Drop gen_fsm in favor of gen_server. 1/4 code from poolboy.erl is gone.
Is it worth to be merged into master?
Change gen_fsm to gen_server behaviour
Fix wrong status
This even seems to pass the quickcheck tests, which is nice. I'll try to do some more review soon.
I'll try to do some benchmarking this week.
Actually, longer EQC runs are exposing some issues, I think there might be a bug lurking.
Any way to reproduce this tests?
Despite this issue, how about the idea? Currently i trying to add new functionality to poolboy, and it is much easier in gen_server version then in original.
The idea is great. As long as performance doesn't suffer, who can argue against simpler code?
Still trying to decide if the quickcheck problems are just line noise, sorry.
But I think we should pursue this branch, just want to make sure this change is solid first.
It looks like the EQC test had a timeout in it that was too short and it'd randomly timeout trying to do a checkout. I think we can ignore the failure and move forward with this.
@Vagabond I don't think there's any performance difference. As long as there are no regressions I'm going to merge this. Any last thoughts?
+1 from me.