Public bug report changed: https://launchpad.net/malone/bugs/32064
Comment: QTimer suffers from same problems as processEvents. It is just further bit more opaque and hairy. For threading model, again, i do not break anything. Calls from non-gui thread are not distinguishable from calls from gui thread, apart from the backtrace/stack (and state of registers). Specifically, non-stack memory state is exactly identical. Qt widgets do not store anything in thread-local storage. It would not even make sense since qt widgets don't support multi-thread access. Sure enough calls to qt methods don't depend on something buried down the stack. They also don't depend on register state. Since gui thread is blocked in a point where it was idle (not doing anything qt related), qt has no chance to run into an unprotected critical section in two threads at the same time. Even if we treat every qt widget method as a critical section, we can still be sure of this. The threads do not run at once uncontrolled. -- kubuntu-bugs mailing list [email protected] https://lists.ubuntu.com/mailman/listinfo/kubuntu-bugs
