[Development] Are SiCs through #include cleanups considered acceptable?

[Matthew Woehlke]

On 2015-04-09 07:04, André Somers wrote:
> I think it *is* reasonable to remove such indirect includes if they are 
> not needed for Qt itself. However, it is worth looking into how the 
> thing was documented to begin with. In the case of qHash(), the 
> documentation for that one actualy says the include is <QHash>. I think 
> it is *not* reasonable to break that.

Is that relevant? My understanding is that what is being changed is not
that #include <QHash> declares qHash (which seems to me like it should -
and probably, must, since QHash itself needs qHash - continue to
happen), but that unrelated headers that incidentally used to include
qhash.h would newly include only qhashfunctions.h instead.

I don't object to that.

I also support the notion of a separate header *just* for qHash. Bonus
points if we get <QHashFunctions> :-).



Slightly off-topic: <QStringList> included <QObject>? Ick, that's one
change I'm glad to see.

-- 
Matthew