> On Sun, 6 Nov 2016, René J.V. Bertin wrote:
> (Q_WS_MAC) which I thought is no longer defined
I was right.
And when I say some of the Mac adaptations look (almost) like quick hacks:
```
#ifdef Q_OS_MAC
#include <cstring>
#include <windows.h>
#include <shellapi.h>
/**
* Native Win32 method for starting a process. This is required in order to
* launch the installer with User Account Control enabled.
<snip>
#else // Q_OS_MAC
<snip>
#ifdef Q_WS_MAC
QDir bundleDir(qApp->applicationDirPath());
#endif
<snip>
#endif // Q_OS_WIN
```
On Sunday November 06 2016 14:26:41 Friedrich W. H. Kossebau wrote:
> When it comes to Calligra and MacPorts, I do have not seen anyone working on
> macOS-related stuff. So if you have patches/improvements, there might be
> noone to discuss how-to-macOS. Which means if it does not break/screw the
> Linux build (or Windows when it comes to Kexi & Co.), your chance to rule.
That sounds more like it :) (Not that I'm intent on ruling, but sometimes it's
nice to be first ;))
Truth be told, MacPorts provides a very nice "meta build system" with some
nifty features like port de/activation, so I'm using the same build "scripts"
to build KF5 things on my KUbuntu 14.04 system with its trusty Plasma4 desktop.
I still prefer the stability of that legacy desktop, but appreciate running
select KF5 applications from a parallel prefix (guess what, /opt/local). So,
basically every patch I make for MacPorts is tested on Linux too before I
consider submitting it.
R.
