Make thread names of CoroutineScheduler more distinguishable

[qwwdfsad]

Currently, names of all threads that belong to CoroutineScheduler (Dispatchers.Default, Dispatchers.IO and Dispatchers.IO.limitedParallelism(...) are named DefaultDispatcher-worker-* which might be quite confusing to look at in thread dumps (i.e. it's possible to have dozens of such threads on 8 core machine).
Example of such confusion: https://stackoverflow.com/questions/78502056/why-thread-name-is-defaultdispatcher-worker-even-though-i-specify-function-to-ru

While we cannot leverage Thread.setName because of its performance impact, we still can sweeten the pill:

• Name the thread somehow more generic (for example: CoroutineScheduler-worker-, something like shared-pool-worker-, coroutine-pool-worker- etc.)
• Add an artificial stackframe that indicates that the thread was spawned and/or is parked as part of the "IO" pool if it's possible implementation-wise

[ck2e14]

Agreed. As a relatively new Kotlin dev this definitely tripped me up. I should add that the Kotlin docs are really clear about this thread sharing behaviour, so thanks for that! That said... for semantics' sake:

Is it an overreach of naming if IO dispatcher and Default dispatcher share the pool? Why not call the threads 'shared-pool-worker-x' as qwwdfsad suggests, and then we can just name our scopes such that we know what the dispatcher was with a prefix or something?