https://bz.apache.org/bugzilla/show_bug.cgi?id=56729
--- Comment #16 from Yann Ylavic <[email protected]> --- Thanks Michael for the analyse and patches! (In reply to Michael Kaufmann from comment #13) > > But sometimes check_pipeline() is called *after* the connection has already > entered the keep-alive state. Then the branch "if (ccfg->in_keep_alive) ..." > is taken, and this branch does not check for speculative reads - that's the > bug. > > I don't know exactly how these two cases are triggered. Could you please provide a backtrace (gdb's "bt full"?) of this situation? > > It would be great if this bugfix and the previous bugfixes r1621453 / > r1641376 would be included in a future 2.4.x release. I'll propose a backport for the whole, but first the above trace would help determine the root cause (I'm not sure whether this should happen). -- You are receiving this mail because: You are the assignee for the bug. --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
