. ap_mpm_query() can return APR_EGENERAL if called too early (for debugging a module)
. some hypothetical module which implements the mpm-query hook can bypass the MPM
  with APR_ENOTIMPL


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/trunk@760864 13f79535-47bb-0310-9956-ffa450edef68

  makes no sense with loadable MPMs
axe the unnecessary MPM_DIR build variable
move the stray APACHE_SUBST(MPM_LIB) from configure.in to server/mpm/config.m4, where
  the other MPM_foos are handled


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/trunk@760403 13f79535-47bb-0310-9956-ffa450edef68

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/trunk@760379 13f79535-47bb-0310-9956-ffa450edef68

When a script returns a no-body response code, pretend it was a HEAD
request so as to drop any body the script might erroneously generate.


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/trunk@760167 13f79535-47bb-0310-9956-ffa450edef68

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/trunk@760026 13f79535-47bb-0310-9956-ffa450edef68

mention SIGQUIT/3 not being processed by the signal handler.  

SIGQUIT may unfortunately be a go-to signal for java users.



git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/trunk@759957 13f79535-47bb-0310-9956-ffa450edef68

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/trunk@759954 13f79535-47bb-0310-9956-ffa450edef68

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/trunk@759946 13f79535-47bb-0310-9956-ffa450edef68

consistency with existing APIs (e.g., slotmem)

don't bother changing the MMN, which doesn't reflect recent -dev changes anyway


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/trunk@759924 13f79535-47bb-0310-9956-ffa450edef68

The watchdog has zero interval, leaving to the callback to determine the running loop.

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/trunk@759862 13f79535-47bb-0310-9956-ffa450edef68

This allows to break the blocked callbacks

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/trunk@759860 13f79535-47bb-0310-9956-ffa450edef68

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/trunk@759849 13f79535-47bb-0310-9956-ffa450edef68

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/trunk@759847 13f79535-47bb-0310-9956-ffa450edef68

     returns EINPROGRESS and a subsequent poll() returns only POLLERR.
     Observed on HP-UX. 



git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/trunk@759832 13f79535-47bb-0310-9956-ffa450edef68

the mpm_get_child_pid hook is unnecessary, as was the per-MPM MPM_CHILD_PID() macro which it replaced

axe this new hook, and use ap_get_scoreboard_process() instead


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/trunk@759757 13f79535-47bb-0310-9956-ffa450edef68

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/trunk@759751 13f79535-47bb-0310-9956-ffa450edef68

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/trunk@759746 13f79535-47bb-0310-9956-ffa450edef68

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/trunk@759713 13f79535-47bb-0310-9956-ffa450edef68

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/trunk@759711 13f79535-47bb-0310-9956-ffa450edef68

Windows continues to specify this in the MPM, at least until someone
can separate out the MPM-specific bits.


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/trunk@759699 13f79535-47bb-0310-9956-ffa450edef68

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/trunk@759698 13f79535-47bb-0310-9956-ffa450edef68

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/trunk@759682 13f79535-47bb-0310-9956-ffa450edef68

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/trunk@759681 13f79535-47bb-0310-9956-ffa450edef68

    (hb_server_sort): Use correct variables for sorting.

Noticed by: Rüdiger


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/trunk@759673 13f79535-47bb-0310-9956-ffa450edef68

Now, dunno if that was intended, but at least allow to start the httpd

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/trunk@759648 13f79535-47bb-0310-9956-ffa450edef68

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/trunk@759647 13f79535-47bb-0310-9956-ffa450edef68

  apu-config which is gone in APR 2.0.


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/trunk@759594 13f79535-47bb-0310-9956-ffa450edef68

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/trunk@759518 13f79535-47bb-0310-9956-ffa450edef68

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/trunk@759492 13f79535-47bb-0310-9956-ffa450edef68

Translated by: Nilgün Belma Bugüner <nilgun belgeler.org>
Reviewed by:  Orhan Berent <berent belgeler.org>

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/trunk@759491 13f79535-47bb-0310-9956-ffa450edef68

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/trunk@759484 13f79535-47bb-0310-9956-ffa450edef68

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/trunk@759476 13f79535-47bb-0310-9956-ffa450edef68

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/trunk@759469 13f79535-47bb-0310-9956-ffa450edef68

- Use brigade_write instead of immortal buckets that aren't immortal.
- Read HTTP status code from serf.
- Remove extra logging.


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/trunk@759462 13f79535-47bb-0310-9956-ffa450edef68

Convert the 100ms timed callback to a single cleanup callback that is added
when the request is ready to finish. Basically works, though it has
some issues with flushing and closing the connection.


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/trunk@759460 13f79535-47bb-0310-9956-ffa450edef68

completely async from the original thread.

Right now it uses a 100ms timer to determine when the request is actaully complete,
but I think this can be removed in the long run with a better mechanism.


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/trunk@759414 13f79535-47bb-0310-9956-ffa450edef68

Event MPM.

Add a new MPM Query to determine if an MPM supports this.


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/trunk@759413 13f79535-47bb-0310-9956-ffa450edef68

* modules/proxy/mod_serf.c
    (serf_config_t): Add preservehost member.
    (setup_request): If preservehost is set, use the client provided Host header,
        otherwise use the one from the configuration.
    (is_true): New helper function for decoding true/false strings.
    (add_pass): Change to an argv configuration function, check for not enough
        args, and parse everything after the URI into key/value pairs.
    (create_dir_config): Default to setting preservehost to on.
    (serf_cmds): Change add_pass to a take argv directive.


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/trunk@759395 13f79535-47bb-0310-9956-ffa450edef68

    (hb_list_servers): Allows relative paths for the heartbeat file.


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/trunk@759389 13f79535-47bb-0310-9956-ffa450edef68

Add Clustered proxying support to mod_serf, by using the heartbeats system.

No preconfiguration of cluster members is needed.

Just a config like this:
    SerfCluster sweet heartbeat file=/var/cache/apache/hb.dat
    SerfCluster sour heartbeat file=/var/cache/apache/cluster2.dat
    <Location "/">
      SerfPass cluster://sweet
    </Location>
    <Location "/different_cluster">
      SerfPass cluster://sour
    </Location>

The location of all possible destination servers is provided by a new 
providers interface, that includes configuration checking of the arguments to 
the SerfCluster command, solving one of the worst problems with the mod_proxy 
load balancer subsystem.


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/trunk@759386 13f79535-47bb-0310-9956-ffa450edef68