1 <?xml version="1.0" encoding="ISO-8859-1"?>
2 <!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd">
3 <html xmlns="http://www.w3.org/1999/xhtml" lang="en" xml:lang="en"><head><!--
4 XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
5 This file is generated from xml source: DO NOT EDIT
6 XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
8 <title>Apache 1.3 API notes - Apache HTTP Server</title>
9 <link href="../style/css/manual.css" rel="stylesheet" media="all" type="text/css" title="Main stylesheet" />
10 <link href="../style/css/manual-loose-100pc.css" rel="alternate stylesheet" media="all" type="text/css" title="No Sidebar - Default font size" />
11 <link href="../style/css/manual-print.css" rel="stylesheet" media="print" type="text/css" />
12 <link href="../images/favicon.ico" rel="shortcut icon" /></head>
13 <body id="manual-page"><div id="page-header">
14 <p class="menu"><a href="../mod/">Modules</a> | <a href="../mod/directives.html">Directives</a> | <a href="../faq/">FAQ</a> | <a href="../glossary.html">Glossary</a> | <a href="../sitemap.html">Sitemap</a></p>
15 <p class="apache">Apache HTTP Server Version 2.3</p>
16 <img alt="" src="../images/feather.gif" /></div>
17 <div class="up"><a href="./"><img title="<-" alt="<-" src="../images/left.gif" /></a></div>
19 <a href="http://www.apache.org/">Apache</a> > <a href="http://httpd.apache.org/">HTTP Server</a> > <a href="http://httpd.apache.org/docs/">Documentation</a> > <a href="../">Version 2.3</a> > <a href="./">Developer Documentation</a></div><div id="page-content"><div id="preamble"><h1>Apache 1.3 API notes</h1>
21 <p><span>Available Languages: </span><a href="../en/developer/API.html" title="English"> en </a></p>
24 <div class="warning"><h3>Warning</h3>
25 <p>This document has not been updated to take into account changes made
26 in the 2.0 version of the Apache HTTP Server. Some of the information may
27 still be relevant, but please use it with care.</p>
30 <p>These are some notes on the Apache API and the data structures you have
31 to deal with, <em>etc.</em> They are not yet nearly complete, but hopefully,
32 they will help you get your bearings. Keep in mind that the API is still
33 subject to change as we gain experience with it. (See the TODO file for
34 what <em>might</em> be coming). However, it will be easy to adapt modules
35 to any changes that are made. (We have more modules to adapt than you
38 <p>A few notes on general pedagogical style here. In the interest of
39 conciseness, all structure declarations here are incomplete -- the real
40 ones have more slots that I'm not telling you about. For the most part,
41 these are reserved to one component of the server core or another, and
42 should be altered by modules with caution. However, in some cases, they
43 really are things I just haven't gotten around to yet. Welcome to the
46 <p>Finally, here's an outline, to give you some bare idea of what's coming
47 up, and in what order:</p>
51 <a href="#basics">Basic concepts.</a>
54 <li><a href="#HMR">Handlers, Modules, and
57 <li><a href="#moduletour">A brief tour of a
63 <a href="#handlers">How handlers work</a>
66 <li><a href="#req_tour">A brief tour of the
67 <code>request_rec</code></a></li>
69 <li><a href="#req_orig">Where request_rec structures come
72 <li><a href="#req_return">Handling requests, declining,
73 and returning error codes</a></li>
75 <li><a href="#resp_handlers">Special considerations for
76 response handlers</a></li>
78 <li><a href="#auth_handlers">Special considerations for
79 authentication handlers</a></li>
81 <li><a href="#log_handlers">Special considerations for
82 logging handlers</a></li>
86 <li><a href="#pools">Resource allocation and resource
90 <a href="#config">Configuration, commands and the like</a>
93 <li><a href="#per-dir">Per-directory configuration
96 <li><a href="#commands">Command handling</a></li>
98 <li><a href="#servconf">Side notes --- per-server
99 configuration, virtual servers, <em>etc</em>.</a></li>
104 <div id="quickview"><ul id="toc"><li><img alt="" src="../images/down.gif" /> <a href="#basics">Basic concepts</a></li>
105 <li><img alt="" src="../images/down.gif" /> <a href="#handlers">How handlers work</a></li>
106 <li><img alt="" src="../images/down.gif" /> <a href="#pools">Resource allocation and resource pools</a></li>
107 <li><img alt="" src="../images/down.gif" /> <a href="#config">Configuration, commands and the like</a></li>
109 <div class="top"><a href="#page-header"><img alt="top" src="../images/up.gif" /></a></div>
110 <div class="section">
111 <h2><a name="basics" id="basics">Basic concepts</a></h2>
112 <p>We begin with an overview of the basic concepts behind the API, and how
113 they are manifested in the code.</p>
115 <h3><a name="HMR" id="HMR">Handlers, Modules, and Requests</a></h3>
116 <p>Apache breaks down request handling into a series of steps, more or
117 less the same way the Netscape server API does (although this API has a
118 few more stages than NetSite does, as hooks for stuff I thought might be
119 useful in the future). These are:</p>
122 <li>URI -> Filename translation</li>
123 <li>Auth ID checking [is the user who they say they are?]</li>
124 <li>Auth access checking [is the user authorized <em>here</em>?]</li>
125 <li>Access checking other than auth</li>
126 <li>Determining MIME type of the object requested</li>
127 <li>`Fixups' -- there aren't any of these yet, but the phase is intended
128 as a hook for possible extensions like <code class="directive"><a href="../mod/mod_env.html#setenv">SetEnv</a></code>, which don't really fit well elsewhere.</li>
129 <li>Actually sending a response back to the client.</li>
130 <li>Logging the request</li>
133 <p>These phases are handled by looking at each of a succession of
134 <em>modules</em>, looking to see if each of them has a handler for the
135 phase, and attempting invoking it if so. The handler can typically do one
139 <li><em>Handle</em> the request, and indicate that it has done so by
140 returning the magic constant <code>OK</code>.</li>
142 <li><em>Decline</em> to handle the request, by returning the magic integer
143 constant <code>DECLINED</code>. In this case, the server behaves in all
144 respects as if the handler simply hadn't been there.</li>
146 <li>Signal an error, by returning one of the HTTP error codes. This
147 terminates normal handling of the request, although an ErrorDocument may
148 be invoked to try to mop up, and it will be logged in any case.</li>
151 <p>Most phases are terminated by the first module that handles them;
152 however, for logging, `fixups', and non-access authentication checking,
153 all handlers always run (barring an error). Also, the response phase is
154 unique in that modules may declare multiple handlers for it, via a
155 dispatch table keyed on the MIME type of the requested object. Modules may
156 declare a response-phase handler which can handle <em>any</em> request,
157 by giving it the key <code>*/*</code> (<em>i.e.</em>, a wildcard MIME type
158 specification). However, wildcard handlers are only invoked if the server
159 has already tried and failed to find a more specific response handler for
160 the MIME type of the requested object (either none existed, or they all
163 <p>The handlers themselves are functions of one argument (a
164 <code>request_rec</code> structure. vide infra), which returns an integer,
168 <h3><a name="moduletour" id="moduletour">A brief tour of a module</a></h3>
169 <p>At this point, we need to explain the structure of a module. Our
170 candidate will be one of the messier ones, the CGI module -- this handles
171 both CGI scripts and the <code class="directive"><a href="../mod/mod_alias.html#scriptalias">ScriptAlias</a></code> config file command. It's actually a great deal
172 more complicated than most modules, but if we're going to have only one
173 example, it might as well be the one with its fingers in every place.</p>
175 <p>Let's begin with handlers. In order to handle the CGI scripts, the
176 module declares a response handler for them. Because of <code class="directive"><a href="../mod/mod_alias.html#scriptalias">ScriptAlias</a></code>, it also has handlers for the
177 name translation phase (to recognize <code class="directive"><a href="../mod/mod_alias.html#scriptalias">ScriptAlias</a></code>ed URIs), the type-checking phase (any
178 <code class="directive"><a href="../mod/mod_alias.html#scriptalias">ScriptAlias</a></code>ed request is typed
179 as a CGI script).</p>
181 <p>The module needs to maintain some per (virtual) server information,
182 namely, the <code class="directive"><a href="../mod/mod_alias.html#scriptalias">ScriptAlias</a></code>es in
183 effect; the module structure therefore contains pointers to a functions
184 which builds these structures, and to another which combines two of them
185 (in case the main server and a virtual server both have <code class="directive"><a href="../mod/mod_alias.html#scriptalias">ScriptAlias</a></code>es declared).</p>
187 <p>Finally, this module contains code to handle the <code class="directive"><a href="../mod/mod_alias.html#scriptalias">ScriptAlias</a></code> command itself. This particular
188 module only declares one command, but there could be more, so modules have
189 <em>command tables</em> which declare their commands, and describe where
190 they are permitted, and how they are to be invoked.</p>
192 <p>A final note on the declared types of the arguments of some of these
193 commands: a <code>pool</code> is a pointer to a <em>resource pool</em>
194 structure; these are used by the server to keep track of the memory which
195 has been allocated, files opened, <em>etc.</em>, either to service a
196 particular request, or to handle the process of configuring itself. That
197 way, when the request is over (or, for the configuration pool, when the
198 server is restarting), the memory can be freed, and the files closed,
199 <em>en masse</em>, without anyone having to write explicit code to track
200 them all down and dispose of them. Also, a <code>cmd_parms</code>
201 structure contains various information about the config file being read,
202 and other status information, which is sometimes of use to the function
203 which processes a config-file command (such as <code class="directive"><a href="../mod/mod_alias.html#scriptalias">ScriptAlias</a></code>). With no further ado, the
206 <div class="example"><p><code>
207 /* Declarations of handlers. */<br />
209 int translate_scriptalias (request_rec *);<br />
210 int type_scriptalias (request_rec *);<br />
211 int cgi_handler (request_rec *);<br />
213 /* Subsidiary dispatch table for response-phase <br />
214 * handlers, by MIME type */<br />
216 handler_rec cgi_handlers[] = {<br />
217 <span class="indent">
218 { "application/x-httpd-cgi", cgi_handler },<br />
223 /* Declarations of routines to manipulate the <br />
224 * module's configuration info. Note that these are<br />
225 * returned, and passed in, as void *'s; the server<br />
226 * core keeps track of them, but it doesn't, and can't,<br />
227 * know their internal structure.<br />
230 void *make_cgi_server_config (pool *);<br />
231 void *merge_cgi_server_config (pool *, void *, void *);<br />
233 /* Declarations of routines to handle config-file commands */<br />
235 extern char *script_alias(cmd_parms *, void *per_dir_config, char *fake,
238 command_rec cgi_cmds[] = {<br />
239 <span class="indent">
240 { "ScriptAlias", script_alias, NULL, RSRC_CONF, TAKE2,<br />
241 <span class="indent">"a fakename and a realname"},<br /></span>
246 module cgi_module = {
247 </code></p><pre> STANDARD_MODULE_STUFF,
248 NULL, /* initializer */
249 NULL, /* dir config creator */
250 NULL, /* dir merger */
251 make_cgi_server_config, /* server config */
252 merge_cgi_server_config, /* merge server config */
253 cgi_cmds, /* command table */
254 cgi_handlers, /* handlers */
255 translate_scriptalias, /* filename translation */
256 NULL, /* check_user_id */
257 NULL, /* check auth */
258 NULL, /* check access */
259 type_scriptalias, /* type_checker */
262 NULL /* header parser */
265 </div><div class="top"><a href="#page-header"><img alt="top" src="../images/up.gif" /></a></div>
266 <div class="section">
267 <h2><a name="handlers" id="handlers">How handlers work</a></h2>
268 <p>The sole argument to handlers is a <code>request_rec</code> structure.
269 This structure describes a particular request which has been made to the
270 server, on behalf of a client. In most cases, each connection to the
271 client generates only one <code>request_rec</code> structure.</p>
273 <h3><a name="req_tour" id="req_tour">A brief tour of the request_rec</a></h3>
274 <p>The <code>request_rec</code> contains pointers to a resource pool
275 which will be cleared when the server is finished handling the request;
276 to structures containing per-server and per-connection information, and
277 most importantly, information on the request itself.</p>
279 <p>The most important such information is a small set of character strings
280 describing attributes of the object being requested, including its URI,
281 filename, content-type and content-encoding (these being filled in by the
282 translation and type-check handlers which handle the request,
285 <p>Other commonly used data items are tables giving the MIME headers on
286 the client's original request, MIME headers to be sent back with the
287 response (which modules can add to at will), and environment variables for
288 any subprocesses which are spawned off in the course of servicing the
289 request. These tables are manipulated using the <code>ap_table_get</code>
290 and <code>ap_table_set</code> routines.</p>
293 <p>Note that the <code>Content-type</code> header value <em>cannot</em>
294 be set by module content-handlers using the <code>ap_table_*()</code>
295 routines. Rather, it is set by pointing the <code>content_type</code>
296 field in the <code>request_rec</code> structure to an appropriate
297 string. <em>e.g.</em>,</p>
298 <div class="example"><p><code>
299 r->content_type = "text/html";
303 <p>Finally, there are pointers to two data structures which, in turn,
304 point to per-module configuration structures. Specifically, these hold
305 pointers to the data structures which the module has built to describe
306 the way it has been configured to operate in a given directory (via
307 <code>.htaccess</code> files or <code class="directive"><a href="../mod/core.html#directory"><Directory></a></code> sections), for private data it has built in the
308 course of servicing the request (so modules' handlers for one phase can
309 pass `notes' to their handlers for other phases). There is another such
310 configuration vector in the <code>server_rec</code> data structure pointed
311 to by the <code>request_rec</code>, which contains per (virtual) server
312 configuration data.</p>
314 <p>Here is an abridged declaration, giving the fields most commonly
317 <div class="example"><p><code>
318 struct request_rec {<br />
321 conn_rec *connection;<br />
322 server_rec *server;<br />
324 /* What object is being requested */<br />
327 char *filename;<br />
329 </code></p><pre>char *args; /* QUERY_ARGS, if any */
330 struct stat finfo; /* Set by server core;
331 * st_mode set to zero if no such file */</pre><p><code>
332 char *content_type;<br />
333 char *content_encoding;<br />
335 /* MIME header environments, in and out. Also, <br />
336 * an array containing environment variables to<br />
337 * be passed to subprocesses, so people can write<br />
338 * modules to add to that environment.<br />
340 * The difference between headers_out and <br />
341 * err_headers_out is that the latter are printed <br />
342 * even on error, and persist across internal<br />
343 * redirects (so the headers printed for <br />
344 * <code class="directive"><a href="../mod/core.html#errordocument">ErrorDocument</a></code> handlers will have
348 table *headers_in;<br />
349 table *headers_out;<br />
350 table *err_headers_out;<br />
351 table *subprocess_env;<br />
353 /* Info about the request itself... */<br />
355 </code></p><pre>int header_only; /* HEAD request, as opposed to GET */
356 char *protocol; /* Protocol, as given to us, or HTTP/0.9 */
357 char *method; /* GET, HEAD, POST, <em>etc.</em> */
358 int method_number; /* M_GET, M_POST, <em>etc.</em> */
361 /* Info for logging */<br />
363 char *the_request;<br />
364 int bytes_sent;<br />
366 /* A flag which modules can set, to indicate that<br />
367 * the data being returned is volatile, and clients<br />
368 * should be told not to cache it.<br />
373 /* Various other config info which may change<br />
374 * with .htaccess files<br />
375 * These are config vectors, with one void*<br />
376 * pointer for each module (the thing pointed<br />
377 * to being the module's business).<br />
380 </code></p><pre>void *per_dir_config; /* Options set in config files, <em>etc.</em> */
381 void *request_config; /* Notes on *this* request */</pre><p><code>
387 <h3><a name="req_orig" id="req_orig">Where request_rec structures come from</a></h3>
388 <p>Most <code>request_rec</code> structures are built by reading an HTTP
389 request from a client, and filling in the fields. However, there are a
393 <li>If the request is to an imagemap, a type map (<em>i.e.</em>, a
394 <code>*.var</code> file), or a CGI script which returned a local
395 `Location:', then the resource which the user requested is going to be
396 ultimately located by some URI other than what the client originally
397 supplied. In this case, the server does an <em>internal redirect</em>,
398 constructing a new <code>request_rec</code> for the new URI, and
399 processing it almost exactly as if the client had requested the new URI
402 <li>If some handler signaled an error, and an <code>ErrorDocument</code>
403 is in scope, the same internal redirect machinery comes into play.</li>
405 <li><p>Finally, a handler occasionally needs to investigate `what would
406 happen if' some other request were run. For instance, the directory
407 indexing module needs to know what MIME type would be assigned to a
408 request for each directory entry, in order to figure out what icon to
411 <p>Such handlers can construct a <em>sub-request</em>, using the
412 functions <code>ap_sub_req_lookup_file</code>,
413 <code>ap_sub_req_lookup_uri</code>, and <code>ap_sub_req_method_uri</code>;
414 these construct a new <code>request_rec</code> structure and processes it
415 as you would expect, up to but not including the point of actually sending
416 a response. (These functions skip over the access checks if the
417 sub-request is for a file in the same directory as the original
420 <p>(Server-side includes work by building sub-requests and then actually
421 invoking the response handler for them, via the function
422 <code>ap_run_sub_req</code>).</p>
427 <h3><a name="req_return" id="req_return">Handling requests, declining, and returning
429 <p>As discussed above, each handler, when invoked to handle a particular
430 <code>request_rec</code>, has to return an <code>int</code> to indicate
431 what happened. That can either be</p>
434 <li><code>OK</code> -- the request was handled successfully. This may or
435 may not terminate the phase.</li>
437 <li><code>DECLINED</code> -- no erroneous condition exists, but the module
438 declines to handle the phase; the server tries to find another.</li>
440 <li>an HTTP error code, which aborts handling of the request.</li>
443 <p>Note that if the error code returned is <code>REDIRECT</code>, then
444 the module should put a <code>Location</code> in the request's
445 <code>headers_out</code>, to indicate where the client should be
446 redirected <em>to</em>.</p>
449 <h3><a name="resp_handlers" id="resp_handlers">Special considerations for response
451 <p>Handlers for most phases do their work by simply setting a few fields
452 in the <code>request_rec</code> structure (or, in the case of access
453 checkers, simply by returning the correct error code). However, response
454 handlers have to actually send a request back to the client.</p>
456 <p>They should begin by sending an HTTP response header, using the
457 function <code>ap_send_http_header</code>. (You don't have to do anything
458 special to skip sending the header for HTTP/0.9 requests; the function
459 figures out on its own that it shouldn't do anything). If the request is
460 marked <code>header_only</code>, that's all they should do; they should
461 return after that, without attempting any further output.</p>
463 <p>Otherwise, they should produce a request body which responds to the
464 client as appropriate. The primitives for this are <code>ap_rputc</code>
465 and <code>ap_rprintf</code>, for internally generated output, and
466 <code>ap_send_fd</code>, to copy the contents of some <code>FILE *</code>
467 straight to the client.</p>
469 <p>At this point, you should more or less understand the following piece
470 of code, which is the handler which handles <code>GET</code> requests
471 which have no more specific handler; it also shows how conditional
472 <code>GET</code>s can be handled, if it's desirable to do so in a
473 particular response handler -- <code>ap_set_last_modified</code> checks
474 against the <code>If-modified-since</code> value supplied by the client,
475 if any, and returns an appropriate code (which will, if nonzero, be
476 USE_LOCAL_COPY). No similar considerations apply for
477 <code>ap_set_content_length</code>, but it returns an error code for
480 <div class="example"><p><code>
481 int default_handler (request_rec *r)<br />
483 <span class="indent">
487 if (r->method_number != M_GET) return DECLINED;<br />
488 if (r->finfo.st_mode == 0) return NOT_FOUND;<br />
490 if ((errstatus = ap_set_content_length (r, r->finfo.st_size))<br />
491 ||
492 (errstatus = ap_set_last_modified (r, r->finfo.st_mtime)))<br />
493 return errstatus;<br />
495 f = fopen (r->filename, "r");<br />
497 if (f == NULL) {<br />
498 <span class="indent">
499 log_reason("file permissions deny server access", r->filename, r);<br />
500 return FORBIDDEN;<br />
504 register_timeout ("send", r);<br />
505 ap_send_http_header (r);<br />
507 if (!r->header_only) send_fd (f, r);<br />
508 ap_pfclose (r->pool, f);<br />
514 <p>Finally, if all of this is too much of a challenge, there are a few
515 ways out of it. First off, as shown above, a response handler which has
516 not yet produced any output can simply return an error code, in which
517 case the server will automatically produce an error response. Secondly,
518 it can punt to some other handler by invoking
519 <code>ap_internal_redirect</code>, which is how the internal redirection
520 machinery discussed above is invoked. A response handler which has
521 internally redirected should always return <code>OK</code>.</p>
523 <p>(Invoking <code>ap_internal_redirect</code> from handlers which are
524 <em>not</em> response handlers will lead to serious confusion).</p>
527 <h3><a name="auth_handlers" id="auth_handlers">Special considerations for authentication
529 <p>Stuff that should be discussed here in detail:</p>
532 <li>Authentication-phase handlers not invoked unless auth is
533 configured for the directory.</li>
535 <li>Common auth configuration stored in the core per-dir
536 configuration; it has accessors <code>ap_auth_type</code>,
537 <code>ap_auth_name</code>, and <code>ap_requires</code>.</li>
539 <li>Common routines, to handle the protocol end of things, at
540 least for HTTP basic authentication
541 (<code>ap_get_basic_auth_pw</code>, which sets the
542 <code>connection->user</code> structure field
543 automatically, and <code>ap_note_basic_auth_failure</code>,
544 which arranges for the proper <code>WWW-Authenticate:</code>
545 header to be sent back).</li>
549 <h3><a name="log_handlers" id="log_handlers">Special considerations for logging
551 <p>When a request has internally redirected, there is the question of
552 what to log. Apache handles this by bundling the entire chain of redirects
553 into a list of <code>request_rec</code> structures which are threaded
554 through the <code>r->prev</code> and <code>r->next</code> pointers.
555 The <code>request_rec</code> which is passed to the logging handlers in
556 such cases is the one which was originally built for the initial request
557 from the client; note that the <code>bytes_sent</code> field will only be
558 correct in the last request in the chain (the one for which a response was
561 </div><div class="top"><a href="#page-header"><img alt="top" src="../images/up.gif" /></a></div>
562 <div class="section">
563 <h2><a name="pools" id="pools">Resource allocation and resource pools</a></h2>
564 <p>One of the problems of writing and designing a server-pool server is
565 that of preventing leakage, that is, allocating resources (memory, open
566 files, <em>etc.</em>), without subsequently releasing them. The resource
567 pool machinery is designed to make it easy to prevent this from happening,
568 by allowing resource to be allocated in such a way that they are
569 <em>automatically</em> released when the server is done with them.</p>
571 <p>The way this works is as follows: the memory which is allocated, file
572 opened, <em>etc.</em>, to deal with a particular request are tied to a
573 <em>resource pool</em> which is allocated for the request. The pool is a
574 data structure which itself tracks the resources in question.</p>
576 <p>When the request has been processed, the pool is <em>cleared</em>. At
577 that point, all the memory associated with it is released for reuse, all
578 files associated with it are closed, and any other clean-up functions which
579 are associated with the pool are run. When this is over, we can be confident
580 that all the resource tied to the pool have been released, and that none of
581 them have leaked.</p>
583 <p>Server restarts, and allocation of memory and resources for per-server
584 configuration, are handled in a similar way. There is a <em>configuration
585 pool</em>, which keeps track of resources which were allocated while reading
586 the server configuration files, and handling the commands therein (for
587 instance, the memory that was allocated for per-server module configuration,
588 log files and other files that were opened, and so forth). When the server
589 restarts, and has to reread the configuration files, the configuration pool
590 is cleared, and so the memory and file descriptors which were taken up by
591 reading them the last time are made available for reuse.</p>
593 <p>It should be noted that use of the pool machinery isn't generally
594 obligatory, except for situations like logging handlers, where you really
595 need to register cleanups to make sure that the log file gets closed when
596 the server restarts (this is most easily done by using the function <code><a href="#pool-files">ap_pfopen</a></code>, which also arranges for the
597 underlying file descriptor to be closed before any child processes, such as
598 for CGI scripts, are <code>exec</code>ed), or in case you are using the
599 timeout machinery (which isn't yet even documented here). However, there are
600 two benefits to using it: resources allocated to a pool never leak (even if
601 you allocate a scratch string, and just forget about it); also, for memory
602 allocation, <code>ap_palloc</code> is generally faster than
603 <code>malloc</code>.</p>
605 <p>We begin here by describing how memory is allocated to pools, and then
606 discuss how other resources are tracked by the resource pool machinery.</p>
608 <h3>Allocation of memory in pools</h3>
609 <p>Memory is allocated to pools by calling the function
610 <code>ap_palloc</code>, which takes two arguments, one being a pointer to
611 a resource pool structure, and the other being the amount of memory to
612 allocate (in <code>char</code>s). Within handlers for handling requests,
613 the most common way of getting a resource pool structure is by looking at
614 the <code>pool</code> slot of the relevant <code>request_rec</code>; hence
615 the repeated appearance of the following idiom in module code:</p>
617 <div class="example"><p><code>
618 int my_handler(request_rec *r)<br />
620 <span class="indent">
621 struct my_structure *foo;<br />
624 foo = (foo *)ap_palloc (r->pool, sizeof(my_structure));<br />
629 <p>Note that <em>there is no <code>ap_pfree</code></em> --
630 <code>ap_palloc</code>ed memory is freed only when the associated resource
631 pool is cleared. This means that <code>ap_palloc</code> does not have to
632 do as much accounting as <code>malloc()</code>; all it does in the typical
633 case is to round up the size, bump a pointer, and do a range check.</p>
635 <p>(It also raises the possibility that heavy use of
636 <code>ap_palloc</code> could cause a server process to grow excessively
637 large. There are two ways to deal with this, which are dealt with below;
638 briefly, you can use <code>malloc</code>, and try to be sure that all of
639 the memory gets explicitly <code>free</code>d, or you can allocate a
640 sub-pool of the main pool, allocate your memory in the sub-pool, and clear
641 it out periodically. The latter technique is discussed in the section
642 on sub-pools below, and is used in the directory-indexing code, in order
643 to avoid excessive storage allocation when listing directories with
644 thousands of files).</p>
647 <h3>Allocating initialized memory</h3>
648 <p>There are functions which allocate initialized memory, and are
649 frequently useful. The function <code>ap_pcalloc</code> has the same
650 interface as <code>ap_palloc</code>, but clears out the memory it
651 allocates before it returns it. The function <code>ap_pstrdup</code>
652 takes a resource pool and a <code>char *</code> as arguments, and
653 allocates memory for a copy of the string the pointer points to, returning
654 a pointer to the copy. Finally <code>ap_pstrcat</code> is a varargs-style
655 function, which takes a pointer to a resource pool, and at least two
656 <code>char *</code> arguments, the last of which must be
657 <code>NULL</code>. It allocates enough memory to fit copies of each of
658 the strings, as a unit; for instance:</p>
660 <div class="example"><p><code>
661 ap_pstrcat (r->pool, "foo", "/", "bar", NULL);
664 <p>returns a pointer to 8 bytes worth of memory, initialized to
665 <code>"foo/bar"</code>.</p>
668 <h3><a name="pools-used" id="pools-used">Commonly-used pools in the Apache Web
670 <p>A pool is really defined by its lifetime more than anything else.
671 There are some static pools in http_main which are passed to various
672 non-http_main functions as arguments at opportune times. Here they
676 <dt><code>permanent_pool</code></dt>
677 <dd>never passed to anything else, this is the ancestor of all pools</dd>
679 <dt><code>pconf</code></dt>
682 <li>subpool of permanent_pool</li>
684 <li>created at the beginning of a config "cycle"; exists
685 until the server is terminated or restarts; passed to all
686 config-time routines, either via cmd->pool, or as the
687 "pool *p" argument on those which don't take pools</li>
689 <li>passed to the module init() functions</li>
693 <dt><code>ptemp</code></dt>
696 <li>sorry I lie, this pool isn't called this currently in
697 1.3, I renamed it this in my pthreads development. I'm
698 referring to the use of ptrans in the parent... contrast
699 this with the later definition of ptrans in the
702 <li>subpool of permanent_pool</li>
704 <li>created at the beginning of a config "cycle"; exists
705 until the end of config parsing; passed to config-time
706 routines <em>via</em> cmd->temp_pool. Somewhat of a
707 "bastard child" because it isn't available everywhere.
708 Used for temporary scratch space which may be needed by
709 some config routines but which is deleted at the end of
714 <dt><code>pchild</code></dt>
717 <li>subpool of permanent_pool</li>
719 <li>created when a child is spawned (or a thread is
720 created); lives until that child (thread) is
723 <li>passed to the module child_init functions</li>
725 <li>destruction happens right after the child_exit
726 functions are called... (which may explain why I think
727 child_exit is redundant and unneeded)</li>
731 <dt><code>ptrans</code></dt>
734 <li>should be a subpool of pchild, but currently is a
735 subpool of permanent_pool, see above</li>
737 <li>cleared by the child before going into the accept()
738 loop to receive a connection</li>
740 <li>used as connection->pool</li>
744 <dt><code>r->pool</code></dt>
747 <li>for the main request this is a subpool of
748 connection->pool; for subrequests it is a subpool of
749 the parent request's pool.</li>
751 <li>exists until the end of the request (<em>i.e.</em>,
752 ap_destroy_sub_req, or in child_main after
753 process_request has finished)</li>
755 <li>note that r itself is allocated from r->pool;
756 <em>i.e.</em>, r->pool is first created and then r is
757 the first thing palloc()d from it</li>
762 <p>For almost everything folks do, <code>r->pool</code> is the pool to
763 use. But you can see how other lifetimes, such as pchild, are useful to
764 some modules... such as modules that need to open a database connection
765 once per child, and wish to clean it up when the child dies.</p>
767 <p>You can also see how some bugs have manifested themself, such as
768 setting <code>connection->user</code> to a value from
769 <code>r->pool</code> -- in this case connection exists for the
770 lifetime of <code>ptrans</code>, which is longer than
771 <code>r->pool</code> (especially if <code>r->pool</code> is a
772 subrequest!). So the correct thing to do is to allocate from
773 <code>connection->pool</code>.</p>
775 <p>And there was another interesting bug in <code class="module"><a href="../mod/mod_include.html">mod_include</a></code>
776 / <code class="module"><a href="../mod/mod_cgi.html">mod_cgi</a></code>. You'll see in those that they do this test
777 to decide if they should use <code>r->pool</code> or
778 <code>r->main->pool</code>. In this case the resource that they are
779 registering for cleanup is a child process. If it were registered in
780 <code>r->pool</code>, then the code would <code>wait()</code> for the
781 child when the subrequest finishes. With <code class="module"><a href="../mod/mod_include.html">mod_include</a></code> this
782 could be any old <code>#include</code>, and the delay can be up to 3
783 seconds... and happened quite frequently. Instead the subprocess is
784 registered in <code>r->main->pool</code> which causes it to be
785 cleaned up when the entire request is done -- <em>i.e.</em>, after the
786 output has been sent to the client and logging has happened.</p>
789 <h3><a name="pool-files" id="pool-files">Tracking open files, etc.</a></h3>
790 <p>As indicated above, resource pools are also used to track other sorts
791 of resources besides memory. The most common are open files. The routine
792 which is typically used for this is <code>ap_pfopen</code>, which takes a
793 resource pool and two strings as arguments; the strings are the same as
794 the typical arguments to <code>fopen</code>, <em>e.g.</em>,</p>
796 <div class="example"><p><code>
798 FILE *f = ap_pfopen (r->pool, r->filename, "r");<br />
800 if (f == NULL) { ... } else { ... }<br />
803 <p>There is also a <code>ap_popenf</code> routine, which parallels the
804 lower-level <code>open</code> system call. Both of these routines arrange
805 for the file to be closed when the resource pool in question is
808 <p>Unlike the case for memory, there <em>are</em> functions to close files
809 allocated with <code>ap_pfopen</code>, and <code>ap_popenf</code>, namely
810 <code>ap_pfclose</code> and <code>ap_pclosef</code>. (This is because, on
811 many systems, the number of files which a single process can have open is
812 quite limited). It is important to use these functions to close files
813 allocated with <code>ap_pfopen</code> and <code>ap_popenf</code>, since to
814 do otherwise could cause fatal errors on systems such as Linux, which
815 react badly if the same <code>FILE*</code> is closed more than once.</p>
817 <p>(Using the <code>close</code> functions is not mandatory, since the
818 file will eventually be closed regardless, but you should consider it in
819 cases where your module is opening, or could open, a lot of files).</p>
822 <h3>Other sorts of resources -- cleanup functions</h3>
823 <p>More text goes here. Describe the cleanup primitives in terms of
824 which the file stuff is implemented; also, <code>spawn_process</code>.</p>
826 <p>Pool cleanups live until <code>clear_pool()</code> is called:
827 <code>clear_pool(a)</code> recursively calls <code>destroy_pool()</code>
828 on all subpools of <code>a</code>; then calls all the cleanups for
829 <code>a</code>; then releases all the memory for <code>a</code>.
830 <code>destroy_pool(a)</code> calls <code>clear_pool(a)</code> and then
831 releases the pool structure itself. <em>i.e.</em>,
832 <code>clear_pool(a)</code> doesn't delete <code>a</code>, it just frees
833 up all the resources and you can start using it again immediately.</p>
836 <h3>Fine control -- creating and dealing with sub-pools, with
837 a note on sub-requests</h3>
838 <p>On rare occasions, too-free use of <code>ap_palloc()</code> and the
839 associated primitives may result in undesirably profligate resource
840 allocation. You can deal with such a case by creating a <em>sub-pool</em>,
841 allocating within the sub-pool rather than the main pool, and clearing or
842 destroying the sub-pool, which releases the resources which were
843 associated with it. (This really <em>is</em> a rare situation; the only
844 case in which it comes up in the standard module set is in case of listing
845 directories, and then only with <em>very</em> large directories.
846 Unnecessary use of the primitives discussed here can hair up your code
847 quite a bit, with very little gain).</p>
849 <p>The primitive for creating a sub-pool is <code>ap_make_sub_pool</code>,
850 which takes another pool (the parent pool) as an argument. When the main
851 pool is cleared, the sub-pool will be destroyed. The sub-pool may also be
852 cleared or destroyed at any time, by calling the functions
853 <code>ap_clear_pool</code> and <code>ap_destroy_pool</code>, respectively.
854 (The difference is that <code>ap_clear_pool</code> frees resources
855 associated with the pool, while <code>ap_destroy_pool</code> also
856 deallocates the pool itself. In the former case, you can allocate new
857 resources within the pool, and clear it again, and so forth; in the
858 latter case, it is simply gone).</p>
860 <p>One final note -- sub-requests have their own resource pools, which are
861 sub-pools of the resource pool for the main request. The polite way to
862 reclaim the resources associated with a sub request which you have
863 allocated (using the <code>ap_sub_req_...</code> functions) is
864 <code>ap_destroy_sub_req</code>, which frees the resource pool. Before
865 calling this function, be sure to copy anything that you care about which
866 might be allocated in the sub-request's resource pool into someplace a
867 little less volatile (for instance, the filename in its
868 <code>request_rec</code> structure).</p>
870 <p>(Again, under most circumstances, you shouldn't feel obliged to call
871 this function; only 2K of memory or so are allocated for a typical sub
872 request, and it will be freed anyway when the main request pool is
873 cleared. It is only when you are allocating many, many sub-requests for a
874 single main request that you should seriously consider the
875 <code>ap_destroy_...</code> functions).</p>
877 </div><div class="top"><a href="#page-header"><img alt="top" src="../images/up.gif" /></a></div>
878 <div class="section">
879 <h2><a name="config" id="config">Configuration, commands and the like</a></h2>
880 <p>One of the design goals for this server was to maintain external
881 compatibility with the NCSA 1.3 server --- that is, to read the same
882 configuration files, to process all the directives therein correctly, and
883 in general to be a drop-in replacement for NCSA. On the other hand, another
884 design goal was to move as much of the server's functionality into modules
885 which have as little as possible to do with the monolithic server core. The
886 only way to reconcile these goals is to move the handling of most commands
887 from the central server into the modules.</p>
889 <p>However, just giving the modules command tables is not enough to divorce
890 them completely from the server core. The server has to remember the
891 commands in order to act on them later. That involves maintaining data which
892 is private to the modules, and which can be either per-server, or
893 per-directory. Most things are per-directory, including in particular access
894 control and authorization information, but also information on how to
895 determine file types from suffixes, which can be modified by
896 <code class="directive"><a href="../mod/mod_mime.html#addtype">AddType</a></code> and <code class="directive"><a href="../mod/core.html#defaulttype">DefaultType</a></code> directives, and so forth. In general,
897 the governing philosophy is that anything which <em>can</em> be made
898 configurable by directory should be; per-server information is generally
899 used in the standard set of modules for information like
900 <code class="directive"><a href="../mod/mod_alias.html#alias">Alias</a></code>es and <code class="directive"><a href="../mod/mod_alias.html#redirect">Redirect</a></code>s which come into play before the
901 request is tied to a particular place in the underlying file system.</p>
903 <p>Another requirement for emulating the NCSA server is being able to handle
904 the per-directory configuration files, generally called
905 <code>.htaccess</code> files, though even in the NCSA server they can
906 contain directives which have nothing at all to do with access control.
907 Accordingly, after URI -> filename translation, but before performing any
908 other phase, the server walks down the directory hierarchy of the underlying
909 filesystem, following the translated pathname, to read any
910 <code>.htaccess</code> files which might be present. The information which
911 is read in then has to be <em>merged</em> with the applicable information
912 from the server's own config files (either from the <code class="directive"><a href="../mod/core.html#directory"><Directory></a></code> sections in
913 <code>access.conf</code>, or from defaults in <code>srm.conf</code>, which
914 actually behaves for most purposes almost exactly like <code><Directory
917 <p>Finally, after having served a request which involved reading
918 <code>.htaccess</code> files, we need to discard the storage allocated for
919 handling them. That is solved the same way it is solved wherever else
920 similar problems come up, by tying those structures to the per-transaction
923 <h3><a name="per-dir" id="per-dir">Per-directory configuration structures</a></h3>
924 <p>Let's look out how all of this plays out in <code>mod_mime.c</code>,
925 which defines the file typing handler which emulates the NCSA server's
926 behavior of determining file types from suffixes. What we'll be looking
927 at, here, is the code which implements the <code class="directive"><a href="../mod/mod_mime.html#addtype">AddType</a></code> and <code class="directive"><a href="../mod/mod_mime.html#addencoding">AddEncoding</a></code> commands. These commands can appear in
928 <code>.htaccess</code> files, so they must be handled in the module's
929 private per-directory data, which in fact, consists of two separate
930 tables for MIME types and encoding information, and is declared as
933 <div class="example"><pre>typedef struct {
934 table *forced_types; /* Additional AddTyped stuff */
935 table *encoding_types; /* Added with AddEncoding... */
936 } mime_dir_config;</pre></div>
938 <p>When the server is reading a configuration file, or <code class="directive"><a href="../mod/core.html#directory"><Directory></a></code> section, which includes
939 one of the MIME module's commands, it needs to create a
940 <code>mime_dir_config</code> structure, so those commands have something
941 to act on. It does this by invoking the function it finds in the module's
942 `create per-dir config slot', with two arguments: the name of the
943 directory to which this configuration information applies (or
944 <code>NULL</code> for <code>srm.conf</code>), and a pointer to a
945 resource pool in which the allocation should happen.</p>
947 <p>(If we are reading a <code>.htaccess</code> file, that resource pool
948 is the per-request resource pool for the request; otherwise it is a
949 resource pool which is used for configuration data, and cleared on
950 restarts. Either way, it is important for the structure being created to
951 vanish when the pool is cleared, by registering a cleanup on the pool if
954 <p>For the MIME module, the per-dir config creation function just
955 <code>ap_palloc</code>s the structure above, and a creates a couple of
956 tables to fill it. That looks like this:</p>
958 <div class="example"><p><code>
959 void *create_mime_dir_config (pool *p, char *dummy)<br />
961 <span class="indent">
962 mime_dir_config *new =<br />
963 <span class="indent">
964 (mime_dir_config *) ap_palloc (p, sizeof(mime_dir_config));<br />
967 new->forced_types = ap_make_table (p, 4);<br />
968 new->encoding_types = ap_make_table (p, 4);<br />
975 <p>Now, suppose we've just read in a <code>.htaccess</code> file. We
976 already have the per-directory configuration structure for the next
977 directory up in the hierarchy. If the <code>.htaccess</code> file we just
978 read in didn't have any <code class="directive"><a href="../mod/mod_mime.html#addtype">AddType</a></code>
979 or <code class="directive"><a href="../mod/mod_mime.html#addencoding">AddEncoding</a></code> commands, its
980 per-directory config structure for the MIME module is still valid, and we
981 can just use it. Otherwise, we need to merge the two structures
984 <p>To do that, the server invokes the module's per-directory config merge
985 function, if one is present. That function takes three arguments: the two
986 structures being merged, and a resource pool in which to allocate the
987 result. For the MIME module, all that needs to be done is overlay the
988 tables from the new per-directory config structure with those from the
991 <div class="example"><p><code>
992 void *merge_mime_dir_configs (pool *p, void *parent_dirv, void *subdirv)<br />
994 <span class="indent">
995 mime_dir_config *parent_dir = (mime_dir_config *)parent_dirv;<br />
996 mime_dir_config *subdir = (mime_dir_config *)subdirv;<br />
997 mime_dir_config *new =<br />
998 <span class="indent">
999 (mime_dir_config *)ap_palloc (p, sizeof(mime_dir_config));<br />
1002 new->forced_types = ap_overlay_tables (p, subdir->forced_types,<br />
1003 <span class="indent">
1004 parent_dir->forced_types);<br />
1006 new->encoding_types = ap_overlay_tables (p, subdir->encoding_types,<br />
1007 <span class="indent">
1008 parent_dir->encoding_types);<br />
1016 <p>As a note -- if there is no per-directory merge function present, the
1017 server will just use the subdirectory's configuration info, and ignore
1018 the parent's. For some modules, that works just fine (<em>e.g.</em>, for
1019 the includes module, whose per-directory configuration information
1020 consists solely of the state of the <code>XBITHACK</code>), and for those
1021 modules, you can just not declare one, and leave the corresponding
1022 structure slot in the module itself <code>NULL</code>.</p>
1025 <h3><a name="commands" id="commands">Command handling</a></h3>
1026 <p>Now that we have these structures, we need to be able to figure out how
1027 to fill them. That involves processing the actual <code class="directive"><a href="../mod/mod_mime.html#addtype">AddType</a></code> and <code class="directive"><a href="../mod/mod_mime.html#addencoding">AddEncoding</a></code> commands. To find commands, the server looks in
1028 the module's command table. That table contains information on how many
1029 arguments the commands take, and in what formats, where it is permitted,
1030 and so forth. That information is sufficient to allow the server to invoke
1031 most command-handling functions with pre-parsed arguments. Without further
1032 ado, let's look at the <code class="directive"><a href="../mod/mod_mime.html#addtype">AddType</a></code>
1033 command handler, which looks like this (the <code class="directive"><a href="../mod/mod_mime.html#addencoding">AddEncoding</a></code> command looks basically the same, and won't be
1036 <div class="example"><p><code>
1037 char *add_type(cmd_parms *cmd, mime_dir_config *m, char *ct, char *ext)<br />
1039 <span class="indent">
1040 if (*ext == '.') ++ext;<br />
1041 ap_table_set (m->forced_types, ext, ct);<br />
1047 <p>This command handler is unusually simple. As you can see, it takes
1048 four arguments, two of which are pre-parsed arguments, the third being the
1049 per-directory configuration structure for the module in question, and the
1050 fourth being a pointer to a <code>cmd_parms</code> structure. That
1051 structure contains a bunch of arguments which are frequently of use to
1052 some, but not all, commands, including a resource pool (from which memory
1053 can be allocated, and to which cleanups should be tied), and the (virtual)
1054 server being configured, from which the module's per-server configuration
1055 data can be obtained if required.</p>
1057 <p>Another way in which this particular command handler is unusually
1058 simple is that there are no error conditions which it can encounter. If
1059 there were, it could return an error message instead of <code>NULL</code>;
1060 this causes an error to be printed out on the server's
1061 <code>stderr</code>, followed by a quick exit, if it is in the main config
1062 files; for a <code>.htaccess</code> file, the syntax error is logged in
1063 the server error log (along with an indication of where it came from), and
1064 the request is bounced with a server error response (HTTP error status,
1067 <p>The MIME module's command table has entries for these commands, which
1070 <div class="example"><p><code>
1071 command_rec mime_cmds[] = {<br />
1072 <span class="indent">
1073 { "AddType", add_type, NULL, OR_FILEINFO, TAKE2,<br />
1074 <span class="indent">"a mime type followed by a file extension" },<br /></span>
1075 { "AddEncoding", add_encoding, NULL, OR_FILEINFO, TAKE2,<br />
1076 <span class="indent">
1077 "an encoding (<em>e.g.</em>, gzip), followed by a file extension" },<br />
1084 <p>The entries in these tables are:</p>
1086 <li>The name of the command</li>
1087 <li>The function which handles it</li>
1088 <li>a <code>(void *)</code> pointer, which is passed in the
1089 <code>cmd_parms</code> structure to the command handler ---
1090 this is useful in case many similar commands are handled by
1091 the same function.</li>
1093 <li>A bit mask indicating where the command may appear. There
1094 are mask bits corresponding to each
1095 <code>AllowOverride</code> option, and an additional mask
1096 bit, <code>RSRC_CONF</code>, indicating that the command may
1097 appear in the server's own config files, but <em>not</em> in
1098 any <code>.htaccess</code> file.</li>
1100 <li>A flag indicating how many arguments the command handler
1101 wants pre-parsed, and how they should be passed in.
1102 <code>TAKE2</code> indicates two pre-parsed arguments. Other
1103 options are <code>TAKE1</code>, which indicates one
1104 pre-parsed argument, <code>FLAG</code>, which indicates that
1105 the argument should be <code>On</code> or <code>Off</code>,
1106 and is passed in as a boolean flag, <code>RAW_ARGS</code>,
1107 which causes the server to give the command the raw, unparsed
1108 arguments (everything but the command name itself). There is
1109 also <code>ITERATE</code>, which means that the handler looks
1110 the same as <code>TAKE1</code>, but that if multiple
1111 arguments are present, it should be called multiple times,
1112 and finally <code>ITERATE2</code>, which indicates that the
1113 command handler looks like a <code>TAKE2</code>, but if more
1114 arguments are present, then it should be called multiple
1115 times, holding the first argument constant.</li>
1117 <li>Finally, we have a string which describes the arguments
1118 that should be present. If the arguments in the actual config
1119 file are not as required, this string will be used to help
1120 give a more specific error message. (You can safely leave
1121 this <code>NULL</code>).</li>
1124 <p>Finally, having set this all up, we have to use it. This is ultimately
1125 done in the module's handlers, specifically for its file-typing handler,
1126 which looks more or less like this; note that the per-directory
1127 configuration structure is extracted from the <code>request_rec</code>'s
1128 per-directory configuration vector by using the
1129 <code>ap_get_module_config</code> function.</p>
1131 <div class="example"><p><code>
1132 int find_ct(request_rec *r)<br />
1134 <span class="indent">
1136 char *fn = ap_pstrdup (r->pool, r->filename);<br />
1137 mime_dir_config *conf = (mime_dir_config *)<br />
1138 <span class="indent">
1139 ap_get_module_config(r->per_dir_config, &mime_module);<br />
1143 if (S_ISDIR(r->finfo.st_mode)) {<br />
1144 <span class="indent">
1145 r->content_type = DIR_MAGIC_TYPE;<br />
1150 if((i=ap_rind(fn,'.')) < 0) return DECLINED;<br />
1153 if ((type = ap_table_get (conf->encoding_types, &fn[i])))<br />
1155 <span class="indent">
1156 r->content_encoding = type;<br />
1158 /* go back to previous extension to try to use it as a type */<br />
1159 fn[i-1] = '\0';<br />
1160 if((i=ap_rind(fn,'.')) < 0) return OK;<br />
1165 if ((type = ap_table_get (conf->forced_types, &fn[i])))<br />
1167 <span class="indent">
1168 r->content_type = type;<br />
1178 <h3><a name="servconf" id="servconf">Side notes -- per-server configuration,
1179 virtual servers, <em>etc</em>.</a></h3>
1180 <p>The basic ideas behind per-server module configuration are basically
1181 the same as those for per-directory configuration; there is a creation
1182 function and a merge function, the latter being invoked where a virtual
1183 server has partially overridden the base server configuration, and a
1184 combined structure must be computed. (As with per-directory configuration,
1185 the default if no merge function is specified, and a module is configured
1186 in some virtual server, is that the base configuration is simply
1189 <p>The only substantial difference is that when a command needs to
1190 configure the per-server private module data, it needs to go to the
1191 <code>cmd_parms</code> data to get at it. Here's an example, from the
1192 alias module, which also indicates how a syntax error can be returned
1193 (note that the per-directory configuration argument to the command
1194 handler is declared as a dummy, since the module doesn't actually have
1195 per-directory config data):</p>
1197 <div class="example"><p><code>
1198 char *add_redirect(cmd_parms *cmd, void *dummy, char *f, char *url)<br />
1200 <span class="indent">
1201 server_rec *s = cmd->server;<br />
1202 alias_server_conf *conf = (alias_server_conf *)<br />
1203 <span class="indent">
1204 ap_get_module_config(s->module_config,&alias_module);<br />
1206 alias_entry *new = ap_push_array (conf->redirects);<br />
1208 if (!ap_is_url (url)) return "Redirect to non-URL";<br />
1210 new->fake = f; new->real = url;<br />
1217 <div class="bottomlang">
1218 <p><span>Available Languages: </span><a href="../en/developer/API.html" title="English"> en </a></p>
1219 </div><div id="footer">
1220 <p class="apache">Copyright 2007 The Apache Software Foundation.<br />Licensed under the <a href="http://www.apache.org/licenses/LICENSE-2.0">Apache License, Version 2.0</a>.</p>
1221 <p class="menu"><a href="../mod/">Modules</a> | <a href="../mod/directives.html">Directives</a> | <a href="../faq/">FAQ</a> | <a href="../glossary.html">Glossary</a> | <a href="../sitemap.html">Sitemap</a></p></div>