On 5/6/14, 9:48 AM, Johan Corveleyn wrote: > Given the (non-reproducible) failure I had with upgrade_tests-29 with > 1.7.17 [1], I stopped testing and am a bit reluctant to sign the > release. I won't veto it either, but if I don't have at least a > hypothesis of why this failure happened, I'm not inclined to sign it. > > Whatever happened, it's unlikely that it's a regression (i.e. chances > are that I could also have run into this with 1.8.x or trunk), but > still ... I don't like it. > > Ideas welcome :-). > > [1] http://svn.haxx.se/dev/archive-2014-05/0037.shtml
I think this is a hardware issue on your test setup. See my reply on your other thread: http://svn.haxx.se/dev/archive-2014-05/0061.shtml
