At 02:14 AM 11/24/2004, Yussef Alkhamrichi wrote: >>But we have to decide what to do here. Either 1) the hosts aren't >>being torn down (and we need to), or 2) we need to avoid 'test >>creating' the hosts at initial startup. The gs (global server_rec) >>pointer is volatile upon subsequent config phases. > >I've had a good look at the code and tested some minor changes. But I've a >question: In the method 'asp_net_pre_config' we are making sure that global >conf is persistent. If the global conf is persistent consequently all hosts in >conf->hosts should be created (and thus have ASP.NET configured) exactly once. >I ask this because you mentioned the test creating by the way Apache >functions, I don't see in which phase the testcreating takes place.
If you look here, in function main() for calls to ap_read_config http://svn.apache.org/viewcvs.cgi/httpd/httpd/trunk/server/main.c?rev=106103&view=markup you will see that we read the config, run the pre, open_logs and post config phases once before the for () loop, and each time the server gracefully restarts within that loop. But you will notice the process->pool is created only once, up front, and persists for the lifetime of this process. >A related question is wether or not the data set by apr_pool_userdata_setn >holds during the two passes of configurating Apache as was mentioned before ? Not against the configuration pool. But the process pool lives much longer. This is why we root out the conf->pool's parent, our process pool. Bill