> I ended up using "proper" geometry processing library for the "model" and used Qt to do the rendering (the view).
Somehow I get the feeling you just saved me a ton of headaches in the future :) thx On Sun, May 12, 2019 at 1:50 PM Christian Gagneraud <chg...@gmail.com> wrote: > On Sun, 12 May 2019 at 20:26, Konstantin Shegunov <kshegu...@gmail.com> > wrote: > > > > Hi, > > I'd want to clear the context out of the way, so this is the bug[1] that > got me thinking. > > I appreciate that we want to keep external dependencies to a minimum, > and for a good reason, but can we talk about how feasible it is to pull > something (or parts of it) in Qt, even if only internally, to facilitate > stable fp calculation. > > > > I've seen some complaints (on the forums mostly) about the stability of > the transformations Qt supplies, generally unfounded, however it'd be nice > if we can solve this with generality. I realize Qt is no math library and > support the idea to drop most of the global functions that dealt with math, > in the end they're already in the STL. However, as it is, we have some > linear algebra done (internally mostly), so I think it'd be nice to have > that done "correctly" for the user. > > > > I use to think that Qt could do a better job about FP > precision/stability, but i had to realise that i was using Qt in a way > that it was not designed for. > For example, I tried to use QPainterPath, QLineF, QRectF, ... to do > geometry processing. And i can tell you that QPainterPath is all but > stable when it comes to small values. Highly zoomed-in QGraphicsView > based geometry object yields crazy artifacts. > I then turned on specialised library, Qt is a GUI toolkit, and is > optimised for painting. Qt takes all opportunities to be fast and > efficient in that context, and that includes being "mathematically" > imprecised, painting is all about pixels at the end of the day. > Who cares where exactly a line intersect a polygon, if all you need to > know is if you need to paint a pixel black or red. > > To summarise: > I ended up using "proper" geometry processing library for the "model" > and used Qt to do the rendering (the view). > It came at a cost, but in the long term was a big win. > > Chris > > > I'd appreciate feelings and opinions on the matter, as I'm but one > person and my view is rather limited. > > > > [1]: https://bugreports.qt.io/browse/QTBUG-75146 > > _______________________________________________ > > Development mailing list > > Development@qt-project.org > > https://lists.qt-project.org/listinfo/development > _______________________________________________ > Development mailing list > Development@qt-project.org > https://lists.qt-project.org/listinfo/development >
_______________________________________________ Development mailing list Development@qt-project.org https://lists.qt-project.org/listinfo/development