diff --git a/aw_server/csv_export.py b/aw_server/csv_export.py index ab5a46d5..721424a5 100644 --- a/aw_server/csv_export.py +++ b/aw_server/csv_export.py @@ -12,8 +12,10 @@ import json from datetime import timedelta, timezone from typing import Any, Dict, Iterable, Iterator, List +from urllib.parse import quote from aw_core.models import Event +from werkzeug.http import dump_options_header MAX_CSV_DATA_COLUMNS = 32 @@ -107,9 +109,12 @@ def events_to_csv(events: List[Event]) -> Iterator[str]: def content_disposition(filename: str) -> str: - """Quoted attachment filename with header-breaking characters replaced.""" + """WSGI-safe attachment with an ASCII fallback and RFC 5987 Unicode name.""" safe = "".join( "_" if (ord(c) < 32 or 127 <= ord(c) < 160 or c in '"\\;') else c for c in filename ) - return f'attachment; filename="{safe}"' + options = {"filename": "".join(c if c.isascii() else "_" for c in safe)} + if not safe.isascii(): + options["filename*"] = "UTF-8''" + quote(safe, safe="!#$&+-.^_`|~") + return dump_options_header("attachment", options) diff --git a/aw_server/rest.py b/aw_server/rest.py index 91acf045..3c3a6e8a 100644 --- a/aw_server/rest.py +++ b/aw_server/rest.py @@ -462,9 +462,7 @@ def get(self): payload = {"buckets": buckets_export} response = make_response(json.dumps(payload)) filename = "aw-buckets-export.json" - response.headers["Content-Disposition"] = "attachment; filename={}".format( - filename - ) + response.headers["Content-Disposition"] = content_disposition(filename) return response @@ -478,9 +476,7 @@ def get(self, bucket_id): payload = {"buckets": {bucket_export["id"]: bucket_export}} response = make_response(json.dumps(payload)) filename = "aw-bucket-export_{}.json".format(bucket_export["id"]) - response.headers["Content-Disposition"] = "attachment; filename={}".format( - filename - ) + response.headers["Content-Disposition"] = content_disposition(filename) return response diff --git a/tests/test_server.py b/tests/test_server.py index 28776dde..9eecbc70 100644 --- a/tests/test_server.py +++ b/tests/test_server.py @@ -3,8 +3,10 @@ import json import random from datetime import datetime, timedelta +from urllib.parse import quote import pytest +from werkzeug.http import parse_options_header @pytest.fixture() @@ -207,9 +209,9 @@ def test_export_bucket_csv(flask_client, bucket): r = flask_client.get(f"/api/0/buckets/{bucket}/export/csv") assert r.status_code == 200 assert r.mimetype == "text/csv" - assert ( - r.headers["Content-Disposition"] - == f'attachment; filename="aw-events-export-{bucket}.csv"' + assert parse_options_header(r.headers["Content-Disposition"]) == ( + "attachment", + {"filename": f"aw-events-export-{bucket}.csv"}, ) lines = r.get_data(as_text=True).split("\r\n") # Newest first, union of data keys, formula-looking values neutralized @@ -495,6 +497,84 @@ def test_export_bucket_csv_prefixes_colliding_data_keys(flask_client, bucket): } +@pytest.mark.parametrize("format", ["json", "csv"]) +@pytest.mark.parametrize( + "bucket_id,safe_id", + [ + ("plain", "plain"), + ("work;secondary", "work_secondary"), + ('quoted"name', "quoted_name"), + ("back\\slash", "back_slash"), + ("line\nbreak", "line_break"), + ("carriage\rreturn", "carriage_return"), + ("tab\tname", "tab_name"), + ("control\x01\x7f\x85", "control___"), + ("with spaces", "with spaces"), + ("café", "café"), + ("工作", "工作"), + ("percent%name", "percent%name"), + ], +) +def test_export_attachment_filename(flask_client, bucket_id, safe_id, format): + route = "/api/0/buckets/" + quote(bucket_id, safe="") + created = flask_client.post( + route, json={"client": "test", "type": "test", "hostname": "test"} + ) + assert created.status_code == 200 + try: + event = { + "timestamp": "2026-10-01T00:00:00Z", + "duration": 5, + "data": {"label": "unchanged café 工作"}, + } + inserted = flask_client.post(route + "/events", json=event) + assert inserted.status_code == 200 + suffix = "/export" if format == "json" else "/export/csv" + response = flask_client.get(route + suffix, buffered=True) + assert response.status_code == 200 + header = response.headers["Content-Disposition"] + # WSGI servers encode response header values as Latin-1, unlike the + # Flask test client. Require ASCII, including the Unicode fallback. + for key, value in response.headers.to_wsgi_list(): + key.encode("latin-1") + value.encode("latin-1") + header.encode("ascii") + assert all(32 <= ord(c) < 127 for c in header) + prefix = "aw-bucket-export_" if format == "json" else "aw-events-export-" + expected = f"{prefix}{safe_id}.{format}" + disposition, options = parse_options_header(header) + assert disposition == "attachment" + assert options == {"filename": expected} + fallback = parse_options_header(header.split("; filename*=", 1)[0])[1] + assert fallback["filename"] == "".join( + c if c.isascii() else "_" for c in expected + ) + if not expected.isascii(): + assert "filename*=UTF-8''" in header + if format == "json": + payload = json.loads(response.get_data()) + exported = payload["buckets"][bucket_id] + assert exported["id"] == bucket_id + assert exported["events"][0]["data"] == event["data"] + else: + header_row, row = csv.reader(io.StringIO(response.get_data(as_text=True))) + assert header_row == ["id", "timestamp", "duration", "label"] + assert row[-1] == event["data"]["label"] + finally: + deleted = flask_client.delete(route) + assert deleted.status_code == 200 + + +def test_export_all_attachment_filename(flask_client): + response = flask_client.get("/api/0/export") + assert response.status_code == 200 + assert parse_options_header(response.headers["Content-Disposition"]) == ( + "attachment", + {"filename": "aw-buckets-export.json"}, + ) + assert "buckets" in json.loads(response.get_data()) + + def test_csv_header_prefix_repeats_until_unique(): from aw_server.csv_export import _header_names