On Mon, 28 Sep 2026 14:27:52 +0200,
Stefan Sperling <stsp@stsp.name> wrote:
> 
> Not wrong, it's a small leak, but I believe this leak is temporary.
> We only leak it until the interface goes down because qwz_core_deinit()
> will call qwz_free_peers(). So I don't think this is a big problem.
> These timeout cases should be rare edge cases anyway.
> 

Now I lost.

It is not big problem because it leaks in edge case and only on stop.

But, as far as I understnad the code it stops in case of leak, and we never
free that peer.

Anyway, to move forward I fine to drop that commit if you make review simpler.

-- 
wbr, Kirill