On Montag, 23. November 2009, Phil Thompson wrote: > On Mon, 23 Nov 2009 18:08:48 +0530, Prafulla Kiran > > <[email protected]> wrote: > > Hi Phil, > > > > The following piece of code throws up the error messages that Detlev has > > mentioned. Is there something wrong that I am doing here ? > > The same piece of code works fine in PyQt 4.5.* and 4.6.1 > > > > from PyQt4 import QtCore > > > > > > settings_ = QtCore.QSettings("Foo", "Bar") > > > > qString = QtCore.QString("teamtalk.im") > > > > settings_.setValue("URL", QtCore.QVariant(qString)) > > > > This throws up: /QVariant::save: unable to save type 256. /in PyQt > > 4.6.2 > > It's a SIP code generation bug - though why it has chosen now to make > itself known I don't know. > > I've just released SIP v4.9.3 - that should fix it.
Did you update the Windows installers as well? I had a few bug reports from windows users that could have been caused by this issue. > > > Phil Thompson wrote: > >> On Sun, 22 Nov 2009 13:58:00 +0100, detlev <[email protected]> > >> > >> wrote: > >>> Hi, > >>> > >>> after upgrading to PyQt4 4.6.2 (from 4.6.1) I get error messages like > >>> > >>> QVariant::save: unable to save type 256. > >>> QVariant::load: unable to load type 256. > >>> > >>> What does this mean? With 4.6.1 everything was ok. > >> > >> I don't think it has anything to do with v4.6.2. > >> > >> Although any object can be converted to and from a QVariant it is not > >> true > >> that any QVariant can be saved to and loaded from files. > >> > >> My guess is that you have written a value (at some time in the past) > > that > > >> can't be saved and not noticed the Qt error message. Note that QSettings > >> still writes a value in this case. > >> > >> There is a bug (fixed in tonight's snapshot) where an exception isn't > >> raised when QSettings tries to read an unreadable value, but this bug > > has > > >> been there since PyQt v4.5. > >> > >> For v4.7 I will add support for saving and loading objects that can be > >> pickled - but that still won't cover all objects. > > Although the above isn't the correct explanation of this particular > problem, it still applies. > > Phil > _______________________________________________ > PyQt mailing list [email protected] > http://www.riverbankcomputing.com/mailman/listinfo/pyqt > -- Detlev Offenbach [email protected] _______________________________________________ PyQt mailing list [email protected] http://www.riverbankcomputing.com/mailman/listinfo/pyqt
