[358] | 1 | /*
|
---|
| 2 | * Copyright 1999-2001,2004 The Apache Software Foundation.
|
---|
| 3 | *
|
---|
| 4 | * Licensed under the Apache License, Version 2.0 (the "License");
|
---|
| 5 | * you may not use this file except in compliance with the License.
|
---|
| 6 | * You may obtain a copy of the License at
|
---|
| 7 | *
|
---|
| 8 | * http://www.apache.org/licenses/LICENSE-2.0
|
---|
| 9 | *
|
---|
| 10 | * Unless required by applicable law or agreed to in writing, software
|
---|
| 11 | * distributed under the License is distributed on an "AS IS" BASIS,
|
---|
| 12 | * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
---|
| 13 | * See the License for the specific language governing permissions and
|
---|
| 14 | * limitations under the License.
|
---|
| 15 | */
|
---|
| 16 |
|
---|
| 17 | /*
|
---|
| 18 | * $Log: SAXCount.cpp,v $
|
---|
| 19 | * Revision 1.30 2004/09/08 13:55:33 peiyongz
|
---|
| 20 | * Apache License Version 2.0
|
---|
| 21 | *
|
---|
| 22 | * Revision 1.29 2004/09/02 14:59:29 cargilld
|
---|
| 23 | * Add OutOfMemoryException block to samples.
|
---|
| 24 | *
|
---|
| 25 | * Revision 1.28 2003/05/30 09:36:36 gareth
|
---|
| 26 | * Use new macros for iostream.h and std:: issues.
|
---|
| 27 | *
|
---|
| 28 | * Revision 1.27 2002/12/10 13:34:42 tng
|
---|
| 29 | * Samples minor update in usage information.
|
---|
| 30 | *
|
---|
| 31 | * Revision 1.26 2002/11/08 16:19:05 peiyongz
|
---|
| 32 | * no message
|
---|
| 33 | *
|
---|
| 34 | * Revision 1.25 2002/11/07 18:31:04 peiyongz
|
---|
| 35 | * command line option for "locale"
|
---|
| 36 | *
|
---|
| 37 | * Revision 1.24 2002/11/04 14:09:16 tng
|
---|
| 38 | * [Bug 14201] use of ios::nocreate breaks build.
|
---|
| 39 | *
|
---|
| 40 | * Revision 1.23 2002/11/01 22:05:44 tng
|
---|
| 41 | * Samples/Test update: Issue error if the list file failed to open.
|
---|
| 42 | *
|
---|
| 43 | * Revision 1.22 2002/09/27 19:25:10 tng
|
---|
| 44 | * Samples Fix: wrong length in memset
|
---|
| 45 | *
|
---|
| 46 | * Revision 1.21 2002/07/17 18:58:36 tng
|
---|
| 47 | * samples update: for testing special encoding purpose.
|
---|
| 48 | *
|
---|
| 49 | * Revision 1.20 2001/11/13 13:22:35 tng
|
---|
| 50 | * SAXCount fix: restore previous feature to accept multiple input files.
|
---|
| 51 | *
|
---|
| 52 | * Revision 1.19 2001/10/29 17:02:57 tng
|
---|
| 53 | * Fix typo in samples.
|
---|
| 54 | *
|
---|
| 55 | * Revision 1.18 2001/10/25 15:18:33 tng
|
---|
| 56 | * delete the parser before XMLPlatformUtils::Terminate.
|
---|
| 57 | *
|
---|
| 58 | * Revision 1.17 2001/10/19 19:02:43 tng
|
---|
| 59 | * [Bug 3909] return non-zero an exit code when error was encounted.
|
---|
| 60 | * And other modification for consistent help display and return code across samples.
|
---|
| 61 | *
|
---|
| 62 | * Revision 1.16 2001/08/15 12:41:04 tng
|
---|
| 63 | * Initialize the fURI array to zeros, in case, some compilers like AIX xlC_r doesn't reset the memory.
|
---|
| 64 | *
|
---|
| 65 | * Revision 1.15 2001/08/08 12:12:32 tng
|
---|
| 66 | * Print the file name only if doList is on.
|
---|
| 67 | *
|
---|
| 68 | * Revision 1.14 2001/08/03 15:08:17 tng
|
---|
| 69 | * close the list file.
|
---|
| 70 | *
|
---|
| 71 | * Revision 1.13 2001/08/02 17:10:29 tng
|
---|
| 72 | * Allow DOMCount/SAXCount/IDOMCount/SAX2Count to take a file that has a list of xml file as input.
|
---|
| 73 | *
|
---|
| 74 | * Revision 1.12 2001/08/01 19:11:01 tng
|
---|
| 75 | * Add full schema constraint checking flag to the samples and the parser.
|
---|
| 76 | *
|
---|
| 77 | * Revision 1.11 2001/05/11 13:24:57 tng
|
---|
| 78 | * Copyright update.
|
---|
| 79 | *
|
---|
| 80 | * Revision 1.10 2001/05/03 16:00:12 tng
|
---|
| 81 | * Schema: samples update with schema
|
---|
| 82 | *
|
---|
| 83 | * Revision 1.9 2000/10/19 23:52:41 andyh
|
---|
| 84 | * SAXCount: Allow multiple files on command line
|
---|
| 85 | *
|
---|
| 86 | * Revision 1.8 2000/06/16 20:25:38 rahulj
|
---|
| 87 | * Add the -v=always option to force validation checking. Need this
|
---|
| 88 | * option for running the conformance tests.
|
---|
| 89 | *
|
---|
| 90 | * Revision 1.7 2000/05/31 18:39:59 rahulj
|
---|
| 91 | * 'Auto' validation is the default processing mode.
|
---|
| 92 | *
|
---|
| 93 | * Revision 1.6 2000/05/09 00:22:29 andyh
|
---|
| 94 | * Memory Cleanup. XMLPlatformUtils::Terminate() deletes all lazily
|
---|
| 95 | * allocated memory; memory leak checking tools will no longer report
|
---|
| 96 | * that leaks exist. (DOM GetElementsByTagID temporarily removed
|
---|
| 97 | * as part of this.)
|
---|
| 98 | *
|
---|
| 99 | * Revision 1.5 2000/03/03 01:29:31 roddey
|
---|
| 100 | * Added a scanReset()/parseReset() method to the scanner and
|
---|
| 101 | * parsers, to allow for reset after early exit from a progressive parse.
|
---|
| 102 | * Added calls to new Terminate() call to all of the samples. Improved
|
---|
| 103 | * documentation in SAX and DOM parsers.
|
---|
| 104 | *
|
---|
| 105 | * Revision 1.4 2000/03/02 19:53:47 roddey
|
---|
| 106 | * This checkin includes many changes done while waiting for the
|
---|
| 107 | * 1.1.0 code to be finished. I can't list them all here, but a list is
|
---|
| 108 | * available elsewhere.
|
---|
| 109 | *
|
---|
| 110 | * Revision 1.3 2000/02/11 02:39:10 abagchi
|
---|
| 111 | * Removed StrX::transcode
|
---|
| 112 | *
|
---|
| 113 | * Revision 1.2 2000/02/06 07:47:23 rahulj
|
---|
| 114 | * Year 2K copyright swat.
|
---|
| 115 | *
|
---|
| 116 | * Revision 1.1.1.1 1999/11/09 01:09:30 twl
|
---|
| 117 | * Initial checkin
|
---|
| 118 | *
|
---|
| 119 | * Revision 1.7 1999/11/08 20:43:40 rahul
|
---|
| 120 | * Swat for adding in Product name and CVS comment log variable.
|
---|
| 121 | *
|
---|
| 122 | */
|
---|
| 123 |
|
---|
| 124 |
|
---|
| 125 | // ---------------------------------------------------------------------------
|
---|
| 126 | // Includes
|
---|
| 127 | // ---------------------------------------------------------------------------
|
---|
| 128 | #include "SAXCount.hpp"
|
---|
| 129 | #if defined(XERCES_NEW_IOSTREAMS)
|
---|
| 130 | #include <fstream>
|
---|
| 131 | #else
|
---|
| 132 | #include <fstream.h>
|
---|
| 133 | #endif
|
---|
| 134 | #include <xercesc/util/OutOfMemoryException.hpp>
|
---|
| 135 |
|
---|
| 136 | // ---------------------------------------------------------------------------
|
---|
| 137 | // Local helper methods
|
---|
| 138 | // ---------------------------------------------------------------------------
|
---|
| 139 | void usage()
|
---|
| 140 | {
|
---|
| 141 | XERCES_STD_QUALIFIER cout << "\nUsage:\n"
|
---|
| 142 | " SAXCount [options] <XML file | List file>\n\n"
|
---|
| 143 | "This program invokes the SAX Parser, and then prints the\n"
|
---|
| 144 | "number of elements, attributes, spaces and characters found\n"
|
---|
| 145 | "in each XML file, using SAX API.\n\n"
|
---|
| 146 | "Options:\n"
|
---|
| 147 | " -l Indicate the input file is a List File that has a list of xml files.\n"
|
---|
| 148 | " Default to off (Input file is an XML file).\n"
|
---|
| 149 | " -v=xxx Validation scheme [always | never | auto*].\n"
|
---|
| 150 | " -n Enable namespace processing. Defaults to off.\n"
|
---|
| 151 | " -s Enable schema processing. Defaults to off.\n"
|
---|
| 152 | " -f Enable full schema constraint checking. Defaults to off.\n"
|
---|
| 153 | " -locale=ll_CC specify the locale, default: en_US.\n"
|
---|
| 154 | " -? Show this help.\n\n"
|
---|
| 155 | " * = Default if not provided explicitly.\n"
|
---|
| 156 | << XERCES_STD_QUALIFIER endl;
|
---|
| 157 | }
|
---|
| 158 |
|
---|
| 159 |
|
---|
| 160 | // ---------------------------------------------------------------------------
|
---|
| 161 | // Program entry point
|
---|
| 162 | // ---------------------------------------------------------------------------
|
---|
| 163 | int main(int argC, char* argV[])
|
---|
| 164 | {
|
---|
| 165 |
|
---|
| 166 | // Check command line and extract arguments.
|
---|
| 167 | if (argC < 2)
|
---|
| 168 | {
|
---|
| 169 | usage();
|
---|
| 170 | return 1;
|
---|
| 171 | }
|
---|
| 172 |
|
---|
| 173 | const char* xmlFile = 0;
|
---|
| 174 | SAXParser::ValSchemes valScheme = SAXParser::Val_Auto;
|
---|
| 175 | bool doNamespaces = false;
|
---|
| 176 | bool doSchema = false;
|
---|
| 177 | bool schemaFullChecking = false;
|
---|
| 178 | bool doList = false;
|
---|
| 179 | bool errorOccurred = false;
|
---|
| 180 | bool recognizeNEL = false;
|
---|
| 181 | char localeStr[64];
|
---|
| 182 | memset(localeStr, 0, sizeof localeStr);
|
---|
| 183 |
|
---|
| 184 | int argInd;
|
---|
| 185 | for (argInd = 1; argInd < argC; argInd++)
|
---|
| 186 | {
|
---|
| 187 | // Break out on first parm not starting with a dash
|
---|
| 188 | if (argV[argInd][0] != '-')
|
---|
| 189 | break;
|
---|
| 190 |
|
---|
| 191 | // Watch for special case help request
|
---|
| 192 | if (!strcmp(argV[argInd], "-?"))
|
---|
| 193 | {
|
---|
| 194 | usage();
|
---|
| 195 | return 2;
|
---|
| 196 | }
|
---|
| 197 | else if (!strncmp(argV[argInd], "-v=", 3)
|
---|
| 198 | || !strncmp(argV[argInd], "-V=", 3))
|
---|
| 199 | {
|
---|
| 200 | const char* const parm = &argV[argInd][3];
|
---|
| 201 |
|
---|
| 202 | if (!strcmp(parm, "never"))
|
---|
| 203 | valScheme = SAXParser::Val_Never;
|
---|
| 204 | else if (!strcmp(parm, "auto"))
|
---|
| 205 | valScheme = SAXParser::Val_Auto;
|
---|
| 206 | else if (!strcmp(parm, "always"))
|
---|
| 207 | valScheme = SAXParser::Val_Always;
|
---|
| 208 | else
|
---|
| 209 | {
|
---|
| 210 | XERCES_STD_QUALIFIER cerr << "Unknown -v= value: " << parm << XERCES_STD_QUALIFIER endl;
|
---|
| 211 | return 2;
|
---|
| 212 | }
|
---|
| 213 | }
|
---|
| 214 | else if (!strcmp(argV[argInd], "-n")
|
---|
| 215 | || !strcmp(argV[argInd], "-N"))
|
---|
| 216 | {
|
---|
| 217 | doNamespaces = true;
|
---|
| 218 | }
|
---|
| 219 | else if (!strcmp(argV[argInd], "-s")
|
---|
| 220 | || !strcmp(argV[argInd], "-S"))
|
---|
| 221 | {
|
---|
| 222 | doSchema = true;
|
---|
| 223 | }
|
---|
| 224 | else if (!strcmp(argV[argInd], "-f")
|
---|
| 225 | || !strcmp(argV[argInd], "-F"))
|
---|
| 226 | {
|
---|
| 227 | schemaFullChecking = true;
|
---|
| 228 | }
|
---|
| 229 | else if (!strcmp(argV[argInd], "-l")
|
---|
| 230 | || !strcmp(argV[argInd], "-L"))
|
---|
| 231 | {
|
---|
| 232 | doList = true;
|
---|
| 233 | }
|
---|
| 234 | else if (!strcmp(argV[argInd], "-special:nel"))
|
---|
| 235 | {
|
---|
| 236 | // turning this on will lead to non-standard compliance behaviour
|
---|
| 237 | // it will recognize the unicode character 0x85 as new line character
|
---|
| 238 | // instead of regular character as specified in XML 1.0
|
---|
| 239 | // do not turn this on unless really necessary
|
---|
| 240 | recognizeNEL = true;
|
---|
| 241 | }
|
---|
| 242 | else if (!strncmp(argV[argInd], "-locale=", 8))
|
---|
| 243 | {
|
---|
| 244 | // Get out the end of line
|
---|
| 245 | strcpy(localeStr, &(argV[argInd][8]));
|
---|
| 246 | }
|
---|
| 247 | else
|
---|
| 248 | {
|
---|
| 249 | XERCES_STD_QUALIFIER cerr << "Unknown option '" << argV[argInd]
|
---|
| 250 | << "', ignoring it\n" << XERCES_STD_QUALIFIER endl;
|
---|
| 251 | }
|
---|
| 252 | }
|
---|
| 253 |
|
---|
| 254 | //
|
---|
| 255 | // There should at least one parameter left, and that
|
---|
| 256 | // should be the file name(s).
|
---|
| 257 | //
|
---|
| 258 | if (argInd == argC)
|
---|
| 259 | {
|
---|
| 260 | usage();
|
---|
| 261 | return 1;
|
---|
| 262 | }
|
---|
| 263 |
|
---|
| 264 | // Initialize the XML4C2 system
|
---|
| 265 | try
|
---|
| 266 | {
|
---|
| 267 | if (strlen(localeStr))
|
---|
| 268 | {
|
---|
| 269 | XMLPlatformUtils::Initialize(localeStr);
|
---|
| 270 | }
|
---|
| 271 | else
|
---|
| 272 | {
|
---|
| 273 | XMLPlatformUtils::Initialize();
|
---|
| 274 | }
|
---|
| 275 |
|
---|
| 276 | if (recognizeNEL)
|
---|
| 277 | {
|
---|
| 278 | XMLPlatformUtils::recognizeNEL(recognizeNEL);
|
---|
| 279 | }
|
---|
| 280 | }
|
---|
| 281 |
|
---|
| 282 | catch (const XMLException& toCatch)
|
---|
| 283 | {
|
---|
| 284 | XERCES_STD_QUALIFIER cerr << "Error during initialization! Message:\n"
|
---|
| 285 | << StrX(toCatch.getMessage()) << XERCES_STD_QUALIFIER endl;
|
---|
| 286 | return 1;
|
---|
| 287 | }
|
---|
| 288 |
|
---|
| 289 | //
|
---|
| 290 | // Create a SAX parser object. Then, according to what we were told on
|
---|
| 291 | // the command line, set it to validate or not.
|
---|
| 292 | //
|
---|
| 293 | SAXParser* parser = new SAXParser;
|
---|
| 294 |
|
---|
| 295 | parser->setValidationScheme(valScheme);
|
---|
| 296 | parser->setDoNamespaces(doNamespaces);
|
---|
| 297 | parser->setDoSchema(doSchema);
|
---|
| 298 | parser->setValidationSchemaFullChecking(schemaFullChecking);
|
---|
| 299 |
|
---|
| 300 | //
|
---|
| 301 | // Create our SAX handler object and install it on the parser, as the
|
---|
| 302 | // document and error handler.
|
---|
| 303 | //
|
---|
| 304 | SAXCountHandlers handler;
|
---|
| 305 | parser->setDocumentHandler(&handler);
|
---|
| 306 | parser->setErrorHandler(&handler);
|
---|
| 307 |
|
---|
| 308 |
|
---|
| 309 | //
|
---|
| 310 | // Get the starting time and kick off the parse of the indicated
|
---|
| 311 | // file. Catch any exceptions that might propogate out of it.
|
---|
| 312 | //
|
---|
| 313 | unsigned long duration;
|
---|
| 314 |
|
---|
| 315 | XERCES_STD_QUALIFIER ifstream fin;
|
---|
| 316 |
|
---|
| 317 | // the input is a list file
|
---|
| 318 | if (doList)
|
---|
| 319 | fin.open(argV[argInd]);
|
---|
| 320 |
|
---|
| 321 | if (fin.fail()) {
|
---|
| 322 | XERCES_STD_QUALIFIER cerr <<"Cannot open the list file: " << argV[argInd] << XERCES_STD_QUALIFIER endl;
|
---|
| 323 | return 2;
|
---|
| 324 | }
|
---|
| 325 |
|
---|
| 326 | while (true)
|
---|
| 327 | {
|
---|
| 328 | char fURI[1000];
|
---|
| 329 | //initialize the array to zeros
|
---|
| 330 | memset(fURI,0,sizeof(fURI));
|
---|
| 331 |
|
---|
| 332 | if (doList) {
|
---|
| 333 | if (! fin.eof() ) {
|
---|
| 334 | fin.getline (fURI, sizeof(fURI));
|
---|
| 335 | if (!*fURI)
|
---|
| 336 | continue;
|
---|
| 337 | else {
|
---|
| 338 | xmlFile = fURI;
|
---|
| 339 | XERCES_STD_QUALIFIER cerr << "==Parsing== " << xmlFile << XERCES_STD_QUALIFIER endl;
|
---|
| 340 | }
|
---|
| 341 | }
|
---|
| 342 | else
|
---|
| 343 | break;
|
---|
| 344 | }
|
---|
| 345 | else {
|
---|
| 346 | if (argInd < argC)
|
---|
| 347 | {
|
---|
| 348 | xmlFile = argV[argInd];
|
---|
| 349 | argInd++;
|
---|
| 350 | }
|
---|
| 351 | else
|
---|
| 352 | break;
|
---|
| 353 | }
|
---|
| 354 |
|
---|
| 355 | //reset error count first
|
---|
| 356 | handler.resetErrors();
|
---|
| 357 |
|
---|
| 358 | try
|
---|
| 359 | {
|
---|
| 360 | const unsigned long startMillis = XMLPlatformUtils::getCurrentMillis();
|
---|
| 361 | parser->parse(xmlFile);
|
---|
| 362 | const unsigned long endMillis = XMLPlatformUtils::getCurrentMillis();
|
---|
| 363 | duration = endMillis - startMillis;
|
---|
| 364 | }
|
---|
| 365 | catch (const OutOfMemoryException&)
|
---|
| 366 | {
|
---|
| 367 | XERCES_STD_QUALIFIER cerr << "OutOfMemoryException" << XERCES_STD_QUALIFIER endl;
|
---|
| 368 | errorOccurred = true;
|
---|
| 369 | continue;
|
---|
| 370 | }
|
---|
| 371 | catch (const XMLException& e)
|
---|
| 372 | {
|
---|
| 373 | XERCES_STD_QUALIFIER cerr << "\nError during parsing: '" << xmlFile << "'\n"
|
---|
| 374 | << "Exception message is: \n"
|
---|
| 375 | << StrX(e.getMessage()) << "\n" << XERCES_STD_QUALIFIER endl;
|
---|
| 376 | errorOccurred = true;
|
---|
| 377 | continue;
|
---|
| 378 | }
|
---|
| 379 |
|
---|
| 380 | catch (...)
|
---|
| 381 | {
|
---|
| 382 | XERCES_STD_QUALIFIER cerr << "\nUnexpected exception during parsing: '" << xmlFile << "'\n";
|
---|
| 383 | errorOccurred = true;
|
---|
| 384 | continue;
|
---|
| 385 | }
|
---|
| 386 |
|
---|
| 387 |
|
---|
| 388 | // Print out the stats that we collected and time taken
|
---|
| 389 | if (!handler.getSawErrors())
|
---|
| 390 | {
|
---|
| 391 | XERCES_STD_QUALIFIER cout << xmlFile << ": " << duration << " ms ("
|
---|
| 392 | << handler.getElementCount() << " elems, "
|
---|
| 393 | << handler.getAttrCount() << " attrs, "
|
---|
| 394 | << handler.getSpaceCount() << " spaces, "
|
---|
| 395 | << handler.getCharacterCount() << " chars)" << XERCES_STD_QUALIFIER endl;
|
---|
| 396 | }
|
---|
| 397 | else
|
---|
| 398 | errorOccurred = true;
|
---|
| 399 | }
|
---|
| 400 |
|
---|
| 401 | if (doList)
|
---|
| 402 | fin.close();
|
---|
| 403 |
|
---|
| 404 | //
|
---|
| 405 | // Delete the parser itself. Must be done prior to calling Terminate, below.
|
---|
| 406 | //
|
---|
| 407 | delete parser;
|
---|
| 408 |
|
---|
| 409 | // And call the termination method
|
---|
| 410 | XMLPlatformUtils::Terminate();
|
---|
| 411 |
|
---|
| 412 | if (errorOccurred)
|
---|
| 413 | return 4;
|
---|
| 414 | else
|
---|
| 415 | return 0;
|
---|
| 416 |
|
---|
| 417 | }
|
---|
| 418 |
|
---|