David Baker
8798bf42e6
Fix indexeddb logging
...
1. Fix double 'loaded' on sync data logging
2. Move the 'loaded' message into the bit where the data has
actually loaded rather than the promise try block.
3. Add '...' to the 'loading' messages so they're easier to tell
apart from the 'loaded' messages.
2018-03-09 10:16:32 +00:00
David Baker
5a23927e56
Move comment up
2018-03-09 10:09:36 +00:00
Matthew Hodgson
beafd597dd
ensure indexeddb workers are never double-connected
2018-03-09 02:18:19 +00:00
Matthew Hodgson
fbc43b0d58
stupid typo
2018-03-09 00:01:14 +00:00
David Baker
facfcf679d
DeviceList: bring save forward if necessary
...
If save is called with a delay that would want the save to happen
sooner then the save we currently have scheduled, cancel the
current save and schedule a new one for the sooner time.
2018-03-08 15:35:35 +00:00
Luke Barnard
68b230a78f
Add function to cancel and resend key request ( #624 )
2018-03-08 15:01:01 +00:00
David Baker
3d1fcc6f83
One day I'll learn to spell guaranteed
2018-03-08 14:26:48 +00:00
David Baker
a0578efeb9
fix tests
2018-03-08 13:52:48 +00:00
David Baker
727ad5755e
lint
2018-03-08 12:40:01 +00:00
David Baker
4f17352858
Don't do /keys/changes on incremental sync
...
Remove the call to /keys/changes when we do an incremental syn
where the old sync token doesn't match the one in the device list
store. To allow us to do this, always save the device list store
before saving the sync data, so we can safely assume the device
list store is at least as fresh as the sync token in the sync store.
Thread save functions through to allow this, add a delay parameter
so the sync can save the device list immediately and skip the wait,
and add a wantsSave() method so the sync can skip saving the device
list if the sync store isn't going to save anyway.
Fixes https://github.com/vector-im/riot-web/issues/6068
2018-03-08 12:33:08 +00:00
David Baker
0eb72122ce
Merge pull request #623 from matrix-org/dbkr/devicelist_is_too_dirty
...
Don't mark devicelist dirty unnecessarily
2018-03-07 17:55:38 +00:00
David Baker
66e2b3bb70
Don't mark devicelist dirty unnecessarily
...
This was marking the device list as dirty even when nothing had
actually changed, causing unnecessary saves.
2018-03-07 17:32:51 +00:00
David Baker
5f12d858eb
Keep a push processor and re-use it. ( #622 )
...
Because it does some nice caching stuff but that's no good if we
re-create a new one each time.
2018-03-06 19:01:30 +00:00
Luke Barnard
e258d6ca8d
Fix bug in crypto-store-backend; use oncomplete ( #620 )
...
instead of onsuccess for the txn to add an outgoing room key
request.
2018-03-02 12:01:21 +00:00
David Baker
2d25dedbcc
Merge pull request #619 from matrix-org/dbkr/cache_joined_member_count
...
Cache the joined member count for a room state
2018-02-26 16:56:26 +00:00
David Baker
bdf6fcb222
Fix tests
2018-02-26 16:53:17 +00:00
David Baker
7e1cea1ef6
Cache the joined member count for a room state
...
Pushrule evaluation needs the count of joined room members and
filtering the list for joined members takes a nontrivial amount
of time, but caching it is trivial, both code and memory wise.
2018-02-26 16:43:26 +00:00
Richard Lewis
d98c803b54
Merge pull request #618 from matrix-org/rxl881/stickers
...
Fix JS doc
2018-02-23 15:52:42 +00:00
Matthew Hodgson
95238466b5
Merge pull request #617 from matrix-org/dbkr/fix_push_actions_frozen_matrixevents
...
Precompute push actions for state events
2018-02-23 14:32:11 +00:00
Matthew Hodgson
71652043a0
improve doc
2018-02-23 14:31:51 +00:00
David Baker
c9cbaf254b
Precompute push actions for state events
...
State events get froze now, so we can't write to the push actions
cache after having done so: precompute the push actions to work
around this.
2018-02-23 11:06:22 +00:00
David Baker
f9cc5cbd33
Merge pull request #616 from matrix-org/luke/fix-unverified-device-blacklist
...
Fix bug where global "Never send to unverified..." is ignored
2018-02-23 10:11:59 +00:00
Luke Barnard
bcb9405793
Instead of emitting, rely on the promise
2018-02-22 18:41:18 +00:00
Luke Barnard
30cb6f196f
Fix overriding undefined per-room unverified devices setting
...
If the per-room setting for preventing sending keys to unverified
devices is `undefined`, it overrides the global setting (which
could be `true`).
2018-02-22 18:02:39 +00:00
Luke Barnard
856ef01632
Emit a crypto.initComplete once crypto is initialised
...
So that the app can call crypto-dependent functions at
the correct point in time.
2018-02-22 17:59:37 +00:00
David Baker
6f95554655
Merge pull request #615 from matrix-org/dbkr/intern_legacy_membership
...
Intern legacy top-level 'membership' field
2018-02-22 17:01:38 +00:00
David Baker
a72f915646
Intern legacy top-level 'membership' field
...
There is a legacy top-level 'membership' field on events. We were
interning the normal one in 'content', but not this one, so the
legacy field was still keeping a copy of the string 'join' / 'leave'
etc in memory for every member event.
This make the interning code know about this field.
2018-02-22 16:04:35 +00:00
Matthew Hodgson
94605417f6
Merge pull request #598 from matrix-org/t3chguy/implicit_rr_redaction
...
Don't synthesize RR for m.room.redaction as causes the RR to go missing.
2018-02-20 18:23:06 +00:00
David Baker
1e017df128
Merge pull request #613 from matrix-org/dbkr/dates_on_demand
...
Make Events create Dates on demand
2018-02-20 18:08:45 +00:00
David Baker
ec27bb5131
Merge pull request #612 from matrix-org/dbkr/dont_clone_events
...
Stop cloning events when adding to state
2018-02-20 18:08:35 +00:00
David Baker
9637fc098a
comment
2018-02-20 18:01:01 +00:00
David Baker
874020ced7
Make Events create Dates on demand
...
My test account had 37MB (shallow) of Date objects knocking around
in memory. This gets rid of them. I can't see any appreciable
difference in the time taken to switch rooms (where now we recreate
a bunch of Dates that previously would have been cached).
2018-02-20 17:49:25 +00:00
David Baker
a7beedcfb6
Unused imports
2018-02-20 17:11:40 +00:00
David Baker
4351c4dd6f
Merge pull request #611 from matrix-org/dbkr/use_initialisestate
...
De-dup code: use the initialiseState function
2018-02-20 17:10:11 +00:00
David Baker
1ddf7fb96c
Add XXX comment
2018-02-20 17:07:14 +00:00
David Baker
ec5cfe4ee9
Stop cloning events when adding to state
...
As comment hopefully explains.
On my test account:
Before: 394657 MatrixEvents, 53MB shallow size
After: 198863 MatrixEvents, 27MB shallow size
2018-02-20 16:45:20 +00:00
Matthew Hodgson
4fed2ea7bf
log event IDs of ignored calls
2018-02-20 12:38:12 +00:00
Matthew Hodgson
ae14cf4740
typo
2018-02-20 12:38:12 +00:00
David Baker
8aa68b3dc1
lint
2018-02-20 11:09:32 +00:00
David Baker
e810ee7750
Add test for fixed state misbehaviour
2018-02-20 11:06:33 +00:00
David Baker
9a08194597
Factor out calls to getLiveTimeline
2018-02-19 16:17:30 +00:00
David Baker
c77277b60c
rename variable
2018-02-19 16:07:28 +00:00
David Baker
21a324558f
Comments
2018-02-19 16:05:43 +00:00
David Baker
7a31751564
Merge pull request #610 from matrix-org/dbkr/sentinels_on_demand
...
Create sentinel members on-demand
2018-02-16 19:01:48 +00:00
David Baker
b11bacc2e2
Process state events before timeline events
...
Don't ignore them: its valid to send them in a non-limited sync,
they're state events that preceed the timeline events.
2018-02-16 18:52:56 +00:00
David Baker
8c02e7ba67
comment
2018-02-16 17:52:39 +00:00
David Baker
275eb8d434
De-dup code: use the initialiseState function
...
This should behave identically, but the code here appeared to be
identical to the code in initialiseState, so let's use it (it also
has an extra sanity check in there that we only init empty timelines).
2018-02-16 17:49:29 +00:00
David Baker
736d0df38d
Handle null userID
...
The tests inject messages with no sender, so for now let's maintain
behaviour as it was before.
2018-02-16 14:58:19 +00:00
David Baker
89d5d41015
Merge pull request #609 from matrix-org/dbkr/doc_sentinels
...
Some more doc on how sentinels work
2018-02-16 12:34:34 +00:00
David Baker
0e1444c84b
grammar
2018-02-16 11:57:22 +00:00
