Paul "LeoNerd" Evans
5583e29513
Report process open filehandles in metrics
2015-04-01 19:15:23 +01:00
David Baker
c5bf0343e8
Explain how I justified to myself making JsonResource not always send JSON.
2015-04-01 15:13:14 +01:00
David Baker
e9c908ebc0
Completely replace fallback auth for C/S V2:
...
* Now only the auth part goes to fallback, not the whole operation
* Auth fallback is a normal API endpoint, not a static page
* Params like the recaptcha pubkey can just live in the config
Involves a little engineering on JsonResource so its servlets aren't always forced to return JSON. I should document this more, in fact I'll do that now.
2015-04-01 15:05:30 +01:00
Kegan Dougal
813e54bd5b
Fix more AS sender ID thinkos.
...
Specifically, the ASes own user ID wasn't being treated as 'exclusive' so
a human could nab it. Also, the HS would needlessly send user queries to the
AS for its own user ID.
2015-04-01 14:05:24 +01:00
Kegsay
80a620a83a
Merge pull request #116 from matrix-org/application-services-registration-script
...
Application services registration changes (PR #116 )
2015-04-01 13:51:15 +01:00
Kegan Dougal
9fa8bda099
Merge branch 'develop' into application-services-registration-script
2015-04-01 10:19:17 +01:00
David Baker
f129ee1e18
Make docs a bit more true
2015-03-31 18:25:10 +01:00
Kegan Dougal
09cbff174a
Fix thinko whereby events *for the AS specifically* were not passed on.
...
This was caused by not explicitly checking the service.sender field. This
has now been fixed and a regression test has been added.
2015-03-31 16:44:45 +01:00
David Baker
d18e7779ca
Grammar and deduplication
2015-03-31 14:40:02 +01:00
Kegan Dougal
5e88a09a42
Add same user_id char checks as registration.
2015-03-31 14:00:25 +01:00
Kegan Dougal
cf1fa59f4b
Use a sender localpart instead of a user ID.
...
Form the user ID at runtime instead, This gives less room for error in AS
config files since they cannot specify the domain of another HS.
2015-03-31 13:48:03 +01:00
Kegan Dougal
3470cb36a8
Pyflakes
2015-03-31 13:03:31 +01:00
Kegan Dougal
c217504949
Edit SQL schema to use string IDs not ints. Use token as ID. Update tests.
2015-03-31 12:07:56 +01:00
Kegan Dougal
b59aa74556
Fix tests and missing returns on deferreds.
2015-03-31 11:35:45 +01:00
Kegan Dougal
d33ae65efc
Remove more reg/unreg methods. Read config not database for cache.
2015-03-31 11:00:00 +01:00
David Baker
9f642a93ec
pep8
2015-03-31 09:50:44 +01:00
Kegan Dougal
e7887e37a8
Remove appservice REST servlets
2015-03-31 09:32:40 +01:00
Kegan Dougal
af853a4cdb
Add AppServiceConfig
2015-03-31 09:22:31 +01:00
David Baker
4891c4ff72
Update CAPTCHA_SETUP (it continues to ignore fallback, but I guess I should fix it so that doesn't need the key in two different places)
2015-03-30 18:27:42 +01:00
David Baker
46183cc69f
Add original, unmodified CAPTCHA-SETUP from the webclient repo before modifying (captcha setup is now purely on the HS).
2015-03-30 18:18:19 +01:00
David Baker
59bf16eddc
New registration for C/S API v2. Only ReCAPTCHA working currently.
2015-03-30 18:13:10 +01:00
Erik Johnston
9a506a191a
Add note in changelog about change in config option names
2015-03-30 17:24:09 +01:00
Matthew Hodgson
8675ea03de
actually tell users /how/ to turn on registration
2015-03-30 12:05:38 -04:00
Matthew Hodgson
8366fde82f
turn --disable-registration into --enable-registration, given the default is for registration to be disabled by default now. this is backwards incompatible by removing the old --disable-registration arg, but makes for a much more intuitive arg
2015-03-30 12:01:09 -04:00
Paul "LeoNerd" Evans
3e420aebd8
Revert "Add another @cached wrapper, this time on get_presence_state()"
...
This reverts commit ff1fa0fbf8
2015-03-27 16:16:58 +00:00
Paul "LeoNerd" Evans
ff1fa0fbf8
Add another @cached wrapper, this time on get_presence_state()
2015-03-27 15:57:16 +00:00
Erik Johnston
abcd03af02
Merge pull request #115 from matrix-org/allow_registration_for_federation_demo
...
Allow registration in the HSes federation demo
2015-03-27 10:14:29 +00:00
manuroe
5116946ae9
Allow registration in the HSes federation demo
2015-03-27 11:10:52 +01:00
David Baker
6f4f7e4e22
pep8
2015-03-26 14:12:06 +00:00
David Baker
a32e876ef4
Delete pushers when changing password
2015-03-26 13:40:16 +00:00
Paul "LeoNerd" Evans
a198894bf7
Appease pep8
2015-03-26 11:53:58 +00:00
Kegsay
5b999e206e
Merge pull request #106 from matrix-org/application-services-txn-reliability
...
Application services transaction reliability (PR #106 )
2015-03-26 10:30:47 +00:00
Kegan Dougal
32206dde3f
Fixes from PR comments
2015-03-26 10:11:52 +00:00
Kegan Dougal
4edcbcee3b
Merge branch 'develop' into application-services-txn-reliability
...
Conflicts:
synapse/storage/__init__.py
2015-03-26 10:07:59 +00:00
Paul "LeoNerd" Evans
953e40f9dc
Implement the main getEvent cache using Cache() instead of a custom application of LruCache; also unify its two-level structure into just one
2015-03-25 19:12:16 +00:00
David Baker
df4c12c762
pep8 blank lines
2015-03-25 19:08:17 +00:00
David Baker
c1a256cc4c
Allow multiple pushers for a single app ID & pushkey, honouring the 'append' flag in the API.
2015-03-25 19:06:22 +00:00
Paul "LeoNerd" Evans
f173d40a32
Use FrozenEvent's reject_reason to decide whether to return it; don't include allow_rejected in the main getEvents cache key
2015-03-25 19:06:05 +00:00
Paul "LeoNerd" Evans
1b988b051b
Store the rejected reason in (Frozen)Event structs
2015-03-25 19:06:05 +00:00
Paul "LeoNerd" Evans
033a517feb
Indirect invalidations of _get_event_cache via a helper method to keep all uses of the cache lexically within one .py file
2015-03-25 19:06:05 +00:00
Paul "LeoNerd" Evans
9ba6487b3f
Allow a choice of LRU behaviour for Cache() by using LruCache() or OrderedDict()
2015-03-25 19:05:34 +00:00
Paul "LeoNerd" Evans
d6b3ea75d4
Implement the 'key in dict' test for LruCache()
2015-03-25 19:04:59 +00:00
Paul "LeoNerd" Evans
7ab9f91a60
Unit-test that Cache() key eviction is ordered
2015-03-25 18:50:43 +00:00
David Baker
ce2766d19c
Fix tests
2015-03-24 18:56:51 +00:00
David Baker
438a21c87b
Don't test exact equality of the list: as long as it has the fields we expect, that's just fine. I added the user_id (as in database pkey) and it broke: no point testing what that comes out as: it's determined by the db.
2015-03-24 18:21:54 +00:00
David Baker
9aa0224cdf
unused import
2015-03-24 17:25:59 +00:00
David Baker
c7023f2155
1) Pushers are now associated with an access token
...
2) Change places where we mean unauthenticated to 401, not 403, in C/S v2: hack so it stays as 403 in v1 because web client relies on it.
2015-03-24 17:24:15 +00:00
David Baker
d19e79ecc9
Make deleting other access tokens when you change your password actually work
2015-03-24 15:33:48 +00:00
Paul "LeoNerd" Evans
ed008e85a8
Reduce activity timer granularity to avoid too many quick updates (SYN-247)
2015-03-23 17:25:50 +00:00
David Baker
78adccfaf4
pep8 / pyflakes
2015-03-23 14:23:51 +00:00
