[Interest] Qt Quick Controls 1 deprecated but no native styles for Qt Quick Controls 2?
Dimitar Dobrev
dpldobrev at protonmail.com
Fri Dec 7 14:01:25 CET 2018
Electron is so bloated that it cannot serve as an example for anything,
it's simply not worth discussing.
Thank you, however, for the link to this project of KDAB's, it might be
useful.
I'm aware there are technical difficulties but there should be a
solution. For example, they could choose reusing QStyle as a quick
solution and improve - or replace - it later on with GPU optimizations.
Finally, what I meant by saying "won't" was that I haven't seen anybody
saying "will". And this issue has already been brought up at least a
year ago. So it has so far sounded like "we'll do it later" which often
means never unless somebody speaks up. And so I did.
On 7.12.18 13:35, Giuseppe D'Angelo via Interest wrote:
> Hi,
>
> Not a real answer to the Qt Project's direction (as you're also asking
> about that), just a few pointers and remarks here and there.
>
>
> Il 07/12/18 10:09, Dimitar Dobrev via Interest ha scritto:
>> The release notes for Qt 5.12
>> <https://wiki.qt.io/New_Features_in_Qt_5.12> worry me quite a little.
>> They say that Qt Quick Controls 1 is deprecated. There's a single but
>> key reason this is extremely bad news. And this reason is the lack of
>> native styles in Qt Quick Controls 2. This alone renders Qt Quick
>> Controls 2 useless for building decent desktop applications.
> Note that this isn't necessarily true in general, e.g. take a look at
> Electron-based desktop applications, which deliberately not follow any
> native style. (I'm not saying that this is good or bad; I'm just saying
> that those applications exist). QQC2 would be a good match for such
> usecases.
>
>
>
>
>> This in
>> turn means our only option remains Qt Widgets - a piece of technology
>> which is like a horse carriage. Good for its time but useless in the era
>> of automobiles. The very notion of suggesting that for desktop
>> development in 2018 we would be deprived of a simple declarative
>> language for GUI, a flexible scripting language to match, GPU-based
>> optimizations and all other wonderful features Qt Quick has to offer -
>> is ridiculous at best.
> A few technical notes:
>
> 1) You can define QWidget based hiearchies using QML , see for instance
> KDAB's Declarative Widgets project:
>
>> https://github.com/KDAB/DeclarativeWidgets
>
>
> 2) At the very core of native styling in Qt there's QStyle, which uses
> CPU-based rendering, relying on pixel perfect output provided by
> QPainter. (Take a look into a style class, you'll found thousands of
> QPainter draw calls). A native style built for QQC2 will mean one of
> these options:
>
> * re-using QStyle existing code, which means neglecting your GPU-based
> optimizations
>
> * reimplementing native styles just for QQ2 using GPU-accelerated
> primitives, which is super-difficult and surely an investment noone
> wants to make (not to mention it would mean two implementations of
> native styles, who is going to maintain those?)
>
> * port QStyle itself to some abstraction that allows for GPU-accelerated
> rendering (and of course keep CPU-based rendering around, for GPU-less
> environments) which again sounds extremely hard, risks regression all
> over the place, it's probably a Qt 6-worth break, etc.
>
>
>
>> If it's true that Qt Quick Controls 1 is
>> deprecated and Qt Quick Controls 2 won't get native styles any time
>> soon, this simply means Qt has severely regressed in its offerings to
>> developers.
> Why do you say "won't get native styles"? This is kind of a negation of
> the future. Was it announced somewhere and I missed that line?
>
> Thanks,
> --
> Giuseppe D'Angelo | giuseppe.dangelo at kdab.com | Senior Software Engineer
> KDAB (France) S.A.S., a KDAB Group company
> Tel. France +33 (0)4 90 84 08 53, http://www.kdab.com
> KDAB - The Qt, C++ and OpenGL Experts
>
> _______________________________________________
> Interest mailing list
> Interest at lists.qt-project.org
> https://lists.qt-project.org/listinfo/interest
More information about the Interest
mailing list