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 2.0</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 2.0</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 */
388 <h3><a name="req_orig" id="req_orig">Where request_rec structures come from</a></h3>
389 <p>Most <code>request_rec</code> structures are built by reading an HTTP
390 request from a client, and filling in the fields. However, there are a
394 <li>If the request is to an imagemap, a type map (<em>i.e.</em>, a
395 <code>*.var</code> file), or a CGI script which returned a local
396 `Location:', then the resource which the user requested is going to be
397 ultimately located by some URI other than what the client originally
398 supplied. In this case, the server does an <em>internal redirect</em>,
399 constructing a new <code>request_rec</code> for the new URI, and
400 processing it almost exactly as if the client had requested the new URI
403 <li>If some handler signaled an error, and an <code>ErrorDocument</code>
404 is in scope, the same internal redirect machinery comes into play.</li>
406 <li><p>Finally, a handler occasionally needs to investigate `what would
407 happen if' some other request were run. For instance, the directory
408 indexing module needs to know what MIME type would be assigned to a
409 request for each directory entry, in order to figure out what icon to
412 <p>Such handlers can construct a <em>sub-request</em>, using the
413 functions <code>ap_sub_req_lookup_file</code>,
414 <code>ap_sub_req_lookup_uri</code>, and <code>ap_sub_req_method_uri</code>;
415 these construct a new <code>request_rec</code> structure and processes it
416 as you would expect, up to but not including the point of actually sending
417 a response. (These functions skip over the access checks if the
418 sub-request is for a file in the same directory as the original
421 <p>(Server-side includes work by building sub-requests and then actually
422 invoking the response handler for them, via the function
423 <code>ap_run_sub_req</code>).</p>
428 <h3><a name="req_return" id="req_return">Handling requests, declining, and returning
430 <p>As discussed above, each handler, when invoked to handle a particular
431 <code>request_rec</code>, has to return an <code>int</code> to indicate
432 what happened. That can either be</p>
435 <li><code>OK</code> -- the request was handled successfully. This may or
436 may not terminate the phase.</li>
438 <li><code>DECLINED</code> -- no erroneous condition exists, but the module
439 declines to handle the phase; the server tries to find another.</li>
441 <li>an HTTP error code, which aborts handling of the request.</li>
444 <p>Note that if the error code returned is <code>REDIRECT</code>, then
445 the module should put a <code>Location</code> in the request's
446 <code>headers_out</code>, to indicate where the client should be
447 redirected <em>to</em>.</p>
450 <h3><a name="resp_handlers" id="resp_handlers">Special considerations for response
452 <p>Handlers for most phases do their work by simply setting a few fields
453 in the <code>request_rec</code> structure (or, in the case of access
454 checkers, simply by returning the correct error code). However, response
455 handlers have to actually send a request back to the client.</p>
457 <p>They should begin by sending an HTTP response header, using the
458 function <code>ap_send_http_header</code>. (You don't have to do anything
459 special to skip sending the header for HTTP/0.9 requests; the function
460 figures out on its own that it shouldn't do anything). If the request is
461 marked <code>header_only</code>, that's all they should do; they should
462 return after that, without attempting any further output.</p>
464 <p>Otherwise, they should produce a request body which responds to the
465 client as appropriate. The primitives for this are <code>ap_rputc</code>
466 and <code>ap_rprintf</code>, for internally generated output, and
467 <code>ap_send_fd</code>, to copy the contents of some <code>FILE *</code>
468 straight to the client.</p>
470 <p>At this point, you should more or less understand the following piece
471 of code, which is the handler which handles <code>GET</code> requests
472 which have no more specific handler; it also shows how conditional
473 <code>GET</code>s can be handled, if it's desirable to do so in a
474 particular response handler -- <code>ap_set_last_modified</code> checks
475 against the <code>If-modified-since</code> value supplied by the client,
476 if any, and returns an appropriate code (which will, if nonzero, be
477 USE_LOCAL_COPY). No similar considerations apply for
478 <code>ap_set_content_length</code>, but it returns an error code for
481 <div class="example"><p><code>
482 int default_handler (request_rec *r)<br />
484 <span class="indent">
488 if (r->method_number != M_GET) return DECLINED;<br />
489 if (r->finfo.st_mode == 0) return NOT_FOUND;<br />
491 if ((errstatus = ap_set_content_length (r, r->finfo.st_size))<br />
492 ||
493 (errstatus = ap_set_last_modified (r, r->finfo.st_mtime)))<br />
494 return errstatus;<br />
496 f = fopen (r->filename, "r");<br />
498 if (f == NULL) {<br />
499 <span class="indent">
500 log_reason("file permissions deny server access", r->filename, r);<br />
501 return FORBIDDEN;<br />
505 register_timeout ("send", r);<br />
506 ap_send_http_header (r);<br />
508 if (!r->header_only) send_fd (f, r);<br />
509 ap_pfclose (r->pool, f);<br />
515 <p>Finally, if all of this is too much of a challenge, there are a few
516 ways out of it. First off, as shown above, a response handler which has
517 not yet produced any output can simply return an error code, in which
518 case the server will automatically produce an error response. Secondly,
519 it can punt to some other handler by invoking
520 <code>ap_internal_redirect</code>, which is how the internal redirection
521 machinery discussed above is invoked. A response handler which has
522 internally redirected should always return <code>OK</code>.</p>
524 <p>(Invoking <code>ap_internal_redirect</code> from handlers which are
525 <em>not</em> response handlers will lead to serious confusion).</p>
528 <h3><a name="auth_handlers" id="auth_handlers">Special considerations for authentication
530 <p>Stuff that should be discussed here in detail:</p>
533 <li>Authentication-phase handlers not invoked unless auth is
534 configured for the directory.</li>
536 <li>Common auth configuration stored in the core per-dir
537 configuration; it has accessors <code>ap_auth_type</code>,
538 <code>ap_auth_name</code>, and <code>ap_requires</code>.</li>
540 <li>Common routines, to handle the protocol end of things, at
541 least for HTTP basic authentication
542 (<code>ap_get_basic_auth_pw</code>, which sets the
543 <code>connection->user</code> structure field
544 automatically, and <code>ap_note_basic_auth_failure</code>,
545 which arranges for the proper <code>WWW-Authenticate:</code>
546 header to be sent back).</li>
550 <h3><a name="log_handlers" id="log_handlers">Special considerations for logging
552 <p>When a request has internally redirected, there is the question of
553 what to log. Apache handles this by bundling the entire chain of redirects
554 into a list of <code>request_rec</code> structures which are threaded
555 through the <code>r->prev</code> and <code>r->next</code> pointers.
556 The <code>request_rec</code> which is passed to the logging handlers in
557 such cases is the one which was originally built for the initial request
558 from the client; note that the <code>bytes_sent</code> field will only be
559 correct in the last request in the chain (the one for which a response was
562 </div><div class="top"><a href="#page-header"><img alt="top" src="../images/up.gif" /></a></div>
563 <div class="section">
564 <h2><a name="pools" id="pools">Resource allocation and resource pools</a></h2>
565 <p>One of the problems of writing and designing a server-pool server is
566 that of preventing leakage, that is, allocating resources (memory, open
567 files, <em>etc.</em>), without subsequently releasing them. The resource
568 pool machinery is designed to make it easy to prevent this from happening,
569 by allowing resource to be allocated in such a way that they are
570 <em>automatically</em> released when the server is done with them.</p>
572 <p>The way this works is as follows: the memory which is allocated, file
573 opened, <em>etc.</em>, to deal with a particular request are tied to a
574 <em>resource pool</em> which is allocated for the request. The pool is a
575 data structure which itself tracks the resources in question.</p>
577 <p>When the request has been processed, the pool is <em>cleared</em>. At
578 that point, all the memory associated with it is released for reuse, all
579 files associated with it are closed, and any other clean-up functions which
580 are associated with the pool are run. When this is over, we can be confident
581 that all the resource tied to the pool have been released, and that none of
582 them have leaked.</p>
584 <p>Server restarts, and allocation of memory and resources for per-server
585 configuration, are handled in a similar way. There is a <em>configuration
586 pool</em>, which keeps track of resources which were allocated while reading
587 the server configuration files, and handling the commands therein (for
588 instance, the memory that was allocated for per-server module configuration,
589 log files and other files that were opened, and so forth). When the server
590 restarts, and has to reread the configuration files, the configuration pool
591 is cleared, and so the memory and file descriptors which were taken up by
592 reading them the last time are made available for reuse.</p>
594 <p>It should be noted that use of the pool machinery isn't generally
595 obligatory, except for situations like logging handlers, where you really
596 need to register cleanups to make sure that the log file gets closed when
597 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
598 underlying file descriptor to be closed before any child processes, such as
599 for CGI scripts, are <code>exec</code>ed), or in case you are using the
600 timeout machinery (which isn't yet even documented here). However, there are
601 two benefits to using it: resources allocated to a pool never leak (even if
602 you allocate a scratch string, and just forget about it); also, for memory
603 allocation, <code>ap_palloc</code> is generally faster than
604 <code>malloc</code>.</p>
606 <p>We begin here by describing how memory is allocated to pools, and then
607 discuss how other resources are tracked by the resource pool machinery.</p>
609 <h3>Allocation of memory in pools</h3>
610 <p>Memory is allocated to pools by calling the function
611 <code>ap_palloc</code>, which takes two arguments, one being a pointer to
612 a resource pool structure, and the other being the amount of memory to
613 allocate (in <code>char</code>s). Within handlers for handling requests,
614 the most common way of getting a resource pool structure is by looking at
615 the <code>pool</code> slot of the relevant <code>request_rec</code>; hence
616 the repeated appearance of the following idiom in module code:</p>
618 <div class="example"><p><code>
619 int my_handler(request_rec *r)<br />
621 <span class="indent">
622 struct my_structure *foo;<br />
625 foo = (foo *)ap_palloc (r->pool, sizeof(my_structure));<br />
630 <p>Note that <em>there is no <code>ap_pfree</code></em> --
631 <code>ap_palloc</code>ed memory is freed only when the associated resource
632 pool is cleared. This means that <code>ap_palloc</code> does not have to
633 do as much accounting as <code>malloc()</code>; all it does in the typical
634 case is to round up the size, bump a pointer, and do a range check.</p>
636 <p>(It also raises the possibility that heavy use of
637 <code>ap_palloc</code> could cause a server process to grow excessively
638 large. There are two ways to deal with this, which are dealt with below;
639 briefly, you can use <code>malloc</code>, and try to be sure that all of
640 the memory gets explicitly <code>free</code>d, or you can allocate a
641 sub-pool of the main pool, allocate your memory in the sub-pool, and clear
642 it out periodically. The latter technique is discussed in the section
643 on sub-pools below, and is used in the directory-indexing code, in order
644 to avoid excessive storage allocation when listing directories with
645 thousands of files).</p>
648 <h3>Allocating initialized memory</h3>
649 <p>There are functions which allocate initialized memory, and are
650 frequently useful. The function <code>ap_pcalloc</code> has the same
651 interface as <code>ap_palloc</code>, but clears out the memory it
652 allocates before it returns it. The function <code>ap_pstrdup</code>
653 takes a resource pool and a <code>char *</code> as arguments, and
654 allocates memory for a copy of the string the pointer points to, returning
655 a pointer to the copy. Finally <code>ap_pstrcat</code> is a varargs-style
656 function, which takes a pointer to a resource pool, and at least two
657 <code>char *</code> arguments, the last of which must be
658 <code>NULL</code>. It allocates enough memory to fit copies of each of
659 the strings, as a unit; for instance:</p>
661 <div class="example"><p><code>
662 ap_pstrcat (r->pool, "foo", "/", "bar", NULL);
665 <p>returns a pointer to 8 bytes worth of memory, initialized to
666 <code>"foo/bar"</code>.</p>
669 <h3><a name="pools-used" id="pools-used">Commonly-used pools in the Apache Web
671 <p>A pool is really defined by its lifetime more than anything else.
672 There are some static pools in http_main which are passed to various
673 non-http_main functions as arguments at opportune times. Here they
677 <dt><code>permanent_pool</code></dt>
678 <dd>never passed to anything else, this is the ancestor of all pools</dd>
680 <dt><code>pconf</code></dt>
683 <li>subpool of permanent_pool</li>
685 <li>created at the beginning of a config "cycle"; exists
686 until the server is terminated or restarts; passed to all
687 config-time routines, either via cmd->pool, or as the
688 "pool *p" argument on those which don't take pools</li>
690 <li>passed to the module init() functions</li>
694 <dt><code>ptemp</code></dt>
697 <li>sorry I lie, this pool isn't called this currently in
698 1.3, I renamed it this in my pthreads development. I'm
699 referring to the use of ptrans in the parent... contrast
700 this with the later definition of ptrans in the
703 <li>subpool of permanent_pool</li>
705 <li>created at the beginning of a config "cycle"; exists
706 until the end of config parsing; passed to config-time
707 routines <em>via</em> cmd->temp_pool. Somewhat of a
708 "bastard child" because it isn't available everywhere.
709 Used for temporary scratch space which may be needed by
710 some config routines but which is deleted at the end of
715 <dt><code>pchild</code></dt>
718 <li>subpool of permanent_pool</li>
720 <li>created when a child is spawned (or a thread is
721 created); lives until that child (thread) is
724 <li>passed to the module child_init functions</li>
726 <li>destruction happens right after the child_exit
727 functions are called... (which may explain why I think
728 child_exit is redundant and unneeded)</li>
732 <dt><code>ptrans</code></dt>
735 <li>should be a subpool of pchild, but currently is a
736 subpool of permanent_pool, see above</li>
738 <li>cleared by the child before going into the accept()
739 loop to receive a connection</li>
741 <li>used as connection->pool</li>
745 <dt><code>r->pool</code></dt>
748 <li>for the main request this is a subpool of
749 connection->pool; for subrequests it is a subpool of
750 the parent request's pool.</li>
752 <li>exists until the end of the request (<em>i.e.</em>,
753 ap_destroy_sub_req, or in child_main after
754 process_request has finished)</li>
756 <li>note that r itself is allocated from r->pool;
757 <em>i.e.</em>, r->pool is first created and then r is
758 the first thing palloc()d from it</li>
763 <p>For almost everything folks do, <code>r->pool</code> is the pool to
764 use. But you can see how other lifetimes, such as pchild, are useful to
765 some modules... such as modules that need to open a database connection
766 once per child, and wish to clean it up when the child dies.</p>
768 <p>You can also see how some bugs have manifested themself, such as
769 setting <code>connection->user</code> to a value from
770 <code>r->pool</code> -- in this case connection exists for the
771 lifetime of <code>ptrans</code>, which is longer than
772 <code>r->pool</code> (especially if <code>r->pool</code> is a
773 subrequest!). So the correct thing to do is to allocate from
774 <code>connection->pool</code>.</p>
776 <p>And there was another interesting bug in <code class="module"><a href="../mod/mod_include.html">mod_include</a></code>
777 / <code class="module"><a href="../mod/mod_cgi.html">mod_cgi</a></code>. You'll see in those that they do this test
778 to decide if they should use <code>r->pool</code> or
779 <code>r->main->pool</code>. In this case the resource that they are
780 registering for cleanup is a child process. If it were registered in
781 <code>r->pool</code>, then the code would <code>wait()</code> for the
782 child when the subrequest finishes. With <code class="module"><a href="../mod/mod_include.html">mod_include</a></code> this
783 could be any old <code>#include</code>, and the delay can be up to 3
784 seconds... and happened quite frequently. Instead the subprocess is
785 registered in <code>r->main->pool</code> which causes it to be
786 cleaned up when the entire request is done -- <em>i.e.</em>, after the
787 output has been sent to the client and logging has happened.</p>
790 <h3><a name="pool-files" id="pool-files">Tracking open files, etc.</a></h3>
791 <p>As indicated above, resource pools are also used to track other sorts
792 of resources besides memory. The most common are open files. The routine
793 which is typically used for this is <code>ap_pfopen</code>, which takes a
794 resource pool and two strings as arguments; the strings are the same as
795 the typical arguments to <code>fopen</code>, <em>e.g.</em>,</p>
797 <div class="example"><p><code>
799 FILE *f = ap_pfopen (r->pool, r->filename, "r");<br />
801 if (f == NULL) { ... } else { ... }<br />
804 <p>There is also a <code>ap_popenf</code> routine, which parallels the
805 lower-level <code>open</code> system call. Both of these routines arrange
806 for the file to be closed when the resource pool in question is
809 <p>Unlike the case for memory, there <em>are</em> functions to close files
810 allocated with <code>ap_pfopen</code>, and <code>ap_popenf</code>, namely
811 <code>ap_pfclose</code> and <code>ap_pclosef</code>. (This is because, on
812 many systems, the number of files which a single process can have open is
813 quite limited). It is important to use these functions to close files
814 allocated with <code>ap_pfopen</code> and <code>ap_popenf</code>, since to
815 do otherwise could cause fatal errors on systems such as Linux, which
816 react badly if the same <code>FILE*</code> is closed more than once.</p>
818 <p>(Using the <code>close</code> functions is not mandatory, since the
819 file will eventually be closed regardless, but you should consider it in
820 cases where your module is opening, or could open, a lot of files).</p>
823 <h3>Other sorts of resources -- cleanup functions</h3>
824 <p>More text goes here. Describe the cleanup primitives in terms of
825 which the file stuff is implemented; also, <code>spawn_process</code>.</p>
827 <p>Pool cleanups live until <code>clear_pool()</code> is called:
828 <code>clear_pool(a)</code> recursively calls <code>destroy_pool()</code>
829 on all subpools of <code>a</code>; then calls all the cleanups for
830 <code>a</code>; then releases all the memory for <code>a</code>.
831 <code>destroy_pool(a)</code> calls <code>clear_pool(a)</code> and then
832 releases the pool structure itself. <em>i.e.</em>,
833 <code>clear_pool(a)</code> doesn't delete <code>a</code>, it just frees
834 up all the resources and you can start using it again immediately.</p>
837 <h3>Fine control -- creating and dealing with sub-pools, with
838 a note on sub-requests</h3>
839 <p>On rare occasions, too-free use of <code>ap_palloc()</code> and the
840 associated primitives may result in undesirably profligate resource
841 allocation. You can deal with such a case by creating a <em>sub-pool</em>,
842 allocating within the sub-pool rather than the main pool, and clearing or
843 destroying the sub-pool, which releases the resources which were
844 associated with it. (This really <em>is</em> a rare situation; the only
845 case in which it comes up in the standard module set is in case of listing
846 directories, and then only with <em>very</em> large directories.
847 Unnecessary use of the primitives discussed here can hair up your code
848 quite a bit, with very little gain).</p>
850 <p>The primitive for creating a sub-pool is <code>ap_make_sub_pool</code>,
851 which takes another pool (the parent pool) as an argument. When the main
852 pool is cleared, the sub-pool will be destroyed. The sub-pool may also be
853 cleared or destroyed at any time, by calling the functions
854 <code>ap_clear_pool</code> and <code>ap_destroy_pool</code>, respectively.
855 (The difference is that <code>ap_clear_pool</code> frees resources
856 associated with the pool, while <code>ap_destroy_pool</code> also
857 deallocates the pool itself. In the former case, you can allocate new
858 resources within the pool, and clear it again, and so forth; in the
859 latter case, it is simply gone).</p>
861 <p>One final note -- sub-requests have their own resource pools, which are
862 sub-pools of the resource pool for the main request. The polite way to
863 reclaim the resources associated with a sub request which you have
864 allocated (using the <code>ap_sub_req_...</code> functions) is
865 <code>ap_destroy_sub_req</code>, which frees the resource pool. Before
866 calling this function, be sure to copy anything that you care about which
867 might be allocated in the sub-request's resource pool into someplace a
868 little less volatile (for instance, the filename in its
869 <code>request_rec</code> structure).</p>
871 <p>(Again, under most circumstances, you shouldn't feel obliged to call
872 this function; only 2K of memory or so are allocated for a typical sub
873 request, and it will be freed anyway when the main request pool is
874 cleared. It is only when you are allocating many, many sub-requests for a
875 single main request that you should seriously consider the
876 <code>ap_destroy_...</code> functions).</p>
878 </div><div class="top"><a href="#page-header"><img alt="top" src="../images/up.gif" /></a></div>
879 <div class="section">
880 <h2><a name="config" id="config">Configuration, commands and the like</a></h2>
881 <p>One of the design goals for this server was to maintain external
882 compatibility with the NCSA 1.3 server --- that is, to read the same
883 configuration files, to process all the directives therein correctly, and
884 in general to be a drop-in replacement for NCSA. On the other hand, another
885 design goal was to move as much of the server's functionality into modules
886 which have as little as possible to do with the monolithic server core. The
887 only way to reconcile these goals is to move the handling of most commands
888 from the central server into the modules.</p>
890 <p>However, just giving the modules command tables is not enough to divorce
891 them completely from the server core. The server has to remember the
892 commands in order to act on them later. That involves maintaining data which
893 is private to the modules, and which can be either per-server, or
894 per-directory. Most things are per-directory, including in particular access
895 control and authorization information, but also information on how to
896 determine file types from suffixes, which can be modified by
897 <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,
898 the governing philosophy is that anything which <em>can</em> be made
899 configurable by directory should be; per-server information is generally
900 used in the standard set of modules for information like
901 <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
902 request is tied to a particular place in the underlying file system.</p>
904 <p>Another requirement for emulating the NCSA server is being able to handle
905 the per-directory configuration files, generally called
906 <code>.htaccess</code> files, though even in the NCSA server they can
907 contain directives which have nothing at all to do with access control.
908 Accordingly, after URI -> filename translation, but before performing any
909 other phase, the server walks down the directory hierarchy of the underlying
910 filesystem, following the translated pathname, to read any
911 <code>.htaccess</code> files which might be present. The information which
912 is read in then has to be <em>merged</em> with the applicable information
913 from the server's own config files (either from the <code class="directive"><a href="../mod/core.html#directory"><Directory></a></code> sections in
914 <code>access.conf</code>, or from defaults in <code>srm.conf</code>, which
915 actually behaves for most purposes almost exactly like <code><Directory
918 <p>Finally, after having served a request which involved reading
919 <code>.htaccess</code> files, we need to discard the storage allocated for
920 handling them. That is solved the same way it is solved wherever else
921 similar problems come up, by tying those structures to the per-transaction
924 <h3><a name="per-dir" id="per-dir">Per-directory configuration structures</a></h3>
925 <p>Let's look out how all of this plays out in <code>mod_mime.c</code>,
926 which defines the file typing handler which emulates the NCSA server's
927 behavior of determining file types from suffixes. What we'll be looking
928 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
929 <code>.htaccess</code> files, so they must be handled in the module's
930 private per-directory data, which in fact, consists of two separate
931 tables for MIME types and encoding information, and is declared as
934 <div class="example"><pre>typedef struct {
935 table *forced_types; /* Additional AddTyped stuff */
936 table *encoding_types; /* Added with AddEncoding... */
937 } mime_dir_config;</pre></div>
939 <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
940 one of the MIME module's commands, it needs to create a
941 <code>mime_dir_config</code> structure, so those commands have something
942 to act on. It does this by invoking the function it finds in the module's
943 `create per-dir config slot', with two arguments: the name of the
944 directory to which this configuration information applies (or
945 <code>NULL</code> for <code>srm.conf</code>), and a pointer to a
946 resource pool in which the allocation should happen.</p>
948 <p>(If we are reading a <code>.htaccess</code> file, that resource pool
949 is the per-request resource pool for the request; otherwise it is a
950 resource pool which is used for configuration data, and cleared on
951 restarts. Either way, it is important for the structure being created to
952 vanish when the pool is cleared, by registering a cleanup on the pool if
955 <p>For the MIME module, the per-dir config creation function just
956 <code>ap_palloc</code>s the structure above, and a creates a couple of
957 tables to fill it. That looks like this:</p>
959 <div class="example"><p><code>
960 void *create_mime_dir_config (pool *p, char *dummy)<br />
962 <span class="indent">
963 mime_dir_config *new =<br />
964 <span class="indent">
965 (mime_dir_config *) ap_palloc (p, sizeof(mime_dir_config));<br />
968 new->forced_types = ap_make_table (p, 4);<br />
969 new->encoding_types = ap_make_table (p, 4);<br />
976 <p>Now, suppose we've just read in a <code>.htaccess</code> file. We
977 already have the per-directory configuration structure for the next
978 directory up in the hierarchy. If the <code>.htaccess</code> file we just
979 read in didn't have any <code class="directive"><a href="../mod/mod_mime.html#addtype">AddType</a></code>
980 or <code class="directive"><a href="../mod/mod_mime.html#addencoding">AddEncoding</a></code> commands, its
981 per-directory config structure for the MIME module is still valid, and we
982 can just use it. Otherwise, we need to merge the two structures
985 <p>To do that, the server invokes the module's per-directory config merge
986 function, if one is present. That function takes three arguments: the two
987 structures being merged, and a resource pool in which to allocate the
988 result. For the MIME module, all that needs to be done is overlay the
989 tables from the new per-directory config structure with those from the
992 <div class="example"><p><code>
993 void *merge_mime_dir_configs (pool *p, void *parent_dirv, void *subdirv)<br />
995 <span class="indent">
996 mime_dir_config *parent_dir = (mime_dir_config *)parent_dirv;<br />
997 mime_dir_config *subdir = (mime_dir_config *)subdirv;<br />
998 mime_dir_config *new =<br />
999 <span class="indent">
1000 (mime_dir_config *)ap_palloc (p, sizeof(mime_dir_config));<br />
1003 new->forced_types = ap_overlay_tables (p, subdir->forced_types,<br />
1004 <span class="indent">
1005 parent_dir->forced_types);<br />
1007 new->encoding_types = ap_overlay_tables (p, subdir->encoding_types,<br />
1008 <span class="indent">
1009 parent_dir->encoding_types);<br />
1017 <p>As a note -- if there is no per-directory merge function present, the
1018 server will just use the subdirectory's configuration info, and ignore
1019 the parent's. For some modules, that works just fine (<em>e.g.</em>, for
1020 the includes module, whose per-directory configuration information
1021 consists solely of the state of the <code>XBITHACK</code>), and for those
1022 modules, you can just not declare one, and leave the corresponding
1023 structure slot in the module itself <code>NULL</code>.</p>
1026 <h3><a name="commands" id="commands">Command handling</a></h3>
1027 <p>Now that we have these structures, we need to be able to figure out how
1028 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
1029 the module's command table. That table contains information on how many
1030 arguments the commands take, and in what formats, where it is permitted,
1031 and so forth. That information is sufficient to allow the server to invoke
1032 most command-handling functions with pre-parsed arguments. Without further
1033 ado, let's look at the <code class="directive"><a href="../mod/mod_mime.html#addtype">AddType</a></code>
1034 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
1037 <div class="example"><p><code>
1038 char *add_type(cmd_parms *cmd, mime_dir_config *m, char *ct, char *ext)<br />
1040 <span class="indent">
1041 if (*ext == '.') ++ext;<br />
1042 ap_table_set (m->forced_types, ext, ct);<br />
1048 <p>This command handler is unusually simple. As you can see, it takes
1049 four arguments, two of which are pre-parsed arguments, the third being the
1050 per-directory configuration structure for the module in question, and the
1051 fourth being a pointer to a <code>cmd_parms</code> structure. That
1052 structure contains a bunch of arguments which are frequently of use to
1053 some, but not all, commands, including a resource pool (from which memory
1054 can be allocated, and to which cleanups should be tied), and the (virtual)
1055 server being configured, from which the module's per-server configuration
1056 data can be obtained if required.</p>
1058 <p>Another way in which this particular command handler is unusually
1059 simple is that there are no error conditions which it can encounter. If
1060 there were, it could return an error message instead of <code>NULL</code>;
1061 this causes an error to be printed out on the server's
1062 <code>stderr</code>, followed by a quick exit, if it is in the main config
1063 files; for a <code>.htaccess</code> file, the syntax error is logged in
1064 the server error log (along with an indication of where it came from), and
1065 the request is bounced with a server error response (HTTP error status,
1068 <p>The MIME module's command table has entries for these commands, which
1071 <div class="example"><p><code>
1072 command_rec mime_cmds[] = {<br />
1073 <span class="indent">
1074 { "AddType", add_type, NULL, OR_FILEINFO, TAKE2,<br />
1075 <span class="indent">"a mime type followed by a file extension" },<br /></span>
1076 { "AddEncoding", add_encoding, NULL, OR_FILEINFO, TAKE2,<br />
1077 <span class="indent">
1078 "an encoding (<em>e.g.</em>, gzip), followed by a file extension" },<br />
1085 <p>The entries in these tables are:</p>
1087 <li>The name of the command</li>
1088 <li>The function which handles it</li>
1089 <li>a <code>(void *)</code> pointer, which is passed in the
1090 <code>cmd_parms</code> structure to the command handler ---
1091 this is useful in case many similar commands are handled by
1092 the same function.</li>
1094 <li>A bit mask indicating where the command may appear. There
1095 are mask bits corresponding to each
1096 <code>AllowOverride</code> option, and an additional mask
1097 bit, <code>RSRC_CONF</code>, indicating that the command may
1098 appear in the server's own config files, but <em>not</em> in
1099 any <code>.htaccess</code> file.</li>
1101 <li>A flag indicating how many arguments the command handler
1102 wants pre-parsed, and how they should be passed in.
1103 <code>TAKE2</code> indicates two pre-parsed arguments. Other
1104 options are <code>TAKE1</code>, which indicates one
1105 pre-parsed argument, <code>FLAG</code>, which indicates that
1106 the argument should be <code>On</code> or <code>Off</code>,
1107 and is passed in as a boolean flag, <code>RAW_ARGS</code>,
1108 which causes the server to give the command the raw, unparsed
1109 arguments (everything but the command name itself). There is
1110 also <code>ITERATE</code>, which means that the handler looks
1111 the same as <code>TAKE1</code>, but that if multiple
1112 arguments are present, it should be called multiple times,
1113 and finally <code>ITERATE2</code>, which indicates that the
1114 command handler looks like a <code>TAKE2</code>, but if more
1115 arguments are present, then it should be called multiple
1116 times, holding the first argument constant.</li>
1118 <li>Finally, we have a string which describes the arguments
1119 that should be present. If the arguments in the actual config
1120 file are not as required, this string will be used to help
1121 give a more specific error message. (You can safely leave
1122 this <code>NULL</code>).</li>
1125 <p>Finally, having set this all up, we have to use it. This is ultimately
1126 done in the module's handlers, specifically for its file-typing handler,
1127 which looks more or less like this; note that the per-directory
1128 configuration structure is extracted from the <code>request_rec</code>'s
1129 per-directory configuration vector by using the
1130 <code>ap_get_module_config</code> function.</p>
1132 <div class="example"><p><code>
1133 int find_ct(request_rec *r)<br />
1135 <span class="indent">
1137 char *fn = ap_pstrdup (r->pool, r->filename);<br />
1138 mime_dir_config *conf = (mime_dir_config *)<br />
1139 <span class="indent">
1140 ap_get_module_config(r->per_dir_config, &mime_module);<br />
1144 if (S_ISDIR(r->finfo.st_mode)) {<br />
1145 <span class="indent">
1146 r->content_type = DIR_MAGIC_TYPE;<br />
1151 if((i=ap_rind(fn,'.')) < 0) return DECLINED;<br />
1154 if ((type = ap_table_get (conf->encoding_types, &fn[i])))<br />
1156 <span class="indent">
1157 r->content_encoding = type;<br />
1159 /* go back to previous extension to try to use it as a type */<br />
1160 fn[i-1] = '\0';<br />
1161 if((i=ap_rind(fn,'.')) < 0) return OK;<br />
1166 if ((type = ap_table_get (conf->forced_types, &fn[i])))<br />
1168 <span class="indent">
1169 r->content_type = type;<br />
1179 <h3><a name="servconf" id="servconf">Side notes -- per-server configuration,
1180 virtual servers, <em>etc</em>.</a></h3>
1181 <p>The basic ideas behind per-server module configuration are basically
1182 the same as those for per-directory configuration; there is a creation
1183 function and a merge function, the latter being invoked where a virtual
1184 server has partially overridden the base server configuration, and a
1185 combined structure must be computed. (As with per-directory configuration,
1186 the default if no merge function is specified, and a module is configured
1187 in some virtual server, is that the base configuration is simply
1190 <p>The only substantial difference is that when a command needs to
1191 configure the per-server private module data, it needs to go to the
1192 <code>cmd_parms</code> data to get at it. Here's an example, from the
1193 alias module, which also indicates how a syntax error can be returned
1194 (note that the per-directory configuration argument to the command
1195 handler is declared as a dummy, since the module doesn't actually have
1196 per-directory config data):</p>
1198 <div class="example"><p><code>
1199 char *add_redirect(cmd_parms *cmd, void *dummy, char *f, char *url)<br />
1201 <span class="indent">
1202 server_rec *s = cmd->server;<br />
1203 alias_server_conf *conf = (alias_server_conf *)<br />
1204 <span class="indent">
1205 ap_get_module_config(s->module_config,&alias_module);<br />
1207 alias_entry *new = ap_push_array (conf->redirects);<br />
1209 if (!ap_is_url (url)) return "Redirect to non-URL";<br />
1211 new->fake = f; new->real = url;<br />
1218 <div class="bottomlang">
1219 <p><span>Available Languages: </span><a href="../en/developer/API.html" title="English"> en </a></p>
1220 </div><div id="footer">
1221 <p class="apache">Copyright 2009 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>
1222 <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>