| Age | Commit message (Collapse) | Author |
|---|
|
This reverts commit b1071cf34b86685c3cdb5004d6112881966a7ce6. Passing
-1 to sync() and, respectively, to SyncJob does not add any timeout;
however, careful reading of the spec reveals that the default value
for the timeout (0) means to return as soon as possible, not as late
as possible. As a consequence, syncLoop() without parameters initiates
a sync polling frenzy, with the client sending a new request as soon as
the previous returns, while the server returns the request as soon as
it practically can, not as soon as another event for the client comes
around.
To fix this, the default value for syncLoop() is changed to 30 seconds.
The recently added msecBetween parameter is abolished; we really don't
want to steer people to classic polling from long polling.
|
|
|
|
As it's observed now, Synapse responds almost immediately on /sync
requests - even if there are no events to return. This downgrades
long-polling to simply polling, and since clients don't expect it,
polling loops become pretty violent. To alleviate that somehow,
syncLoop now accepts the second parameter, msecBetween (500 msecs
by default), to configure waiting between the previous sync response
and the next sync request. This is only for syncLoop();
Connection::sync() fires instantly, as before.
|
|
|
|
Qt 5.15 deprecates binary JSON format in favour of CBOR now used as
a backend for its JSON classes.
|
|
|
|
No functional changes here.
|
|
This is to prevent the jobs from several
resolveServer() / setHomeserver() invocations running in parallel.
|
|
|
|
connectToServer() is left for compatibility but deprecated.
|
|
The code is really defensive now, making sure there's exactly one slash
between the base path and the endpoint. It's still very conservative
about the path composition otherwise (no normalisation etc.).
|
|
(Mostly) fixes #397. Fixes #398. If there's a need to reparent the job
to some other object, it should now be done after callApi()/run() call
(not that there were many touch points before that moment, anyway).
Collateral damage: job-starting methods are no more const (but that
didn't belong them anyway, too).
|
|
Also: use 4-arg connect() to make sure lambdas are disconnected
if the connection is gone.
|
|
|
|
The method grew large and a bit unwieldy over the years.
|
|
|
|
fixing msc2432
|
|
The response in the web browser is quite barebone, just enough
to give feedback that things are alright. Closes #386. Closes #388.
|
|
|
|
|
|
The flows themselves are not facilitated in any way (yet).
|
|
Signed-off-by: Alexey Andreev <aa13q@ya.ru>
|
|
|
|
device_one_time_keys_count
Signed-off-by: Alexey Andreev <aa13q@ya.ru>
|
|
Signed-off-by: Alexey Andreev <aa13q@ya.ru>
|
|
|
|
* BaseJob::initiate() now calls ConnectionData::submit()
without relying on Connection to do that
* ConnectionData::submit() is now the only site where a job enters
Pending state
* No more shortcuts to BaseJob::sendRequest(), even retries are sent
through the ConnectionData submission queue
* Additional validation in BaseJob::initiate() that the request data
device is actually open (because QtNetwork API officially
requires that, even if you can get away passing a closed QBuffer
to it)
|
|
No more "The file is already open" log messages.
|
|
The first part closes #358; the second part is a workaround for non-standard
cases when endpoints without security by the spec turn out to be secured
(in particular, the case of authenticating media servers).
|
|
See the change in connection.cpp for the example of usage.
Also: removed static_asserts: the first one is provided by std::optional,
and the second one is only relevant to edit().
|
|
Xcode 10 doesn't have it, and value() is not quite fitting
mostly-exceptionless Quotient anyway.
|
|
That breaks API all over the place but:
1. The fixes are trivial.
2. More of std:: is used instead of home-baking the same stuff.
|
|
This is mostly internal but clients may see fewer spurious sync failures
upon logging out.
|
|
...in the sense that you can call it twice and expect the second
invocation to be gracefully ignored rather than two loops conflicting
with each other.
|
|
Because https://doc.qt.io/qt-5/moc.html#limitations .
For direct connections that doesn't matter but it very much does
for queued ones. Along with this DirectChatsMap and IgnoredUsersList
have been moved from Connection:: to Quotient::.
|
|
All direct slots connected to finished() will run before success() is
even emitted; so create the room object in the earliest slot connected
to finished(), rather than to success().
|
|
Closes #354.
|
|
Invading into std:: is frowned upon, even though legitimate from the C++ standard perspective. Given that it's possible to pass a hash object to unordered_map, it only takes an alias for std::unordered_map to avoid having to specialize std::hash. And besides, a generic compatibility bridge between qHash and std::hash has been long needed.
std::hash<QString> in converters.h remains for now; it will be dropped separately when the API files get regenerated to use UnorderedMap.
|
|
As before, completely transparent for clients, driven by 529 errors from
the server (but cases of rate limiting are signalled by
BaseJob::rateLimited). That brings changes to BaseJob API: timeouts now
use int64_t and also can be handled in std::chrono terms; aboutToStart()
-> aboutToSendRequest(); started() -> sentRequest(). Closes #292.
|
|
Finally, clients can pre-create job objects and then separately submit
them for execution on a given connection - previously such separation
was a privilege of Connection (others had to use Connection::callApi<>,
which invoked jobs right away).
|
|
To collect all connection-identifying information in a single place.
|
|
The fallback to libQMatrixClient still remains.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
.well-known support
|
|
|
