I am able to reproduce with your code (adapted to run in my env, connected to a local prosody server). Removing the code to make the room persistent makes it easier to repro: I ran 4 times with your code and it got stuck twice and passed 2000 iterations without getting stuck twice. After commenting out the config form code it got stuck 6 times out of 6 (with at most 446 iterations).
This was with Smack 4.4.3, but I don’t think master would be different.
Thanks for jumping on it and making a fix so quickly!
I’ve updated PR #494 to target the 4.4 branch and plan to include this fix in the upcoming Smack 4.4.4 release. While I am also testing the changes, it would be great to hear from you if PR #494 fixes the issue for you. I can’t give an exact ETA of the 4.4.4 release, but probably within the next 14 days.
I also created SMACK-914 to track the issue.
@Flow do you have some artefacts for this PR published somewhere … I’m having har time building Smack. I cannot build it on my mac notebook and now some kotlin no-sense on an ubuntu server I’m trying to use just for building it … ?
I did upload a Smack 4.4.4-SNAPSHOT (4.4.4-20211011.112259-1) which includes the fixes at OSS Sonatypes Snapshot repository at Index of /repositories/snapshots/org/igniterealtime/smack
Yesterday I was able to reproduce the problem after executing the steps more than 200 times, but I repro. I’m in different network conditions now, compared to my previous attempts testing, seems that matters how often you repro
Today I updated with changes from those 4.4.4 snapshots and ran the test 600 times and cannot repro.
So it seems good to me!