Personally speaking, when reviewing, a mission will be played many times and in different ways, to test for competency. End triggers can be problematic when not used with special care, so a mission that does not end will be replayed more so to ensure it's not reviewer incompetence.
Where a mission still fails to end [or encounters other serious problems], as a last resort I'll unpack the .pbo and check the mission in the editor.
I don't think I've ever seen a mission lacking an end trigger - but I've seen many where far, far, too large a radius has been used, or conditions of activation have been in error. Sometimes it's even just the AI being a pain in the ... .
As for Resistance upgrades seriously affecting older missions - I've seen this first hand in my own, so it would certainly be a consideration where a mission appears affected.
Generally, though, reviewers are competent mission builders themselves, and appreciate the intricacies and issues involved.