Den 15. sep. 2011 09:42, skrev ext Jedrzej Nowacki: > > On Wednesday 14. September 2011 14.50.37 ext > [email protected] wrote: > > > (...) > > > Disadvantages: > > > > > > * People have to clone the whole repo for just a few examples. (People > > > that compile Qt from source will have to live with this, if you use the > > > SDK this should not impact these people) * If the API of a module > > > changes and an example breaks it will not be caught by the CI system the > > > moment the change is going into mainline. > > > (...) > > > Another disadvantage is that currently autotest and benchmarks use > demos, at least for QtScript. That was done not because we are lazy > and we didn't bother to reduce size of an use case :-) The idea behind > it was to not create only micro benchmarks that shows performance of a > method. We need a easy way to benchmark full code stack for typical > usage and I'm not sure if there is something better then benchmarking > a real, working application or demo. > > > Cheers, > > Jędrek > >
Yes, examples (not demos) are used for automatic testing. Same deal for QtDeclarative, which Alan and Alex already mentioned. This should be a compelling enough reason to keep them where they are. Regards, Kent _______________________________________________ Qt5-feedback mailing list [email protected] http://lists.qt.nokia.com/mailman/listinfo/qt5-feedback
