fix: handle malformed URLs gracefully during import/export (#1868)

* Handling malformed URLs gracefully

* Allow import of invalid URLs, but adding logging.

---------

Co-authored-by: gpapp <gergely.papp@itworks.hu>

apps/server/src/integrations/export/utils.ts

@@ -1,4 +1,5 @@
 import { jsonToNode } from 'src/collaboration/collaboration.util';
+import { Logger } from '@nestjs/common';
 import { ExportFormat } from './dto/export-dto';
 import { Node } from '@tiptap/pm/model';
 import { validate as isValidUUID } from 'uuid';
@@ -88,7 +89,7 @@ export function replaceInternalLinks(
             // if link and text are same, use page title
             if (markLink === node.text) {
               //@ts-expect-error
-              node.text = getInternalLinkPageName(relativePath);
+              node.text = getInternalLinkPageName(relativePath, currentPagePath);
             }
           }
         }
@@ -99,10 +100,19 @@ export function replaceInternalLinks(
   return doc.toJSON();
 }
 
-export function getInternalLinkPageName(path: string): string {
-  return decodeURIComponent(
-    path?.split('/').pop().split('.').slice(0, -1).join('.'),
-  );
+export function getInternalLinkPageName(path: string, currentFilePath?: string): string {
+  const name = path?.split('/').pop().split('.').slice(0, -1).join('.');
+  try {
+    return decodeURIComponent(name);
+  } catch (err) {
+    if (currentFilePath) {
+      Logger.warn(
+        `URI malformed in page ${currentFilePath}: ${name}. Falling back to raw name.`,
+        'ExportUtils',
+      );
+    }
+    return name;
+  }
 }
 
 export function extractPageSlugId(input: string): string {

apps/server/src/integrations/import/utils/import-formatter.ts

@@ -1,4 +1,5 @@
 import { getEmbedUrlAndProvider } from '@docmost/editor-ext';
+import { Logger } from '@nestjs/common';
 import * as path from 'path';
 import { v7 } from 'uuid';
 import { InsertableBacklink } from '@docmost/db/types/entity.types';
@@ -280,8 +281,18 @@ export async function rewriteInternalLinksToMentionHtml(
     const $a = $(el);
     const raw = $a.attr('href')!;
     if (raw.startsWith('http') || raw.startsWith('/api/')) return;
+    let decodedRaw = raw;
+    try {
+      decodedRaw = decodeURIComponent(raw);
+    } catch (err) {
+      Logger.warn(
+        `URI malformed in page ${currentFilePath}: ${raw}. Falling back to raw path.`,
+        'ImportFormatter',
+      );
+    }
+
     const resolved = normalize(
-      path.join(path.dirname(currentFilePath), decodeURIComponent(raw)),
+      path.join(path.dirname(currentFilePath), decodedRaw),
     );
     const meta = filePathToPageMetaMap.get(resolved);
     if (!meta) return;

apps/server/src/integrations/import/utils/import.utils.ts

@@ -1,3 +1,4 @@
+import { Logger } from '@nestjs/common';
 import { promises as fs } from 'fs';
 import * as path from 'path';
 
@@ -30,8 +31,13 @@ export function resolveRelativeAttachmentPath(
   pageDir: string,
   attachmentCandidates: Map<string, string>,
 ): string | null {
-  const mainRel = decodeURIComponent(raw.replace(/^\.?\/+/, ''));
-  const fallback = path.normalize(path.join(pageDir, mainRel));
+  let mainRel = raw.replace(/^\.?\/+/, '');
+  try {
+    mainRel = decodeURIComponent(mainRel);
+  } catch (err) {
+    Logger.warn(`URI malformed for attachment path: ${mainRel}. Falling back to raw path.`, 'ImportUtils');
+  }
+  const fallback = path.normalize(path.join(pageDir, mainRel)).split(path.sep).join('/');
 
   if (attachmentCandidates.has(mainRel)) {
     return mainRel;