rework ThreadSafeAskFee() / askFee() functions
- remove unused parameter from ThreadSafeAskFee(), which also results in the removal of an orphan translation-string
remove unneeded flag from MSG_INFORMATION and fix an indentation
rework ThreadSafeAskFee() / askFee() functions
remove unneeded flag from MSG_INFORMATION and fix an indentation
65 | bool payFee = false; 66 | 67 | QMetaObject::invokeMethod(guiref, "askFee", GUIUtil::blockingGUIThreadConnection(), 68 | Q_ARG(qint64, nFeeRequired), 69 | - Q_ARG(bool*, &payFee)); 70 | + Q_ARG(bool&, payFee));
Have you really tested that this works? I remember having quite some difficulties getting askFee to return a value over the thread boundary, and ended up with a pointer as the only option. AFAIK references are considered read-only by Qt.
Yes, I sent a transaction (to myself in that test-case) and got asked the "do you want to pay ... bla bla" message, clicked Yes and got a tray-notification of that Tx. Can you try that code on Linux? Or do you want me to leave it a pointer?
Edit: Seems to be more sane to revert that changel as the Qt docs speak of Q_ARG() takes a type name and a const reference of that type; Q_RETURN_ARG() takes a type name and a non-const reference.
Automatic sanity-testing: PASSED, see http://jenkins.bluematt.me/pull-tester/e66911cc20c6bcd9200a39c6c3a4027b3102a8ce for binaries and test log.
- remove unused parameter from ThreadSafeAskFee(), which also results in
the removal of an orphan translation-string
Automatic sanity-testing: PASSED, see http://jenkins.bluematt.me/pull-tester/15d8ff20294d20e6b11ec54450943eb7c5282ae0 for binaries and test log.
ACK