On dimanche 27 octobre 2019 03:07:01 CET Friedrich W. H. Kossebau wrote: > Am Freitag, 25. Oktober 2019, 00:35:46 CET schrieb David Faure: > > On jeudi 24 octobre 2019 22:24:55 CEST Friedrich W. H. Kossebau wrote: > > > BTW, we want to also set > > > > > > -DQT_DEPRECATED_WARNINGS_SINCE=0x060000 > > > > > > otherwise the deprecations done with QT_DEPRECATED_VERSION in Qt's API > > > will > > > not emit warnings, because QT_DEPRECATED_WARNINGS_SINCE defaults to > > > QT_DISABLE_DEPRECATED_BEFORE if set. > > > > Oh, good to know. Done now for all of KF5. > > And the same would need to be done also for the KF variant, i.e. setting > -DKF_DEPRECATED_WARNINGS_SINCE=0x060000 > otherwise the use of KF_DISABLE_DEPRECATED_BEFORE_AND_AT will also result in > warnings for API deprecated only in newer versions.
Hmm, true. > Quite some explicit setup. But I made ECMGenerateExportHeader follow what Qt > does, as consistency rules here IMHO. > > Both QT_DEPRECATED_WARNINGS_SINCE & DKF_DEPRECATED_WARNINGS_SINCE could be > though set via KDEFrameworkCompilerSettings perhaps, that would match what > is done for -Wzero-as-null-pointer-constant already. That would make sense, good idea. I didn't want the DISABLE stuff to come from ECM, so we keep some control over it (and it's a good thing I did, see commit dd98b2b7175b in kimageformats) But for warnings, not much harm can happen, so that's a good idea. -- David Faure, fa...@kde.org, http://www.davidfaure.fr Working on KDE Frameworks 5