Using URL component port in url_preview_url_blacklist leads to TypeError

[lxp]

Description

If url_preview_url_blacklist uses the port URL component an TypeError is thrown.

Steps to reproduce

• Use an url_preview_url_blacklist with the URL component port in your homeserver.yaml.

url_preview_url_blacklist:
  # blacklist any URL with an explicit port
  - port: '*'

• When an URL preview for an URL explicit port set is requested, the following exception is logged:

2022-03-29 18:49:55,029 - synapse.http.server - 100 - ERROR - GET-685 - Failed handle request via 'PreviewUrlResource': <XForwardedForRequest at 0x7fa7fb6a1d00 method='GET' uri='/_matrix/media/r0/preview_url?url=https%3A%2F%2Fgoogle.com%3A1234%2F&ts=1648572540000' clientproto='HTTP/1.0' site='8008'>
Traceback (most recent call last):
File "/usr/local/lib/python3.9/site-packages/synapse/http/server.py", line 268, in _async_render_wrapper
callback_return = await self._async_render(request)
File "/usr/local/lib/python3.9/site-packages/synapse/http/server.py", line 296, in _async_render
callback_return = await raw_callback_return
File "/usr/local/lib/python3.9/site-packages/synapse/rest/media/v1/preview_url_resource.py", line 212, in _async_render_GET
if not fnmatch.fnmatch(getattr(url_tuple, attrib), pattern):
File "/usr/local/lib/python3.9/fnmatch.py", line 40, in fnmatch
name = os.path.normcase(name)
File "/usr/local/lib/python3.9/posixpath.py", line 54, in normcase
return os.fspath(s)
TypeError: expected str, bytes or os.PathLike object, not int

• The same also applies to a regex configuration leading to a slightly different exception:

url_preview_url_blacklist:
  # blacklist any URL with an explicit port
  - port: '^.*$'

2022-03-29 18:52:00,582 - synapse.http.server - 100 - ERROR - GET-9 - Failed handle request via 'PreviewUrlResource': <XForwardedForRequest at 0x7f33e13238b0 method='GET' uri='/_matrix/media/r0/preview_url?url=https%3A%2F%2Fgoogle.com%3A1234%2F&ts=1648572720000' clientproto='HTTP/1.0' site='8008'>
Traceback (most recent call last):
File "/usr/local/lib/python3.9/site-packages/synapse/http/server.py", line 268, in _async_render_wrapper
callback_return = await self._async_render(request)
File "/usr/local/lib/python3.9/site-packages/synapse/http/server.py", line 296, in _async_render
callback_return = await raw_callback_return
File "/usr/local/lib/python3.9/site-packages/synapse/rest/media/v1/preview_url_resource.py", line 208, in _async_render_GET
if not re.match(pattern, getattr(url_tuple, attrib)):
File "/usr/local/lib/python3.9/re.py", line 191, in match
return _compile(pattern, flags).match(string)
TypeError: expected string or bytes-like object

• The problem seems to appear because urlsplit returns an int for the port but it is not converted appropriately before applying the matching function.

Version information

• Homeserver: Not matrix.org

• Version: 1.55.2

• Install method: Docker

• Platform: Container

[babolivier]

The problem seems to appear because urlsplit returns an int for the port but it is not converted appropriately before applying the matching function.

Yes I believe your diagnostic is correct here.