On Thu, Nov 22, 2018 at 3:43 PM Max Kirillov <max@xxxxxxxxxx> wrote: > also edited the test to include only push_plain case, > and repeat it several times, to avoid running irrelevant > cases, the failure never happened again. as I explained previously[1] and as odd as it might seem the push_plain case ONLY fails if your run them together with the other tests that return errors with compressed input frankly I don't understand how one could affect the other as they should be running in independent processes but it happens fairly consistently in NetBSD (tested 7.1, 7.2, 8) with only one CPU (tested i386 and amd64) Carlo [1] https://public-inbox.org/git/20181119213924.GA2318@xxxxxxxxxxxxxxxxxxxxx/T/#m041e9703432c39dcb04fe10e86fc53d5254474b4
