Fwd: Re: Qt5 and 4.2 release?

[Daniel J Sebald]

On 08/08/2016 03:19 PM, Mike Miller wrote:
> On Mon, Aug 08, 2016 at 12:27:19 -0700, Rik wrote:
> > All,
> >
> > Should we elevate this bug report to a blocker for the 4.2 release
> > (https://savannah.gnu.org/bugs/?40252)?

This post sets the Qt4 lifespan to somewhere in 2017:

http://perezmeyer.blogspot.com.es/2015/05/qt4s-status-and-qt4s-webkit-removal-in.html

and we're about 80% of that way since Qt 5 was announced.


> Sure. I'd be happy to take another stab at this. Should we drop support
> for Qt 5 or allow both 4 and 5? If both, Should we allow configure to be
> run with options like
>
>    --with-qt=5
>    --with-qt=4
>    --with-qt (default)
>    --without-qt
>
> and the default behavior would be to try Qt 5 first and fall back to Qt
> 4? I think getting the configure rewrite done right would be the hardest
> part of supporting both major versions of Qt, since the pkg-config and
> library names are all different.

I'd say support both Qt 4 and 5, but I hesitate suggesting both.  The
problem is once we support both, then programmers have to be aware they
are using functions and definitions that compile under both...but the
programmer may be working with Qt 5 installed so how to know the mods
compile under Qt 4 as well?

In that bug report I listed what I had to do to manually build under Qt
5 (well, partially).  It didn't seem too difficult to configure, I just
didn't want to mess around with the build files.

Given what Sebastian said about Mac and the fact it looks like only
another year of guaranteed Qt4 presence, dropping Qt4 is tolerable by me
if supporting both seems more confusing to configure than it is worth.

Dan