1 /* ====================================================================
2 * The Apache Software License, Version 1.1
4 * Copyright (c) 2000 The Apache Software Foundation. All rights
7 * Redistribution and use in source and binary forms, with or without
8 * modification, are permitted provided that the following conditions
11 * 1. Redistributions of source code must retain the above copyright
12 * notice, this list of conditions and the following disclaimer.
14 * 2. Redistributions in binary form must reproduce the above copyright
15 * notice, this list of conditions and the following disclaimer in
16 * the documentation and/or other materials provided with the
19 * 3. The end-user documentation included with the redistribution,
20 * if any, must include the following acknowledgment:
21 * "This product includes software developed by the
22 * Apache Software Foundation (http://www.apache.org/)."
23 * Alternately, this acknowledgment may appear in the software itself,
24 * if and wherever such third-party acknowledgments normally appear.
26 * 4. The names "Apache" and "Apache Software Foundation" must
27 * not be used to endorse or promote products derived from this
28 * software without prior written permission. For written
29 * permission, please contact apache@apache.org.
31 * 5. Products derived from this software may not be called "Apache",
32 * nor may "Apache" appear in their name, without prior written
33 * permission of the Apache Software Foundation.
35 * THIS SOFTWARE IS PROVIDED ``AS IS'' AND ANY EXPRESSED OR IMPLIED
36 * WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
37 * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
38 * DISCLAIMED. IN NO EVENT SHALL THE APACHE SOFTWARE FOUNDATION OR
39 * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
40 * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
41 * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF
42 * USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
43 * ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
44 * OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT
45 * OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
47 * ====================================================================
49 * This software consists of voluntary contributions made by many
50 * individuals on behalf of the Apache Software Foundation. For more
51 * information on the Apache Software Foundation, please see
52 * <http://www.apache.org/>.
54 * Portions of this software are based upon public domain software
55 * originally written at the National Center for Supercomputing Applications,
56 * University of Illinois, Urbana-Champaign.
60 #include "apr_file_io.h"
61 #include "apr_strings.h"
63 #define APR_WANT_STRFUNC
66 #define WANT_BASENAME_MATCH
69 #include "http_core.h"
70 #include "http_config.h"
71 #include "http_request.h"
74 /* mod_speling.c - by Alexei Kosut <akosut@organic.com> June, 1996
76 * This module is transparent, and simple. It attempts to correct
77 * misspellings of URLs that users might have entered, namely by checking
78 * capitalizations. If it finds a match, it sends a redirect.
80 * 08-Aug-1997 <Martin.Kraemer@Mch.SNI.De>
81 * o Upgraded module interface to apache_1.3a2-dev API (more NULL's in
83 * o Integrated tcsh's "spelling correction" routine which allows one
84 * misspelling (character insertion/omission/typo/transposition).
85 * Rewrote it to ignore case as well. This ought to catch the majority
86 * of misspelled requests.
87 * o Commented out the second pass where files' suffixes are stripped.
88 * Given the better hit rate of the first pass, this rather ugly
89 * (request index.html, receive index.db ?!?!) solution can be
91 * o wrote a "kind of" html page for mod_speling
93 * Activate it with "CheckSpelling On"
96 AP_MODULE_DECLARE_DATA module speling_module;
103 * Create a configuration specific to this module for a server or directory
104 * location, and fill it with the default settings.
106 * The API says that in the absence of a merge function, the record for the
107 * closest ancestor is used exclusively. That's what we want, so we don't
108 * bother to have such a function.
111 static void *mkconfig(apr_pool_t *p)
113 spconfig *cfg = apr_pcalloc(p, sizeof(spconfig));
120 * Respond to a callback to create configuration record for a server or
123 static void *create_mconfig_for_server(apr_pool_t *p, server_rec *s)
129 * Respond to a callback to create a config record for a specific directory.
131 static void *create_mconfig_for_directory(apr_pool_t *p, char *dir)
137 * Handler for the CheckSpelling directive, which is FLAG.
139 static const char *set_speling(cmd_parms *cmd, void *mconfig, int arg)
141 spconfig *cfg = (spconfig *) mconfig;
148 * Define the directives specific to this module. This structure is referenced
149 * later by the 'module' structure.
151 static const command_rec speling_cmds[] =
153 AP_INIT_FLAG("CheckSpelling", set_speling, NULL, OR_OPTIONS,
154 "whether or not to fix miscapitalized/misspelled requests"),
160 SP_MISCAPITALIZED = 1,
161 SP_TRANSPOSITION = 2,
168 static const char *sp_reason_str[] =
172 "transposed characters",
175 "mistyped character",
185 * spdist() is taken from Kernighan & Pike,
186 * _The_UNIX_Programming_Environment_
187 * and adapted somewhat to correspond better to psychological reality.
188 * (Note the changes to the return values)
190 * According to Pollock and Zamora, CACM April 1984 (V. 27, No. 4),
191 * page 363, the correct order for this is:
192 * OMISSION = TRANSPOSITION > INSERTION > SUBSTITUTION
193 * thus, it was exactly backwards in the old version. -- PWP
195 * This routine was taken out of tcsh's spelling correction code
196 * (tcsh-6.07.04) and re-converted to apache data types ("char" type
197 * instead of tcsh's NLS'ed "Char"). Plus it now ignores the case
198 * during comparisons, so is a "approximate strcasecmp()".
199 * NOTE that is still allows only _one_ real "typo",
200 * it does NOT try to correct multiple errors.
203 static sp_reason spdist(const char *s, const char *t)
205 for (; apr_tolower(*s) == apr_tolower(*t); t++, s++) {
207 return SP_MISCAPITALIZED; /* exact match (sans case) */
212 if (s[1] && t[1] && apr_tolower(*s) == apr_tolower(t[1])
213 && apr_tolower(*t) == apr_tolower(s[1])
214 && strcasecmp(s + 2, t + 2) == 0) {
215 return SP_TRANSPOSITION; /* transposition */
217 if (strcasecmp(s + 1, t + 1) == 0) {
218 return SP_SIMPLETYPO; /* 1 char mismatch */
221 if (strcasecmp(s + 1, t) == 0) {
222 return SP_EXTRACHAR; /* extra character */
225 if (*t && strcasecmp(s, t + 1) == 0) {
226 return SP_MISSINGCHAR; /* missing character */
228 return SP_VERYDIFFERENT; /* distance too large to fix. */
231 static int sort_by_quality(const void *left, const void *rite)
233 return (int) (((misspelled_file *) left)->quality)
234 - (int) (((misspelled_file *) rite)->quality);
237 static int check_speling(request_rec *r)
240 char *good, *bad, *postgood, *url;
242 int filoc, dotloc, urlen, pglen;
243 apr_array_header_t *candidates = NULL;
246 cfg = ap_get_module_config(r->per_dir_config, &speling_module);
251 /* We only want to worry about GETs */
252 if (r->method_number != M_GET) {
256 /* We've already got a file of some kind or another */
257 if (r->proxyreq || (r->finfo.filetype != 0)) {
261 /* This is a sub request - don't mess with it */
267 * The request should end up looking like this:
268 * r->uri: /correct-url/mispelling/more
269 * r->filename: /correct-file/mispelling r->path_info: /more
271 * So we do this in steps. First break r->filename into two pieces
274 filoc = ap_rind(r->filename, '/');
276 * Don't do anything if the request doesn't contain a slash, or
279 if (filoc == -1 || strcmp(r->uri, "/") == 0) {
283 /* good = /correct-file */
284 good = apr_pstrndup(r->pool, r->filename, filoc);
285 /* bad = mispelling */
286 bad = apr_pstrdup(r->pool, r->filename + filoc + 1);
287 /* postgood = mispelling/more */
288 postgood = apr_pstrcat(r->pool, bad, r->path_info, NULL);
290 urlen = strlen(r->uri);
291 pglen = strlen(postgood);
293 /* Check to see if the URL pieces add up */
294 if (strcmp(postgood, r->uri + (urlen - pglen))) {
298 /* url = /correct-url */
299 url = apr_pstrndup(r->pool, r->uri, (urlen - pglen));
301 /* Now open the directory and do ourselves a check... */
302 if (apr_dir_open(&dir, good, r->pool) != APR_SUCCESS) {
303 /* Oops, not a directory... */
307 candidates = apr_array_make(r->pool, 2, sizeof(misspelled_file));
309 dotloc = ap_ind(bad, '.');
311 dotloc = strlen(bad);
314 while (apr_dir_read(&dirent, APR_FINFO_DIRENT, dir) == APR_SUCCESS) {
318 * If we end up with a "fixed" URL which is identical to the
319 * requested one, we must have found a broken symlink or some such.
320 * Do _not_ try to redirect this, it causes a loop!
322 if (strcmp(bad, dirent.name) == 0) {
328 * miscapitalization errors are checked first (like, e.g., lower case
329 * file, upper case request)
331 else if (strcasecmp(bad, dirent.name) == 0) {
332 misspelled_file *sp_new;
334 sp_new = (misspelled_file *) apr_array_push(candidates);
335 sp_new->name = apr_pstrdup(r->pool, dirent.name);
336 sp_new->quality = SP_MISCAPITALIZED;
340 * simple typing errors are checked next (like, e.g.,
341 * missing/extra/transposed char)
343 else if ((q = spdist(bad, dirent.name)) != SP_VERYDIFFERENT) {
344 misspelled_file *sp_new;
346 sp_new = (misspelled_file *) apr_array_push(candidates);
347 sp_new->name = apr_pstrdup(r->pool, dirent.name);
352 * The spdist() should have found the majority of the misspelled
353 * requests. It is of questionable use to continue looking for
354 * files with the same base name, but potentially of totally wrong
355 * type (index.html <-> index.db).
356 * I would propose to not set the WANT_BASENAME_MATCH define.
357 * 08-Aug-1997 <Martin.Kraemer@Mch.SNI.De>
359 * However, Alexei replied giving some reasons to add it anyway:
360 * > Oh, by the way, I remembered why having the
361 * > extension-stripping-and-matching stuff is a good idea:
363 * > If you're using MultiViews, and have a file named foobar.html,
364 * > which you refer to as "foobar", and someone tried to access
365 * > "Foobar", mod_speling won't find it, because it won't find
366 * > anything matching that spelling. With the extension-munging,
367 * > it would locate "foobar.html". Not perfect, but I ran into
368 * > that problem when I first wrote the module.
371 #ifdef WANT_BASENAME_MATCH
373 * Okay... we didn't find anything. Now we take out the hard-core
374 * power tools. There are several cases here. Someone might have
375 * entered a wrong extension (.htm instead of .html or vice
376 * versa) or the document could be negotiated. At any rate, now
377 * we just compare stuff before the first dot. If it matches, we
378 * figure we got us a match. This can result in wrong things if
379 * there are files of different content types but the same prefix
380 * (e.g. foo.gif and foo.html) This code will pick the first one
381 * it finds. Better than a Not Found, though.
383 int entloc = ap_ind(dirent.name, '.');
385 entloc = strlen(dirent.name);
388 if ((dotloc == entloc)
389 && !strncasecmp(bad, dirent.name, dotloc)) {
390 misspelled_file *sp_new;
392 sp_new = (misspelled_file *) apr_array_push(candidates);
393 sp_new->name = apr_pstrdup(r->pool, dirent.name);
394 sp_new->quality = SP_VERYDIFFERENT;
401 if (candidates->nelts != 0) {
402 /* Wow... we found us a mispelling. Construct a fixed url */
405 misspelled_file *variant = (misspelled_file *) candidates->elts;
408 ref = apr_table_get(r->headers_in, "Referer");
410 qsort((void *) candidates->elts, candidates->nelts,
411 sizeof(misspelled_file), sort_by_quality);
414 * Conditions for immediate redirection:
415 * a) the first candidate was not found by stripping the suffix
416 * AND b) there exists only one candidate OR the best match is not
418 * then return a redirection right away.
420 if (variant[0].quality != SP_VERYDIFFERENT
421 && (candidates->nelts == 1
422 || variant[0].quality != variant[1].quality)) {
424 nuri = ap_escape_uri(r->pool, apr_pstrcat(r->pool, url,
426 r->path_info, NULL));
427 if (r->parsed_uri.query)
428 nuri = apr_pstrcat(r->pool, nuri, "?", r->parsed_uri.query, NULL);
430 apr_table_setn(r->headers_out, "Location",
431 ap_construct_url(r->pool, nuri, r));
433 ap_log_rerror(APLOG_MARK, APLOG_NOERRNO | APLOG_INFO, APR_SUCCESS,
435 ref ? "Fixed spelling: %s to %s from %s"
436 : "Fixed spelling: %s to %s",
439 return HTTP_MOVED_PERMANENTLY;
442 * Otherwise, a "[300] Multiple Choices" list with the variants is
448 apr_pool_t *sub_pool;
449 apr_array_header_t *t;
450 apr_array_header_t *v;
453 if (r->main == NULL) {
459 notes = r->main->notes;
462 if (apr_pool_create(&sub_pool, p) != APR_SUCCESS)
465 t = apr_array_make(sub_pool, candidates->nelts * 8 + 8,
467 v = apr_array_make(sub_pool, candidates->nelts * 5,
470 /* Generate the response text. */
472 *(const char **)apr_array_push(t) =
473 "The document name you requested (<code>";
474 *(const char **)apr_array_push(t) = ap_escape_html(sub_pool, r->uri);
475 *(const char **)apr_array_push(t) =
476 "</code>) could not be found on this server.\n"
477 "However, we found documents with names similar "
478 "to the one you requested.<p>"
479 "Available documents:\n<ul>\n";
481 for (i = 0; i < candidates->nelts; ++i) {
485 reason = sp_reason_str[(int) (variant[i].quality)];
486 /* The format isn't very neat... */
487 vuri = apr_pstrcat(sub_pool, url, variant[i].name, r->path_info,
488 (r->parsed_uri.query != NULL) ? "?" : "",
489 (r->parsed_uri.query != NULL)
490 ? r->parsed_uri.query : "",
492 *(const char **)apr_array_push(v) = "\"";
493 *(const char **)apr_array_push(v) = ap_escape_uri(sub_pool, vuri);
494 *(const char **)apr_array_push(v) = "\";\"";
495 *(const char **)apr_array_push(v) = reason;
496 *(const char **)apr_array_push(v) = "\"";
498 *(const char **)apr_array_push(t) = "<li><a href=\"";
499 *(const char **)apr_array_push(t) = ap_escape_uri(sub_pool, vuri);
500 *(const char **)apr_array_push(t) = "\">";
501 *(const char **)apr_array_push(t) = ap_escape_html(sub_pool, vuri);
502 *(const char **)apr_array_push(t) = "</a> (";
503 *(const char **)apr_array_push(t) = reason;
504 *(const char **)apr_array_push(t) = ")\n";
507 * when we have printed the "close matches" and there are
508 * more "distant matches" (matched by stripping the suffix),
509 * then we insert an additional separator text to suggest
510 * that the user LOOK CLOSELY whether these are really the
513 if (i > 0 && i < candidates->nelts - 1
514 && variant[i].quality != SP_VERYDIFFERENT
515 && variant[i + 1].quality == SP_VERYDIFFERENT) {
516 *(const char **)apr_array_push(t) =
517 "</ul>\nFurthermore, the following related "
518 "documents were found:\n<ul>\n";
521 *(const char **)apr_array_push(t) = "</ul>\n";
523 /* If we know there was a referring page, add a note: */
525 *(const char **)apr_array_push(t) =
526 "Please consider informing the owner of the "
528 *(const char **)apr_array_push(t) = ap_escape_uri(sub_pool, ref);
529 *(const char **)apr_array_push(t) = "\">referring page</a> "
530 "about the broken link.\n";
534 /* Pass our apr_table_t to http_protocol.c (see mod_negotiation): */
535 apr_table_setn(notes, "variant-list", apr_array_pstrcat(p, t, 0));
537 apr_table_mergen(r->subprocess_env, "VARIANTS",
538 apr_array_pstrcat(p, v, ','));
540 apr_pool_destroy(sub_pool);
542 ap_log_rerror(APLOG_MARK, APLOG_NOERRNO | APLOG_INFO, 0, r,
543 ref ? "Spelling fix: %s: %d candidates from %s"
544 : "Spelling fix: %s: %d candidates",
545 r->uri, candidates->nelts, ref);
547 return HTTP_MULTIPLE_CHOICES;
554 static void register_hooks(apr_pool_t *p)
556 ap_hook_fixups(check_speling,NULL,NULL,APR_HOOK_LAST);
559 module AP_MODULE_DECLARE_DATA speling_module =
561 STANDARD20_MODULE_STUFF,
562 create_mconfig_for_directory, /* create per-dir config */
563 NULL, /* merge per-dir config */
564 create_mconfig_for_server, /* server config */
565 NULL, /* merge server config */
566 speling_cmds, /* command apr_table_t */
567 register_hooks /* register hooks */