1 | /**********************************************************************
|
---|
2 | *
|
---|
3 | * cgiutils.cpp -- general cgi utilities
|
---|
4 | * Copyright (C) 1999 The New Zealand Digital Library Project
|
---|
5 | *
|
---|
6 | * A component of the Greenstone digital library software
|
---|
7 | * from the New Zealand Digital Library Project at the
|
---|
8 | * University of Waikato, New Zealand.
|
---|
9 | *
|
---|
10 | * This program is free software; you can redistribute it and/or modify
|
---|
11 | * it under the terms of the GNU General Public License as published by
|
---|
12 | * the Free Software Foundation; either version 2 of the License, or
|
---|
13 | * (at your option) any later version.
|
---|
14 | *
|
---|
15 | * This program is distributed in the hope that it will be useful,
|
---|
16 | * but WITHOUT ANY WARRANTY; without even the implied warranty of
|
---|
17 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
---|
18 | * GNU General Public License for more details.
|
---|
19 | *
|
---|
20 | * You should have received a copy of the GNU General Public License
|
---|
21 | * along with this program; if not, write to the Free Software
|
---|
22 | * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
|
---|
23 | *
|
---|
24 | * $Id: cgiutils.cpp 614 1999-09-21 21:41:17Z sjboddie $
|
---|
25 | *
|
---|
26 | *********************************************************************/
|
---|
27 |
|
---|
28 | /*
|
---|
29 | $Log$
|
---|
30 | Revision 1.10 1999/09/21 21:41:17 sjboddie
|
---|
31 | fixed a couple of problems in what I committed last
|
---|
32 |
|
---|
33 | Revision 1.9 1999/09/21 11:30:39 sjboddie
|
---|
34 | added ability to escape out '-' in saved args (currently
|
---|
35 | with a 'Z' ???)
|
---|
36 |
|
---|
37 | Revision 1.8 1999/09/07 04:56:53 sjboddie
|
---|
38 | added GPL notice
|
---|
39 |
|
---|
40 | Revision 1.7 1999/08/25 22:27:13 sjboddie
|
---|
41 | prevented cgi_safe from converting '+' and '-'. It was causing problems
|
---|
42 | with query strings containing spaces. The space was being converted
|
---|
43 | to '+', then %2b, then %xx2b over time when saved in compressed args.
|
---|
44 | I hope this won't cause problems elsewhere...
|
---|
45 |
|
---|
46 | Revision 1.6 1999/07/11 01:05:19 rjmcnab
|
---|
47 | Stored origin of cgiarg with argument.
|
---|
48 |
|
---|
49 | Revision 1.5 1999/06/26 01:08:36 rjmcnab
|
---|
50 | Added encoding and decoding of multibyte compresesd arguments.
|
---|
51 |
|
---|
52 | Revision 1.4 1999/06/08 22:03:43 sjboddie
|
---|
53 | query string is now made cgi safe before being added to compressed args
|
---|
54 |
|
---|
55 | Revision 1.3 1999/02/08 01:28:00 rjmcnab
|
---|
56 |
|
---|
57 | Got the receptionist producing something using the statusaction.
|
---|
58 |
|
---|
59 | Revision 1.2 1999/02/05 10:42:43 rjmcnab
|
---|
60 |
|
---|
61 | Continued working on receptionist
|
---|
62 |
|
---|
63 | Revision 1.1 1999/01/08 08:40:56 rjmcnab
|
---|
64 |
|
---|
65 | Moved from lib directory.
|
---|
66 |
|
---|
67 | Revision 1.1 1999/01/08 03:57:45 rjmcnab
|
---|
68 |
|
---|
69 | Initial revision
|
---|
70 |
|
---|
71 | */
|
---|
72 |
|
---|
73 |
|
---|
74 | #include "cgiutils.h"
|
---|
75 |
|
---|
76 |
|
---|
77 | static unsigned short hexdigit (unsigned short c) {
|
---|
78 | if (c >= '0' && c <= '9') return (c-'0');
|
---|
79 | if (c >= 'a' && c <= 'f') return (c-'a'+10);
|
---|
80 | if (c >= 'A' && c <= 'F') return (c-'A'+10);
|
---|
81 | return c;
|
---|
82 | }
|
---|
83 |
|
---|
84 |
|
---|
85 | static void c2hex (unsigned short c, text_t &t) {
|
---|
86 | t.clear();
|
---|
87 |
|
---|
88 | if (c >= 256) {
|
---|
89 | t = "20"; // ' '
|
---|
90 | return;
|
---|
91 | }
|
---|
92 |
|
---|
93 | unsigned short o1, o2;
|
---|
94 |
|
---|
95 | o1 = (c/16) % 16;
|
---|
96 | o2 = c % 16;
|
---|
97 | if (o1 >= 10) o1 += 'a' - 10;
|
---|
98 | else o1 += '0';
|
---|
99 | if (o2 >= 10) o2 += 'a' - 10;
|
---|
100 | else o2 += '0';
|
---|
101 |
|
---|
102 | t.push_back(o1);
|
---|
103 | t.push_back(o2);
|
---|
104 | }
|
---|
105 |
|
---|
106 | // convert %xx and + to their appropriate equivalents
|
---|
107 | void decode_cgi_arg (text_t &argstr) {
|
---|
108 | text_t::iterator in = argstr.begin();
|
---|
109 | text_t::iterator out = in;
|
---|
110 | text_t::iterator end = argstr.end();
|
---|
111 |
|
---|
112 | while (in != end) {
|
---|
113 | if (*in == '+') *out = ' ';
|
---|
114 |
|
---|
115 | else if (*in == '%') {
|
---|
116 | unsigned short c = '%';
|
---|
117 | in++;
|
---|
118 | if (in != end) {
|
---|
119 | c = hexdigit (*in);
|
---|
120 | in++;
|
---|
121 | }
|
---|
122 | if (in != end && c < 16) { // sanity check on the previous character
|
---|
123 | c = c*16 + hexdigit (*in);
|
---|
124 | }
|
---|
125 |
|
---|
126 | *out = c;
|
---|
127 | } else *out = *in;
|
---|
128 |
|
---|
129 | if (in != end) in++;
|
---|
130 | out++;
|
---|
131 | }
|
---|
132 |
|
---|
133 | // remove the excess characters
|
---|
134 | argstr.erase (out, end);
|
---|
135 | }
|
---|
136 |
|
---|
137 |
|
---|
138 | // split up the cgi arguments
|
---|
139 | void split_cgi_args (text_t argstr, cgiargsclass &args) {
|
---|
140 | args.clear();
|
---|
141 |
|
---|
142 | text_t::iterator here = argstr.begin();
|
---|
143 | text_t::iterator end = argstr.end();
|
---|
144 | text_t key, value;
|
---|
145 |
|
---|
146 | // extract out the key=value pairs
|
---|
147 | while (here != end) {
|
---|
148 | // get the next key and value pair
|
---|
149 | here = getdelimitstr (here, end, '=', key);
|
---|
150 | here = getdelimitstr (here, end, '&', value);
|
---|
151 |
|
---|
152 | // convert %xx and + to their appropriate equivalents
|
---|
153 | decode_cgi_arg (value);
|
---|
154 |
|
---|
155 | value.setencoding(1); // other encoding
|
---|
156 | // store this key=value pair
|
---|
157 | if (!key.empty()) args.setarg (key, value, cgiarg_t::cgi_arg);
|
---|
158 | }
|
---|
159 | }
|
---|
160 |
|
---|
161 | text_t minus_safe (const text_t &intext) {
|
---|
162 |
|
---|
163 | text_t outtext;
|
---|
164 |
|
---|
165 | text_t::const_iterator here = intext.begin ();
|
---|
166 | text_t::const_iterator end = intext.end ();
|
---|
167 |
|
---|
168 | while (here != end) {
|
---|
169 | if (*here == '-') outtext += "Z-";
|
---|
170 | else outtext.push_back (*here);
|
---|
171 | here ++;
|
---|
172 | }
|
---|
173 | outtext = cgi_safe (outtext);
|
---|
174 | return outtext;
|
---|
175 | }
|
---|
176 |
|
---|
177 | text_t cgi_safe (const text_t &intext) {
|
---|
178 | text_t outtext;
|
---|
179 |
|
---|
180 | text_t::const_iterator here = intext.begin ();
|
---|
181 | text_t::const_iterator end = intext.end ();
|
---|
182 | unsigned short c;
|
---|
183 | text_t ttmp;
|
---|
184 |
|
---|
185 | while (here != end) {
|
---|
186 | c = *here;
|
---|
187 | if (((c >= 'a') && (c <= 'z')) ||
|
---|
188 | ((c >= 'A') && (c <= 'Z')) ||
|
---|
189 | ((c >= '0') && (c <= '9')) ||
|
---|
190 | (c == '+') || (c == '%') || (c == '-')) {
|
---|
191 | // alphanumeric character
|
---|
192 | outtext.push_back(c);
|
---|
193 | } else if (c == ' ') {
|
---|
194 | // space
|
---|
195 | outtext.push_back('+');
|
---|
196 | } else {
|
---|
197 | // everything else
|
---|
198 | outtext.push_back('%');
|
---|
199 | c2hex(c, ttmp);
|
---|
200 | outtext += ttmp;
|
---|
201 | }
|
---|
202 |
|
---|
203 | here++;
|
---|
204 | }
|
---|
205 |
|
---|
206 | return outtext;
|
---|
207 | }
|
---|
208 |
|
---|
209 |
|
---|
210 |
|
---|
211 |
|
---|
212 | static text_t::const_iterator get_next_save_arg (text_t::const_iterator first,
|
---|
213 | text_t::const_iterator last,
|
---|
214 | text_t &argname) {
|
---|
215 | first = getdelimitstr (first, last, '-', argname);
|
---|
216 | return first;
|
---|
217 | }
|
---|
218 |
|
---|
219 |
|
---|
220 | // check_save_conf_str checks the configuration string for
|
---|
221 | // the saved args and makes sure it does not conflict with
|
---|
222 | // the information about the arguments. If an error is encountered
|
---|
223 | // it will return false and the program should not produce any
|
---|
224 | // output.
|
---|
225 | bool check_save_conf_str (const text_t &saveconf,
|
---|
226 | const cgiargsinfoclass &argsinfo,
|
---|
227 | ostream &logout) {
|
---|
228 | outconvertclass text_t2ascii;
|
---|
229 |
|
---|
230 | text_tset argsset;
|
---|
231 | text_t::const_iterator saveconfhere = saveconf.begin ();
|
---|
232 | text_t::const_iterator saveconfend = saveconf.end ();
|
---|
233 | text_t argname;
|
---|
234 | const cgiarginfo *info;
|
---|
235 |
|
---|
236 | // first check to make sure all saved arguments can be saved
|
---|
237 |
|
---|
238 | while (saveconfhere != saveconfend) {
|
---|
239 | saveconfhere = get_next_save_arg (saveconfhere, saveconfend, argname);
|
---|
240 |
|
---|
241 | if (!argname.empty()) {
|
---|
242 | // save the argument name for later
|
---|
243 | argsset.insert (argname);
|
---|
244 |
|
---|
245 | // check the argument
|
---|
246 | info = argsinfo.getarginfo (argname);
|
---|
247 | if (info == NULL) {
|
---|
248 | logout << text_t2ascii << "Error: the cgi argument \"" << argname
|
---|
249 | << "\" is used in the configuration string for the\n"
|
---|
250 | << "saved arguments but does not exist as a valid argument.\n\n";
|
---|
251 | return false;
|
---|
252 | }
|
---|
253 | if (info->savedarginfo == cgiarginfo::mustnot) {
|
---|
254 | logout << text_t2ascii << "Error: the cgi argument \"" << argname
|
---|
255 | << "\" is used in the configuration string for the\n"
|
---|
256 | << "saved arguments but has been specified as an argument whose\n"
|
---|
257 | << "state must not be saved.\n\n";
|
---|
258 | return false;
|
---|
259 | }
|
---|
260 | }
|
---|
261 | }
|
---|
262 |
|
---|
263 |
|
---|
264 | // next check that all saved arguments that should be saved
|
---|
265 | // are saved
|
---|
266 | cgiargsinfoclass::const_iterator argsinfohere = argsinfo.begin ();
|
---|
267 | cgiargsinfoclass::const_iterator argsinfoend = argsinfo.end ();
|
---|
268 |
|
---|
269 | while (argsinfohere != argsinfoend) {
|
---|
270 | if (((*argsinfohere).second.savedarginfo == cgiarginfo::must) &&
|
---|
271 | (argsset.find((*argsinfohere).second.shortname) == argsset.end())) {
|
---|
272 | logout << text_t2ascii << "Error: the cgi argument \""
|
---|
273 | << (*argsinfohere).second.shortname << "\" was specified as needing to\n"
|
---|
274 | << "be save but was not listed in the saved arguments.\n\n";
|
---|
275 | return false;
|
---|
276 | }
|
---|
277 |
|
---|
278 | argsinfohere++;
|
---|
279 | }
|
---|
280 |
|
---|
281 | return true; // made it, no clashes
|
---|
282 | }
|
---|
283 |
|
---|
284 |
|
---|
285 | // create_save_conf_str will create a configuration string
|
---|
286 | // based on the information in argsinfo. This method of configuration
|
---|
287 | // is not recomended as small changes can produce large changes in
|
---|
288 | // the resulting configuration string (for instance a totally different
|
---|
289 | // ordering). Only arguments which "must" be saved are included in
|
---|
290 | // the resulting string.
|
---|
291 | text_t create_save_conf_str (const cgiargsinfoclass &argsinfo,
|
---|
292 | ostream &/*logout*/) {
|
---|
293 | cgiargsinfoclass::const_iterator argsinfohere = argsinfo.begin ();
|
---|
294 | cgiargsinfoclass::const_iterator argsinfoend = argsinfo.end ();
|
---|
295 | text_t saveconf;
|
---|
296 | bool first = true;
|
---|
297 |
|
---|
298 | while (argsinfohere != argsinfoend) {
|
---|
299 | // save this argument if it must be saved
|
---|
300 | if ((*argsinfohere).second.savedarginfo == cgiarginfo::must) {
|
---|
301 | if (!first) saveconf.push_back ('-');
|
---|
302 | else first = false;
|
---|
303 | saveconf += (*argsinfohere).second.shortname;
|
---|
304 | }
|
---|
305 |
|
---|
306 | argsinfohere++;
|
---|
307 | }
|
---|
308 |
|
---|
309 | return saveconf;
|
---|
310 | }
|
---|
311 |
|
---|
312 |
|
---|
313 | // expand_save_args will expand the saved arguments based
|
---|
314 | // on saveconf placing the results in args if they are not
|
---|
315 | // already defined. If it encounters an error it will return false
|
---|
316 | // and output more information to logout.
|
---|
317 | bool expand_save_args (const cgiargsinfoclass &argsinfo,
|
---|
318 | const text_t &saveconf,
|
---|
319 | cgiargsclass &args,
|
---|
320 | ostream &logout) {
|
---|
321 | outconvertclass text_t2ascii;
|
---|
322 |
|
---|
323 | text_t *arg_e = args.getarg("e");
|
---|
324 | if (arg_e == NULL) return true; // no compressed arguments
|
---|
325 | if (arg_e->empty()) return true; // no compressed arguments
|
---|
326 |
|
---|
327 | text_t argname, argvalue;
|
---|
328 | const cgiarginfo *argnameinfo;
|
---|
329 |
|
---|
330 | text_t::const_iterator saveconfhere = saveconf.begin();
|
---|
331 | text_t::const_iterator saveconfend = saveconf.end();
|
---|
332 |
|
---|
333 | text_t::iterator arg_ehere = arg_e->begin();
|
---|
334 | text_t::iterator arg_eend = arg_e->end();
|
---|
335 | while (saveconfhere != saveconfend && arg_ehere != arg_eend) {
|
---|
336 | saveconfhere = get_next_save_arg (saveconfhere, saveconfend, argname);
|
---|
337 |
|
---|
338 | if (!argname.empty()) {
|
---|
339 | // found another entry
|
---|
340 | argnameinfo = argsinfo.getarginfo (argname);
|
---|
341 |
|
---|
342 | if (argnameinfo == NULL) {
|
---|
343 | // no information about the argument could be found
|
---|
344 | // we can't keep going because we don't know whether
|
---|
345 | // this argument is a single or multiple character value
|
---|
346 | logout << text_t2ascii << "Error: the cgi argument \"" << argname
|
---|
347 | << "\" was specified as being a compressed argument\n"
|
---|
348 | << "but no information about it could be found within the "
|
---|
349 | << "cgiargsinfoclass.\n";
|
---|
350 | return false;
|
---|
351 |
|
---|
352 | } else {
|
---|
353 |
|
---|
354 | // found the argument information
|
---|
355 | if (argnameinfo->multiplechar) {
|
---|
356 | text_t::const_iterator sav = arg_ehere;
|
---|
357 | arg_ehere = getdelimitstr (arg_ehere, arg_eend, '-', argvalue);
|
---|
358 | // replace any '-' chars escaped with 'Z'
|
---|
359 | if (*(arg_ehere-2) == 'Z') {
|
---|
360 | argvalue.clear();
|
---|
361 | arg_ehere = (findchar (arg_ehere, arg_eend, '-')) + 1;
|
---|
362 | while (sav != (arg_ehere-1)) {
|
---|
363 | if (!((*sav == 'Z') && (*(sav+1) == '-'))) argvalue.push_back (*sav);
|
---|
364 | sav ++;
|
---|
365 | }
|
---|
366 | }
|
---|
367 |
|
---|
368 | argvalue.setencoding(1); // other encoding
|
---|
369 | if (!argvalue.empty()) args.setdefaultarg (argname, argvalue, cgiarg_t::compressed_arg);
|
---|
370 | } else {
|
---|
371 | args.setdefaultcarg (argname,*arg_ehere, cgiarg_t::compressed_arg);
|
---|
372 | arg_ehere++;
|
---|
373 | }
|
---|
374 | }
|
---|
375 | }
|
---|
376 | }
|
---|
377 |
|
---|
378 | return true;
|
---|
379 | }
|
---|
380 |
|
---|
381 |
|
---|
382 | // adds the default values for those arguments which have not
|
---|
383 | // been specified
|
---|
384 | void add_default_args (const cgiargsinfoclass &argsinfo,
|
---|
385 | cgiargsclass &args,
|
---|
386 | ostream &/*logout*/) {
|
---|
387 | cgiargsinfoclass::const_iterator argsinfohere = argsinfo.begin ();
|
---|
388 | cgiargsinfoclass::const_iterator argsinfoend = argsinfo.end ();
|
---|
389 |
|
---|
390 | while (argsinfohere != argsinfoend) {
|
---|
391 | if ((*argsinfohere).second.defaultstatus != cgiarginfo::none) {
|
---|
392 | args.setdefaultarg ((*argsinfohere).second.shortname,
|
---|
393 | (*argsinfohere).second.argdefault, cgiarg_t::default_arg);
|
---|
394 | }
|
---|
395 | argsinfohere++;
|
---|
396 | }
|
---|
397 | }
|
---|
398 |
|
---|
399 |
|
---|
400 | // compress_save_args will compress the arguments and return
|
---|
401 | // them in compressed_args. If an error was encountered
|
---|
402 | // compressed_args will be set to to "", an error will be
|
---|
403 | // written to logout, and the function will return false.
|
---|
404 | bool compress_save_args (const cgiargsinfoclass &argsinfo,
|
---|
405 | const text_t &saveconf,
|
---|
406 | cgiargsclass &args,
|
---|
407 | text_t &compressed_args,
|
---|
408 | outconvertclass &outconvert,
|
---|
409 | ostream &logout) {
|
---|
410 | outconvertclass text_t2ascii;
|
---|
411 |
|
---|
412 | compressed_args.clear();
|
---|
413 |
|
---|
414 | text_t argname, argvalue;
|
---|
415 | const cgiarginfo *argnameinfo;
|
---|
416 |
|
---|
417 | text_t::const_iterator saveconfhere = saveconf.begin();
|
---|
418 | text_t::const_iterator saveconfend = saveconf.end();
|
---|
419 |
|
---|
420 | while (saveconfhere != saveconfend) {
|
---|
421 | saveconfhere = get_next_save_arg (saveconfhere, saveconfend, argname);
|
---|
422 |
|
---|
423 | if (!argname.empty()) {
|
---|
424 | // found another entry
|
---|
425 | argnameinfo = argsinfo.getarginfo (argname);
|
---|
426 |
|
---|
427 | if (argnameinfo == NULL) {
|
---|
428 | // no information about the argument could be found
|
---|
429 | // we can't keep going because we don't know whether
|
---|
430 | // this argument is a single or multiple character value
|
---|
431 | logout << text_t2ascii << "Error: the cgi argument \"" << argname
|
---|
432 | << "\" was specified as being a compressed argument\n"
|
---|
433 | << "but no information about it could be found within the "
|
---|
434 | << "cgiargsinfoclass.\n";
|
---|
435 | compressed_args.clear();
|
---|
436 | return false;
|
---|
437 |
|
---|
438 | } else {
|
---|
439 | // found the argument information
|
---|
440 | if (argnameinfo->multiplechar) {
|
---|
441 | // multiple character argument -- sort out any '-' chars
|
---|
442 | compressed_args += minus_safe (outconvert.convert(args[argname]));
|
---|
443 | if (saveconfhere != saveconfend) compressed_args.push_back ('-');
|
---|
444 |
|
---|
445 | } else {
|
---|
446 | // single character argument
|
---|
447 | if (args[argname].size() == 0) {
|
---|
448 | logout << text_t2ascii << "Error: the cgi argument \"" << argname
|
---|
449 | << "\" was specified as being a compressed argument which\n"
|
---|
450 | << "should have a one character value but it was empty.\n\n";
|
---|
451 | compressed_args.clear ();
|
---|
452 | return false;
|
---|
453 |
|
---|
454 | } else if (args[argname].size() > 1) {
|
---|
455 | logout << text_t2ascii << "Error: the cgi argument \"" << argname
|
---|
456 | << "\" was specified as being a compressed argument which\n"
|
---|
457 | << "should have a one character value but it had multiple characters.\n\n";
|
---|
458 | compressed_args.clear ();
|
---|
459 | return false;
|
---|
460 | }
|
---|
461 |
|
---|
462 | // everything is ok
|
---|
463 | compressed_args += args[argname];
|
---|
464 | }
|
---|
465 | }
|
---|
466 | }
|
---|
467 | }
|
---|
468 |
|
---|
469 | return true;
|
---|
470 | }
|
---|
471 |
|
---|
472 |
|
---|
473 | // args_tounicode converts any arguments which are not in unicode
|
---|
474 | // to unicode using inconvert
|
---|
475 | void args_tounicode (cgiargsclass &args, inconvertclass &inconvert) {
|
---|
476 | cgiargsclass::iterator here = args.begin();
|
---|
477 | cgiargsclass::iterator end = args.end();
|
---|
478 |
|
---|
479 | while (here != end) {
|
---|
480 | if ((*here).second.value.getencoding() > 0) {
|
---|
481 | (*here).second.value = inconvert.convert((*here).second.value);
|
---|
482 | }
|
---|
483 |
|
---|
484 | here++;
|
---|
485 | }
|
---|
486 | }
|
---|