aboutsummaryrefslogtreecommitdiff
AgeCommit message (Collapse)Author
2018-01-09BaseJob: change initial job status from NoError to PendingKitsune Ral
Otherwise isJobRunning() is of no great use.
2018-01-09Fix an occasional crash when resolving a server nameKitsune Ral
Closes #149.
2018-01-09Avatar: Use QPointer<> and isJobRunning()Kitsune Ral
2018-01-09Avatar: Mark get() as const operationKitsune Ral
With all the liberty that pimpl idiom gives it's easy to get away without proper const's; but let's be consistent :)
2018-01-09isJobRunning() facility functionKitsune Ral
2018-01-09Avatar: Use Connection::getThumbnail instead of callApi<>Kitsune Ral
2018-01-06Add a missing #includeKitsune Ral
std::bind belongs to <functional>, don't rely on implicitly including it from other standard headers.
2017-12-30BaseJob: do not suppress SSL errorsKitsune Ral
It is the application's responsibility to properly display the error and get confirmation from the user about it.
2017-12-30QMatrixClient::NetworkAccessManager (singleton Qt NAM that remembers ignored ↵Kitsune Ral
SSL errors) Closes #145.
2017-12-29Don't use generic lambdas in initialisersKitsune Ral
This seems to upset MSVC.
2017-12-29Fix building with Microsoft's standard libraryKitsune Ral
2017-12-29Allow to customize and connect to the used QNetworkAccessManager from clientsKitsune Ral
We don't really want to let the world know that we use the only instance of NAM; instead, we provide a point of customisation to whatever NAM is created in ConnectionData, in the form of a static customizeNetworkAccess() method that gets a function to run on a/the created NAM. This function can do additional configuration on NAM (such as setting a proxy factory, network configuration, or caching) and/or connect to its signals, such as sslErrors() and proxyAuthenticationRequired(). Closes #143.
2017-12-29Use std::unique_ptr instead of bare pointers; cleanup stale declarationsKitsune Ral
2017-12-28NetworkSettings: add setupApplicationProxy() accessible from QML; proxyPort ↵Kitsune Ral
is quint16 now setupApplicationProxy(), so that clients in QML could apply changes in proxy settings without going down to C++. quint16, because this is a type actually used by QNetworkProxy.
2017-12-28Switch from QPixmap to QImage; add convenience avatar() overloads to Room ↵Kitsune Ral
and User The switch is necessary because MediaThumbnailJob is supposed to return something that can be worked on in non-GUI threads (as is the case of QML image providers), and QPixmap is not supposed for usage out of the main thread.
2017-12-27Introduce StateEventBase - a non-template base for StateEvent<>Kitsune Ral
This will hold common logic for all state events, including the newly introduced repeatsState() that returns true when prev_content repeats content. This will be used to address QMatrixClient/Quaternion#245.
2017-12-27StateEvent<>: Look for prev_content in unsigned, not top-levelKitsune Ral
2017-12-26NetworkSettings: store proxy configurationKitsune Ral
No credentials, just type, host, and port.
2017-12-26Settings: factored out common declarations/definitions to macrosKitsune Ral
It cannot be made in templates because property and methods names are involved.
2017-12-25CMakeLists: Add -W to the mixKitsune Ral
2017-12-25qmake: Use QMAKE_CXX_FLAGS_WARN_ON rather than QMAKE_CXX_FLAGSKitsune Ral
Because the damn thing puts QMAKE_CXX_FLAGS before what's put by CONFIG, and CONFIG injects -W -Wall that cancels -Wno-unused-parameter.
2017-12-25Connection: change the attempt number propagated from BaseJobKitsune Ral
2017-12-25BaseJob: further minor code cleanupKitsune Ral
2017-12-25qmake: Use the compiler specified in the environment; revert to CONFIG += c++14Kitsune Ral
qmake doesn't follow standard CC and CXX environment variables and its CONFIG produces CXX_FLAGS added after QMAKE_CXX_FLAGS. Big thank you to The Qt Company :-|
2017-12-25qmake: Use -std=C++14, not -std=gnu++1y substituted for CONFIG += c++14Kitsune Ral
2017-12-25Polished warnings configurationKitsune Ral
The whole thing should go without or almost without warnings with GCC and with Clang (MSVC is another story and I don't care about it much).
2017-12-25Room: Remove C++14 code from the header fileKitsune Ral
This will provide some backwards-compatibility to clients that are not ready to move _their_ code to C++14 (at least, it will allow them to not add C++14 requirement to their makefiles as of yet).
2017-12-25BaseJob: consolidate job timeout configurationKitsune Ral
This prepares the backend to enable timeouts/retry intervals configurable from clients.
2017-12-25Connection: Be more accurate with lambda capturesKitsune Ral
2017-12-25Fix CIKitsune Ral
Old GCC was still invoked; qmake-based builds weren't switched to C++14.
2017-12-25Now really switching to the new toolchain: C++14, GCC/Clang 5, Qt 5.6Kitsune Ral
Also a bit of code tightening with some C++14 (but not only) things.
2017-12-25ReceiptEvent: use QVector instead of std::vectorKitsune Ral
Because we should practice what we preach in CONTRIBUTING.md.
2017-12-25CONTRIBUTING.md refreshed and slightly extendedKitsune Ral
2017-12-25Use pimpl for AvatarsKitsune Ral
2017-12-25Code cleanupKitsune Ral
2017-12-25Qt 5.6 or newer is required from now onKitsune Ral
2017-12-21Connection::logout: supply receiver to connect(job,...)Kitsune Ral
2017-12-20Connection::accessToken - return QByteArray instead of QStringKitsune Ral
Because that's what it really is.
2017-12-20Clean away legacy settings when possibleKitsune Ral
Otherwise they stick around when deleting actual settings (e.g. at logout).
2017-12-19Don't use LoginJob::homeServer() anymoreKitsune Ral
Because the respective parameter in /login result is deprecated. Closes #137.
2017-12-16Connection: proactively create a Room object upon joining a roomKitsune Ral
So that clients could start using this object without waiting for the next sync.
2017-12-16CMakeLists.txt: Make -Wreturn-type an errorKitsune Ral
Because a missing return in a non-void function is always an error.
2017-12-16Fix an assertion failure when redacting an unknown eventKitsune Ral
Closes #135.
2017-12-14Another fix for CI on OSXKitsune Ral
2017-12-14Add a missing header to fixing building on OSXKitsune Ral
2017-12-14Declare SyncRoomData as C++-movableKitsune Ral
This fixes compilation with older compilers that try to instantiate a copy constructor (and fail because unique_ptr) but actually is more proper in general. Also: do not advertise the class as Q_MOVABLE_TYPE: this was useful for QList/QVector when SyncRoomData was copyable; now it isn't, and Qt containers can't deal with non-copyable classes at all.
2017-12-14One more fix for older compilersKitsune Ral
2017-12-14Fixed building with older compilersKitsune Ral
2017-12-14That virtual ~Event() mentioned in the previous commit messageKitsune Ral
2017-12-14Move all internal event pointers to std::unique_ptr<>Kitsune Ral
This causes the following changes along the way: - Owning<> template is decommissioned. - event.h has been rearranged, and Event/RoomEvent::fromJson static methods have been replaced with an external makeEvent<> function template. A side effect of that is that one cannot use a factory with a type other than the one it's defined for (i.e. you cannot call makeEvent<TypingEvent>) but that feature has been out of use for long anyway. - Room::doAddNewMessageEvents() and Room::doAddHistoricalMessageEvents() have been removed, giving place to Room::onAddNewTimelineEvents() and Room::onAddHistoricalTimelineEvents(). The most important difference is that all code that must be executed now resides in addNewMessageEvents() (it moved from Room to Room::Private) and classes inheriting from Room are not obliged to call the overridden function from the overriding function (they can do it but those functions have empty bodies in Room). This was a long overdue change, and owning pointers simply mandated it. Room::onAddNewTimelineEvents/onAddHistoricalTimelineEvents should not do anything with the passed range in terms of ownership, it's just a way to allow the derived class to update his data in due course. - Room::Private::dropDuplicateEvents() and Room::Private::insertEvents(), notably, have been updated to work with owning pointers. insertEvents() move()s pointers to the timeline, while dropDuplicateEvents uses remove_if instead of stable_partition and doesn't explicitly delete event objects. Also, a bugfix: Event accidentally had not virtual destructor for quite a long time. According to the standard, deleting an object through a pointer to a base class without a virtual destructor leads to UB. So the fact that libqmatrixclient clients even worked all these months is mere coincidence and compiler authors good will :-D
generated by cgit v1.2.3 (git 2.25.1) at 2024-11-16 09:50:18 +0000