On 8/22/2024 5:22 PM, Greg Wilkins via jetty-users wrote:
> With previous non-deterministic failures in classloading like this, it
> has often walked,quacked and looked like a race... but in the end has
> often turned out to be an ordering issue. If an unordered collection is
> being used somewhere, then when it is initialized can often affect the
> iteration order. Even listing items from the file system can have
> strange ordering behaviours.
Little out of my depth here, sorry; do you mean the ordering of the
specific object being iterated over that is returning the
InaccessibleObjectException exception?
I don't mean anything specifically, as I've not looked in detail at this specific issue.
I just browsed and saw some concern about a possible "race condition" in the start up of jetty - which sparked my interest. So I was just suggesting that non-deterministic iteration order can often present as a non-deterministic failure (especially with loading and initialization) that can look like a race. So all I'm saying is don't assume a race because sometimes it works and sometimes it fails. It is just a general comment I'm making.
Simone is our JPMS expert and has looked at this, so I'm not going to delve further... unless there actually is a real suggestion of a race, non determinism or other jetty issue.... but I warn you... any work I do on JPMS related stuff comes with a 45+ minute rant about what a waste of time/effort/electrons the whole thing is, how it is just make-work that does nothing for you other than fail itself rather than find/fix any real dependency problems in the code. :)
cheers
-- 