Henry: > > This list aims to be specific enough to describe what actually would > > run in the continuous integration system, hence it mentions Ubuntu. > > In the release documentation or product documentation, we should be > > more generic, "extrapolate", and include the other non-reference > > platforms that have been verified.
Thiago: > I think that's doing it the other way around. The release documentation > should list exactly where it was tested, but the reference platforms should > be generic enough. Yes. The release documentation should list where it was tested -- but it can also document where it is reasonable to expect Qt to work (in other words it is OK to "extrapolate"). > For example, if someone writes a feature that compiles on Ubuntu but fails > to compile on Fedora, it's still in need of fixing. Yes, we would like it to be fixed. Should a contribution be reverted if it turns out that it brings up a bug in Fedora that is hard to work around? What are the distributions where we _require_ such bugs to be fixed (=reference configurations), as opposed to platform configurations where we would like Qt to work on? If we don't specify the reference configurations based on what runs in the CI system, then shouldn't we document them explicitly in some other way? Best regards, Henry _______________________________________________ Qt5-feedback mailing list [email protected] http://lists.qt.nokia.com/mailman/listinfo/qt5-feedback
