Should the test module be cached during manual runs?

[rschristian]

Just tried to naively debug an object and ran into #2798 again, maxing out CPU usage & crashing the tab. Unfortunately due to the default caching policy, this was quite difficult to fix as Firefox would immediately load the cached version of the test module, hit that log, and once again CPU usage would skyrocket until the tab crashed.

And so I ask: is there even any benefit to caching the test module? Dependencies make sense, but the test module itself should never cause an issue to my knowledge and would avoid situations like this where WTR creates essentially a poisoned cache entry.