[Development] QTBUG-61258 and Qt3D status

[anton]

I was considering using Qt and
I even did a prototype of a mostly working tool half a year ago.

But looking at the strange development culture and quality attitude
I think I give it up.

I will rewrite my software with another tool again... too much surprises
in Qt.

Your story is only one of similar ones.

This is sad, because I really would like to use Qt and I like KDE as one
example.

Greetings
Anton

Am 14.12.18 um 14:07 schrieb Massimo Callegari:
> Hi devs,
> I hate to write these emails, but Jira and Gerrit don't really work for
> me. (and at this point I suspect I definitely have BAD luck)
> 
> The question this time is: what's the deal with Qt3D ?
> Since a year or so, submitting Qt3D issues to Jira or even contributing
> on Gerrit is like writing to /dev/null. Not even a "please close" is
> considered.
> I take all the efforts are going on Qt3D Studio, but what about all the
> rest ?
> 
> Please let me explain what I'm doing.
> 
> I run a quite popular open source project called Q Light Controller+
> (https://www.qlcplus.org)
> It's a software to control stage lighting and it's entirely based on Qt.
> Over the years I adopted more and more Qt modules and (here's my bad
> luck) I stumbled on every possible Qt bug.
> At some point I wanted to find a solution to preview in 3D and in real
> time the lighting simulation.
> I followed the Qt3D 2.0 developments for years and when I felt it was
> mature enough, I started to code the 3D preview.
> Since I don't have the necessary OGL skills for what I needed, I was
> willing to pay someone to do the job.
> I contacted KDAB for an official quotation and they didn't even reply to
> me. So I crowdfunded the feature among my users and paid someone else to
> develop the 3D techniques.
> The bright side is that with Qt3D as it is, we've been able to achieve
> this (https://youtu.be/eI_NfA_vyA0) and this (https://youtu.be/yoQVzYR-NwM).
> On the other hand, performance sucks and since Qt 5.11, a few
> regressions started to kick in.
> The blocking one for me is QTBUG-69721. I invested time and effort to
> adopt the DAE format, cause it supports named+nested meshes, and it's
> XML, so it can be tracked on GitHub.
> Since 5.11, meshes can no longer be picked on macOS, so it means I
> cannot release any version of my software there. (also cause of broken
> video playback on core context QTBUG-51064)
> 
> So I evaluated glTF, being another nice open format...and found a disaster.
> The current scene import plugin does not conform to the Khronos 1.0
> reference samples. Even the "wine" example bundled in Qt3D doesn't work!
> So basically Qt is bundling a nearly useless feature.
> Plus, Blender 2.80 beta has started to roll, and it finally supports
> native export to glTF....2.0....not supported by Qt3D.
> So I spent a few days and got preliminary glTF 2.0 support. With PBR
> materials it looks pretty awesome! (see screenshots in QTBUG-61258)
> I submitted my code to Gerrit
> (https://codereview.qt-project.org/#/c/247080/) and invited as many
> reviewers as possible.
> Almost 2 weeks passed and I got ZERO comments/reviews by any Qt
> Company/KDAB developer.
> I mean, don't you want it? Not even in Qt3D Studio?
> 
> This is very discouraging for developers like me who spend their time
> trying to improve Qt.
> If this attitude keeps going, I will end up not contributing at all in
> the future, and I suppose I'm not alone on this.
> 
> So, once again, what's the deal with Qt3D? 
> 
> Please explain, cause to me it looks in a quite abandoned state right
> now. (management-wise speaking)
> Thanks,
> Massimo
> 
> _______________________________________________
> Development mailing list
> Development at qt-project.org
> https://lists.qt-project.org/listinfo/development
>