Hi, I'm still exchanging patches and test results with James, but I will
commit the current development status to branches/2.2 and trunk/ because
I feel I'm not too far from fixing the long running bug #57501.
I found another sore point in the child process exit handler where the
Rivet channel could be unregistered several times (one for each virtual
host) whereas the Rivet Channel is now unique per child process unless
SeparateChannels is on. This condition is now handled consistently.
I wonder if this could be the culprit for the segfaults Brice was
observing during his tests. Brice, when I commit this stuff in
branches/2.2 would you give it a try? Thanks
-- Massimo
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
