rng2cpp.cpp build fail with qt5.5

[Boudewijn Rempt]

On Thu, 28 May 2015, Friedrich W. H. Kossebau wrote:

> Hi Treeve,
>
> thanks for reporting the error and working on a patch :)
>
> Am Donnerstag, 28. Mai 2015, 13:31:29 schrieb Treeve Jelbert:
>> thanks, the following diff fixes it
>> 
>> diff --git a/devtools/rng2cpp/rng2cpp.cpp
>> b/devtools/rng2cpp/rng2cpp.cpp
>> index 3dc3214..de65ac4 100644
>> --- a/devtools/rng2cpp/rng2cpp.cpp
>> +++ b/devtools/rng2cpp/rng2cpp.cpp
>> @@ -23,12 +23,10 @@
>>   #include <QtGlobal>
>>   #include <QSharedPointer>
>>   #include <QStringList>
>> +#include <QSet>
>> +#include <QVector>
>
> Aha, this is why there was an offset of +2 in the build log you posted :)
> But, why did you add those two lines? Are they left overs from some 
> experiment, or do you think they are needed to fix this build problem you 
> reported?

I think the extra includes might be what does the trick: the Qt people 
removed some transitive #includes in the latest Qt5, blithely disregarding 
that that'd break code all over the world.

>
>> -#if QT_VERSION < 0x040800
>> -#   define assert(cond, what) (Q_ASSERT_X(cond, "",
>> qPrintable(QString(what))))
>> -#else
>> -#   define assert(cond, what) (Q_ASSERT_X(cond, "", qPrintable(what)))
>> -#endif
>> +#   define assert(cond, what) Q_ASSERT_X(cond, "", qPrintable(what))
>>
>>   static const QString ns = "writeodf";
>
> Can you explain why this fixes it? Is it because QT_VERSION no longer works 
> with Qt 5.5.0, or is it the removed outer () in the replacement the culprit? 
> Why exactly did it fail before? Can you tell how all the macros are resolved 
> in the end with Qt 5.5.0
>
> Cheers
> Friedrich
> _______________________________________________
> calligra-devel mailing list
> calligra-devel at kde.org
> https://mail.kde.org/mailman/listinfo/calligra-devel