Re: The mod_python wiki has materialized!

2006-10-12 Thread Graham Dumpleton
Its late here now, so time for me to sleep. You being on the PMC might also carry more weight in getting the Confluence wiki space setup. If we follow suggested convention, should be called MODPYTHON to match JIRA tag. You being on the PMC means you can probably be given admin rights over the

Re: httpd 2.2 cache - disable and enable

2006-10-12 Thread Bart van der Schans
Matthieu Estrade wrote: Colm MacCarthaigh a écrit : I agree on this point, but people using this directive choose to loose time on this regexp check. The problem i have is many people wants to enable cache on .gif and .jpeg and .pdf, but they don't know the exact path. Wouldn't it be easier

Re: Issue with persistent http proxy backend connection

2006-10-12 Thread Ruediger Pluem
On 10/11/2006 11:09 PM, Jeff Trawick wrote: On 10/11/06, Ruediger Pluem [EMAIL PROTECTED] wrote: 2. Although GET is mentioned to be idempotent in RFC 2616 (9.1.2) along with some other methods it is not a good idea to regard a GET / HEAD with query parameters to be idempotent.

how to retrieve the contents of the POST method through the C API?

2006-10-12 Thread Manish Chakravarty
Hi All, I want to retrieve the contents of a (non HTML) POST request. (Eg some XML file or some images) I am developing on apache 1.3.34. Could someone please tell me how to go about it? I just want the contents of the messgae minus the header. -- Warm Regards, Manish Chakravarty

Re: Issue with persistent http proxy backend connection

2006-10-12 Thread Jeff Trawick
On 10/12/06, Ruediger Pluem [EMAIL PROTECTED] wrote: On 10/11/2006 11:09 PM, Jeff Trawick wrote: On 10/11/06, Ruediger Pluem [EMAIL PROTECTED] wrote: 2. Although GET is mentioned to be idempotent in RFC 2616 (9.1.2) along with some other methods it is not a good idea to regard a GET /

Re: Issue with persistent http proxy backend connection

2006-10-12 Thread Ruediger Pluem
On 10/11/2006 11:09 PM, Jeff Trawick wrote: On 10/11/06, Ruediger Pluem [EMAIL PROTECTED] wrote: 3. Sometimes servers (including httpd) include a keep-alive header in their response with the parameter timeout set. This can give a hint when the backend will close its persistent

Re: Issue with persistent http proxy backend connection

2006-10-12 Thread Mladen Turk
Ruediger Pluem wrote: But yes, in theory their remains a race here if the proxy reuses the connection faster then the backend needs time to close the socket after sending the response. Does the problem exists with the current trunk as well? The way how trunk detects the closed backend has

Re: Issue with persistent http proxy backend connection

2006-10-12 Thread Jeff Trawick
On 10/12/06, Ruediger Pluem [EMAIL PROTECTED] wrote: On 10/11/2006 11:09 PM, Jeff Trawick wrote: On 10/11/06, Ruediger Pluem [EMAIL PROTECTED] wrote: 3. Sometimes servers (including httpd) include a keep-alive header in their response with the parameter timeout set. This can give a

Re: The mod_python wiki has materialized!

2006-10-12 Thread Graham Dumpleton
Anyone had any thoughts on how we are going to use the wiki? From prior comments it looks like we can't use it for the mod_python documentation if we intend to then ship a snapshot of the documentation with a release. I am not sure we are precluded from still using it for the documentation, it

Re: The mod_python wiki has materialized!

2006-10-12 Thread Graham Dumpleton
I spoke too soon. The ASF has: http://cwiki.apache.org/confluence/homepage.action Now I am confused as to what roles MoinMoin fills versus the Confluence wiki. Can anyone fill me in on the roles of each? Graham On 12/10/2006, at 9:49 PM, Graham Dumpleton wrote: Anyone had any thoughts on

Re: httpd 2.2 cache - disable and enable

2006-10-12 Thread Matthieu Estrade
Bart van der Schans a écrit : Matthieu Estrade wrote: Colm MacCarthaigh a écrit : I agree on this point, but people using this directive choose to loose time on this regexp check. The problem i have is many people wants to enable cache on .gif and .jpeg and .pdf, but they don't know the

Re: The mod_python wiki has materialized!

2006-10-12 Thread Jim Gallacher
Graham Dumpleton wrote: Anyone had any thoughts on how we are going to use the wiki? Sections 4, 5 and 6 (API, Apache Configuration Directives and Standard Handlers) of the current docs stay with in the source distribution. Everything else would be a candidate for the wiki. (We should likely

Segfault creating detached thread in module

2006-10-12 Thread Gregor Melhorn
Hi all, I've got a problem with creating a detached thread in the post init hook. static void register_hooks(apr_pool_t *p) { ap_hook_post_config(spread_init, NULL, NULL, APR_HOOK_MIDDLE); } void* APR_THREAD_FUNC doit(apr_thread_t *thd, void *data){ server_rec *s = (server_rec *)

Re: The mod_python wiki has materialized!

2006-10-12 Thread Graham Dumpleton
Doing some digging into the Confluence wiki site, it seems we may be better off getting a wiki space created in there for mod_python which would be specifically for developing the official documentation. This could have restricted write access for core developers. That wiki space can then be

Share data between modules

2006-10-12 Thread [EMAIL PROTECTED]
Hi, I'd like to share an apr_hash_t between different modules. In my apache 2.2 I have three different modules: each one should parse an xml file at startup and create a hash table. Because the xml configuration file is the same for all my modules, I'd like to share my apr_hash_t. So my idea was

Re: [mod_python] Python 2.5 binaries for Windows

2006-10-12 Thread Nicolas Lehuen
Hi Paul,I have built binaries for Python 2.5 and Apache 2.0 and 2.2 :http://nicolas.lehuen.com/download/mod_pythonDo make sure that you download the version that matches your Apache version, as they are incompatible. I've ran the unit test successfully except for the two known failures related to

Re: httpd 2.2 cache - disable and enable

2006-10-12 Thread Bart van der Schans
Matthieu Estrade wrote: Bart van der Schans a écrit : Matthieu Estrade wrote: It could be great to do it by type, but it can have a big side effect. Usually, resources not identified by extension are resources delivering content based on dynamic information. If these dynamic information

Win32 server register_cleanup failures.

2006-10-12 Thread Graham Dumpleton
On 12/10/2006, at 10:47 PM, Nicolas Lehuen wrote:I've ran the unit test successfully except for the two known failures related to server.register_cleanup and apache.register_cleanup, so those binaries can be officially released.Can you remind me what Apache actually does on Win32 for those

Re: The mod_python wiki has materialized!

2006-10-12 Thread Jim Gallacher
+1 Sounds like a good plan. Jim Graham Dumpleton wrote: Doing some digging into the Confluence wiki site, it seems we may be better off getting a wiki space created in there for mod_python which would be specifically for developing the official documentation. This could have restricted write

Re: Share data between modules

2006-10-12 Thread Nick Kew
On Thursday 12 October 2006 13:41, [EMAIL PROTECTED] wrote: So my idea was that one of the three modules parses the file, creates the hash_table_t and populates it. Then each child of these modules should be able to read (only read e not write) the hash table. Is it possible in your opinion?

Re: Win32 server register_cleanup failures.

2006-10-12 Thread Nicolas Lehuen
It crashes the Apache 2.2 process, and works correctly with Apache 2.0.If I get some time soon (not very probable) I'll try again building a full debug version of Python + Apache 2.2 + mod_python 3.2.10 in order to track down where the bug appears. Regards,Nicolas2006/10/12, Graham Dumpleton

Re: Issue with persistent http proxy backend connection

2006-10-12 Thread Ruediger Pluem
On 12.10.2006 13:26, Mladen Turk wrote: Ruediger Pluem wrote: But yes, in theory their remains a race here if the proxy reuses the connection faster then the backend needs time to close the socket after sending the response. Does the problem exists with the current trunk as well? The

Re: Share data between modules

2006-10-12 Thread [EMAIL PROTECTED]
Hi Nick, On Thursday 12 October 2006 13:41, [EMAIL PROTECTED] wrote: So my idea was that one of the three modules parses the file, creates the hash_table_t and populates it. Then each child of these modules should be able to read (only read e not write) the hash table. Is it possible in

RE: Developing C++ modules with apache 1.3.34

2006-10-12 Thread Kratzer, James \(Xetron\)
Hi, Here is what I'm doing with Apache 2.0. I use autoconf and libtool to create my makefiles and to build my apache modules. I then put all of your C++ functions is a separate source file having a cpp extension and I wrap the c++ functions with extern C declarations. I then call the c++

Re: Share data between modules

2006-10-12 Thread Nick Kew
On Thursday 12 October 2006 14:37, [EMAIL PROTECTED] wrote: Hi Nick, On Thursday 12 October 2006 13:41, [EMAIL PROTECTED] wrote: So my idea was that one of the three modules parses the file, creates the hash_table_t and populates it. Then each child of these modules should be able to

Re: Share data between modules

2006-10-12 Thread [EMAIL PROTECTED]
Maybe I have to use it In module A I declare apr_hash_t; in the post config I creates and populates it. Then I write a funcion called my_function and in the register_hooks I write: APR_REGISTER_OPTIONAL_FN(my_function); Then in my file.h I write: APR_DECLARE_OPTIONAL_FN(..., my_function,

Re: Issue with persistent http proxy backend connection

2006-10-12 Thread Jim Jagielski
Ruediger Pluem wrote: I do not think that this matters all too much, because the backend closes the connection *immediately* after sending out the response. So the socket connection check on proxy side that is executed before reusing a connection will detect that this socket has been closed

Re: Issue with persistent http proxy backend connection

2006-10-12 Thread Jim Jagielski
Jeff Trawick wrote: On 10/12/06, Ruediger Pluem [EMAIL PROTECTED] wrote: On 10/11/2006 11:09 PM, Jeff Trawick wrote: On 10/11/06, Ruediger Pluem [EMAIL PROTECTED] wrote: 2. Although GET is mentioned to be idempotent in RFC 2616 (9.1.2) along with some other methods it is

Re: how to retrieve the contents of the POST method through the C API?

2006-10-12 Thread Eric Covener
On 10/12/06, Manish Chakravarty [EMAIL PROTECTED] wrote: I want to retrieve the contents of a (non HTML) POST request. (Eg some XML file or some images) I am developing on apache 1.3.34. Could someone please tell me how to go about it? I just want the contents of the messgae minus the header.

Re: Segfault creating detached thread in module

2006-10-12 Thread Gregor Melhorn
UPDATE: It very likely is a memory management bug. I found out I can create a detached thread without problems, if it is finished before the init function exits! Running the code below works if sleep(1) in doit() is commented out. Else it segfaults immediately. Any guesses on what to change

Re: httpd 2.2 cache - disable and enable

2006-10-12 Thread Brian Akins
wrote: Wouldn't it be easier to do a match on mime type, like ExpiresByType? we do not know the mime type in quick_handler. In theory, you could disable/enable the CACHE_SAVE filter by mime type, but that seems a little messy because we would check in quick_handler to see if its cached.

Re: httpd 2.2 cache - disable and enable

2006-10-12 Thread Brian Akins
Matthieu Estrade wrote: IMHO, a regexp based cache enable or disable could be very usefull for a default caching policy shipped with httpd. We could do per default caching only on all images, css and all static content. Some random thoughts: Personally, I think the cache rules matching

Bad file descriptor: apr_socket_accept: (client socket) on graceful-stop

2006-10-12 Thread Tal Cohen
Hi, I am new to this group, so please be gentle :) Recently, I installed apache 2.2.3 on a Crux 2.1 (Linux Kernel 2.6.11.7). Now, every time I do an apachectl graceful or apachectl graceful-stop I get [error] (9)Bad file descriptor: apr_socket_accept: (client socket) in the apache

Re: Issue with persistent http proxy backend connection

2006-10-12 Thread Henrik Nordstrom
tor 2006-10-12 klockan 13:19 +0200 skrev Ruediger Pluem: I do not think that this matters all too much, because the backend closes the connection *immediately* after sending out the response. To help this, perhaps there should be a check just before sending the response as well, and send