On Tue, Sep 01, 2020 at 03:11:59PM +0200, Ruediger Pluem wrote: > Your point is that there is no case where multiple threads work on the same > apr_memcache_conn_t at the same time, correct? > If yes, I agree that this is the case and that I missed this and hence my > point is mood.
I think that's right. I don't want to disappear too far down the rabbit hole for this, given that the *only* symptom we are seeing is the APR pool check failing. Is there any way in which APR is going to guarantee that the internal pool state is always consistent across threads if one thread is simultaneously inside e.g. apr_palloc()/apr_palloc_debug() and another calls apr_pool_find()? I can't see it. Regards, Joe
