| Age | Commit message (Collapse) | Author |
|---|
|
Due to a missing return statement, a retry with auth case led to the job
being finished and pending at the same time, with no good consequences.
|
|
Longer running on RHEL/CentOS 8 leads to crashes that no more occur
with disabled pipelining.
|
|
|
|
This is a Quotient part of #328 fix; QtOlm part is pending.
|
|
|
|
A part of the fix for #398.
|
|
|
|
* 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)
|
|
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).
|
|
|
|
|
|
|
|
Compilers warn on using 'using ...::operator""s' because they
think we're redefining the reserved suffix.
|
|
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.
|
|
Notably, recovered Status::fromHttpCode() that was introduced in
5722ceaf4bd10c29f1091e3dc5a87f5650ea8c71 but fell victim of a careless
merge (so much for introducing non-topical changes in feature branches).
|
|
Closes #344.
|
|
Closes #326.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
# Conflicts:
# CMakeLists.txt
# lib/avatar.cpp
# lib/connection.cpp
# lib/connection.h
# lib/connectiondata.cpp
# lib/csapi/account-data.cpp
# lib/csapi/account-data.h
# lib/csapi/capabilities.cpp
# lib/csapi/capabilities.h
# lib/csapi/content-repo.cpp
# lib/csapi/create_room.cpp
# lib/csapi/filter.cpp
# lib/csapi/joining.cpp
# lib/csapi/keys.cpp
# lib/csapi/list_joined_rooms.cpp
# lib/csapi/notifications.cpp
# lib/csapi/openid.cpp
# lib/csapi/presence.cpp
# lib/csapi/pushrules.cpp
# lib/csapi/registration.cpp
# lib/csapi/room_upgrades.cpp
# lib/csapi/room_upgrades.h
# lib/csapi/search.cpp
# lib/csapi/users.cpp
# lib/csapi/versions.cpp
# lib/csapi/whoami.cpp
# lib/csapi/{{base}}.cpp.mustache
# lib/events/accountdataevents.h
# lib/events/eventcontent.h
# lib/events/roommemberevent.cpp
# lib/events/stateevent.cpp
# lib/jobs/basejob.cpp
# lib/jobs/basejob.h
# lib/networkaccessmanager.cpp
# lib/networksettings.cpp
# lib/room.cpp
# lib/room.h
# lib/settings.cpp
# lib/settings.h
# lib/syncdata.cpp
# lib/user.cpp
# lib/user.h
# lib/util.cpp
|
|
JsonParseError
|
|
|
|
Resetting the code to IncorrectRequestError has been a part of the cause for the incorrect Quaternion behaviour on expired tokens.
|
|
See https://github.com/QMatrixClient/Quaternion/issues/566 for details.
|
|
|
|
Previously slashes in eventIds (that come plenty in v3 due to base64 encoding) were not properly encoded - they are now.
|
|
|
|
1. It should work with non-started jobs now (Closes #289).
2. It should allow clients to still handle `finished()` instead of
cutting them off right before emitting the signal.
|
|
|
|
A new recommended (and localisable) way of getting a piece of raw
response to display next to error messages as "details".
BaseJob::rawData() returns exactly the trimmed piece of data, no
"truncated" suffix there anymore.
|
|
|
|
|
|
|
|
If the payload is too large, an attempt to allocate a QString out of
QByteArray may end with qBadAlloc(). So by default the data emitted in
case of error are trimmed to 64KiB, and this can be overridden to a
different value (or switched off entirely with <n>=0) by adding -
DTRIM_RAW_DATA=<n> to CPP_FLAGS.
|
|
|
|
|
|
Closes #210.
|
|
Closes #207.
|
|
Running a request in background, aside from some tweaks on the network
layer (see QNetworkRequest::BackgroundRequestAttribute), allows to
distinguish jobs not immediately caused by user interaction (such as
fetching thumbnails). This can be used to show or not show certain
notifications in UI of clients.
Error reporting has been extended with more methods:
errorCaption() - a human-readable phrase calculated from the status
code; intended to be shown as a dialog caption and in similar
situations.
errorRawData() - former errorDetails(), returns the raw response from
the server.
errorUrl() - returns a URL that may be useful with the error (e.g. for
the upcoming "consent not given" error, this will have the policy URL).
Connection::resultFailed() - a new signal emitted when _any_
BaseJob::failure() is emitted (enables centralised error handling
across all network requests in clients).
As a part of matching changes in Connection, callApi has an overload
that allows to specify the policy; a custom enum instead of bool has
been chosen for the parameter type, to avoid clashes with (arbitrary)
types of job parameters.
|
|
The newly introduced statusChanged() signal allows clients to be
notified about the job state transitions.
|
|
There is no more BaseJob::kill() for a very long time; we have abandon()
instead, with a bit different meaning. Which, by the way, now emits
finished() because that's what finished() has been for.
|
|
Connection
(Note: this commit breaks back-compatibility.)
BaseJob::gotReply() had a bold assumption that whenever there's no reasonable JSON body in case of error, that meant IncorrectRequestError (which of course was incorrect). This led to syncs dying in Quaternion because it (correctly) considered IncorrectRequestError to be grave enough to not retry. Closes #206.
Also: don't dump the whole payload to logs, limit with initial 500 characters (closes #205)
|
|
|
