summary refs log tree commit diff
path: root/synapse/rest
diff options
context:
space:
mode:
Diffstat (limited to 'synapse/rest')
-rw-r--r--synapse/rest/client/v1/login.py13
-rw-r--r--synapse/rest/client/v2_alpha/auth.py42
-rw-r--r--synapse/rest/consent/consent_resource.py15
-rw-r--r--synapse/rest/media/v1/_base.py122
-rw-r--r--synapse/rest/media/v1/media_repository.py48
-rw-r--r--synapse/rest/media/v1/preview_url_resource.py61
6 files changed, 169 insertions, 132 deletions
diff --git a/synapse/rest/client/v1/login.py b/synapse/rest/client/v1/login.py
index 0010699d31..f6b4a85e40 100644
--- a/synapse/rest/client/v1/login.py
+++ b/synapse/rest/client/v1/login.py
@@ -27,7 +27,7 @@ from twisted.web.client import PartialDownloadError
 
 from synapse.api.errors import Codes, LoginError, SynapseError
 from synapse.http.server import finish_request
-from synapse.http.servlet import parse_json_object_from_request
+from synapse.http.servlet import RestServlet, parse_json_object_from_request
 from synapse.types import UserID
 from synapse.util.msisdn import phone_number_to_msisdn
 
@@ -83,6 +83,7 @@ class LoginRestServlet(ClientV1RestServlet):
     PATTERNS = client_path_patterns("/login$")
     SAML2_TYPE = "m.login.saml2"
     CAS_TYPE = "m.login.cas"
+    SSO_TYPE = "m.login.sso"
     TOKEN_TYPE = "m.login.token"
     JWT_TYPE = "m.login.jwt"
 
@@ -105,6 +106,10 @@ class LoginRestServlet(ClientV1RestServlet):
         if self.saml2_enabled:
             flows.append({"type": LoginRestServlet.SAML2_TYPE})
         if self.cas_enabled:
+            flows.append({"type": LoginRestServlet.SSO_TYPE})
+
+            # we advertise CAS for backwards compat, though MSC1721 renamed it
+            # to SSO.
             flows.append({"type": LoginRestServlet.CAS_TYPE})
 
             # While its valid for us to advertise this login type generally,
@@ -384,11 +389,11 @@ class SAML2RestServlet(ClientV1RestServlet):
         defer.returnValue((200, {"status": "not_authenticated"}))
 
 
-class CasRedirectServlet(ClientV1RestServlet):
-    PATTERNS = client_path_patterns("/login/cas/redirect", releases=())
+class CasRedirectServlet(RestServlet):
+    PATTERNS = client_path_patterns("/login/(cas|sso)/redirect")
 
     def __init__(self, hs):
-        super(CasRedirectServlet, self).__init__(hs)
+        super(CasRedirectServlet, self).__init__()
         self.cas_server_url = hs.config.cas_server_url.encode('ascii')
         self.cas_service_url = hs.config.cas_service_url.encode('ascii')
 
diff --git a/synapse/rest/client/v2_alpha/auth.py b/synapse/rest/client/v2_alpha/auth.py
index a8d8ed6590..fa73bdf3a1 100644
--- a/synapse/rest/client/v2_alpha/auth.py
+++ b/synapse/rest/client/v2_alpha/auth.py
@@ -21,7 +21,7 @@ from synapse.api.constants import LoginType
 from synapse.api.errors import SynapseError
 from synapse.api.urls import CLIENT_V2_ALPHA_PREFIX
 from synapse.http.server import finish_request
-from synapse.http.servlet import RestServlet
+from synapse.http.servlet import RestServlet, parse_string
 
 from ._base import client_v2_patterns
 
@@ -131,16 +131,12 @@ class AuthRestServlet(RestServlet):
         self.auth_handler = hs.get_auth_handler()
         self.registration_handler = hs.get_handlers().registration_handler
 
-    @defer.inlineCallbacks
     def on_GET(self, request, stagetype):
-        yield
-        if stagetype == LoginType.RECAPTCHA:
-            if ('session' not in request.args or
-                    len(request.args['session']) == 0):
-                raise SynapseError(400, "No session supplied")
-
-            session = request.args["session"][0]
+        session = parse_string(request, "session")
+        if not session:
+            raise SynapseError(400, "No session supplied")
 
+        if stagetype == LoginType.RECAPTCHA:
             html = RECAPTCHA_TEMPLATE % {
                 'session': session,
                 'myurl': "%s/auth/%s/fallback/web" % (
@@ -155,13 +151,11 @@ class AuthRestServlet(RestServlet):
 
             request.write(html_bytes)
             finish_request(request)
-            defer.returnValue(None)
+            return None
         elif stagetype == LoginType.TERMS:
-            session = request.args['session'][0]
-
             html = TERMS_TEMPLATE % {
                 'session': session,
-                'terms_url': "%s/_matrix/consent?v=%s" % (
+                'terms_url': "%s_matrix/consent?v=%s" % (
                     self.hs.config.public_baseurl,
                     self.hs.config.user_consent_version,
                 ),
@@ -176,25 +170,25 @@ class AuthRestServlet(RestServlet):
 
             request.write(html_bytes)
             finish_request(request)
-            defer.returnValue(None)
+            return None
         else:
             raise SynapseError(404, "Unknown auth stage type")
 
     @defer.inlineCallbacks
     def on_POST(self, request, stagetype):
-        yield
+
+        session = parse_string(request, "session")
+        if not session:
+            raise SynapseError(400, "No session supplied")
+
         if stagetype == LoginType.RECAPTCHA:
-            if ('g-recaptcha-response' not in request.args or
-                    len(request.args['g-recaptcha-response'])) == 0:
-                raise SynapseError(400, "No captcha response supplied")
-            if ('session' not in request.args or
-                    len(request.args['session'])) == 0:
-                raise SynapseError(400, "No session supplied")
+            response = parse_string(request, "g-recaptcha-response")
 
-            session = request.args['session'][0]
+            if not response:
+                raise SynapseError(400, "No captcha response supplied")
 
             authdict = {
-                'response': request.args['g-recaptcha-response'][0],
+                'response': response,
                 'session': session,
             }
 
@@ -242,7 +236,7 @@ class AuthRestServlet(RestServlet):
             else:
                 html = TERMS_TEMPLATE % {
                     'session': session,
-                    'terms_url': "%s/_matrix/consent?v=%s" % (
+                    'terms_url': "%s_matrix/consent?v=%s" % (
                         self.hs.config.public_baseurl,
                         self.hs.config.user_consent_version,
                     ),
diff --git a/synapse/rest/consent/consent_resource.py b/synapse/rest/consent/consent_resource.py
index e0f7de5d5c..ad525b22e1 100644
--- a/synapse/rest/consent/consent_resource.py
+++ b/synapse/rest/consent/consent_resource.py
@@ -142,10 +142,10 @@ class ConsentResource(Resource):
         userhmac = None
         has_consented = False
         public_version = username == ""
-        if not public_version or not self.hs.config.user_consent_at_registration:
-            userhmac = parse_string(request, "h", required=True, encoding=None)
+        if not public_version:
+            userhmac_bytes = parse_string(request, "h", required=True, encoding=None)
 
-            self._check_hash(username, userhmac)
+            self._check_hash(username, userhmac_bytes)
 
             if username.startswith('@'):
                 qualified_user_id = username
@@ -155,13 +155,18 @@ class ConsentResource(Resource):
             u = yield self.store.get_user_by_id(qualified_user_id)
             if u is None:
                 raise NotFoundError("Unknown user")
+
             has_consented = u["consent_version"] == version
+            userhmac = userhmac_bytes.decode("ascii")
 
         try:
             self._render_template(
                 request, "%s.html" % (version,),
-                user=username, userhmac=userhmac, version=version,
-                has_consented=has_consented, public_version=public_version,
+                user=username,
+                userhmac=userhmac,
+                version=version,
+                has_consented=has_consented,
+                public_version=public_version,
             )
         except TemplateNotFound:
             raise NotFoundError("Unknown policy version")
diff --git a/synapse/rest/media/v1/_base.py b/synapse/rest/media/v1/_base.py
index 76e479afa3..efe42a429d 100644
--- a/synapse/rest/media/v1/_base.py
+++ b/synapse/rest/media/v1/_base.py
@@ -16,6 +16,7 @@
 import logging
 import os
 
+from six import PY3
 from six.moves import urllib
 
 from twisted.internet import defer
@@ -48,26 +49,21 @@ def parse_media_id(request):
         return server_name, media_id, file_name
     except Exception:
         raise SynapseError(
-            404,
-            "Invalid media id token %r" % (request.postpath,),
-            Codes.UNKNOWN,
+            404, "Invalid media id token %r" % (request.postpath,), Codes.UNKNOWN
         )
 
 
 def respond_404(request):
     respond_with_json(
-        request, 404,
-        cs_error(
-            "Not found %r" % (request.postpath,),
-            code=Codes.NOT_FOUND,
-        ),
-        send_cors=True
+        request,
+        404,
+        cs_error("Not found %r" % (request.postpath,), code=Codes.NOT_FOUND),
+        send_cors=True,
     )
 
 
 @defer.inlineCallbacks
-def respond_with_file(request, media_type, file_path,
-                      file_size=None, upload_name=None):
+def respond_with_file(request, media_type, file_path, file_size=None, upload_name=None):
     logger.debug("Responding with %r", file_path)
 
     if os.path.isfile(file_path):
@@ -97,31 +93,26 @@ def add_file_headers(request, media_type, file_size, upload_name):
         file_size (int): Size in bytes of the media, if known.
         upload_name (str): The name of the requested file, if any.
     """
+
     def _quote(x):
         return urllib.parse.quote(x.encode("utf-8"))
 
     request.setHeader(b"Content-Type", media_type.encode("UTF-8"))
     if upload_name:
         if is_ascii(upload_name):
-            disposition = ("inline; filename=%s" % (_quote(upload_name),)).encode("ascii")
+            disposition = "inline; filename=%s" % (_quote(upload_name),)
         else:
-            disposition = (
-                "inline; filename*=utf-8''%s" % (_quote(upload_name),)).encode("ascii")
+            disposition = "inline; filename*=utf-8''%s" % (_quote(upload_name),)
 
-        request.setHeader(b"Content-Disposition", disposition)
+        request.setHeader(b"Content-Disposition", disposition.encode('ascii'))
 
     # cache for at least a day.
     # XXX: we might want to turn this off for data we don't want to
     # recommend caching as it's sensitive or private - or at least
     # select private. don't bother setting Expires as all our
     # clients are smart enough to be happy with Cache-Control
-    request.setHeader(
-        b"Cache-Control", b"public,max-age=86400,s-maxage=86400"
-    )
-
-    request.setHeader(
-        b"Content-Length", b"%d" % (file_size,)
-    )
+    request.setHeader(b"Cache-Control", b"public,max-age=86400,s-maxage=86400")
+    request.setHeader(b"Content-Length", b"%d" % (file_size,))
 
 
 @defer.inlineCallbacks
@@ -153,6 +144,7 @@ class Responder(object):
     Responder is a context manager which *must* be used, so that any resources
     held can be cleaned up.
     """
+
     def write_to_consumer(self, consumer):
         """Stream response into consumer
 
@@ -186,9 +178,18 @@ class FileInfo(object):
         thumbnail_method (str)
         thumbnail_type (str): Content type of thumbnail, e.g. image/png
     """
-    def __init__(self, server_name, file_id, url_cache=False,
-                 thumbnail=False, thumbnail_width=None, thumbnail_height=None,
-                 thumbnail_method=None, thumbnail_type=None):
+
+    def __init__(
+        self,
+        server_name,
+        file_id,
+        url_cache=False,
+        thumbnail=False,
+        thumbnail_width=None,
+        thumbnail_height=None,
+        thumbnail_method=None,
+        thumbnail_type=None,
+    ):
         self.server_name = server_name
         self.file_id = file_id
         self.url_cache = url_cache
@@ -197,3 +198,74 @@ class FileInfo(object):
         self.thumbnail_height = thumbnail_height
         self.thumbnail_method = thumbnail_method
         self.thumbnail_type = thumbnail_type
+
+
+def get_filename_from_headers(headers):
+    """
+    Get the filename of the downloaded file by inspecting the
+    Content-Disposition HTTP header.
+
+    Args:
+        headers (twisted.web.http_headers.Headers): The HTTP
+            request headers.
+
+    Returns:
+        A Unicode string of the filename, or None.
+    """
+    content_disposition = headers.get(b"Content-Disposition", [b''])
+
+    # No header, bail out.
+    if not content_disposition[0]:
+        return
+
+    # dict of unicode: bytes, corresponding to the key value sections of the
+    # Content-Disposition header.
+    params = {}
+    parts = content_disposition[0].split(b";")
+    for i in parts:
+        # Split into key-value pairs, if able
+        # We don't care about things like `inline`, so throw it out
+        if b"=" not in i:
+            continue
+
+        key, value = i.strip().split(b"=")
+        params[key.decode('ascii')] = value
+
+    upload_name = None
+
+    # First check if there is a valid UTF-8 filename
+    upload_name_utf8 = params.get("filename*", None)
+    if upload_name_utf8:
+        if upload_name_utf8.lower().startswith(b"utf-8''"):
+            upload_name_utf8 = upload_name_utf8[7:]
+            # We have a filename*= section. This MUST be ASCII, and any UTF-8
+            # bytes are %-quoted.
+            if PY3:
+                try:
+                    # Once it is decoded, we can then unquote the %-encoded
+                    # parts strictly into a unicode string.
+                    upload_name = urllib.parse.unquote(
+                        upload_name_utf8.decode('ascii'), errors="strict"
+                    )
+                except UnicodeDecodeError:
+                    # Incorrect UTF-8.
+                    pass
+            else:
+                # On Python 2, we first unquote the %-encoded parts and then
+                # decode it strictly using UTF-8.
+                try:
+                    upload_name = urllib.parse.unquote(upload_name_utf8).decode('utf8')
+                except UnicodeDecodeError:
+                    pass
+
+    # If there isn't check for an ascii name.
+    if not upload_name:
+        upload_name_ascii = params.get("filename", None)
+        if upload_name_ascii and is_ascii(upload_name_ascii):
+            # Make sure there's no %-quoted bytes. If there is, reject it as
+            # non-valid ASCII.
+            if b"%" not in upload_name_ascii:
+                upload_name = upload_name_ascii.decode('ascii')
+
+    # This may be None here, indicating we did not find a matching name.
+    return upload_name
diff --git a/synapse/rest/media/v1/media_repository.py b/synapse/rest/media/v1/media_repository.py
index d6c5f07af0..e117836e9a 100644
--- a/synapse/rest/media/v1/media_repository.py
+++ b/synapse/rest/media/v1/media_repository.py
@@ -14,14 +14,12 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-import cgi
 import errno
 import logging
 import os
 import shutil
 
-from six import PY3, iteritems
-from six.moves.urllib import parse as urlparse
+from six import iteritems
 
 import twisted.internet.error
 import twisted.web.http
@@ -34,14 +32,18 @@ from synapse.api.errors import (
     NotFoundError,
     SynapseError,
 )
-from synapse.http.matrixfederationclient import MatrixFederationHttpClient
 from synapse.metrics.background_process_metrics import run_as_background_process
 from synapse.util import logcontext
 from synapse.util.async_helpers import Linearizer
 from synapse.util.retryutils import NotRetryingDestination
-from synapse.util.stringutils import is_ascii, random_string
+from synapse.util.stringutils import random_string
 
-from ._base import FileInfo, respond_404, respond_with_responder
+from ._base import (
+    FileInfo,
+    get_filename_from_headers,
+    respond_404,
+    respond_with_responder,
+)
 from .config_resource import MediaConfigResource
 from .download_resource import DownloadResource
 from .filepath import MediaFilePaths
@@ -62,7 +64,7 @@ class MediaRepository(object):
     def __init__(self, hs):
         self.hs = hs
         self.auth = hs.get_auth()
-        self.client = MatrixFederationHttpClient(hs)
+        self.client = hs.get_http_client()
         self.clock = hs.get_clock()
         self.server_name = hs.hostname
         self.store = hs.get_datastore()
@@ -397,39 +399,9 @@ class MediaRepository(object):
             yield finish()
 
         media_type = headers[b"Content-Type"][0].decode('ascii')
-
+        upload_name = get_filename_from_headers(headers)
         time_now_ms = self.clock.time_msec()
 
-        content_disposition = headers.get(b"Content-Disposition", None)
-        if content_disposition:
-            _, params = cgi.parse_header(content_disposition[0].decode('ascii'),)
-            upload_name = None
-
-            # First check if there is a valid UTF-8 filename
-            upload_name_utf8 = params.get("filename*", None)
-            if upload_name_utf8:
-                if upload_name_utf8.lower().startswith("utf-8''"):
-                    upload_name = upload_name_utf8[7:]
-
-            # If there isn't check for an ascii name.
-            if not upload_name:
-                upload_name_ascii = params.get("filename", None)
-                if upload_name_ascii and is_ascii(upload_name_ascii):
-                    upload_name = upload_name_ascii
-
-            if upload_name:
-                if PY3:
-                    upload_name = urlparse.unquote(upload_name)
-                else:
-                    upload_name = urlparse.unquote(upload_name.encode('ascii'))
-                try:
-                    if isinstance(upload_name, bytes):
-                        upload_name = upload_name.decode("utf-8")
-                except UnicodeDecodeError:
-                    upload_name = None
-        else:
-            upload_name = None
-
         logger.info("Stored remote media in file %r", fname)
 
         yield self.store.store_cached_remote_media(
diff --git a/synapse/rest/media/v1/preview_url_resource.py b/synapse/rest/media/v1/preview_url_resource.py
index 91d1dafe64..d0ecf241b6 100644
--- a/synapse/rest/media/v1/preview_url_resource.py
+++ b/synapse/rest/media/v1/preview_url_resource.py
@@ -13,7 +13,6 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-import cgi
 import datetime
 import errno
 import fnmatch
@@ -44,15 +43,19 @@ from synapse.http.server import (
 )
 from synapse.http.servlet import parse_integer, parse_string
 from synapse.metrics.background_process_metrics import run_as_background_process
+from synapse.rest.media.v1._base import get_filename_from_headers
 from synapse.util.async_helpers import ObservableDeferred
 from synapse.util.caches.expiringcache import ExpiringCache
 from synapse.util.logcontext import make_deferred_yieldable, run_in_background
-from synapse.util.stringutils import is_ascii, random_string
+from synapse.util.stringutils import random_string
 
 from ._base import FileInfo
 
 logger = logging.getLogger(__name__)
 
+_charset_match = re.compile(br"<\s*meta[^>]*charset\s*=\s*([a-z0-9-]+)", flags=re.I)
+_content_type_match = re.compile(r'.*; *charset="?(.*?)"?(;|$)', flags=re.I)
+
 
 class PreviewUrlResource(Resource):
     isLeaf = True
@@ -223,15 +226,25 @@ class PreviewUrlResource(Resource):
             with open(media_info['filename'], 'rb') as file:
                 body = file.read()
 
-            # clobber the encoding from the content-type, or default to utf-8
-            # XXX: this overrides any <meta/> or XML charset headers in the body
-            # which may pose problems, but so far seems to work okay.
-            match = re.match(
-                r'.*; *charset="?(.*?)"?(;|$)',
-                media_info['media_type'],
-                re.I
-            )
-            encoding = match.group(1) if match else "utf-8"
+            encoding = None
+
+            # Let's try and figure out if it has an encoding set in a meta tag.
+            # Limit it to the first 1kb, since it ought to be in the meta tags
+            # at the top.
+            match = _charset_match.search(body[:1000])
+
+            # If we find a match, it should take precedence over the
+            # Content-Type header, so set it here.
+            if match:
+                encoding = match.group(1).decode('ascii')
+
+            # If we don't find a match, we'll look at the HTTP Content-Type, and
+            # if that doesn't exist, we'll fall back to UTF-8.
+            if not encoding:
+                match = _content_type_match.match(
+                    media_info['media_type']
+                )
+                encoding = match.group(1) if match else "utf-8"
 
             og = decode_and_calc_og(body, media_info['uri'], encoding)
 
@@ -323,31 +336,7 @@ class PreviewUrlResource(Resource):
                 media_type = "application/octet-stream"
             time_now_ms = self.clock.time_msec()
 
-            content_disposition = headers.get(b"Content-Disposition", None)
-            if content_disposition:
-                _, params = cgi.parse_header(content_disposition[0],)
-                download_name = None
-
-                # First check if there is a valid UTF-8 filename
-                download_name_utf8 = params.get("filename*", None)
-                if download_name_utf8:
-                    if download_name_utf8.lower().startswith("utf-8''"):
-                        download_name = download_name_utf8[7:]
-
-                # If there isn't check for an ascii name.
-                if not download_name:
-                    download_name_ascii = params.get("filename", None)
-                    if download_name_ascii and is_ascii(download_name_ascii):
-                        download_name = download_name_ascii
-
-                if download_name:
-                    download_name = urlparse.unquote(download_name)
-                    try:
-                        download_name = download_name.decode("utf-8")
-                    except UnicodeDecodeError:
-                        download_name = None
-            else:
-                download_name = None
+            download_name = get_filename_from_headers(headers)
 
             yield self.store.store_local_media(
                 media_id=file_id,