From 1fe0f6e20b4b01f66615aa8da9376f1daeb31991 Mon Sep 17 00:00:00 2001
From: Raffael Meyer <14891507+barredterra@users.noreply.github.com>
Date: Tue, 16 Jul 2024 14:15:43 +0200
Subject: [PATCH] fix: handle invalid images in HTML (#27095)

* fix: handle invalid image

* refactor: simplify broken image replacement

* fix: add alt attribute to broken image
---
 frappe/core/doctype/file/test_file.py | 21 ++++++++++++++++++++-
 frappe/core/doctype/file/utils.py     | 12 +++++++++++-
 2 files changed, 31 insertions(+), 2 deletions(-)

diff --git a/frappe/core/doctype/file/test_file.py b/frappe/core/doctype/file/test_file.py
index 6c9b9f5872..03a8927dac 100644
--- a/frappe/core/doctype/file/test_file.py
+++ b/frappe/core/doctype/file/test_file.py
@@ -17,7 +17,7 @@ from frappe.core.api.file import (
 	unzip_file,
 )
 from frappe.core.doctype.file.exceptions import FileTypeNotAllowed
-from frappe.core.doctype.file.utils import get_extension
+from frappe.core.doctype.file.utils import get_corrupted_image_msg, get_extension
 from frappe.desk.form.utils import add_comment
 from frappe.exceptions import ValidationError
 from frappe.tests.utils import FrappeTestCase, change_settings
@@ -768,6 +768,25 @@ class TestFileUtils(FrappeTestCase):
 		)
 		self.assertRegex(communication.content, r"<img src=\"(.*)/files/pix\.png(.*)\">")
 
+	def test_broken_image(self):
+		"""Ensure that broken inline images don't cause errors."""
+		is_private = not frappe.get_meta("Communication").make_attachments_public
+		communication = frappe.get_doc(
+			doctype="Communication",
+			communication_type="Communication",
+			communication_medium="Email",
+			content='<div class="ql-editor read-mode"><img src="data:image/png;filename=pix.png;base64,iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAQAAAC1HAwCAAAAC0lEQVR42mNkYAAAAAYAAjCB0C8AAAAASUVORK5CY="></div>',
+			recipients="to <to@test.com>",
+			cc=None,
+			bcc=None,
+			sender="sender@test.com",
+		).insert(ignore_permissions=True)
+
+		self.assertFalse(
+			frappe.db.exists("File", {"attached_to_name": communication.name, "is_private": is_private})
+		)
+		self.assertIn(f'<img src="#broken-image" alt="{get_corrupted_image_msg()}">', communication.content)
+
 	def test_create_new_folder(self):
 		folder = create_new_folder("test_folder", "Home")
 		self.assertTrue(folder.is_folder)
diff --git a/frappe/core/doctype/file/utils.py b/frappe/core/doctype/file/utils.py
index 578c760389..d16f76b090 100644
--- a/frappe/core/doctype/file/utils.py
+++ b/frappe/core/doctype/file/utils.py
@@ -2,6 +2,7 @@ import hashlib
 import mimetypes
 import os
 import re
+from binascii import Error as BinasciiError
 from io import BytesIO
 from typing import TYPE_CHECKING, Optional
 from urllib.parse import unquote
@@ -234,7 +235,12 @@ def extract_images_from_html(doc: "Document", content: str, is_private: bool = F
 			content = content.encode("utf-8")
 		if b"," in content:
 			content = content.split(b",")[1]
-		content = safe_b64decode(content)
+
+		try:
+			content = safe_b64decode(content)
+		except BinasciiError:
+			frappe.flags.has_dataurl = True
+			return f'<img src="#broken-image" alt="{get_corrupted_image_msg()}"'
 
 		if "filename=" in headers:
 			filename = headers.split("filename=")[-1]
@@ -273,6 +279,10 @@ def extract_images_from_html(doc: "Document", content: str, is_private: bool = F
 	return content
 
 
+def get_corrupted_image_msg():
+	return _("Image: Corrupted Data Stream")
+
+
 def get_random_filename(content_type: str | None = None) -> str:
 	extn = None
 	if content_type: