On Wed, Mar 7, 2018 at 4:39 PM, Mark Thomas <ma...@apache.org> wrote:
> > The issue (re)started occurring in that test when I added that new > parser. > > IMO, it's simply that it's not as responsive for dealing with the > connector > > pause that the test is testing. > > Sounds about right. I'd still like to understand exactly what is going > on - if only for my own peace of mind. The fix is probably a longer > delay but I'll dig into it some more - I might find another edge case or > two to fix along the way. > Yes, it might be a bug, I have no idea right now although the lower responsiveness to this sort of event is expected. Otherwise, the said parser has been stable in testing for me (like this h2spec test I bothered you with). Rémy
