docmost.com",
"version": "0.80.0",
"version": "0.80.1",
"private": true,
"scripts": {
"build": "nx run-many -t build",
@@ -62,7 +62,7 @@
"cross-env": "^10.1.0",
"date-fns": "^4.1.0",
"diff": "8.0.3",
"dompurify": "^3.3.3",
"dompurify": "3.4.1",
"fractional-indexing-jittered": "^1.0.0",
"highlight.js": "^11.11.1",
"image-dimensions": "^2.5.0",
@@ -72,7 +72,7 @@
"ms": "3.0.0-canary.1",
"qrcode": "^1.5.4",
"rfc6902": "5.2.0",
"uuid": "^13.0.0",
"uuid": "^14.0.0",
"y-indexeddb": "^9.0.12",
"y-prosemirror": "1.3.7",
"yjs": "^13.6.30"
@@ -102,9 +102,9 @@
"y-prosemirror": "1.3.7",
"glob": "13.0.6",
"ws": "8.20.0",
"dompurify": "3.3.3",
"dompurify": "3.4.1",
"tmp": "0.2.5",
"hono": "4.12.12",
"hono": "4.12.14",
"mermaid": "11.13.0",
"nanoid@^3": "3.3.8",
"socket.io-parser": "4.2.6",
@@ -123,16 +123,17 @@
"flatted": "3.4.2",
"picomatch@<2.3.2": "2.3.2",
"picomatch@>=4.0.0 <4.0.4": "4.0.4",
"fastify": "5.8.3",
"fastify": "5.8.5",
"yaml@>=1.0.0 <1.10.3": "1.10.3",
"yaml@>=2.0.0 <2.8.3": "2.8.3",
"path-to-regexp@^8": "8.4.0",
"brace-expansion@^5": "5.0.5",
"@xmldom/xmldom": "0.8.12",
"@xmldom/xmldom": "0.8.13",
"handlebars": "4.7.9",
"axios": "1.15.0",
"langsmith": "0.5.18",
"follow-redirects": "1.16.0"
"langsmith": "0.5.19",
"follow-redirects": "1.16.0",
"protobufjs": "7.5.5"
},
"neverBuiltDependencies": []
}
+884 -894
View File
File diff suppressed because it is too large Load Diff