[Qgis-developer] Plans for QGIS 3.0 and QWebView.... *sigh*

[Matthias Kuhn]

On 09/14/2016 09:48 AM, Nyall Dawson wrote:
> On 14 September 2016 at 17:37, Sebastiaan Couwenberg <sebastic at xs4all.nl> wrote:
> 
> By the way Bas - If I'm coming across as annoyed it's because I am.
> 
> But it's not at you or our distro partners! I'm annoyed at Qt upstream
> for both the web view situation and for the predicament that
> https://bugreports.qt.io/browse/QTBUG-47192 has put us in (a broken
> application on Qt 5.5 and 5.6 which we can't work around).
> 
> I totally realise that you're also a volunteer and put in difficult
> situations because of the choices we as upstream make. Thanks for all
> the hours you've put into supporting this project behind the scenes.
> It's VERY much appreciated.
> 
> Nyall
> 

I fully support this statement and would like to apologize for any
provoking comments I have done.

Please let me explain the situation from my side:

We have a very high amount of users which use QGIS for their daily work,
professional and for pleasure. They expect as little interruption as
possible doing their job. They don't actually care about Qt versions at
all. They work on a wide variety of platforms.

We have also a very large plugin ecosystem which needs to care about Qt
versions because they depend on the API. But they also are happiest if
there is as little change as possible.

QGIS 2.0 is not too long ago when there was a major API change. Doing
this too often will make people very unhappy.

There is also Qt upstream which started to develop Qt 5 quite some time
ago but didn't manage to make a clear plan for things like QtWebKit.
Lack of knowledge about the path of action there makes it not very easy
to take decisions on our end.

Then there are also distributions coming with a wide variety of
different versions of all kind of dependencies which we try to support
as good as possible from some 4 year old ones to the brandnew bleeding
edge ones.

Then there is us, putting a vast amount of volunteer time (besides the
funded time) into planning and implementing an upgrade that is as
comprehensive, stable and smooth as possible taking into account all the
different parties mentioned above.

In this scenario it is sometimes possible that there is a little
frustration sitting too easy if there is some pressure from one of these
parties to support something that is only on the short term radar (for
us at least), a problem not created by us and no real proper solution
available - just like support for QtWebKit in Qt4 on one Debian release.

Thank you very much for the time you have invested, it is very valuable
and much appreciated.

Kind regards
Matthias