Actually, I prefer this option, too. The changes in my fork is substantial. Merging now will inevitably delay release of 0.18 for who knows how long. Splitting the changes cost time, and risk new bugs, and worse, incomplete features.
What are these implications for you? You have 50k lines of difficult to split code. In 3-4 mo is it 70k lines?
The idea was to make a "quick" FreeCAD 0.18 release that ports to QT5/PY3.
For the rest of us, we should start a NEW thread about working towards an accelerated release.
As packagers we need to support @realthunder's branch (I'm not willing to call it a fork). As testers we need to test @realthunder's branch.
I'll float again the idea of an "official" LinkStage3 branch. It would make it easier, but it won't happen overnight, to off load the test-build burden from @realthunder. Maybe @realthunder prefers more control over when there is a release?
PLEASE, PLEASE, PLEASE discuss a new release in a new topic.