Cliff Woolley wrote:
> Nevermind, scratch this off the list of suspects.
Thanks for checking into it. I also put a breakpoint on
apr_brigade_partition; it is never called in the new_features_1_3 /
mod_include problem.
A couple of recent discoveries: I wasn't hitting the seg fault 100% of
the time. Sometimes it worked fine, even with the "Accept-Language:
ru". But when it doesn't work, the debugger sees a SIGPIPE shortly
before the seg fault. That may be why no one has been able to re-create
it under controlled conditions. Perhaps the common denominator between
the Russian and Chinese clients is crummy lines.
Greg
