Hi everybody, https://bz.apache.org/bugzilla/show_bug.cgi?id=60956 seems to be a really interesting bug noticed by Frank Meier, full of data and really well written. My limited understanding of how mpm-event and the output filters chain work makes me think that start_lingering_close_nonblocking should be called by a worker thread, but there is probably a better solution.
Any ideas? Thanks! Luca
