It's important to be able to estimate the risk involved. That's called engineering. I know that the barbs and tubing attached to a cpu block when a pump has failed to start are definitely hot to the touch. As such I would little faith in the integrity of a push fit on the cpu block when the pump is not running. That you are indifferent and I am not is reflected in my bothering to research this and you telling me not to bother. Each to their own, it's my few minutes not yours.
Tubing becomes far more flexible when hot. Elongation comfortably triple what it was when cold. Load fails at 35 kg with warm water, fails at 30 with hot (sub boiling) water. Experimental difficulty in keeping the damned thing hot for long enough to load, so the 5kg decrease in capacity is probably at around 60 degrees. Noticeably hot when it splashed out, but not enough to scold.
This concludes the testing I can do, as I have no barbs. Compression fittings showing decreased load when hot supports the notion that push fitted tubing will suffer similar. I believe the decreased load carrying of push fitted barbs is worth investigating if one is not going to use clamps in addition to wall tension. I'm not going to investigate this as 1/I don't use barbs and 2/I don't have any barbs.
p.s. as a potential use, cutting threads in the top of a case and suspending it from a desk as a means of reducing vibrations. Possibly as a convenient fill port as well.
I'm not telling you to not do your 'tests' but your pushing parts to their limit and then saying they’re not safe when they obviously weren’t designed to accommodate those scenarios. If my pump fails then the PC WILL shut down before the water heats up enough to remove the tubing from the barb. All I'm saying is with 'my setup' it has been fine which then gives me every right to advise people about 'my' experience with 7/16 tubing and 1/2 fat boys. You do seem to be trying to scare monger though with your vigorous 'experiments' and deter people away from barbs.
What you've said about heat and the tubes elongation is correct but again you're talking about temperatures that you're simply not going to get (seatbelts / 500mph). But everyone’s system and setup is different. I have a TFC Quad Rad and a Heatkiller on CPU and Koolance on 4870x2. Temps for the CPU full load don't go above 60 and GPU full load at 38. I've touched the tubes and they aren’t even warm and that is after hours of gaming.
You obviously know more about all of this than I do (when it comes to the experimental side of things) but I'm just giving my opinion of my experiences so far with water cooling.
I’d love to know the fate of my PC then? Please tell me how it will die…
RoEy