2011/4/5 Daniel Mack <zonque at gmail.com>: > Was anyone able to reproduce this? The test is called "connect-stress" > in the mainline tree. As I said on IRC, I was able to reproduce this on Ubuntu, but it needed two connect-stress running in parallel to get the daemon to fail. How are you coming along with investigating the source of the problem? May be a git bisect is helpful here to determine where the bug originated. I guess we are looking for a place where a message get pushed onto the queue with an uninitialized memchunk parameter. > Daniel Maarten
