pin <voidpin%protonmail.com@localhost> writes: >> If you are planning to import as qt6 while not changing qt5, and people >> can build/install them both at the same time, then I don't think it's >> critical exactly when. >> >> (If you are talking about replacing 5.15 with 6.x, then that's much >> scarier and we should have some kind of bulk builds with them staged, >> as I expect a lot would break.) > > I was talking about the former, maybe should have made it clear from the start. > > We simply can not replace Qt5 with Qt6, there're are too many packages > that have not migrated yet, a lot would break. There're are, though > some that provide support for both. In that case, adding the qt6 packages won't really break anything. I will try to build them on mac. > LXQt's next release maybe Qt6 hence, my interest. Sure, and in general more things will require it. > To think about, though. Would it be possible to drop Qt4? I honestly > don't know, so no need to get angry :) We went through this before and went probably further than we should have. If there are packages that depend on qt4 and their upstreams have releases that have moved to qt5 or whatever, then updating them certainly makes sense. Overall, though, I am very negative on "let's get rid of X" as a goal for package removals, as it tends to lead to deleting things that don't really meet the deletion criteria. If for some package we arrive at "this is very unmaintained upstream, problematic somehow, and we genuinely believe there are no users", that's something that can be proposed for deletion on pkgsrc-users. But please no "I want to remove qt4-libs" and from there recursing to deleting anything that uses it. I also don't think what's up with qt4 should have any bearing on qt6. A simple "find-depends x11/qt4-libs" yields 244 lines, and 155 of those are from pkgsrc proper. So it's pretty clearly out of the question to rm qt4 now.
Attachment:
signature.asc
Description: PGP signature