Comment 34 for bug 1811817

We have had multiple attempts at this now. When I first reviewed this, my build with the patch didn't fix the problem, whereas other builds did. Then it was accepted, and manual verification showed that it was not fixed. I think we need something in the package to make sure this doesn't happen again. I understand the arguments against dh-autoreconf, and that might introduce other changes we won't be aware of (a diff of a rendered configure file is very hard to read). So maybe we need something else.

Some ideas that come to mind:
a) a dep8 test. The [test plan] as is could be made into a dep8 test;
b) some sort of build-time check. Something simple like making sure the final binary is linked with libssl1.1.