Re: [Development] QQuaternion issues with new 5.5 API
On Friday 24 April 2015 14:32:22 Gunnar Sletta wrote: > > While implementing qHash() overloads for gui/math3d classes, > > Why are we doing this? a QHash makes very little sense me.. Well, there are are bug reports asking for qHash(QPoint), so it seems to make sense to _someone_ :) Esp. for types that don't have a natural total order (=most), and thus don't / shouldn't have operator<, adding a qHash() overload, as long as it can be done (it can't for QPointF) is the only way to enable associative containers of any kind for that type. Thanks, Marc -- Marc Mutz | Senior Software Engineer KDAB (Deutschland) GmbH & Co.KG, a KDAB Group Company Tel: +49-30-521325470 KDAB - The Qt Experts ___ Development mailing list Development@qt-project.org http://lists.qt-project.org/mailman/listinfo/development
Re: [Development] QQuaternion issues with new 5.5 API
On Friday 24 April 2015 14:32:22 Gunnar Sletta wrote: > > While implementing qHash() overloads for gui/math3d classes, > > Why are we doing this? a QHash makes very little sense me.. That's besides the point. Marc could have said "I've just noticed" and the rest of his arguments would still be valid. -- Thiago Macieira - thiago.macieira (AT) intel.com Software Architect - Intel Open Source Technology Center ___ Development mailing list Development@qt-project.org http://lists.qt-project.org/mailman/listinfo/development
Re: [Development] QQuaternion issues with new 5.5 API
2015-04-24 16:33 GMT+04:00 Marc Mutz : > Hi, > > While implementing qHash() overloads for gui/math3d classes, I found that > QQuaternion gained several methods for 5.5 which I don't like: > > *EulerAngles(): > > They are missing a QEulerAngles class. Instead, they deal with (float, > float, > float) and QVector3D. One function even returns three floats as out- > parameters. I think my (partial) work on QDate/Time has shown just how much > compilers don't like out parameters. The question here is how general such > a > QEulerAngles class should be... > > *AxisAndAngle()/*Axes(): > > Same here, to a lesser extent. What bugs me most is the return-by-out- > parameters, not so much that there's no QAxis3D class. > > There are several steps forward: > > - Create QEulerAngles as a float-only class > - ditto, but as a template > - ditto, but also add Q*Angle classes that have DEG/RAD hard-coded as > template > arguments, with explicit conversions between, then use that in > QEulerAngles. > [this likely won't happen for 5.5, though] > - remove the methods in question for 5.5 and try again in 5.6 > > Any opinions about which ones to take? > Sure I wanted to do it via introducing QAngle (which is also useful for many other classes, even outside math3d), but that's definitely a 6.0 material; and then there is no much sense in introducing it just for QQuaternion. Konstantin ___ Development mailing list Development@qt-project.org http://lists.qt-project.org/mailman/listinfo/development
Re: [Development] QQuaternion issues with new 5.5 API
2015-04-24 16:32 GMT+04:00 Gunnar Sletta : > > > On 24 Apr 2015, at 14:33, Marc Mutz wrote: > > > > Hi, > > > > While implementing qHash() overloads for gui/math3d classes, > > Why are we doing this? a QHash makes very little sense me.. > Completely agreed. Konstantin ___ Development mailing list Development@qt-project.org http://lists.qt-project.org/mailman/listinfo/development
Re: [Development] QQuaternion issues with new 5.5 API
> On 24 Apr 2015, at 14:33, Marc Mutz wrote: > > Hi, > > While implementing qHash() overloads for gui/math3d classes, Why are we doing this? a QHash makes very little sense me.. - Gunnar > I found that > QQuaternion gained several methods for 5.5 which I don't like: > > *EulerAngles(): > > They are missing a QEulerAngles class. Instead, they deal with (float, float, > float) and QVector3D. One function even returns three floats as out- > parameters. I think my (partial) work on QDate/Time has shown just how much > compilers don't like out parameters. The question here is how general such a > QEulerAngles class should be... > > *AxisAndAngle()/*Axes(): > > Same here, to a lesser extent. What bugs me most is the return-by-out- > parameters, not so much that there's no QAxis3D class. > > There are several steps forward: > > - Create QEulerAngles as a float-only class > - ditto, but as a template > - ditto, but also add Q*Angle classes that have DEG/RAD hard-coded as template > arguments, with explicit conversions between, then use that in QEulerAngles. > [this likely won't happen for 5.5, though] > - remove the methods in question for 5.5 and try again in 5.6 > > Any opinions about which ones to take? > > Thanks, > Marc > > -- > Marc Mutz | Senior Software Engineer > KDAB (Deutschland) GmbH & Co.KG, a KDAB Group Company > Tel: +49-30-521325470 > KDAB - The Qt Experts > ___ > Development mailing list > Development@qt-project.org > http://lists.qt-project.org/mailman/listinfo/development ___ Development mailing list Development@qt-project.org http://lists.qt-project.org/mailman/listinfo/development
Re: [Development] QQuaternion issues with new 5.5 API
On Friday 24 April 2015 14:33:27 Marc Mutz wrote: > There are several steps forward: > > - Create QEulerAngles as a float-only class > - ditto, but as a template > - ditto, but also add Q*Angle classes that have DEG/RAD hard-coded as > template arguments, with explicit conversions between, then use that in > QEulerAngles. [this likely won't happen for 5.5, though] > - remove the methods in question for 5.5 and try again in 5.6 > > Any opinions about which ones to take? Of course - do nothing is an option, too, but clearly not the one that I'd prefer :) -- Marc Mutz | Senior Software Engineer KDAB (Deutschland) GmbH & Co.KG, a KDAB Group Company Tel: +49-30-521325470 KDAB - The Qt Experts ___ Development mailing list Development@qt-project.org http://lists.qt-project.org/mailman/listinfo/development
