How to return an error code/page from output filter?

[John Zhang]

When I encounter certain error conditions in my output filter, I would like to 
return a specific error code/page (eg 503 - service not available).  However, 
the following code results in a 500 beeing returned to the browser:

apr_status_t my_out_filter(ap_filter_t *f, apr_bucket_brigade *bb)
{

if (I_should_return_503)
{
ap_remove_output_filter(f);
// f-r-status = HTTP_SERVICE_UNAVAILABLE;
return HTTP_SERVICE_UNAVAILABLE;
}

}

In fact, no matter what error code I return, the browser always gets a 500 
error.  Also, if I uncomment the f-r-... line, the browser gets an empty 
(200) page.

Any suggestions?

Thanks in advance,
John

Can you set the request timeout dynamically within a module?

[John Zhang]

Hi Everyone,
We have an output filter.  Some requests take a VERY long time for the 
filter to process (I know many of you have questions here.  But the logic is in 
the page).  We would like to know if there is a way for the module (our output 
filter) to tell apache to not terminate THIS request.  We do not want to 
increase the default timeout.

   I looked at the server_rec struct, it has a timeout member (default = 300 
secs, yes this is a very long time).  Is this the one that controls the request 
timeout? If yes, I guess it controls every request.  And is it a good idea for 
a request to change the server_rec struct?  Even if it is OK to change it, will 
it take affect for the current request?

  Or I am looking at a wrong place/parameter.

Thanks in advance!
John

Re: Static compilation of a module

[John Zhang]

--- On Thu, 7/31/08, Brian J. France [EMAIL PROTECTED] wrote:

 At work we have moved away from using static libraries to
 shared  
 libraries because of duplicate symbols issues.

I have a smilar question: I am using apxs to compile my .so module on Linux.  
The module requires another lib.  When I used -c -l TheLib, I got a huge 
static .so file.  How do I generate a dynamic .so file? I tried -Wc,-static 
-Wl,-static, it gives a .a file instead.

Issue when saving a copy of the postdata in input filter

[John Zhang]

I have a module that saves a copy of the postdata in an input filter.  Here is 
how it works: In the ap_hook_fixups function, I check to see if I am interested 
in this request (examine url).  If yes, I then check if r-method_number == 
M_POST.  If yes, then I add my input filter function to capture the postdata.

All worked fine until someone did this:

conf:
Directory /test
DirectoryIndex index.php index.php3 index.html index.htm
/Directory

index.htm:
htmlbody
form action= method=post
  input type=text name=test/
  input type=submit/
 /form
/body/html

Now, load the above page with http://localhost/test/ (NOTE, without index.htm), 
and then sumbit the form.  By the time I decided if I should add my input 
filter for index.htm, the test r-method_number == M_POST is false (it is a 
M_GET).

Some debugging showed that the referer header is not null in this case, so I 
modified the test from
r-method_number == M_POST 
to
r-method_number == M_POST || apr_table_get(r-headers_in, Referer)

and in my input filter, I check to see if the 1st bucket is EOS before saving a 
copy of the postdata (because a non-null referer does not mean it is really a 
POST request.

Is this the right way?  I would like it to work on apaceh 2.*

Thanks
John

Re: enable override (allow htaccess use)

[John Zhang]

--- On Tue, 6/10/08, Andy Grant [EMAIL PROTECTED] wrote:

From: Andy Grant [EMAIL PROTECTED]
Subject: enable override (allow htaccess use)
To: modules-dev@httpd.apache.org
Date: Tuesday, June 10, 2008, 9:40 AM

How do you enable/disable the use of your directives in htaccess files?
I've read about context and overrides, but can't find them in reference
to developing a module, just in reference to using them (i.e.
AllowOverride FileInfo). Thanks. 

In your command table, when you specify your directive.  The 4th argument of 
the macros (eg AP_INIT_TAKE12(...)) specifies where/how your directive can be 
placed. See http_config.h for the meanings.

Here is one define from the .h file
#define OR_LIMIT 1   /** *.conf inside Directory or Location
and .htaccess when AllowOverride Limit */

Re: How do I know the character encoding?

[John Zhang]

--- On Sat, 5/3/08, Nick Kew [EMAIL PROTECTED] wrote:

 From: Nick Kew [EMAIL PROTECTED]
 Subject: Re: How do I know the character encoding?
 To: modules-dev@httpd.apache.org
 Date: Saturday, May 3, 2008, 5:17 AM

  In my output filter, I need to parse the document to
 search for
  certain patterns.
  
  Where can I get the information about the (character)
 encoding so
  that I can parse the document correctly?  Eg the
 document may contain
  unicode characters and are encoded in a special
 encoding. 
 
 See http://apache.webthing.com/mod_xml2enc/
 
 If your filter uses libxml2, just use mod_xml2enc alongside
 it.
 If not, you can still use the charset detection and
 transcoding.
 
Thanks Nick!
I've never used the modules you mentioned.  I will have a look at them.
Thanks again!
John

How do I know the character encoding?

[John Zhang]

In my output filter, I need to parse the document to search for certain 
patterns.

Where can I get the information about the (character) encoding so that I can 
parse the document correctly?  Eg the document may contain unicode characters 
and are encoded in a special encoding. 

Thanks,
John

perform a redirect in output filter

[John Zhang]

I am using apache 2.0/2.2, in my output filter, I
would like to have the ability to redirect.  Some
tests indicated that I can redirect to a page on the
same site, but not to an external site.

Here is my code for redirect:
char* location = 
apr_table_get(r-headers_out, Location);
//Note this Location is set by my filter during the
process, so I know I need to do a redirect.
if(location  location[0])
{
  //r-status = 301;
  r-method = apr_pstrdup(r-pool, GET);
  r-method_number = M_GET;
  apr_table_unset(r-headers_in, Content-Length);
  ap_remove_output_filter(my_filter);
  ap_internal_redirect_handler(location, r);
  return APR_SUCCESS;
}

//tests: (if my site is http://localhost/foo)
and when http://localhost/foo/bar.html (my test
page)is loaded I want to redirect it to
a) /foo/foo2/existing.html (works)
b) http://localhost/foo/foo2/existing.html (works)
c) http://www.google.com (does not work), instead I
get http://localhost/foo displayed on the browser.

Is there anything I did is incorrect?  I noticed the
method is called *internal_redirect*, but did not
know another one for external redirect.

Thanks,
John

Configure question -- allow access only from local machine

[John Zhang]

We have a situation that we would like to restrict the
access to certain folders only to requests from the
local machine.  Here is why:
When a page is processed by our filter, the filter
(based on page logic) may request pages (just like a
regular web page request) that should never go to the
browser.  We put these pages in a folder.  And would
like to use apache config to restrict the access to
only the local machine.  Here is the config
LocationMatch /secrete-stuff/
Order Deny,Allow
Deny from all
Allow from 127.0.0.1
#Allow from localhost
/LocationMatch

The issue we face:
When our filter issues the request, we use the
hostname from the original request. eg, original
request
http://1.2.3.4/index.html
our filter might issue
http://1.2.3.4/something/secrete-stuff/server.js

In order to make the above directive work, we will
have to put the ip (1.2.3.4) in the Allow section. 
However, we are planning to deply many servers, it
would be very hard for us to edit each config file. 
So we are wondering if there are anyway we can achieve
the same result without make ip-specific changes.

Thanks in advanvce for your help.
John

Re: Is there are any way to know if the request is regular (http) or SSL (https) in a module?

[John Zhang]

--- Graham Dumpleton [EMAIL PROTECTED]
wrote:

 On 18/12/2007, Sander Temme [EMAIL PROTECTED]
 wrote:
 
  On Dec 17, 2007, at 6:36 PM, Eric Covener wrote:
 
   I would like to know the request type in my
 module
   (handler/filter), is there any way to know that
 (HTTP
   vs HTTPS)?
  
   apr_table_get(r-subprocess_env, HTTPS)  might
 be  what you want
 
  That gets set in the Fixup hook, relatively late
 in the request
  processing process.  You could call
 ap_http_scheme(r) to run the
  request scheme hook, which will return https if
 SSL is disabled or
  optional on the vhost that accepted the request.
 
 Or better still, use the function registered by the
 ssl module for the purpose:
 
 #if AP_SERVER_MAJORVERSION_NUMBER = 2
 APR_DECLARE_OPTIONAL_FN(int, ssl_is_https, (conn_rec
 *));
 static APR_OPTIONAL_FN_TYPE(ssl_is_https)
 *wsgi_is_https = NULL;
 #endif
 
...
 
if (!wsgi_is_https)
 wsgi_is_https =
 APR_RETRIEVE_OPTIONAL_FN(ssl_is_https);
 
 if (wsgi_is_https 
 wsgi_is_https(r-connection))
 apr_table_set(r-subprocess_env, HTTPS,
 1);
 
 Ie., HTTPS is set by the above sort of process, but
 as pointed out
 that is done only in fixup phase, but you can call
 the ssl_is_https()
 function direct if in earlier phase.
 
Thanks everyone!  All very useful to me!

Re: Memory issue in apache modules

[John Zhang]

--- Joe Lewis [EMAIL PROTECTED] wrote:
 Your buckets can still be created using the
 request-pool .  My created
 buckets in my output filter are done that way.  Have
 you tried it?  You
 still use the request-connection-bucket_alloc for
 the other parameter,
 but the request-pool for the memory pool to be
 used.
Thanks Joe!

Is there are any way to know if the request is regular (http) or SSL (https) in a module?

[John Zhang]

I would like to know the request type in my module
(handler/filter), is there any way to know that (HTTP
vs HTTPS)?

Thanks,
John

Re: Memory issue in apache modules

[John Zhang]

--- Ray Morris [EMAIL PROTECTED] wrote:

  I am using the apr_* functions to allocate
  memory (most of the time from the request-pool).
 
If there are few places where you allocate from
 othr than the reqquest pool, I'd look at those
 first.

I used the bucket/brigade for my data that requires
the use of request-connection-bucket_alloc.  I read
a very old post stating that when keep alive is on
(which is true in my case), then the memory associated
with the connection will not be freed?  But no answer
to that post.  Is that a valid concern?  If so , how
can I reclaim the connection-related memory or use a
different memory pool?

Thanks,
John

Re: Memory issue in apache modules

[John Zhang]

--- Eric Covener [EMAIL PROTECTED] wrote:
 
 It should level out as each thread in the process
 has had a chance to
 run and allocate some memory for a range of
 requests.  MaxMemFree can
 be used to return the heap memory that would
 otherwise be used by
 subsequent requests on that thread.
 
Thanks Eric,
But I see consistent memory increase when I load
the SAME page repeatedly.  So something is wrong with
my module.  But I do not know what.

Thanks again.
John

Re: How does set status code in filter work

[John Zhang]

Thanks a lot, Joe!

--- Joe Lewis [EMAIL PROTECTED] wrote:

 John Zhang wrote:
  Joe,
  Thanks for you rreply, but I do not quite
  understand Your filter should run before the
 content
  type is set.  Could you elaberated on that? My
 filter
  is an output filter.

 
 The filter types are AP_FTYPE_ (followed by one of
 the following )
 RESOURCE, CONTENT_SET, PROTOCOL, TRANSCODE,
 CONNECTION or NETWORK.  The
 two most common ones are AP_FTYPE_RESOURCE, and
 AP_FTYPE_CONTENT_SET. 
 The difference between the two is one (RESOURCE)
 runs prior to the
 Content-Type header being set.  (Meaning they can
 change any of those
 headers).  The other (CONTENT_SET) runs after that
 header is set, and
 are not supposed to change that header (I expect
 there are other
 headers, such as the status, that may also be
 effected here, but I do
 not know for sure).  It is in your register_hook
 function that you
 declare when you want your filter to be run that
 those are specified.
 
 
  Also, if I issue a redirect in my (output)
 filter
  to (say /errors/HTTP_FORBIDDEN.html) and I am a
 filter
  for .html, will I be called in loading the error
 page?

 
 Error pages do not run in the standard filter chain
 by default.  If you
 want your filter running on the error HTML files in
 addition to others,
 you would have to use an error_filter hook as well.
 
 Joe
 -- 
 Joseph Lewis http://sharktooth.org/
 Divide the fire, and you will sooner put it out. -
 Publius Syrus
 

Re: How does set status code in filter work

[John Zhang]

Joe,
Thanks for you rreply, but I do not quite
understand Your filter should run before the content
type is set.  Could you elaberated on that? My filter
is an output filter.

Also, if I issue a redirect in my (output) filter
to (say /errors/HTTP_FORBIDDEN.html) and I am a filter
for .html, will I be called in loading the error page?

Thanks,
John

--- Joe Lewis [EMAIL PROTECTED] wrote:

 Your filter should run before the content type is
 set.  Then, you can do 
 the usual redirect to /errors/HTTP_FORBIDDEN.html or
 whatever document 
 you should be returning.  Yes, a filter should only
 be working on the 
 through-put, so you should be handling that. 
 Perhaps someone knows 
 about the pre-content-set filter types - I'm
 definitely not the expert.
 
 Joe
 
 John Zhang wrote:
  In an output filter.  If something is wrong, and I
 set
  the request object's status to an error code, then
 am
  I responsible for providing the proper error page
  content?  What is the correct way to handle errors
 in
  an output filter?
 
  Thanks,
  John

 
 

Re: How to issue a subrequest in my module

[John Zhang]

--- Joe Lewis wrote:

 John Zhang wrote:
  I have a module (act as a filter or handler). 
 While
  processing a request, I encountered a url (similar
 to
  an include, eg /foo/bar.xyz, or abc.txt).  How can
 I 
  1) ask apache to process this request and 
  2) give me the result back (I am waiting for
 it
  before I can continue)?

 I did is available on the web :
 

http://sharktooth.org/tutorials/whitepapers/apache_includes.html
 
Joe,
Thanks for the reply.  I will look into your
sample and see if I can do something similar.
Thanks Again!
John

How to issue a subrequest in my module

[John Zhang]

I have a module (act as a filter or handler).  While
processing a request, I encountered a url (similar to
an include, eg /foo/bar.xyz, or abc.txt).  How can I 
1) ask apache to process this request and 
2) give me the result back (I am waiting for it
before I can continue)?

I got an answer for a similar question, and have
looked at the mod_include, but have not figured out
how to do this: the mod_include has a include_ctx, and
it does a lot of work using that (eg mapping the
include uri to an absolute path).

Thanks in advance!
John

How to request a file in apache module?

[John Zhang]

I have the follow need:
While processing a request in my module (handler or
output filter), I need to ask apache to give me the
content of another page.  Is there any way to do that,
or do I have to find the file and read it just within
my module?

Thanks in Advance for any help!
John

Re: apache 2.0.61 do not read config settings

[John Zhang]

--- Joe Lewis wrote:

 John Zhang wrote:
  --- Joe Lewis wrote:

  Comment out the IfModule pieces surrounding your
  directive just to eliminate one more thing.
  
 
  Commenting out the IfModule does make it work!
 
  Now, how do I figure out the EXACT module name?
 
 When you compile your module using apxs, there could
 (or should) be a -n 
 option.  This is specifying the module name and is
 required in -g, but 
 guessed in a -i mode.  Just type man apxs for
 more information.  Use 
 that name when testing an IfModule.
 
Thanks for the help.  I never used apxs.  My module
was   compiled using Visual studio.  Through debugging
the apache code here is what I found out:
In apache 2.2x, I can say
LoadModule my_module the location of my
module/mod_my.so
IfModule my_module
...
/IfModule

but in 2.0, I have to change the IfModule to
IfModule mod_my.c

With IfModule my_module, the entire block is skipped
because it cannot find a model with name my_module.

Not sure if this the right way.  Any comments would be
preciated.

Thanks,
John

Re: apache 2.0.61 do not read config settings

[John Zhang]

--- William A. Rowe, Jr. wrote:

 Is your cmd array listed in your module data
 section?
 
yes.  Everything worked fine in apache 2.2.  The
structure looked like this:

static const command_rec mod_my_cmds[] = {
 AP_INIT_TAKEx(MyName, my_func, NULL, RSRC_CONF, my
comments,
...,
{NULL}
};

module AP_MODULE_DECLARE_DATA my_module = {
STANDARD20_MODULE_STUFF,
...
mod_my_cmds,
register_hooks
};

my_func is declared and defined somewhere.  But
appears never called.

Re: apache 2.0.61 do not read config settings

[John Zhang]

--- Joe Lewis wrote:
 Comment out the IfModule pieces surrounding your
 directive just to
 eliminate one more thing.  How do we know that the
 problem is the
 directive itself, when it could be that apxs
 compiled it with a
 different my_mod.c definition and the IfModule is
 what is really
 preventing your directive from being called?
 
 If that doesn't reveal anything additional, try
 providing us with the
 my_func function (and at least the declaration).
 
Thanks Joe,
Commenting out the IfModule does make it work! 
Now, how do I figure out the EXACT module name?  I am
still in the early stage of using apache.

Here are some of the config stuff

LoadModule my_module C:/foo/.../mod_my.so
IfModule my_module
   MyName foo
/IfModule

mod_my.so is compiled against apache 2.2 first (and
the above config structure worked).  When I tried
apache 2.0, I merely pointing the *.lib and *.h files
to the apache 2.0 folders (to generate the mod_my.so
for apache 2.0).

Thanks again for your help any further suggestions.
John

Re: Where is the right place to setup a connection pool?

[John Zhang]

Thanks Nick!
I am actually reading your book.  Will your
example on Thread Safety (4.5.1) work (setup the pool
in init_child hook with thread mutex)?  I do not need
a global mutex, right?  I have not read the late
chapters yet.

With Regards,
John

--- Nick Kew [EMAIL PROTECTED] wrote:

 On Wed, 31 Oct 2007 13:41:59 -0700 (PDT)
 John Zhang [EMAIL PROTECTED] wrote:
 
  Thanks in advance for any help.
 
 Use an apr_reslist to manage the pool.
 See mod_dbd for an example.
 
 -- 
 Nick Kew
 
 Application Development with Apache - the Apache
 Modules Book
 http://www.apachetutor.org/
 

apache 2.0.61 do not read config settings

[John Zhang]

I have a module for 2.2x, and have some config
settings like

IfModule my_module
...
MyName foo
...
/IfModule

and I can get the configurations per the AP_INIT_TAKEx
macro.  However, when I compiled my module against
apache 2.0.61, my functions for setting the parameters
are not called.

Anything specif I have to do to make the functions to
be called (so the config parameters are read).

Thanks,
John

Where is the right place to setup a connection pool?

[John Zhang]

I am writing a module talking to a remote server that
handles the real work.  I need to build/keep a pool of
connections for reuse almong the different Apache
worker threads.  Should I build that in the
ap_hook_post_config or ap_hook_child_init function?

When apache creates multiple processes, will each
process have its own copy of the connection pool
automatically (taken care by apache) or I will have to
do more work?

Thanks in advance for any help.

John