🐛 Handle unicode characters when redacting email attachment filenames

froide/account/services.py

@@ -1,6 +1,7 @@
 import hashlib
 import hmac
 import re
+import unicodedata
 from datetime import timedelta
 from typing import Dict, Optional
 from urllib.parse import urlencode
@@ -313,7 +314,7 @@ def check_against_blocklist(cls, user, save=True) -> bool:
                 user.save()
         return blocklisted
 
-    def apply_name_redaction(self, content, replacement=""):
+    def apply_name_redaction(self, content, replacement="", unicode=True):
         if not self.user.private:
             return content
 
@@ -328,6 +329,15 @@ def apply_name_redaction(self, content, replacement=""):
             *self.user.first_name.split(),
             *self.user.last_name.split(),
         ]
+
+        if not unicode:
+            user_asciiish_name = (
+                unicodedata.normalize("NFKD", self.user.get_full_name())
+                .encode("ascii", "ignore")
+                .decode()
+            )
+            needles.extend(user_asciiish_name.split())
+
         if self.user.organization_name:
             needles.append(self.user.organization_name)
 

froide/account/tests/test_account.py

@@ -854,3 +854,13 @@ def test_multipart_name_redaction():
     repl = "NAME"
     redacted_name = account_service.apply_name_redaction(name, repl)
     assert redacted_name == "Reply-NAME-NAME-NAME.pdf"
+
+
+@pytest.mark.django_db
+def test_unicode_name_redaction():
+    user = User.objects.create(first_name="Älex", last_name="Eğçamplé", private=True)
+    account_service = AccountService(user)
+    name = "reply-alex-egcample.pdf"
+    repl = "NAME"
+    redacted_name = account_service.apply_name_redaction(name, repl, unicode=False)
+    assert redacted_name == "reply-NAME-NAME.pdf"

froide/foirequest/services.py

@@ -527,7 +527,9 @@ def add_attachments(self, foirequest, message, attachments):
 
             # Translators: replacement for person name in filename
             repl = str(_("NAME"))
-            att.name = account_service.apply_name_redaction(att.name, repl)
+            att.name = account_service.apply_name_redaction(
+                att.name, repl, unicode=False
+            )
             att.name = re.sub(r"[^A-Za-z0-9_\.\-]", "", att.name)
             att.name = att.name[:250]