I defer the review to @tcberner and @adridg who were already looking at it before. Meanwhile, I've landed the space->tab changes in pyqt.mk separately in rP458034.

lgtm, but the PORTREVISION bump can be dropped, as this change is just cosmetic.

In this case, you don't need to bump PORTREVISION: the file being patched is not installed, the generated code should remain mostly the same and the build isn't really changed in the platforms where it was already working before.

In D13702#286800, @tcberner wrote:

Make eigen only a build depend in

• graphics/movit [1]
• biology/iqtree [2]

[1] http://libregraphicsworld.org/blog/entry/introducing-movit-free-library-for-gpu-side-video-processing
[2] http://www.iqtree.org/doc/Compilation-Guide

It'd be good if you could take a look at the 2 ports that claim to have a runtime dependency on Eigen: unless they're actually building programs themselves, it doesn't make much sense for them to runtime-depend on a header-only library.

no objections

lgtm, no need to update the patch here if you fix the typo I pointed out.

Personally I don't see a huge improvement here, nor do I see why boolean CMake variables should be treated differently from the rest, but no objections if you feel like adding this.

I'm fine as long as @tcberner is fine :)

I agree. Even though there's been only one change since 3.9.3, an exp-run wouldn't hurt to avoid headaches.

lgtm from me once tcberner's fine with the patch.

thanks for working on this!