Re: Proper declaration for apr_wait_for_io_or_timeout() (was: Re: NWGNUmakefile update for APR1.2?)
On 11/11/05, Brad Nicholes <[EMAIL PROTECTED]> wrote: > > >>> On 11/11/2005 at 1:40:32 pm, in message > <[EMAIL PROTECTED]>, Jeff > Trawick > <[EMAIL PROTECTED]> wrote: > > On 11/11/05, Brad Nicholes <[EMAIL PROTECTED]> wrote: > >> Reposting to the APR list with a new subject line. Does this need > to > >> be taken care of in APR 1.2.x before httpd 2.2 ships? > > > > apr_wait_for_io_or_timeout() is private APR function (not API). Is > > the caller outside of APR? > > Yes, HTTPD. trunk/server/core_filters.c Simply shocking ;) That code needs to be fixed to call a real API :(
Re: svn commit: r327872 - in /httpd/httpd/trunk: CHANGES include/httpd.h server/core_filters.c
On 10/23/05, [EMAIL PROTECTED] <[EMAIL PROTECTED]> wrote:
> Author: brianp
> Date: Sun Oct 23 15:38:35 2005
> New Revision: 327872
>
> URL: http://svn.apache.org/viewcvs?rev=327872&view=rev
> Log:
> New version of ap_core_output_filter that does nonblocking writes
...
> +static apr_status_t send_brigade_blocking(apr_socket_t *s,
> + apr_bucket_brigade *bb,
> + apr_size_t *bytes_written,
> + conn_rec *c)
> +{
...
> +if (APR_STATUS_IS_EAGAIN(rv)) {
> +rv = apr_wait_for_io_or_timeout(NULL, s, 0);
Whoops, not an API. This needs to use apr_poll().
Re: bug in mod_dav.c
On 11/05/2005 11:41 PM, Ruediger Pluem wrote: > [..cut..] > Good catch. Thank you. I also think that it should be err2->status instead > of err->status. I just checked in a patch to the trunk > (r331041, http://svn.apache.org/viewcvs.cgi?rev=331041&view=rev). > Let's wait if there are any objections. If not I will try to get it > backported. Now backported to 2.2.x (r332657, http://svn.apache.org/viewcvs.cgi?rev=332657&view=rev). and proposed for backport to 2.0.x (r332661, http://svn.apache.org/viewcvs.cgi?rev=332662&view=rev). Regards Rüdiger
Re: ap_get_server_port differences
Here's the quick workaround module. Just load the modules and it tries
to "revert" to 2.0 behavior
#include "httpd.h"
#include "http_config.h"
#include "ap_config.h"
#include "apr_strings.h"
#include "http_protocol.h"
module AP_MODULE_DECLARE_DATA port_module;
typedef struct {
apr_port_t port;
char *port_str;
} port_config;
static int port(request_rec *r)
{
port_config *conf;
if((conf = ap_get_module_config(r->server->module_config,
&port_module)) != NULL) {
if(conf->port_str) {
/*looking at server/core.c, ap_get_server_port This looks
to be the best place
* UseCanonicalName MUST be off, or this is ignored*/
r->parsed_uri.port = conf->port;
r->parsed_uri.port_str = conf->port_str;
}
}
return DECLINED;
}
static int post_config(apr_pool_t * pconf, apr_pool_t * plog,
apr_pool_t * ptemp, server_rec *s)
{
server_rec *virt;
port_config *port;
for(virt = s; virt; virt = virt->next) {
port = apr_pcalloc(pconf, sizeof(port_config));
if(virt->port) {
port->port = virt->port;
port->port_str = apr_psprintf(pconf, "%u", virt->port);
ap_set_module_config(virt->module_config, &port_module, port);
}
}
return OK;
}
static void register_hooks(apr_pool_t * p)
{
ap_hook_post_read_request(port, NULL, NULL, APR_HOOK_REALLY_FIRST);
ap_hook_post_config(post_config, NULL, NULL, APR_HOOK_MIDDLE);
}
/* Dispatch list for API hooks */
module AP_MODULE_DECLARE_DATA port_module = {
STANDARD20_MODULE_STUFF,
NULL, /* create per-dirconfig structures */
NULL,/* merge per-dirconfig structures */
NULL, /* create per-server config structures */
NULL, /* merge per-server config structures */
NULL,/* table of config file commands */
register_hooks /* register hooks */
};
--
Brian Akins
Lead Systems Engineer
CNN Internet Technologies
ap_get_server_port differences
in 2.1.9 ap_get_server_port uses the following: port = r->parsed_uri.port_str ? r->parsed_uri.port : r->connection->local_addr->port ? r->connection->local_addr->port : r->server->port ? r->server->port : ap_default_port(r); in 2.0.55 it uses: port = r->parsed_uri.port_str ? r->parsed_uri.port : r->server->port ? r->server->port : ap_default_port(r); This just bit me bad! I was used to being able to declare things like: ServerName www.domain.com:80 and having redirects (mod_dir, etc) to pick up on this via ap_construct_url What, if any, is the work around for 2.1.9? I really don't want to use "UseCanonicalName On" -- Brian Akins Lead Systems Engineer CNN Internet Technologies
mod_proxy balancer maxattempts bug??
If I have the following:
BalancerMember http://10.165.80.90:80 max=256 timeout=3
BalancerMember http://10.165.30.89:80 max=256 timeout=3
BalancerMember http://10.165.80.88:80 max=256 timeout=3
BalancerMember http://10.165.80.89:80 max=256 timeout=3
BalancerMember http://10.165.30.91:80 max=256 timeout=3
BalancerMember http://10.165.30.92:80 max=256 timeout=3
and later:
ProxyPass / balancer://fill/ timeout=10 maxattempts=3 lbmethod=byrequests
if all 6 members are unreachable, but still "OK", proxy tries each
member before giving up. should it only try 3?
maybe some different logic needs to be mod_proxy?
access_status = ap_proxy_pre_request(&worker, &balancer, r, conf, &url);
if (access_status != OK) {
/*
* Only return if access_status is not HTTP_SERVICE_UNAVAILABLE
* This gives other modules the chance to hook into the
* request_status hook and decide what to do in this situation.
*/
if (access_status != HTTP_SERVICE_UNAVAILABLE)
return access_status;
/*
* Ensure that balancer is NULL if worker is NULL to prevent
* potential problems in the post_request hook.
*/
if (!worker)
balancer = NULL;
goto cleanup;
}
if (balancer && balancer->max_attempts_set && !max_attempts)
max_attempts = balancer->max_attempts;
--
Brian Akins
Lead Systems Engineer
CNN Internet Technologies
Re: Building Apache 2.0.55
Paul Smedley wrote: Hi All, I'm trying to get Apache2 2.0.55 building and running on OS/2 using GCC 3.3.5. There's a current build of Apache using GCC 2.8.1 but that GCC is getting a bit long in the tooth so I'd like to update things :) Anyway, I have it building ok - using the unix network_io and poll code instead of the OS/2 code as GCC 3.3.5 on OS/2 provides a much more unix like behaviour. But what is GOOD about unix-like behavior? This is sadly the same nonsense as trying to build Win32 via cygwin. You have wrappers that work *almost* like unix, but you've added another layer of abstraction, and another library's bugs... [Thu Nov 10 19:51:18 2005] [error] (38)Socket operation on non-socket: apr_accept Ta da..., and boom. Once you can accept, can it poll/select correctly, or will you have yet-more quirks? You've evidently barely scratched the surface of making apr/unix play OS/2. The better solution, and nothing against attacking OS/2 using gcc, would be to modify the build system to correctly detect OS/2 and use that API, internally. The beauty of APR is that most everything is abstracted out and programs such as httpd won't know/care about the difference (and will be informed of supported/unsupported APIs.) Once apr is built correctly, there's very little to do on the httpd side. It has almost no platform specifics, and those are generally picked out correctly by autoconf. Please don't reply again on this thread. Jeff asked politely for you to continue on dev@apr.apache.org where issues of portability are addressed; please do. Bill
Re: Building Apache 2.0.55
HI Jeff, On Thu, 10 Nov 2005 11:51:26 UTC, Jeff Trawick <[EMAIL PROTECTED]> wrote: > On 11/10/05, Paul Smedley <[EMAIL PROTECTED]> wrote: > > Hi All, > > I'm trying to get Apache2 2.0.55 building and running on OS/2 using > > GCC 3.3.5. There's a current build of Apache using GCC 2.8.1 but that > > GCC is getting a bit long in the tooth so I'd like to update things :) > > > > Anyway, I have it building ok - using the unix network_io and poll > > code instead of the OS/2 code as GCC 3.3.5 on OS/2 provides a much > > more unix like behaviour. > > > [Thu Nov 10 19:51:18 2005] [error] (38)Socket operation on non-socket: > > apr_accept > > apparently it doesn't provide as much as is necessary... does it > still work with the OS/2 networking? do APR test programs work better > or worse with the unix networking implementation? > > you are essentially deciding to be an APR developer by switching the > networking implementation to something that nobody else has used > before > > (APR questions to dev@apr.apache.org) Yeah I did try the OS/2 networking implementation first and it shows in error_log: (OS 10047)Address family not supported by protocol family: make_sock: could not bind to address 0.0.0.0:81 [Fri Nov 11 17:46:47 2005] [alert] no listening sockets available, shutting down and there's no output to screen. -- Cheers, Paul.
