It took me surprisingly long to realize that reverb not only puts early reflections and reverb tail onto your material, but actually alters the sound as a whole!
You might have known that already, but I was thinking for a long time that a good reverb is supposed to leave the original sound as “untouched” as possible, and that if it didn’t, the reverb just wasn’t a good one. That’s not the case. Whether reverb alters frequencies a lot or not is a matter of what kind of reverb we’re talking about and not so much about quality. Furthermore I learnt that reverb is not some kind of add-on, it’s actually something that the sound signal goes through with changes being made. (The fact that reverb is often handled by a parallel effect channel, suggests otherwise, but it’s kind of deceptive).
Which leads to a simple, yet important architectural decision: not to put the only master EQ in your mix after the master reverb (at least not without a good reason), because as soon as you alter the reverb, you alter the overall sound so much that your carefully selected EQ settings will no longer be appropriate for the mix. In order to have a more “stable” mix, put the reverb towards the end of the effects chain.
This has another advantage as well, because EQing the reverb-ed signal would make you more likely to cut away or manipulate parts of the sound that are integral to the reverb impression. Unless you really know what you’re doing that’s probably not something you’re after.
My approach: I have one EQ in the master mix bus as an insert BEFORE my overall reverb. That means I use the master EQ to level out the signal before it goes into the master reverb. When I change reverb settings or even the reverb/impulse response itself (I occasionally do), chances are high that those EQ settings will still be valid and need not be reconsidered.