Comment 3 for bug 2029149

My only concern with the build-dep is that theoretically someone could build the clevis package on a system with a version of cryptsetup lower than 2.6.0. If we add a build-dep, that package would fail to build when realistically it could still work without that feature enabled. I think my main issue here is not that this compile time check should never be able to disable the feature as much as it was incorrectly disabled in this case. Would your preference be to have clevis fail to build in certain cases where it could theoretically work in some cases with that feature disabled? I can see that potentially avoiding problems from a clarity perspective but it seems like a less flexible solution.