On Sat, 5 Nov 2011 10:05:45 +0100, Detlev Offenbach <det...@die-offenbachs.de> wrote: > Am 04.11.2011 um 18:04 schrieb Phil Thompson: > >> On Fri, 4 Nov 2011 11:23:07 +0100, Detlev Offenbach >> <det...@die-offenbachs.de> wrote: >>> Hello, >>> >>> I am defining a signal in one of my classes like this. >>> >>> tabRelocateRequested = pyqtSignal(int, int, int) >>> >>> When this signal is emitted, the first parameter is bigger than an int >>> (e.g. by using id()). This causes the signal receiver to get a first >>> parameter, that is different to the value given in the emit call. How >> can I >>> correct my code (code i within eric4/5). >> >> Use long instead of int? > > Python3 (forgot to mention that) doesn't know about long anymore. Maybe > the pyqtSignal code should work with long instead of int?
The current behaviour is intentional as I considered this to be the least surprising when using a signal defined in Python in C++ code. Thinking about it now this is probably a very minor use case. At the moment a Python v2 int (actually implemented as a C++ long), a Python v2 long and a Python v3 int are all mapped to a C++ int. Maybe a better mapping should be Python v2 int to a C++ long, and a Python v2 long and a Python v3 int to a C++ long long. However it would still be possible to lose data for very big values - the only way to avoid that would be to pass the Python object itself which would make it impossible to reuse the class defining the signal from C++ code. As a work around you can always specify the C++ type as a string, eg. "long long". Thoughts? Phil _______________________________________________ PyQt mailing list PyQt@riverbankcomputing.com http://www.riverbankcomputing.com/mailman/listinfo/pyqt
