Devkits =/= retail to a degree. Testing Terraria in the broader sense is not at all like testing a "normal" game - functionally infinite variables, if that makes sense.
Your admittedly impressive edge case farms being one thing that a more mainline test play is not going to find... and even if they make a basic trap arena, they are not testing necessarily every possible permutation of things... and then there are some lines where its just hitting up against hardware issues. Not saying that is applicable to this particular setting, but just as an illustrative example.
So, often its a function of that - other times, who even knows? A lot falls to test cases - for example, the stuff I found Friday on the XB1 build on my retail version... it was clear how/why that slipped past in the hindsight, so those test cases have been adjusted to prevent that moving forward and the hotfix was put in for XB1 friday and it was fixed in advance for PS4.
RL itself is a small team, so we cannot look at 100% of every build across every platform. Have to rely on others to assist there, and they do.
Not making excuses, this is a daunting game to test... but the endgoal has to remain getting the cleanest experience possible.