On Tue, 26 Oct 2010 20:49:32 +0200 Sorin Manolache <sor...@gmail.com> wrote:
> On Tue, Oct 26, 2010 at 20:36, Mike Meyer > <mwm-keyword-apache.b9a...@mired.org> wrote: > > On Tue, 26 Oct 2010 10:03:11 +0100 > > Nick Kew <n...@apache.org> wrote: > > > >> On 26 Oct 2010, at 03:52, Mike Meyer wrote: > >> > Anyone got some hints on how I ought to set this up? Maybe there's a > >> > writeup somewhere I missed on config processing Possibly there's some > >> > facility that I ought to be looking at to replace the mmaped file? > >> > >> One possible suggestion. post_config and child_init get passed > >> the first server_rec. The server_rec in any later hook is the > >> virtual host. Could your problem be setting data in one server > >> then trying to retrieve it from another? > > > > That indeed appears to be the case (Sorin got this also, pointing out > > that it comes from using a virtual server). > > > > Since I want this to work properly with multiple virtual servers mapping > > different files, I'd like this to work. To me, the obvious place to > > put this would be the config merge hook, except that doesn't get > > passed a copy of the server rec, and the pool it gets passed is > > apparently different (or possibly reset) between invocations, so the > > trick of using pool_userdata to skip doing this during the config > > check doesn't work. > > > > I don't see a hook that's obviously convenient for this. Maybe there's > > one with a non-obvious name? Barring that, it would seem that I have > > to have both my post_config hook and child_init hook walk through the > > servers checking their config to see if my module is enabled, and > > setting them up appropriately. > > I did something similar. As you say, I had to walk through the servers > in both hooks. I can lookup up the code tomorrow if you want. I'll ask if I have trouble figuring it out, or nobody provides a pointer. Thanks > I do not know the pool_userdata trick. What I did was to > apr_pool_cleanup_register a callback on the config pool that undid > what the post_config callback did. The config pool is cleaned up > between the two invocations of the post_config callback. That's documented here: http://wiki.apache.org/httpd/ModuleLife I don't believe any of the pool's but s->process->pool survive to the second invocation. <mike -- Mike Meyer <m...@mired.org> http://www.mired.org/consulting.html Independent Network/Unix/Perforce consultant, email for more information. O< ascii ribbon campaign - stop html mail - www.asciiribbon.org
