[108] | 1 | /**********************************************************************
|
---|
| 2 | *
|
---|
| 3 | * cgiutils.cpp -- general cgi utilities
|
---|
| 4 | * Copyright (C) 1999 The New Zealand Digital Library Project
|
---|
| 5 | *
|
---|
[533] | 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.
|
---|
[108] | 9 | *
|
---|
[533] | 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 | *
|
---|
[108] | 24 | *********************************************************************/
|
---|
| 25 |
|
---|
| 26 | #include "cgiutils.h"
|
---|
[12513] | 27 | #include "fileutil.h"
|
---|
[3151] | 28 | #include "gsdlunicode.h"
|
---|
[11998] | 29 | #include "fileutil.h"
|
---|
[3217] | 30 | #include "unitool.h" // in mg, for output_utf8_char
|
---|
[18882] | 31 | #include <cstdlib>
|
---|
[22796] | 32 | #include <time.h>
|
---|
[108] | 33 |
|
---|
[12513] | 34 | #if defined(GSDL_USE_OBJECTSPACE)
|
---|
| 35 | # include <ospace\std\iostream>
|
---|
| 36 | # include <ospace\std\fstream>
|
---|
| 37 | #elif defined(GSDL_USE_IOS_H)
|
---|
| 38 | # include <iostream.h>
|
---|
| 39 | # include <fstream.h>
|
---|
| 40 | #else
|
---|
| 41 | # include <iostream>
|
---|
| 42 | # include <fstream>
|
---|
| 43 | #endif
|
---|
[11998] | 44 |
|
---|
[12513] | 45 |
|
---|
[108] | 46 | static unsigned short hexdigit (unsigned short c) {
|
---|
| 47 | if (c >= '0' && c <= '9') return (c-'0');
|
---|
| 48 | if (c >= 'a' && c <= 'f') return (c-'a'+10);
|
---|
| 49 | if (c >= 'A' && c <= 'F') return (c-'A'+10);
|
---|
| 50 | return c;
|
---|
| 51 | }
|
---|
| 52 |
|
---|
| 53 |
|
---|
| 54 | static void c2hex (unsigned short c, text_t &t) {
|
---|
| 55 | t.clear();
|
---|
| 56 |
|
---|
| 57 | if (c >= 256) {
|
---|
| 58 | t = "20"; // ' '
|
---|
| 59 | return;
|
---|
| 60 | }
|
---|
| 61 |
|
---|
| 62 | unsigned short o1, o2;
|
---|
| 63 |
|
---|
| 64 | o1 = (c/16) % 16;
|
---|
| 65 | o2 = c % 16;
|
---|
| 66 | if (o1 >= 10) o1 += 'a' - 10;
|
---|
| 67 | else o1 += '0';
|
---|
| 68 | if (o2 >= 10) o2 += 'a' - 10;
|
---|
| 69 | else o2 += '0';
|
---|
| 70 |
|
---|
| 71 | t.push_back(o1);
|
---|
| 72 | t.push_back(o2);
|
---|
| 73 | }
|
---|
| 74 |
|
---|
[12513] | 75 | static text_t::iterator getline (text_t::iterator first,
|
---|
| 76 | text_t::iterator last,
|
---|
| 77 | bool include_crlf) {
|
---|
| 78 | while (first != last) {
|
---|
| 79 | if (((first+1) != last) && (*first == 13) && (*(first+1) == 10)) {
|
---|
| 80 | // found <CRLF>
|
---|
| 81 | if (include_crlf) first += 2;
|
---|
| 82 | break;
|
---|
| 83 | }
|
---|
[12794] | 84 |
|
---|
[12513] | 85 | first++;
|
---|
| 86 | }
|
---|
[12794] | 87 |
|
---|
[12513] | 88 | return first;
|
---|
| 89 | }
|
---|
| 90 |
|
---|
| 91 | static void process_post_section (text_t &argname, text_t &argdata, text_t &filename, text_t &filedata,
|
---|
| 92 | text_t &filetype, bool &isfile, text_t &argstr,
|
---|
| 93 | fileupload_tmap &fileuploads, const text_t &gsdlhome) {
|
---|
| 94 |
|
---|
| 95 | if (!argname.empty()) {
|
---|
| 96 |
|
---|
| 97 | if (!isfile) {
|
---|
| 98 | // argdata includes a trailing <CRLF> that we must remove
|
---|
| 99 | if ((argdata.size() > 1) && (*(argdata.end()-2) == 13) && (*(argdata.end()-1) == 10)) {
|
---|
| 100 | argdata.erase(argdata.end()-2, argdata.end());
|
---|
| 101 | }
|
---|
| 102 | if (!argstr.empty()) argstr += "&";
|
---|
| 103 | argstr += argname + "=" + argdata;
|
---|
| 104 |
|
---|
| 105 | } else if (!filename.empty()) {
|
---|
| 106 | // filedata includes a trailing <CRLF> that we must remove
|
---|
| 107 | if ((filedata.size() > 1) && (*(filedata.end()-2) == 13) && (*(filedata.end()-1) == 10)) {
|
---|
| 108 | filedata.erase(filedata.end()-2, filedata.end());
|
---|
| 109 | }
|
---|
| 110 |
|
---|
[22791] | 111 | // create tmp_name for storing the file on disk, using the current timestamp
|
---|
| 112 | text_t tmp_name(time(NULL));
|
---|
[12513] | 113 | tmp_name = filename_cat(gsdlhome, "tmp", tmp_name);
|
---|
[12794] | 114 |
|
---|
[12513] | 115 | char *tmp_name_c = tmp_name.getcstr();
|
---|
| 116 |
|
---|
| 117 | // write the file data to disk
|
---|
| 118 | outconvertclass out;
|
---|
| 119 | ofstream filestream(tmp_name_c, ios::out | ios::binary);
|
---|
| 120 | filestream << out << filedata;
|
---|
| 121 | filestream.close();
|
---|
| 122 | delete tmp_name_c;
|
---|
| 123 |
|
---|
| 124 | // populate the fields of a fileupload_t and put it in the
|
---|
| 125 | // fileuploads map
|
---|
| 126 | fileupload_t fu;
|
---|
| 127 | // note that filename currently may or may not include the path since
|
---|
| 128 | // some browsers (e.g. IE) include the path while others
|
---|
| 129 | // (e.g. mozilla) do not. we should probably remove the path from
|
---|
| 130 | // this field here to get a consistent value across all browsers.
|
---|
| 131 | text_t::iterator slash = findlastchar(filename.begin(), filename.end(), '\\');
|
---|
| 132 | if (slash != filename.end()) {
|
---|
| 133 | filename = substr(slash+1, filename.end());
|
---|
| 134 | }
|
---|
| 135 | fu.name = filename;
|
---|
| 136 | fu.type = filetype;
|
---|
| 137 | // size has yet to be implemented
|
---|
[12579] | 138 | fu.size = filedata.size();
|
---|
[12794] | 139 |
|
---|
[12513] | 140 | fu.tmp_name = tmp_name;
|
---|
| 141 | fileuploads[argname] = fu;
|
---|
| 142 | }
|
---|
| 143 | }
|
---|
| 144 | isfile = false;
|
---|
| 145 | argname.clear();
|
---|
| 146 | argdata.clear();
|
---|
| 147 | filename.clear();
|
---|
| 148 | filedata.clear();
|
---|
| 149 | filetype.clear();
|
---|
| 150 | }
|
---|
| 151 |
|
---|
| 152 | // parse data obtained through a CGI POST request
|
---|
| 153 | text_t parse_post_data (text_t &content_type, text_t &raw_post_data,
|
---|
| 154 | fileupload_tmap &fileuploads, const text_t &gsdlhome) {
|
---|
| 155 |
|
---|
| 156 | text_t argstr;
|
---|
| 157 |
|
---|
| 158 | text_t::iterator content_type_begin = content_type.begin();
|
---|
| 159 | text_t::iterator content_type_end = content_type.end();
|
---|
| 160 | if (findword(content_type_begin, content_type_end, "multipart/form-data") == content_type_end) {
|
---|
| 161 | // a simple post request
|
---|
[12794] | 162 |
|
---|
[12513] | 163 | return raw_post_data;
|
---|
| 164 |
|
---|
| 165 | } else {
|
---|
| 166 | // multipart/form data - may contain one or more uploaded files
|
---|
| 167 |
|
---|
| 168 | /*
|
---|
| 169 | content_type should look something like the following
|
---|
| 170 | multipart/form-data; boundary=---------------------------7d411e1a50330
|
---|
| 171 |
|
---|
| 172 | while raw_post_data will be as follows
|
---|
| 173 | -----------------------------7d43e73450330CRLF
|
---|
| 174 | Content-Disposition: form-data; name="e"<CRLF>
|
---|
| 175 | <CRLF>
|
---|
| 176 | d-0testss--1-0-00---4----0--0-110--1en-Zz-1---10-about-0--00031-001utfZz-8-0<CRLF>
|
---|
| 177 | -----------------------------7d43e73450330<CRLF>
|
---|
| 178 | Content-Disposition: form-data; name="afile"; filename="C:\somedoc.doc"<CRLF>
|
---|
| 179 | Content-Type: application/msword<CRLF>
|
---|
| 180 | <CRLF>
|
---|
| 181 | <Content of file><CRLF>
|
---|
| 182 |
|
---|
| 183 | */
|
---|
| 184 |
|
---|
| 185 | // first get the boundary from content-type
|
---|
| 186 | text_t::iterator boundary_begin = findword(content_type_begin, content_type_end, "boundary=");
|
---|
[12794] | 187 | if (boundary_begin+9 < content_type_end)
|
---|
| 188 | {
|
---|
| 189 | // skip over "boundary=" part of string
|
---|
| 190 | boundary_begin += 9;
|
---|
| 191 | }
|
---|
[12513] | 192 | else {
|
---|
| 193 | // error
|
---|
[12796] | 194 | cerr << "Error: malformed boundary? '" << content_type << "'" << endl;
|
---|
[12513] | 195 | return "";
|
---|
| 196 | }
|
---|
| 197 | text_t boundary = substr(boundary_begin, getline(boundary_begin, content_type_end, false));
|
---|
| 198 | int boundary_len = boundary.size();
|
---|
| 199 |
|
---|
[12794] | 200 |
|
---|
[12513] | 201 | text_t argname, argdata, filename, filedata, filetype;
|
---|
| 202 | bool isfile = false;
|
---|
| 203 | text_t::iterator data_here = raw_post_data.begin();
|
---|
| 204 | text_t::iterator data_end = raw_post_data.end();
|
---|
| 205 | while (data_here != data_end) {
|
---|
| 206 |
|
---|
| 207 | // get the next available line (including the trailing <CRLF>
|
---|
| 208 | text_t line = substr(data_here, getline(data_here, data_end, true));
|
---|
[12794] | 209 |
|
---|
[12513] | 210 | data_here += line.size();
|
---|
| 211 | text_t::iterator line_begin = line.begin();
|
---|
| 212 | text_t::iterator line_end = line.end();
|
---|
| 213 | if (findword(line_begin, line_end, boundary) != line_end) {
|
---|
| 214 | // we've found a boundary
|
---|
| 215 | process_post_section(argname, argdata, filename, filedata, filetype,
|
---|
| 216 | isfile, argstr, fileuploads, gsdlhome);
|
---|
| 217 |
|
---|
| 218 | } else if (findword(line_begin, line_end, "Content-Disposition: form-data") != line_end) {
|
---|
| 219 | // we've found the the beginning of a new section
|
---|
| 220 | argname.clear();
|
---|
| 221 | argdata.clear();
|
---|
| 222 |
|
---|
| 223 | // get the name of this piece of form data
|
---|
| 224 | text_t::iterator it = findword(line_begin, line_end, "name=\"");
|
---|
| 225 | if (it == line_end) break; // error - this shouldn't happen
|
---|
| 226 | it = findchar(it, line_end, '"');
|
---|
| 227 | if ((it != line_end) && (it+1 != line_end)) {
|
---|
| 228 | argname = substr(it+1, findchar(it+1, line_end, '"'));
|
---|
| 229 | }
|
---|
| 230 |
|
---|
| 231 | // if this piece of form data contains filename="" it's a file
|
---|
| 232 | // upload and needs to be treated special
|
---|
| 233 | it = (findword(line_begin, line_end, "filename=\""));
|
---|
| 234 | if (it != line_end) {
|
---|
| 235 | // we've found a file upload
|
---|
| 236 | isfile = true;
|
---|
| 237 | it = findchar(it, line_end, '"');
|
---|
| 238 | if ((it != line_end) && (it+1 != line_end)) {
|
---|
| 239 | filename = substr(it+1, findchar(it+1, line_end, '"'));
|
---|
| 240 | }
|
---|
| 241 |
|
---|
| 242 | // the next line is the content-type of this section
|
---|
| 243 | line = substr(data_here, getline(data_here, data_end, true));
|
---|
| 244 | data_here += line.size();
|
---|
| 245 | line_begin = line.begin();
|
---|
| 246 | line_end = line.end();
|
---|
| 247 | it = (findword(line_begin, line_end, "Content-Type: "));
|
---|
| 248 | if (it != line_end) {
|
---|
| 249 | filetype = substr(it+14, getline(it, line_end, false));
|
---|
| 250 | }
|
---|
| 251 | }
|
---|
| 252 |
|
---|
| 253 | // eat up the next line as it's just a <CRLF> on it's own
|
---|
| 254 | data_here += 2;
|
---|
| 255 |
|
---|
| 256 | } else {
|
---|
| 257 | if (isfile) filedata += line;
|
---|
| 258 | else argdata += line;
|
---|
| 259 | }
|
---|
[12794] | 260 |
|
---|
[12513] | 261 | }
|
---|
| 262 |
|
---|
| 263 | // process last section
|
---|
| 264 | process_post_section(argname, argdata, filename, filedata, filetype,
|
---|
| 265 | isfile, argstr, fileuploads, gsdlhome);
|
---|
| 266 |
|
---|
| 267 | return argstr;
|
---|
| 268 | }
|
---|
| 269 | }
|
---|
| 270 |
|
---|
[108] | 271 | // convert %xx and + to their appropriate equivalents
|
---|
[3217] | 272 | // IE 6.0 and later use "%u" followed by 4 hex digits... MS IIS extension!
|
---|
[13461] | 273 | // NOTE: this method is crap. It assumes the input encoding is utf-8. If it
|
---|
| 274 | // actually was, then this returns utf-8, and needs to_uni on the
|
---|
| 275 | // result to get it back to unicode. If the encoding wasn't utf-8, then the
|
---|
| 276 | // output may be crap. Seems to work for 8 bit encodings.
|
---|
| 277 | // Really, this should be given the encoding, and should always return unicode.
|
---|
[108] | 278 | void decode_cgi_arg (text_t &argstr) {
|
---|
| 279 | text_t::iterator in = argstr.begin();
|
---|
| 280 | text_t::iterator out = in;
|
---|
| 281 | text_t::iterator end = argstr.end();
|
---|
| 282 |
|
---|
| 283 | while (in != end) {
|
---|
| 284 | if (*in == '+') *out = ' ';
|
---|
| 285 |
|
---|
| 286 | else if (*in == '%') {
|
---|
| 287 | unsigned short c = '%';
|
---|
[3151] | 288 | ++in;
|
---|
| 289 | if (in != end) { // this is an encoding...
|
---|
| 290 | if (*in == 'u') { // convert %uHHHH to unicode then current encoding
|
---|
| 291 | // this assumes a short int is at least 16 bits...
|
---|
| 292 | ++in;
|
---|
| 293 | if (in != end)
|
---|
| 294 | c=hexdigit(*in++) << 12;
|
---|
| 295 | if (in != end)
|
---|
| 296 | c+=hexdigit(*in++) << 8;
|
---|
| 297 | if (in != end)
|
---|
| 298 | c+=hexdigit(*in++) << 4;
|
---|
| 299 | if (in != end)
|
---|
| 300 | c+=hexdigit(*in);
|
---|
| 301 | /* BAD!! The following assumes the interface is using utf-8. But
|
---|
| 302 | at this point we don't know what encoding we are using, unless
|
---|
| 303 | we can parse it out of the string we are currently decoding... */
|
---|
| 304 | text_t uni=" ";
|
---|
| 305 | uni[0]=c;
|
---|
| 306 | text_t utf8=to_utf8(uni);
|
---|
| 307 | int last_byte=utf8.size()-1;
|
---|
[9620] | 308 | for (int i=0;i<last_byte;++i)
|
---|
[3151] | 309 | *out++ = utf8[i];
|
---|
| 310 | c=utf8[last_byte];
|
---|
| 311 | } else { // convert %HH to hex value
|
---|
| 312 | c = hexdigit (*in);
|
---|
| 313 | ++in;
|
---|
| 314 | if (in != end && c < 16) { // sanity check on the previous character
|
---|
| 315 | c = c*16 + hexdigit (*in);
|
---|
| 316 | }
|
---|
| 317 | }
|
---|
[108] | 318 | }
|
---|
| 319 | *out = c;
|
---|
| 320 | } else *out = *in;
|
---|
| 321 |
|
---|
[9620] | 322 | if (in != end) ++in;
|
---|
| 323 | ++out;
|
---|
[108] | 324 | }
|
---|
| 325 |
|
---|
| 326 | // remove the excess characters
|
---|
| 327 | argstr.erase (out, end);
|
---|
[13461] | 328 |
|
---|
[108] | 329 | }
|
---|
| 330 |
|
---|
| 331 |
|
---|
| 332 | // split up the cgi arguments
|
---|
[776] | 333 | void split_cgi_args (const cgiargsinfoclass &argsinfo, text_t argstr,
|
---|
| 334 | cgiargsclass &args) {
|
---|
[108] | 335 | args.clear();
|
---|
| 336 |
|
---|
[7432] | 337 | text_t::const_iterator here = argstr.begin();
|
---|
| 338 | text_t::const_iterator end = argstr.end();
|
---|
[108] | 339 | text_t key, value;
|
---|
[11998] | 340 |
|
---|
[108] | 341 | // extract out the key=value pairs
|
---|
| 342 | while (here != end) {
|
---|
| 343 | // get the next key and value pair
|
---|
| 344 | here = getdelimitstr (here, end, '=', key);
|
---|
| 345 | here = getdelimitstr (here, end, '&', value);
|
---|
| 346 |
|
---|
| 347 | // convert %xx and + to their appropriate equivalents
|
---|
[614] | 348 | decode_cgi_arg (value);
|
---|
[108] | 349 | value.setencoding(1); // other encoding
|
---|
| 350 | // store this key=value pair
|
---|
[764] | 351 | if (!key.empty()) {
|
---|
[2426] | 352 |
|
---|
| 353 | // if arg occurs multiple times (as is the case with multiple
|
---|
| 354 | // checkboxes using the same name) we'll create a comma separated
|
---|
| 355 | // list of all the values (this uses a hack that encodes naturally
|
---|
| 356 | // occurring commas as %2C - values will therefore need to be decoded
|
---|
| 357 | // again before use) - it should use an array instead
|
---|
[2417] | 358 | const cgiarginfo *info = argsinfo.getarginfo (key);
|
---|
[12562] | 359 | if (info==NULL) {
|
---|
| 360 | // If info is NULL, we can't tell if the arg is multiple value or not
|
---|
| 361 | // Because we need to have dynamically named arguments multivalued, we
|
---|
| 362 | // will always assume multiplevalue = true
|
---|
| 363 | // If the arg is not multi valued, then you need to decode the commas.
|
---|
| 364 | if (args.getarg(key)==NULL) {
|
---|
| 365 | args.setarg (key, encode_commas(value), cgiarg_t::cgi_arg);
|
---|
| 366 | }
|
---|
| 367 | else {
|
---|
| 368 | text_t newvalue = args[key];
|
---|
[2417] | 369 |
|
---|
[12562] | 370 | newvalue += "," + encode_commas(value);
|
---|
[22404] | 371 | newvalue.setencoding(1); // other encoding
|
---|
[12562] | 372 | args.setarg (key, newvalue, cgiarg_t::cgi_arg);
|
---|
| 373 | }
|
---|
[2417] | 374 | }
|
---|
[12562] | 375 | else {
|
---|
| 376 | if (info->multiplevalue) {
|
---|
| 377 |
|
---|
| 378 | text_t newvalue = args[key];
|
---|
| 379 | if (args.lookupcgiarg(key).source == cgiarg_t::cgi_arg) newvalue += ",";
|
---|
| 380 | newvalue += encode_commas(value);
|
---|
[22404] | 381 | newvalue.setencoding(1); // other encoding
|
---|
[12562] | 382 | args.setarg (key, newvalue, cgiarg_t::cgi_arg);
|
---|
| 383 |
|
---|
| 384 | } else {
|
---|
| 385 | args.setarg (key, value, cgiarg_t::cgi_arg);
|
---|
| 386 | }
|
---|
| 387 | }
|
---|
[764] | 388 | }
|
---|
[108] | 389 | }
|
---|
| 390 | }
|
---|
| 391 |
|
---|
[2426] | 392 | text_t encode_commas (const text_t &intext) {
|
---|
| 393 |
|
---|
| 394 | text_t outtext;
|
---|
| 395 |
|
---|
| 396 | text_t::const_iterator here = intext.begin ();
|
---|
| 397 | text_t::const_iterator end = intext.end ();
|
---|
| 398 |
|
---|
| 399 | while (here != end) {
|
---|
| 400 | if (*here == ',') outtext += "%2C";
|
---|
| 401 | else outtext.push_back (*here);
|
---|
[9620] | 402 | ++here;
|
---|
[2426] | 403 | }
|
---|
| 404 | return outtext;
|
---|
| 405 | }
|
---|
| 406 |
|
---|
| 407 | text_t decode_commas (const text_t &intext) {
|
---|
| 408 |
|
---|
| 409 | text_t outtext;
|
---|
| 410 |
|
---|
| 411 | text_t::const_iterator here = intext.begin ();
|
---|
| 412 | text_t::const_iterator end = intext.end ();
|
---|
| 413 |
|
---|
[22942] | 414 | // for loop
|
---|
| 415 | int intext_len = intext.size();
|
---|
| 416 | for(int i = 0; i < intext_len; i++) {
|
---|
| 417 | if ((i+2)<intext_len) {
|
---|
| 418 | if(intext[i] == '%' && intext[i+1] == '2'
|
---|
| 419 | && (intext[i+2] == 'C' || intext[i+2] == 'c')) {
|
---|
| 420 | i += 2;
|
---|
| 421 | outtext.push_back(',');
|
---|
| 422 | continue;
|
---|
| 423 | }
|
---|
| 424 | }
|
---|
| 425 | outtext.push_back (intext[i]);
|
---|
| 426 | }
|
---|
| 427 | return outtext;
|
---|
[2426] | 428 | }
|
---|
| 429 |
|
---|
[13456] | 430 | // set utf8 to true if input is in utf-8, otherwise expects input in unicode
|
---|
| 431 | text_t minus_safe (const text_t &intext, bool utf8) {
|
---|
[607] | 432 |
|
---|
| 433 | text_t outtext;
|
---|
| 434 |
|
---|
| 435 | text_t::const_iterator here = intext.begin ();
|
---|
| 436 | text_t::const_iterator end = intext.end ();
|
---|
| 437 |
|
---|
| 438 | while (here != end) {
|
---|
[1504] | 439 | if (*here == '-') outtext += "Zz-";
|
---|
[607] | 440 | else outtext.push_back (*here);
|
---|
[9620] | 441 | ++here;
|
---|
[607] | 442 | }
|
---|
[13456] | 443 | if (utf8) {
|
---|
| 444 | outtext = cgi_safe_utf8 (outtext);
|
---|
| 445 | } else {
|
---|
| 446 | outtext = cgi_safe_unicode (outtext);
|
---|
| 447 | }
|
---|
[607] | 448 | return outtext;
|
---|
| 449 | }
|
---|
| 450 |
|
---|
[13456] | 451 | // takes utf-8 input
|
---|
| 452 | text_t cgi_safe_utf8 (const text_t &intext) {
|
---|
[108] | 453 | text_t outtext;
|
---|
| 454 |
|
---|
| 455 | text_t::const_iterator here = intext.begin ();
|
---|
| 456 | text_t::const_iterator end = intext.end ();
|
---|
| 457 | unsigned short c;
|
---|
| 458 | text_t ttmp;
|
---|
| 459 |
|
---|
| 460 | while (here != end) {
|
---|
| 461 | c = *here;
|
---|
| 462 | if (((c >= 'a') && (c <= 'z')) ||
|
---|
| 463 | ((c >= 'A') && (c <= 'Z')) ||
|
---|
[474] | 464 | ((c >= '0') && (c <= '9')) ||
|
---|
[13456] | 465 | (c == '%') || (c == '-')) {
|
---|
[108] | 466 | // alphanumeric character
|
---|
| 467 | outtext.push_back(c);
|
---|
| 468 | } else if (c == ' ') {
|
---|
| 469 | // space
|
---|
[150] | 470 | outtext.push_back('+');
|
---|
[13456] | 471 | } else if (c > 255) { // not utf-8 character
|
---|
| 472 | cerr << "WARNING: expected utf-8 char, but got unicode!!\n";
|
---|
| 473 | } else {
|
---|
| 474 | // everything else
|
---|
| 475 | outtext.push_back('%');
|
---|
| 476 | c2hex(c, ttmp);
|
---|
| 477 | outtext += ttmp;
|
---|
| 478 | }
|
---|
| 479 |
|
---|
| 480 | ++here;
|
---|
| 481 | }
|
---|
| 482 |
|
---|
| 483 | return outtext;
|
---|
| 484 | }
|
---|
| 485 | // takes unicode input
|
---|
| 486 | text_t cgi_safe_unicode (const text_t &intext) {
|
---|
| 487 | text_t outtext;
|
---|
| 488 |
|
---|
| 489 | text_t::const_iterator here = intext.begin ();
|
---|
| 490 | text_t::const_iterator end = intext.end ();
|
---|
| 491 | unsigned short c;
|
---|
| 492 | text_t ttmp;
|
---|
| 493 |
|
---|
| 494 | while (here != end) {
|
---|
| 495 | c = *here;
|
---|
| 496 | if (((c >= 'a') && (c <= 'z')) ||
|
---|
| 497 | ((c >= 'A') && (c <= 'Z')) ||
|
---|
| 498 | ((c >= '0') && (c <= '9')) ||
|
---|
| 499 | (c == '%') || (c == '-')) {
|
---|
| 500 | // alphanumeric character
|
---|
| 501 | outtext.push_back(c);
|
---|
| 502 | } else if (c == ' ') {
|
---|
| 503 | // space
|
---|
| 504 | outtext.push_back('+');
|
---|
[13103] | 505 | } else if (c > 127) { // unicode character
|
---|
[3217] | 506 | unsigned char buf[3]; // up to 3 bytes
|
---|
| 507 | buf[0]='\0';buf[1]='\0';buf[2]='\0';
|
---|
| 508 | output_utf8_char(c,buf, buf+2);
|
---|
| 509 | outtext.push_back('%');
|
---|
| 510 | c2hex(buf[0], ttmp);
|
---|
| 511 | outtext += ttmp;
|
---|
| 512 | outtext.push_back('%');
|
---|
| 513 | c2hex(buf[1], ttmp);
|
---|
| 514 | outtext += ttmp;
|
---|
| 515 | if (buf[2]) {
|
---|
| 516 | outtext.push_back('%');
|
---|
| 517 | c2hex(buf[2], ttmp);
|
---|
| 518 | outtext += ttmp;
|
---|
| 519 | }
|
---|
[108] | 520 | } else {
|
---|
| 521 | // everything else
|
---|
| 522 | outtext.push_back('%');
|
---|
| 523 | c2hex(c, ttmp);
|
---|
| 524 | outtext += ttmp;
|
---|
| 525 | }
|
---|
| 526 |
|
---|
[9620] | 527 | ++here;
|
---|
[108] | 528 | }
|
---|
| 529 |
|
---|
| 530 | return outtext;
|
---|
| 531 | }
|
---|
[155] | 532 |
|
---|
| 533 |
|
---|
| 534 |
|
---|
| 535 |
|
---|
| 536 | static text_t::const_iterator get_next_save_arg (text_t::const_iterator first,
|
---|
| 537 | text_t::const_iterator last,
|
---|
| 538 | text_t &argname) {
|
---|
| 539 | first = getdelimitstr (first, last, '-', argname);
|
---|
| 540 | return first;
|
---|
| 541 | }
|
---|
| 542 |
|
---|
| 543 |
|
---|
| 544 | // check_save_conf_str checks the configuration string for
|
---|
| 545 | // the saved args and makes sure it does not conflict with
|
---|
| 546 | // the information about the arguments. If an error is encountered
|
---|
| 547 | // it will return false and the program should not produce any
|
---|
| 548 | // output.
|
---|
| 549 | bool check_save_conf_str (const text_t &saveconf,
|
---|
| 550 | const cgiargsinfoclass &argsinfo,
|
---|
| 551 | ostream &logout) {
|
---|
| 552 | outconvertclass text_t2ascii;
|
---|
| 553 |
|
---|
| 554 | text_tset argsset;
|
---|
| 555 | text_t::const_iterator saveconfhere = saveconf.begin ();
|
---|
| 556 | text_t::const_iterator saveconfend = saveconf.end ();
|
---|
| 557 | text_t argname;
|
---|
| 558 | const cgiarginfo *info;
|
---|
| 559 |
|
---|
| 560 | // first check to make sure all saved arguments can be saved
|
---|
| 561 |
|
---|
| 562 | while (saveconfhere != saveconfend) {
|
---|
| 563 | saveconfhere = get_next_save_arg (saveconfhere, saveconfend, argname);
|
---|
| 564 |
|
---|
| 565 | if (!argname.empty()) {
|
---|
| 566 | // save the argument name for later
|
---|
| 567 | argsset.insert (argname);
|
---|
| 568 |
|
---|
| 569 | // check the argument
|
---|
| 570 | info = argsinfo.getarginfo (argname);
|
---|
| 571 | if (info == NULL) {
|
---|
| 572 | logout << text_t2ascii << "Error: the cgi argument \"" << argname
|
---|
| 573 | << "\" is used in the configuration string for the\n"
|
---|
| 574 | << "saved arguments but does not exist as a valid argument.\n\n";
|
---|
| 575 | return false;
|
---|
| 576 | }
|
---|
| 577 | if (info->savedarginfo == cgiarginfo::mustnot) {
|
---|
| 578 | logout << text_t2ascii << "Error: the cgi argument \"" << argname
|
---|
| 579 | << "\" is used in the configuration string for the\n"
|
---|
| 580 | << "saved arguments but has been specified as an argument whose\n"
|
---|
| 581 | << "state must not be saved.\n\n";
|
---|
| 582 | return false;
|
---|
| 583 | }
|
---|
| 584 | }
|
---|
| 585 | }
|
---|
| 586 |
|
---|
| 587 |
|
---|
| 588 | // next check that all saved arguments that should be saved
|
---|
| 589 | // are saved
|
---|
| 590 | cgiargsinfoclass::const_iterator argsinfohere = argsinfo.begin ();
|
---|
| 591 | cgiargsinfoclass::const_iterator argsinfoend = argsinfo.end ();
|
---|
| 592 |
|
---|
| 593 | while (argsinfohere != argsinfoend) {
|
---|
| 594 | if (((*argsinfohere).second.savedarginfo == cgiarginfo::must) &&
|
---|
| 595 | (argsset.find((*argsinfohere).second.shortname) == argsset.end())) {
|
---|
| 596 | logout << text_t2ascii << "Error: the cgi argument \""
|
---|
| 597 | << (*argsinfohere).second.shortname << "\" was specified as needing to\n"
|
---|
| 598 | << "be save but was not listed in the saved arguments.\n\n";
|
---|
| 599 | return false;
|
---|
| 600 | }
|
---|
| 601 |
|
---|
[9620] | 602 | ++argsinfohere;
|
---|
[155] | 603 | }
|
---|
| 604 |
|
---|
| 605 | return true; // made it, no clashes
|
---|
| 606 | }
|
---|
| 607 |
|
---|
| 608 |
|
---|
| 609 | // create_save_conf_str will create a configuration string
|
---|
| 610 | // based on the information in argsinfo. This method of configuration
|
---|
| 611 | // is not recomended as small changes can produce large changes in
|
---|
| 612 | // the resulting configuration string (for instance a totally different
|
---|
| 613 | // ordering). Only arguments which "must" be saved are included in
|
---|
| 614 | // the resulting string.
|
---|
| 615 | text_t create_save_conf_str (const cgiargsinfoclass &argsinfo,
|
---|
| 616 | ostream &/*logout*/) {
|
---|
| 617 | cgiargsinfoclass::const_iterator argsinfohere = argsinfo.begin ();
|
---|
| 618 | cgiargsinfoclass::const_iterator argsinfoend = argsinfo.end ();
|
---|
| 619 | text_t saveconf;
|
---|
| 620 | bool first = true;
|
---|
| 621 |
|
---|
| 622 | while (argsinfohere != argsinfoend) {
|
---|
| 623 | // save this argument if it must be saved
|
---|
| 624 | if ((*argsinfohere).second.savedarginfo == cgiarginfo::must) {
|
---|
| 625 | if (!first) saveconf.push_back ('-');
|
---|
| 626 | else first = false;
|
---|
| 627 | saveconf += (*argsinfohere).second.shortname;
|
---|
| 628 | }
|
---|
| 629 |
|
---|
[9620] | 630 | ++argsinfohere;
|
---|
[155] | 631 | }
|
---|
| 632 |
|
---|
| 633 | return saveconf;
|
---|
| 634 | }
|
---|
| 635 |
|
---|
| 636 |
|
---|
| 637 | // expand_save_args will expand the saved arguments based
|
---|
| 638 | // on saveconf placing the results in args if they are not
|
---|
| 639 | // already defined. If it encounters an error it will return false
|
---|
| 640 | // and output more information to logout.
|
---|
| 641 | bool expand_save_args (const cgiargsinfoclass &argsinfo,
|
---|
| 642 | const text_t &saveconf,
|
---|
| 643 | cgiargsclass &args,
|
---|
| 644 | ostream &logout) {
|
---|
| 645 | outconvertclass text_t2ascii;
|
---|
| 646 |
|
---|
| 647 | text_t *arg_e = args.getarg("e");
|
---|
| 648 | if (arg_e == NULL) return true; // no compressed arguments
|
---|
| 649 | if (arg_e->empty()) return true; // no compressed arguments
|
---|
| 650 |
|
---|
| 651 | text_t argname, argvalue;
|
---|
| 652 | const cgiarginfo *argnameinfo;
|
---|
| 653 |
|
---|
| 654 | text_t::const_iterator saveconfhere = saveconf.begin();
|
---|
| 655 | text_t::const_iterator saveconfend = saveconf.end();
|
---|
| 656 |
|
---|
[11259] | 657 | text_t::iterator arg_ebegin = arg_e->begin();
|
---|
[155] | 658 | text_t::iterator arg_eend = arg_e->end();
|
---|
[11259] | 659 | text_t::iterator arg_ehere = arg_ebegin;
|
---|
[155] | 660 | while (saveconfhere != saveconfend && arg_ehere != arg_eend) {
|
---|
| 661 | saveconfhere = get_next_save_arg (saveconfhere, saveconfend, argname);
|
---|
| 662 |
|
---|
| 663 | if (!argname.empty()) {
|
---|
| 664 | // found another entry
|
---|
| 665 | argnameinfo = argsinfo.getarginfo (argname);
|
---|
| 666 |
|
---|
| 667 | if (argnameinfo == NULL) {
|
---|
| 668 | // no information about the argument could be found
|
---|
| 669 | // we can't keep going because we don't know whether
|
---|
| 670 | // this argument is a single or multiple character value
|
---|
| 671 | logout << text_t2ascii << "Error: the cgi argument \"" << argname
|
---|
| 672 | << "\" was specified as being a compressed argument\n"
|
---|
| 673 | << "but no information about it could be found within the "
|
---|
| 674 | << "cgiargsinfoclass.\n";
|
---|
| 675 | return false;
|
---|
| 676 |
|
---|
| 677 | } else {
|
---|
[294] | 678 |
|
---|
[155] | 679 | // found the argument information
|
---|
| 680 | if (argnameinfo->multiplechar) {
|
---|
[607] | 681 | text_t::const_iterator sav = arg_ehere;
|
---|
[155] | 682 | arg_ehere = getdelimitstr (arg_ehere, arg_eend, '-', argvalue);
|
---|
[11259] | 683 | if (distance(arg_ebegin, arg_ehere) > 2) {
|
---|
| 684 | // replace any '-' chars escaped with 'Zz'
|
---|
| 685 | bool first = true;
|
---|
| 686 | while ((*(arg_ehere-3) == 'Z') && (*(arg_ehere-2) == 'z')) {
|
---|
| 687 | if (first) argvalue.clear();
|
---|
[21997] | 688 |
|
---|
| 689 | // Hey, here's a wild idea. Why don't we check that there is
|
---|
| 690 | // another hyphen in the cgiarge before we get a pointer to it and
|
---|
| 691 | // add one. That way we are far less likely to wander off into
|
---|
| 692 | // random memory merrily parsing arguments that are then lovingly
|
---|
| 693 | // spewed all over the HTML page returned at the usage logs.
|
---|
| 694 | text_t::iterator minus_itr = findchar (arg_ehere, arg_eend, '-');
|
---|
| 695 | if (minus_itr == arg_eend)
|
---|
| 696 | {
|
---|
| 697 | logout << text_t2ascii << "Error: the cgi argument \"" << argname << "\" was specified as being a compressed argument but we have run out of cgiarge to decompress!\n";
|
---|
| 698 | return false;
|
---|
| 699 | }
|
---|
| 700 | arg_ehere = minus_itr + 1;
|
---|
| 701 |
|
---|
[11259] | 702 | while (sav != (arg_ehere-1)) {
|
---|
| 703 | if (!((*sav == 'Z') && (*(sav+1) == 'z') && (*(sav+2) == '-')) &&
|
---|
| 704 | !((*(sav-1) == 'Z') && (*sav == 'z') && (*(sav+1) == '-'))) argvalue.push_back (*sav);
|
---|
| 705 | ++sav;
|
---|
| 706 | }
|
---|
| 707 | first = false;
|
---|
[607] | 708 | }
|
---|
| 709 | }
|
---|
[294] | 710 | argvalue.setencoding(1); // other encoding
|
---|
[366] | 711 | if (!argvalue.empty()) args.setdefaultarg (argname, argvalue, cgiarg_t::compressed_arg);
|
---|
[155] | 712 | } else {
|
---|
[366] | 713 | args.setdefaultcarg (argname,*arg_ehere, cgiarg_t::compressed_arg);
|
---|
[9620] | 714 | ++arg_ehere;
|
---|
[155] | 715 | }
|
---|
| 716 | }
|
---|
| 717 | }
|
---|
| 718 | }
|
---|
| 719 |
|
---|
| 720 | return true;
|
---|
| 721 | }
|
---|
| 722 |
|
---|
| 723 |
|
---|
| 724 | // adds the default values for those arguments which have not
|
---|
| 725 | // been specified
|
---|
| 726 | void add_default_args (const cgiargsinfoclass &argsinfo,
|
---|
| 727 | cgiargsclass &args,
|
---|
| 728 | ostream &/*logout*/) {
|
---|
| 729 | cgiargsinfoclass::const_iterator argsinfohere = argsinfo.begin ();
|
---|
| 730 | cgiargsinfoclass::const_iterator argsinfoend = argsinfo.end ();
|
---|
| 731 |
|
---|
| 732 | while (argsinfohere != argsinfoend) {
|
---|
| 733 | if ((*argsinfohere).second.defaultstatus != cgiarginfo::none) {
|
---|
| 734 | args.setdefaultarg ((*argsinfohere).second.shortname,
|
---|
[366] | 735 | (*argsinfohere).second.argdefault, cgiarg_t::default_arg);
|
---|
[155] | 736 | }
|
---|
[9620] | 737 | ++argsinfohere;
|
---|
[155] | 738 | }
|
---|
| 739 | }
|
---|
| 740 |
|
---|
[12513] | 741 | void add_fileupload_args (const cgiargsinfoclass &argsinfo,
|
---|
| 742 | cgiargsclass &args,
|
---|
| 743 | fileupload_tmap &fileuploads,
|
---|
| 744 | ostream &logout) {
|
---|
| 745 |
|
---|
| 746 | const cgiarginfo *info = argsinfo.getarginfo("a");
|
---|
| 747 | fileupload_tmap::const_iterator this_file = fileuploads.begin();
|
---|
| 748 | fileupload_tmap::const_iterator end_file = fileuploads.end();
|
---|
| 749 | while (this_file != end_file) {
|
---|
| 750 | const cgiarginfo *info = argsinfo.getarginfo((*this_file).first);
|
---|
| 751 | if (info != NULL) {
|
---|
[12794] | 752 |
|
---|
[12513] | 753 | if ((*info).fileupload && (file_exists((*this_file).second.tmp_name))) {
|
---|
[12794] | 754 |
|
---|
[12513] | 755 | args.setargfile((*this_file).first, (*this_file).second);
|
---|
| 756 | }
|
---|
| 757 | }
|
---|
| 758 | this_file++;
|
---|
| 759 | }
|
---|
| 760 | }
|
---|
[155] | 761 |
|
---|
| 762 | // compress_save_args will compress the arguments and return
|
---|
| 763 | // them in compressed_args. If an error was encountered
|
---|
| 764 | // compressed_args will be set to to "", an error will be
|
---|
| 765 | // written to logout, and the function will return false.
|
---|
| 766 | bool compress_save_args (const cgiargsinfoclass &argsinfo,
|
---|
| 767 | const text_t &saveconf,
|
---|
| 768 | cgiargsclass &args,
|
---|
| 769 | text_t &compressed_args,
|
---|
[294] | 770 | outconvertclass &outconvert,
|
---|
[155] | 771 | ostream &logout) {
|
---|
| 772 | outconvertclass text_t2ascii;
|
---|
| 773 |
|
---|
| 774 | compressed_args.clear();
|
---|
| 775 |
|
---|
| 776 | text_t argname, argvalue;
|
---|
| 777 | const cgiarginfo *argnameinfo;
|
---|
| 778 |
|
---|
| 779 | text_t::const_iterator saveconfhere = saveconf.begin();
|
---|
| 780 | text_t::const_iterator saveconfend = saveconf.end();
|
---|
| 781 |
|
---|
| 782 | while (saveconfhere != saveconfend) {
|
---|
| 783 | saveconfhere = get_next_save_arg (saveconfhere, saveconfend, argname);
|
---|
| 784 |
|
---|
| 785 | if (!argname.empty()) {
|
---|
| 786 | // found another entry
|
---|
| 787 | argnameinfo = argsinfo.getarginfo (argname);
|
---|
| 788 |
|
---|
| 789 | if (argnameinfo == NULL) {
|
---|
| 790 | // no information about the argument could be found
|
---|
| 791 | // we can't keep going because we don't know whether
|
---|
| 792 | // this argument is a single or multiple character value
|
---|
| 793 | logout << text_t2ascii << "Error: the cgi argument \"" << argname
|
---|
| 794 | << "\" was specified as being a compressed argument\n"
|
---|
| 795 | << "but no information about it could be found within the "
|
---|
| 796 | << "cgiargsinfoclass.\n";
|
---|
| 797 | compressed_args.clear();
|
---|
| 798 | return false;
|
---|
| 799 |
|
---|
| 800 | } else {
|
---|
| 801 | // found the argument information
|
---|
| 802 | if (argnameinfo->multiplechar) {
|
---|
[607] | 803 | // multiple character argument -- sort out any '-' chars
|
---|
[9674] | 804 | if (args["w"]=="utf-16be") // browsers don't like \0 in urls...
|
---|
[13456] | 805 | compressed_args += minus_safe (args[argname], false);
|
---|
[3670] | 806 | else
|
---|
[13456] | 807 | compressed_args += minus_safe (outconvert.convert(args[argname]), true);
|
---|
| 808 |
|
---|
[155] | 809 | if (saveconfhere != saveconfend) compressed_args.push_back ('-');
|
---|
| 810 |
|
---|
| 811 | } else {
|
---|
| 812 | // single character argument
|
---|
| 813 | if (args[argname].size() == 0) {
|
---|
| 814 | logout << text_t2ascii << "Error: the cgi argument \"" << argname
|
---|
| 815 | << "\" was specified as being a compressed argument which\n"
|
---|
| 816 | << "should have a one character value but it was empty.\n\n";
|
---|
| 817 | compressed_args.clear ();
|
---|
| 818 | return false;
|
---|
| 819 |
|
---|
| 820 | } else if (args[argname].size() > 1) {
|
---|
| 821 | logout << text_t2ascii << "Error: the cgi argument \"" << argname
|
---|
| 822 | << "\" was specified as being a compressed argument which\n"
|
---|
| 823 | << "should have a one character value but it had multiple characters.\n\n";
|
---|
| 824 | compressed_args.clear ();
|
---|
| 825 | return false;
|
---|
| 826 | }
|
---|
| 827 |
|
---|
| 828 | // everything is ok
|
---|
| 829 | compressed_args += args[argname];
|
---|
| 830 | }
|
---|
| 831 | }
|
---|
| 832 | }
|
---|
| 833 | }
|
---|
| 834 |
|
---|
| 835 | return true;
|
---|
| 836 | }
|
---|
| 837 |
|
---|
| 838 |
|
---|
| 839 | // args_tounicode converts any arguments which are not in unicode
|
---|
| 840 | // to unicode using inconvert
|
---|
| 841 | void args_tounicode (cgiargsclass &args, inconvertclass &inconvert) {
|
---|
| 842 | cgiargsclass::iterator here = args.begin();
|
---|
| 843 | cgiargsclass::iterator end = args.end();
|
---|
| 844 |
|
---|
| 845 | while (here != end) {
|
---|
[366] | 846 | if ((*here).second.value.getencoding() > 0) {
|
---|
[21961] | 847 | // Call reset() before converting each argument, to prevent problems when converting the last
|
---|
| 848 | // argument left the converter in a bad state
|
---|
| 849 | inconvert.reset();
|
---|
[366] | 850 | (*here).second.value = inconvert.convert((*here).second.value);
|
---|
[155] | 851 | }
|
---|
| 852 |
|
---|
[9620] | 853 | ++here;
|
---|
[155] | 854 | }
|
---|
| 855 | }
|
---|
[873] | 856 |
|
---|
| 857 | // fcgienv will be loaded with environment name-value pairs
|
---|
| 858 | // if using fastcgi (had to do this as getenv doesn't work
|
---|
| 859 | // with our implementation of fastcgi). if fcgienv is empty
|
---|
| 860 | // we'll simply use getenv
|
---|
| 861 | text_t gsdl_getenv (const text_t &name, text_tmap &fcgienv) {
|
---|
| 862 | if (fcgienv.empty()) {
|
---|
| 863 | char *n = name.getcstr();
|
---|
| 864 | char *v = getenv(n);
|
---|
[7432] | 865 | delete []n;
|
---|
[873] | 866 | if (v != NULL) return v;
|
---|
[7432] | 867 | return g_EmptyText;
|
---|
[873] | 868 |
|
---|
| 869 | } else return fcgienv[name];
|
---|
| 870 | }
|
---|