On Tue, Apr 26, 2016 at 5:01 PM, Yann Ylavic <ylavic....@gmail.com> wrote:
> On Tue, Apr 26, 2016 at 10:48 PM, William A Rowe Jr <wr...@rowe-clan.net> > wrote: > > On Tue, Apr 26, 2016 at 3:47 PM, William A Rowe Jr <wr...@rowe-clan.net> > > wrote: > >> > >> Reviewing the applicable patch, the client, request and protocol now > >> appear > > > > ... the client, vhost and protocol now appear ... > > > >> to persist for idle connections, but the request is still NULL? > >> > >> Is this by design? Or is something missing from the patch? > > Is it different behaviour than 2.4.18 (it shouldn't)? > > We are discussing different issues/solutions in PR 59333... > AIUI the previous behavior was to retain the prior request until the new keep-alive request began processing the post-read-request phase, until that point the previous request on that pipeline should still appear in the connection's status. I'm now already following that thread since this afternoon and have to compare notes against 2.4.17-.18-.20-post-patch behaviors. My gut instinct is to propose scoreboard.c for a full svn revert back to a working state, but will help if folks demonstrate some effort toward preserving ABI and functional state compatibility... at least over the next few days. Cheers, Bill