From 0458bad5f24f2a87acd04f3723a416f227488070 Mon Sep 17 00:00:00 2001
From: shamoon <4887959+shamoon@users.noreply.github.com>
Date: Sat, 22 Aug 2026 06:15:53 -0700
Subject: [PATCH] Fix: append charset to file response for text files (#13759)

---
 src/documents/tests/test_api_documents.py | 50 +++++++++++++++++++++++
 src/documents/views.py                    |  3 ++
 2 files changed, 53 insertions(+)

diff --git a/src/documents/tests/test_api_documents.py b/src/documents/tests/test_api_documents.py
index 0acd5e937..b67b87772 100644
--- a/src/documents/tests/test_api_documents.py
+++ b/src/documents/tests/test_api_documents.py
@@ -497,6 +497,56 @@ class TestDocumentApi(DirectoriesMixin, ConsumeTaskMixin, APITestCase):
         )
         response.close()
 
+    @override_settings(FILENAME_FORMAT="")
+    def test_serve_text_file_declares_utf8_charset(self) -> None:
+        """
+        GIVEN:
+            - A UTF-8 encoded text document
+        WHEN:
+            - The file is served for preview or download
+        THEN:
+            - The Content-Type declares the UTF-8 charset, so the browser does
+              not fall back to its locale default and mangle non-ASCII text
+        """
+        doc = Document.objects.create(
+            title="none",
+            filename="my_document.txt",
+            mime_type="text/plain",
+        )
+        Path(doc.source_path).write_bytes("für Grüße München".encode())
+
+        for endpoint in ("preview", "download"):
+            with self.subTest(endpoint=endpoint):
+                response = self.client.get(f"/api/documents/{doc.pk}/{endpoint}/")
+                self.assertEqual(response.status_code, status.HTTP_200_OK)
+                self.assertEqual(response["Content-Type"], "text/plain; charset=utf-8")
+                self.assertEqual(
+                    read_streaming_response(response).decode("utf-8"),
+                    "für Grüße München",
+                )
+
+    @override_settings(FILENAME_FORMAT="")
+    def test_serve_pdf_file_has_no_charset(self) -> None:
+        """
+        GIVEN:
+            - A PDF document
+        WHEN:
+            - The file is served for preview
+        THEN:
+            - No charset is added to the binary content type
+        """
+        doc = Document.objects.create(
+            title="none",
+            filename="my_document.pdf",
+            mime_type="application/pdf",
+        )
+        Path(doc.source_path).write_bytes(b"This is a test")
+
+        response = self.client.get(f"/api/documents/{doc.pk}/preview/")
+        self.assertEqual(response.status_code, status.HTTP_200_OK)
+        self.assertEqual(response["Content-Type"], "application/pdf")
+        response.close()
+
     def test_document_actions_not_existing_file(self) -> None:
         doc = Document.objects.create(
             title="none",
diff --git a/src/documents/views.py b/src/documents/views.py
index 80232cd73..e1b9f194c 100644
--- a/src/documents/views.py
+++ b/src/documents/views.py
@@ -4778,6 +4778,9 @@ def serve_file(
         # Support browser previewing csv files by using text mime type
         if mime_type in {"application/csv", "text/csv"} and disposition == "inline":
             mime_type = "text/plain"
+        # Tell browsers to use UTF-8 for the text files we parse as UTF-8
+        if mime_type in {"text/plain", "text/csv", "application/csv"}:
+            mime_type = f"{mime_type}; charset=utf-8"
 
     response = FileResponse(file_handle, content_type=mime_type)
     # Firefox is not able to handle unicode characters in filename field