Diego wrote: > Unfortunately John Layt missed the Qt 4.6 feature freeze: > https://bugs.kde.org/show_bug.cgi?id=171925#c16 > > So the problem is that these features won't get merged until Qt 4.7, > which is quite far away from release... That really sucks. :-/ It's really crappy to turn down fine work done by contributors for ~9 more months just because it missed the feature freeze (which might not even have been publicly announced anywhere) by a mere week. I think Nokia really needs to show more flexibility there after their boasting about openness and community involvement. (That said, KDE has its problems there as well. Strict freezes suck! There needs to be more flexibility for small, 100% completed features.) > Shipping these patches in Fedora Qt packages would break the "ABI or API > Deviations" of the "staying close to upstream projects" policy rule, > right? https://fedoraproject.org/wiki/PackageMaintainers/WhyUpstream Backporting API additions (assuming they're accepted for 4.7) is fine: it only adds things, it doesn't change or remove anything, so our Qt would still be binary compatible with binaries built against the official one (we don't really care about the other way round), and the API will be in 4.7, so it wouldn't be Fedora-specific. (That said, there IS a risk that the API and/or the ABI will change upstream until the final 4.7 release, we'd have to talk to both John Layt and Qt Software to ascertain the API is the final one.) I'd be fine with that. I'm more worried about the added translatable strings, as they'll stay untranslated. :-( But we might just decide we don't care (like we did for the panel autohide backport to KDE 4.1 and the strings that introduced). That said, I'd prefer this getting merged into the kde-qt git branch (the new qt-copy), that way we'd pick it up automatically (and several other KDE distros would, too). But the KDE folks might complain about the translation issue as well. Kevin Kofler
