A few weeks ago we looked at problems with connecting signals to slots across threads. A result was a documentation change explaining that, when connecting to a Python callable (as opposed to a SLOT()), the slot was executed in the thread in which the connect() call was made.
An argument was made (I think it was Matt Newell) that the behaviour should follow the Qt behaviour in the case where the Python callable was a method of a QObject sub-class. I decided against this partly because I was lazy (more code to write and get wrong) and partly because the explanation (and therefore the understanding) of the behaviour was more complicated. I've now decided that the current behaviour can be a bit of a pain in certain circumstances, and so it has been changed in tonight's snapshot. Phil _______________________________________________ PyQt mailing list [email protected] http://www.riverbankcomputing.com/mailman/listinfo/pyqt
