[KDE/Mac] [Interest] Fwd: OSX/CI: Found umpteen uses of Q_WS_MAC in Qt5 itself

[Ian Wadham]

On 22/01/2015, at 4:46 AM, Marko Käning wrote:
> On 21 Jan 2015, at 11:55 , René J.V. Bertin <rjvbertin at gmail.com> wrote:
>>>> And if junk DNA, should it be cleaned out, or tested on a case-by-case basis if in reality it ought to be activated??
>>> 
>>> They're never defined, so junk, but they can't just be removed. One
>>> should check what they're actually wrapping -- if they were meant to
>>> be Q_OS_*, or perhaps exposing functionality that has been lost when
>>> porting to QPA, etc.
>> 
>> Right, so indeed one has to check them on a case-by-case basis. Sounds like fun, esp. for the occurrences in header files...
>> Is that on the TODO list of member(s) of the Qt dev team, or is it going to be left to community members with too much time (and other resources) on their hands?
> 
> thanks for posting this to the interest list! :)
> 
> Well, this is not OUR job here on KDE-MAC...
> 
> Do we agree on that? ;)

Absolutely.  The buck should stop wherever the discontinued Q_WS_MAC
macro is used, whether in a KF 5 or a Qt 5 module.  In Qt 5 (for Mac) it is
likely to be an integrity problem and in KF 5 a porting problem.

I think it was overlooked in the KF 5 Porting Notes, although one or two KDE
core developers knew about it back in July last year, when it came up in email
and ReviewBoard posts.

On the positive side, Laurent has incorporated the Q_WS_MAC fix into one of his
porting scripts now.  I must ask him if it can be run on previously ported code.

But I agree with René: we can be helpful on this one, for example in looking
at tricky cases.  And so can the MacPorts developers, who know more than us
about Apple OS X.  But I do not think KDE-Mac and MacPorts guys should
have to do all the legwork.

Then again, Qt must have a few Apple Mac experts… :-)  So let's all co-operate!

It is the best way IME to solve compatibility problems between software systems.

All the best, Ian W.