Original Redmine Comment
Author Name: Stefan Tauner (@stefanct)
Original Date: 2017-09-07T09:16:58Z

Wilson Snyder wrote:

From https://www.veripool.org/boards/28/topics/1626?r=2332#message-2332

aaaand we are back :P Apparently there is a bug, where not enough cleanup is done... following commit fixes this.
uastw-embsys/Verilog-Perl@0ef1bbb

Stefan, nit - please file issues so I don't miss patches.

NP, I pondered about it but thought keeping it all together in the thread is better...

Anyhow when I patch and run 42_dumpcheck.t a bunch of times I sometimes get "Attempt to free unreferenced scalar: SV 0x23e4b30, Perl interpreter: .... during global destruction." so something is still unstable. Can you run it in a loop and take a look?

I can reproduce what you see in about 5% of the runs. However, it is not this patch that is producing this problem. I have git bisected this and can conclude that it was introduced already in 3d2ecdb Support for buses and concats in Netlist, msg1626.
Will look into it further now-ish...

Original Redmine Comment
Author Name: Stefan Tauner (@stefanct)
Original Date: 2017-09-07T16:10:53Z

That one hurts a bit... one work day led to the attached (and also pushed) patch :'(
The other patch is nevertheless needed to fix the other problem. Sorry!

Original Redmine Comment
Author Name: Wilson Snyder (@wsnyder)
Original Date: 2017-09-08T01:23:20Z

Excellent work, again.

Pushed both to git towards 3.441. I'll wait a day for this to settle then push out a release.

Original Redmine Comment
Author Name: Wilson Snyder (@wsnyder)
Original Date: 2017-09-09T01:56:27Z

In 3.442.