On a recent trip through a CPU profile taken while Signal Desktop
churned through a large backlog of messages, it was clear that
console.log was a major source of time spent, primarily the sort
operation required after every new entry is added to the Backbone
collection. So, three different techniques to combat this:
1) Reduce the maximum number of entries in the collection from 5k to 2k
2) No more logging of add/update/remove queue in MessageReceiver
3) No more log entries in Message.handleDataMessage main codepath
FREEBIE
* Fetch all conversations on startup of app, not on inbox load
A recent change to fetch conversations less didn't take into account all
that can happen in the app without the inbox loaded. That only happens
when the window is shown, and messages can come in with the app in the
background. In that case, the conversation wouldn't have been loaded
from the database, but would be saved to the database anyway, losing
data.
This change fetches all conversations as soon as the the data store is
ready for a fetch. It also introduces failsafe throws to ensure that
synchronous ConversationController accesses don't happen until the
initial fetch is complete. A new getUnsafe() method was required to
account for some of the model setup that happens during that initial
conversation fetch.
Fixes#1428
FREEBIE
* Fix tests: ConversationController.load() required before get()
FREEBIE
* Fetch conversations once, clean up ConversationController API
Race conditions around re-fetching have caused some problems recently,
so this removes the need to re-fetch conversations. They are fetched
once or saved once, and that is it. All interaction goes through the
ConversationController, which is the central source of truth.
We have two rules for Conversations:
1. If a conversation is in the ConversationController it doesn't need
to be fetched, but its initial fetch/save might be in progress. You
can wait for that fetch/save with conversation.initialPromise.
2. If a conversation is not already in the ConversationController, it's
not yet in the database. It needs to be added to the
ConversationController and saved to the database.
FREEBIE
* Remove Conversation.fetch() call in Message.handleDataMessage()
FREEBIE
* ConversationController.API cleanup: Fix two missing spots
FREEBIE
If the key has changed, saveIdentity will archive sibling sessions, but not the
session for the device it was called on. Therefore we have to archive that one
by hand.
Also switch from saving the identity of an OutgoingIdentityKeyError to just
triggering a profile fetch, mostly for consistency, simplicity, and DRYness.
// FREEBIE
We also show more errors than we used to in the MessageDetail screen
to help make it clear what is happening, and why the user would need to
re-send.
FREEBIE
Check for early read receipts for an incoming messages *after*
processing the expireTimer on that message. Then we can set
expirationStartTime appropriately if an early receipt is found.
Closes#950
// FREEBIE
This reverts commit a768b94471.
d2ddfc7 was enough to fix#989. Removing unregistered members from the
group (as opposed to silently ignorning them) creates greater potential
for getting out of sync with the member lists on other devices.
// FREEBIE
Previously, updateNumbers would throw an Error, so the whole group
update was discarded.
Signal-Android handles this the same way in
GroupMessageProcessor.handleGroupUpdate().
Closes#1056
Don't set lastMessage, let it update itself as needed, such as when
first rendering a conversation list item, and when its messages are
sent, received, or destroyed.
When initialized, or when expiration-related attributes change, expiring
messages will set timers to self-destruct. On self-destruct they trigger
'expired' events so that frontend listeners can clean up any collections
and views referencing them.
At startup, load all messages pending expiration so they can start their
timers even if they haven't been loaded in the frontend yet.
Todo: Remove expired conversation snippets from the left pane.
In some cases, we have already received a read receipt for an incoming
message by the time we go to create a notification about it. In this
case, we should skip the notification.
// FREEBIE
In the case of a double send (same message encrypted and sent twice due
to key conflict bug), we would mark the first instance read twice rather
than marking both instances read. Fix by searching for matching messages
that have not yet been marked read.
// FREEBIE
For messages that failed to send due to network errors, this change
allows retrying them directly from the main conversation view rather
than only from the message detail view.
// FREEBIE
Previously, if a message was sent in between the receive time of an
incoming message and the time it is actually added to the conversation's
message collection (which only occurs later after several async
callbacks), the incoming message would be inserted not-at-the-end of the
collection since it is ordered by receive time. This tricked the front
end into assuming the message was an older message instead of a new one.
Fixes#490
// FREEBIE
Remove individual messages from Notifications when marked read.
Previously this was only done from the conversation model when marking
the entire conversation as read.
Fixes#717
// FREEBIE
Don't save the change until we successfully process the message, but
make it first so that the user sees the error disappear when the new key
is accepted.
// FREEBIE
If the replay failed due to a bad mac or other decryption error for some
other reason we still want to clear the conflict. If it failed because
it's still in conflict then the newly returned error will reflect that
and be saved.
// FREEBIE
Group updates in which nothing change should still display 'Updated the
group'. Previously they would display as empty message bubbles. Fixed by
ensuring that the 'group_update' attribute is set on the model, even if
it is an empty object.
// FREEBIE
Previously, when processing a backlog of sync messages and their
delivery receipts, we would fail to mark some messages as delivered even
though we got a receipt. This was due to an async race condition between
saving a sync message and fetching it after the receipt arrives.
Fix by re-ordering idb requests such that we save the message first and
fetch it after.
Fixes#479
// FREEBIE
Previously, libtextsecure would send a sync message automatically
when appropriate. This fails if any recipient has a key conflict
or if our network connection fails mid-send.
Instead, when appropriate, return a the DataMessage encoded as an array
buffer for later syncing. This lets the application choose when to send
it, which we now do after any successful send to a recipient, rather
than after all recipients are successfully sent to.
Eventually we should move the DataMessage protobuf construction and
group sending logic to the application layer entirely, in which case
we wouldn't need libtextsecure to construct the sync message either.
Fixes#408
Only load the most recent messages when initially rendering a
conversation. Scrolling to the top of a message list loads older
messages.
This required some slight refactoring of how we insert message elements
into the dom. If the message is added to the end of the collection,
append it at the end. Otherwise, assume it is an older message and
prepend it.
When adding elements to the top, reset the scrollPosition to its
previous distance from scrollHeight. This keeps the current set of
elements fixed in the viewport.
// FREEBIE
Create a cleaner seperation between generating notifications
and updating frontend conversation views. The former is now
handled by `conversation.notify` while the latter is achieved
by triggering an event on the conversation model, which will
only be acted on if there are any views listening for it.
Additionally, instead of re-fetching the entire message history,
which is overkill, just add or update the new/modified message.
This will help speed up the newmessage event handler and also
help avoid unnecessary re-rendering when resolving key conflicts.
// FREEBIE
Follow up to ddd2e67eb5
but for incoming messages.
* Conflict state sometimes failed to be removed even though the
conflict was resolved.
* Messages failed to re-render after a conflict. We want to
re-render only the error state on outgoing messages, to avoid
flickering attachments. On incoming messages, we need to call
render to populate the message text, avatar, etc...
// FREEBIE
Previously, with a mix of text and media messages in conflict,
asynchronous callbacks aligned so as to fail to remove some of
the conflict objects on messages.
Fix by serializing conflict processing, but making sure to move
on through the message list even if some conflict resolutions fail.
Fixes#370
// FREEBIE
For non text messages (ex: media messages and group updates), the
lastMessage field was being populated with empty string, resulting in an
empty message preview in the conversation list. Instead, display 'Media
message' or 'Updated the group', etc...
// FREEBIE
