Merge branch 'master' of github.com:matrix-org/synapse into sql_refactor

Conflicts:
	tests/rest/test_presence.py
	tests/rest/test_rooms.py
	tests/utils.py

1 files changed, 157 insertions, 3 deletions

diff --git a/synapse/http/server.py b/synapse/http/server.py
index bad2738bde..c28d9a33f9 100644
--- a/synapse/http/server.py
+++ b/synapse/http/server.py
@@ -17,16 +17,23 @@
 from syutil.jsonutil import (
     encode_canonical_json, encode_pretty_printed_json
 )
-from synapse.api.errors import cs_exception, CodeMessageException
+from synapse.api.errors import (
+    cs_exception, SynapseError, CodeMessageException, Codes, cs_error
+)
+from synapse.util.stringutils import random_string
 
 from twisted.internet import defer, reactor
+from twisted.protocols.basic import FileSender
 from twisted.web import server, resource
 from twisted.web.server import NOT_DONE_YET
 from twisted.web.util import redirectTo
 
+import base64
 import collections
+import json
 import logging
-
+import os
+import re
 
 logger = logging.getLogger(__name__)
 
@@ -125,7 +132,11 @@ class JsonResource(HttpServer, resource.Resource):
                 {"error": "Unrecognized request"}
             )
         except CodeMessageException as e:
-            logger.exception(e)
+            if isinstance(e, SynapseError):
+                logger.error("%s SynapseError: %s - %s", request, e.code,
+                             e.msg)
+            else:
+                logger.exception(e)
             self._send_response(
                 request,
                 e.code,
@@ -140,6 +151,14 @@ class JsonResource(HttpServer, resource.Resource):
             )
 
     def _send_response(self, request, code, response_json_object):
+        # could alternatively use request.notifyFinish() and flip a flag when
+        # the Deferred fires, but since the flag is RIGHT THERE it seems like
+        # a waste.
+        if request._disconnected:
+            logger.warn(
+                "Not sending response to request %s, already disconnected.",
+                request)
+            return
 
         if not self._request_user_agent_is_curl(request):
             json_bytes = encode_canonical_json(response_json_object)
@@ -176,6 +195,141 @@ class RootRedirect(resource.Resource):
         return resource.Resource.getChild(self, name, request)
 
 
+class ContentRepoResource(resource.Resource):
+    """Provides file uploading and downloading.
+
+    Uploads are POSTed to wherever this Resource is linked to. This resource
+    returns a "content token" which can be used to GET this content again. The
+    token is typically a path, but it may not be. Tokens can expire, be one-time
+    uses, etc.
+
+    In this case, the token is a path to the file and contains 3 interesting
+    sections:
+        - User ID base64d (for namespacing content to each user)
+        - random 24 char string
+        - Content type base64d (so we can return it when clients GET it)
+
+    """
+    isLeaf = True
+
+    def __init__(self, directory, auth):
+        resource.Resource.__init__(self)
+        self.directory = directory
+        self.auth = auth
+
+        if not os.path.isdir(self.directory):
+            os.mkdir(self.directory)
+            logger.info("ContentRepoResource : Created %s directory.",
+                        self.directory)
+
+    @defer.inlineCallbacks
+    def map_request_to_name(self, request):
+        # auth the user
+        auth_user = yield self.auth.get_user_by_req(request)
+
+        # namespace all file uploads on the user
+        prefix = base64.urlsafe_b64encode(
+            auth_user.to_string()
+        ).replace('=', '')
+
+        # use a random string for the main portion
+        main_part = random_string(24)
+
+        # suffix with a file extension if we can make one. This is nice to
+        # provide a hint to clients on the file information. We will also reuse
+        # this info to spit back the content type to the client.
+        suffix = ""
+        if request.requestHeaders.hasHeader("Content-Type"):
+            content_type = request.requestHeaders.getRawHeaders(
+                "Content-Type")[0]
+            suffix = "." + base64.urlsafe_b64encode(content_type)
+            if (content_type.split("/")[0].lower() in
+                    ["image", "video", "audio"]):
+                file_ext = content_type.split("/")[-1]
+                # be a little paranoid and only allow a-z
+                file_ext = re.sub("[^a-z]", "", file_ext)
+                suffix += "." + file_ext
+
+        file_path = os.path.join(self.directory, prefix + main_part + suffix)
+        logger.info("User %s is uploading a file to path %s",
+                    auth_user.to_string(),
+                    file_path)
+
+        # keep trying to make a non-clashing file, with a sensible max attempts
+        attempts = 0
+        while os.path.exists(file_path):
+            main_part = random_string(24)
+            file_path = os.path.join(self.directory,
+                                     prefix + main_part + suffix)
+            attempts += 1
+            if attempts > 25:  # really? Really?
+                raise SynapseError(500, "Unable to create file.")
+
+        defer.returnValue(file_path)
+
+    def render_GET(self, request):
+        # no auth here on purpose, to allow anyone to view, even across home
+        # servers.
+
+        # TODO: A little crude here, we could do this better.
+        filename = request.path.split(self.directory + "/")[1]
+        # be paranoid
+        filename = re.sub("[^0-9A-z.-_]", "", filename)
+
+        file_path = self.directory + "/" + filename
+        if os.path.isfile(file_path):
+            # filename has the content type
+            base64_contentype = filename.split(".")[1]
+            content_type = base64.urlsafe_b64decode(base64_contentype)
+            logger.info("Sending file %s", file_path)
+            f = open(file_path, 'rb')
+            request.setHeader('Content-Type', content_type)
+            d = FileSender().beginFileTransfer(f, request)
+
+            # after the file has been sent, clean up and finish the request
+            def cbFinished(ignored):
+                f.close()
+                request.finish()
+            d.addCallback(cbFinished)
+        else:
+            respond_with_json_bytes(
+                request,
+                404,
+                json.dumps(cs_error("Not found", code=Codes.NOT_FOUND)),
+                send_cors=True)
+
+        return server.NOT_DONE_YET
+
+    def render_POST(self, request):
+        self._async_render(request)
+        return server.NOT_DONE_YET
+
+    @defer.inlineCallbacks
+    def _async_render(self, request):
+        try:
+            fname = yield self.map_request_to_name(request)
+
+            # TODO I have a suspcious feeling this is just going to block
+            with open(fname, "wb") as f:
+                f.write(request.content.read())
+
+            respond_with_json_bytes(request, 200,
+                                    json.dumps({"content_token": fname}),
+                                    send_cors=True)
+
+        except CodeMessageException as e:
+            logger.exception(e)
+            respond_with_json_bytes(request, e.code,
+                                    json.dumps(cs_exception(e)))
+        except Exception as e:
+            logger.error("Failed to store file: %s" % e)
+            respond_with_json_bytes(
+                request,
+                500,
+                json.dumps({"error": "Internal server error"}),
+                send_cors=True)
+
+
 def respond_with_json_bytes(request, code, json_bytes, send_cors=False):
     """Sends encoded JSON in response to the given request.