Age | Commit message (Collapse) | Author |
|
|
|
|
|
|
|
|
|
Merge to latest commit.
|
|
This is to allow connecting to before-destruction of one specific room, rather than any room under a connection (for which Connection::aboutToDeleteRoom() still exists).
|
|
It's borrowed from Qt; namespacing basically forced client writers to put "using namespace QMatrixClient" before using qAsConst.
|
|
Doxygen/Qt style is preferred from now on.
|
|
|
|
|
|
The Spec wasn't entirely consistent on this until recently but floats actually are used in the wild, rather than strings.
|
|
|
|
|
|
Closes #240.
|
|
|
|
|
|
Breaking changes:
* guest_can_join is no more a thing - neither in Connection::createRoom, nor even in CreateRoomJob (it turned out that Synapse didn't really process this flag);
* LoginJob has changed its list of arguments. If you use Connection to do logins (and you really should), you shouldn't be affected.
* GetPublicRoomsJob now returns PublicRoomsResponse instead of providing all the response parts within the job
Watch other changes in the diff.
|
|
Also: general cleanup.
|
|
Settings::get() and SettingsGroup::get() allow concise and efficient
type-specific interface to value(); instead of wrapping your
defaultValue into a QVariant, just pass it to get() instead of value().
|
|
|
|
|
|
|
|
Wasn't used in the code in the end.
|
|
|
|
|
|
Speeds up lookup of user(s) in a direct chat room.
Also: "The other one's" avatar is used to set the avatar of direct
chats only, not any room with 2 participants.
|
|
Deletion of Invite rooms was a clear bug; as for left rooms, it makes sense to keep
them in direct chat maps because they may be re-joined later on.
|
|
|
|
Closes #230.
|
|
|
|
Slots are Q_INVOKABLE by definition.
|
|
|
|
|
|
Not that it felt right but Riot does it and so should we. Closes #220.
|
|
Redaction events are now processed before, not after/along with normal
events. This complicates logistics a bit but makes sure to-be-redacted
events get redacted even before the client is notified about new
arrivals. Also, redaction events are unconditionally out of the timeline
for the moment (the non-starter KEEP_REDACTIONS_IN_TIMELINE macro is
gone). This reinstates #196 for now, which will be addressed separately.
|
|
|
|
Closes #227.
|
|
|
|
|
|
|
|
|
|
* TimelineItem and a newly introduced PendingEventItem are now
inheriting from the common EventItemBase class
* PendingEventItem has its own status and annotation, serving to track
transition of the item through pending states
|
|
...and therefore deprecated for use in clients _yet_.
|
|
|
|
|
|
It's present on the vast minority of events so better be embedded into
JSON instead.
|
|
It's not always a JSON object; it's either a SetTweakAction object or a string. (Unfortunately, QVariant isn't a template type and there's no easy way to put Q_DECLARE_METATYPE(SetTweakAction) in a Mustache template, so adding one is left as an exercise to the user now.)
|
|
Most of the actual change was in the API definition files (thanks to @anoadragon453), plus a converter from a QJsonObject map to the URL query string.
|
|
See https://github.com/matrix-org/matrix-doc/pull/1457 for the background.
|
|
This is at least some (actually, not even that bad) identification of a message. Ideally it would probably be to return some handler that would allow to track the end-to-end status of the event - from getting sent to landing in the timeline. Right now the experience is crippled - transactionId always exists but only works for message events and Room has no way to give the event status by transactionId. pendingEvent* signals are somewhat helping, though.
|