| 1 | /* $Id: filesplitter.cpp 42414 2012-07-26 15:06:15Z vboxsync $ */
|
|---|
| 2 | /** @file
|
|---|
| 3 | * File splitter - Splits a text file according to ###### markers in it.
|
|---|
| 4 | */
|
|---|
| 5 |
|
|---|
| 6 | /*
|
|---|
| 7 | * Copyright (C) 2006-2012 Oracle Corporation
|
|---|
| 8 | *
|
|---|
| 9 | * This file is part of VirtualBox Open Source Edition (OSE), as
|
|---|
| 10 | * available from http://www.virtualbox.org. This file is free software;
|
|---|
| 11 | * you can redistribute it and/or modify it under the terms of the GNU
|
|---|
| 12 | * General Public License (GPL) as published by the Free Software
|
|---|
| 13 | * Foundation, in version 2 as it comes in the "COPYING" file of the
|
|---|
| 14 | * VirtualBox OSE distribution. VirtualBox OSE is distributed in the
|
|---|
| 15 | * hope that it will be useful, but WITHOUT ANY WARRANTY of any kind.
|
|---|
| 16 | */
|
|---|
| 17 |
|
|---|
| 18 |
|
|---|
| 19 | /*******************************************************************************
|
|---|
| 20 | * Header Files *
|
|---|
| 21 | *******************************************************************************/
|
|---|
| 22 | #include <sys/types.h>
|
|---|
| 23 | #include <sys/stat.h>
|
|---|
| 24 | #include <stdio.h>
|
|---|
| 25 | #include <stdlib.h>
|
|---|
| 26 | #include <errno.h>
|
|---|
| 27 |
|
|---|
| 28 | #include <iprt/string.h>
|
|---|
| 29 | #include <iprt/stdarg.h>
|
|---|
| 30 |
|
|---|
| 31 |
|
|---|
| 32 | /*******************************************************************************
|
|---|
| 33 | * Defined Constants And Macros *
|
|---|
| 34 | *******************************************************************************/
|
|---|
| 35 | #ifndef S_ISDIR
|
|---|
| 36 | # define S_ISDIR(a_fMode) ( (S_IFMT & (a_fMode)) == S_IFDIR )
|
|---|
| 37 | #endif
|
|---|
| 38 |
|
|---|
| 39 |
|
|---|
| 40 | /**
|
|---|
| 41 | * Calculates the line number for a file position.
|
|---|
| 42 | *
|
|---|
| 43 | * @returns Line number.
|
|---|
| 44 | * @param pcszContent The file content.
|
|---|
| 45 | * @param pcszPos The current position.
|
|---|
| 46 | */
|
|---|
| 47 | static unsigned long lineNumber(const char *pcszContent, const char *pcszPos)
|
|---|
| 48 | {
|
|---|
| 49 | unsigned long cLine = 0;
|
|---|
| 50 | while ( *pcszContent
|
|---|
| 51 | && (uintptr_t)pcszContent < (uintptr_t)pcszPos)
|
|---|
| 52 | {
|
|---|
| 53 | pcszContent = strchr(pcszContent, '\n');
|
|---|
| 54 | if (!pcszContent)
|
|---|
| 55 | break;
|
|---|
| 56 | ++cLine;
|
|---|
| 57 | ++pcszContent;
|
|---|
| 58 | }
|
|---|
| 59 |
|
|---|
| 60 | return cLine;
|
|---|
| 61 | }
|
|---|
| 62 |
|
|---|
| 63 |
|
|---|
| 64 | /**
|
|---|
| 65 | * Writes an error message.
|
|---|
| 66 | *
|
|---|
| 67 | * @returns RTEXITCODE_FAILURE.
|
|---|
| 68 | * @param pcszFormat Error message.
|
|---|
| 69 | * @param ... Format argument referenced in the message.
|
|---|
| 70 | */
|
|---|
| 71 | static int printErr(const char *pcszFormat, ...)
|
|---|
| 72 | {
|
|---|
| 73 | va_list va;
|
|---|
| 74 |
|
|---|
| 75 | fprintf(stderr, "filesplitter: ");
|
|---|
| 76 | va_start(va, pcszFormat);
|
|---|
| 77 | vfprintf(stderr, pcszFormat, va);
|
|---|
| 78 | va_end(va);
|
|---|
| 79 |
|
|---|
| 80 | return RTEXITCODE_FAILURE;
|
|---|
| 81 | }
|
|---|
| 82 |
|
|---|
| 83 |
|
|---|
| 84 | /**
|
|---|
| 85 | * Opens the makefile list for writing.
|
|---|
| 86 | *
|
|---|
| 87 | * @returns Exit code.
|
|---|
| 88 | * @param pcszPath The path to the file.
|
|---|
| 89 | * @param pcszVariableName The make variable name.
|
|---|
| 90 | * @param ppFile Where to return the file stream.
|
|---|
| 91 | */
|
|---|
| 92 | static int openMakefileList(const char *pcszPath, const char *pcszVariableName, FILE **ppFile)
|
|---|
| 93 | {
|
|---|
| 94 | *ppFile = NULL;
|
|---|
| 95 |
|
|---|
| 96 | FILE *pFile= fopen(pcszPath, "w");
|
|---|
| 97 | if (!pFile)
|
|---|
| 98 | return printErr("Failed to open \"%s\" for writing the file list");
|
|---|
| 99 |
|
|---|
| 100 | if (fprintf(pFile, "%s := \\\n", pcszVariableName) <= 0)
|
|---|
| 101 | {
|
|---|
| 102 | fclose(pFile);
|
|---|
| 103 | return printErr("Error writing to the makefile list.\n");
|
|---|
| 104 | }
|
|---|
| 105 |
|
|---|
| 106 | *ppFile = pFile;
|
|---|
| 107 | return 0;
|
|---|
| 108 | }
|
|---|
| 109 |
|
|---|
| 110 |
|
|---|
| 111 | /**
|
|---|
| 112 | * Adds the given file to the makefile list.
|
|---|
| 113 | *
|
|---|
| 114 | * @returns Exit code.
|
|---|
| 115 | * @param pFile The file stream of the makefile list.
|
|---|
| 116 | * @param pszFilename The file name to add.
|
|---|
| 117 | */
|
|---|
| 118 | static int addFileToMakefileList(FILE *pFile, char *pszFilename)
|
|---|
| 119 | {
|
|---|
| 120 | if (pFile)
|
|---|
| 121 | {
|
|---|
| 122 | char *pszSlash = pszFilename;
|
|---|
| 123 | while ((pszSlash = strchr(pszSlash, '\\')) != NULL)
|
|---|
| 124 | *pszSlash++ = '/';
|
|---|
| 125 |
|
|---|
| 126 | if (fprintf(pFile, "\t%s \\\n", pszFilename) <= 0)
|
|---|
| 127 | return printErr("Error adding file to makefile list.\n");
|
|---|
| 128 | }
|
|---|
| 129 | return 0;
|
|---|
| 130 | }
|
|---|
| 131 |
|
|---|
| 132 |
|
|---|
| 133 | /**
|
|---|
| 134 | * Closes the makefile list.
|
|---|
| 135 | *
|
|---|
| 136 | * @returns Exit code derived from @a rc.
|
|---|
| 137 | * @param pFile The file stream of the makefile list.
|
|---|
| 138 | * @param rc The current exit code.
|
|---|
| 139 | */
|
|---|
| 140 | static int closeMakefileList(FILE *pFile, int rc)
|
|---|
| 141 | {
|
|---|
| 142 | fprintf(pFile, "\n\n");
|
|---|
| 143 | if (fclose(pFile))
|
|---|
| 144 | return printErr("Error closing the file list file: %s\n", strerror(errno));
|
|---|
| 145 | return rc;
|
|---|
| 146 | }
|
|---|
| 147 |
|
|---|
| 148 |
|
|---|
| 149 | /**
|
|---|
| 150 | * Reads in a file.
|
|---|
| 151 | *
|
|---|
| 152 | * @returns Exit code.
|
|---|
| 153 | * @param pcszFile The path to the file.
|
|---|
| 154 | * @param ppszFile Where to return the buffer.
|
|---|
| 155 | * @param pcchFile Where to return the file size.
|
|---|
| 156 | */
|
|---|
| 157 | static int readFile(const char *pcszFile, char **ppszFile, size_t *pcchFile)
|
|---|
| 158 | {
|
|---|
| 159 | FILE *pFile;
|
|---|
| 160 | struct stat FileStat;
|
|---|
| 161 | int rc;
|
|---|
| 162 |
|
|---|
| 163 | if (stat(pcszFile, &FileStat))
|
|---|
| 164 | return printErr("Failed to stat \"%s\": %s\n", pcszFile, strerror(errno));
|
|---|
| 165 |
|
|---|
| 166 | pFile = fopen(pcszFile, "r");
|
|---|
| 167 | if (!pFile)
|
|---|
| 168 | return printErr("Failed to open \"%s\": %s\n", pcszFile, strerror(errno));
|
|---|
| 169 |
|
|---|
| 170 | *ppszFile = (char *)malloc(FileStat.st_size + 1);
|
|---|
| 171 | if (*ppszFile)
|
|---|
| 172 | {
|
|---|
| 173 | errno = 0;
|
|---|
| 174 | size_t cbRead = fread(*ppszFile, 1, FileStat.st_size, pFile);
|
|---|
| 175 | if ( cbRead <= (size_t)FileStat.st_size
|
|---|
| 176 | && (cbRead > 0 || !ferror(pFile)) )
|
|---|
| 177 | {
|
|---|
| 178 | if (ftell(pFile) == FileStat.st_size) /* (\r\n vs \n in the DOS world) */
|
|---|
| 179 | {
|
|---|
| 180 | (*ppszFile)[cbRead] = '\0';
|
|---|
| 181 | if (pcchFile)
|
|---|
| 182 | *pcchFile = (size_t)cbRead;
|
|---|
| 183 |
|
|---|
| 184 | fclose(pFile);
|
|---|
| 185 | return 0;
|
|---|
| 186 | }
|
|---|
| 187 | }
|
|---|
| 188 |
|
|---|
| 189 | rc = printErr("Error reading \"%s\": %s\n", pcszFile, strerror(errno));
|
|---|
| 190 | free(*ppszFile);
|
|---|
| 191 | *ppszFile = NULL;
|
|---|
| 192 | }
|
|---|
| 193 | else
|
|---|
| 194 | rc = printErr("Failed to allocate %lu bytes\n", (unsigned long)(FileStat.st_size + 1));
|
|---|
| 195 | fclose(pFile);
|
|---|
| 196 | return rc;
|
|---|
| 197 | }
|
|---|
| 198 |
|
|---|
| 199 |
|
|---|
| 200 | /**
|
|---|
| 201 | * Checks whether the sub-file already exists and has the exact
|
|---|
| 202 | * same content.
|
|---|
| 203 | *
|
|---|
| 204 | * @returns @c true if the existing file matches exactly, otherwise @c false.
|
|---|
| 205 | * @param pcszFilename The path to the file.
|
|---|
| 206 | * @param pcszSubContent The content to write.
|
|---|
| 207 | * @param cchSubContent The length of the content.
|
|---|
| 208 | */
|
|---|
| 209 | static bool compareSubFile(const char *pcszFilename, const char *pcszSubContent, size_t cchSubContent)
|
|---|
| 210 | {
|
|---|
| 211 | struct stat FileStat;
|
|---|
| 212 | if (stat(pcszFilename, &FileStat))
|
|---|
| 213 | return false;
|
|---|
| 214 | if ((size_t)FileStat.st_size < cchSubContent)
|
|---|
| 215 | return false;
|
|---|
| 216 |
|
|---|
| 217 | size_t cchExisting;
|
|---|
| 218 | char *pszExisting;
|
|---|
| 219 | int rc = readFile(pcszFilename, &pszExisting, &cchExisting);
|
|---|
| 220 | if (rc)
|
|---|
| 221 | return false;
|
|---|
| 222 |
|
|---|
| 223 | bool fRc = cchExisting == cchSubContent
|
|---|
| 224 | && !memcmp(pcszSubContent, pszExisting, cchSubContent);
|
|---|
| 225 | free(pszExisting);
|
|---|
| 226 |
|
|---|
| 227 | return fRc;
|
|---|
| 228 | }
|
|---|
| 229 |
|
|---|
| 230 |
|
|---|
| 231 | /**
|
|---|
| 232 | * Writes out a sub-file.
|
|---|
| 233 | *
|
|---|
| 234 | * @returns exit code.
|
|---|
| 235 | * @param pcszFilename The path to the sub-file.
|
|---|
| 236 | * @param pcszSubContent The content of the file.
|
|---|
| 237 | * @param cchSubContent The size of the content.
|
|---|
| 238 | */
|
|---|
| 239 | static int writeSubFile(const char *pcszFilename, const char *pcszSubContent, size_t cchSubContent)
|
|---|
| 240 | {
|
|---|
| 241 | FILE *pFile = fopen(pcszFilename, "w");
|
|---|
| 242 | if (!pFile)
|
|---|
| 243 | return printErr("Failed to open \"%s\" for writing: %s\n", pcszFilename, strerror(errno));
|
|---|
| 244 |
|
|---|
| 245 | errno = 0;
|
|---|
| 246 | int rc = 0;
|
|---|
| 247 | if (fwrite(pcszSubContent, cchSubContent, 1, pFile) != 1)
|
|---|
| 248 | rc = printErr("Error writing \"%s\": %s\n", pcszFilename, strerror(errno));
|
|---|
| 249 |
|
|---|
| 250 | errno = 0;
|
|---|
| 251 | int rc2 = fclose(pFile);
|
|---|
| 252 | if (rc2 == EOF)
|
|---|
| 253 | rc = printErr("Error closing \"%s\": %s\n", pcszFilename, strerror(errno));
|
|---|
| 254 | return rc;
|
|---|
| 255 | }
|
|---|
| 256 |
|
|---|
| 257 |
|
|---|
| 258 | /**
|
|---|
| 259 | * Does the actual file splitting.
|
|---|
| 260 | *
|
|---|
| 261 | * @returns exit code.
|
|---|
| 262 | * @param pcszOutDir Path to the output directory.
|
|---|
| 263 | * @param pcszContent The content to split up.
|
|---|
| 264 | * @param pFileList The file stream of the makefile list. Can be NULL.
|
|---|
| 265 | */
|
|---|
| 266 | static int splitFile(const char *pcszOutDir, const char *pcszContent, FILE *pFileList)
|
|---|
| 267 | {
|
|---|
| 268 | static char const s_szBeginMarker[] = "\n// ##### BEGINFILE \"";
|
|---|
| 269 | static char const s_szEndMarker[] = "\n// ##### ENDFILE";
|
|---|
| 270 | const size_t cchBeginMarker = sizeof(s_szBeginMarker) - 1;
|
|---|
| 271 | const char *pcszSearch = pcszContent;
|
|---|
| 272 | size_t const cchOutDir = strlen(pcszOutDir);
|
|---|
| 273 | unsigned long cFilesWritten = 0;
|
|---|
| 274 | unsigned long cFilesUnchanged = 0;
|
|---|
| 275 | int rc = 0;
|
|---|
| 276 |
|
|---|
| 277 | do
|
|---|
| 278 | {
|
|---|
| 279 | /* find begin marker */
|
|---|
| 280 | const char *pcszBegin = strstr(pcszSearch, s_szBeginMarker);
|
|---|
| 281 | if (!pcszBegin)
|
|---|
| 282 | break;
|
|---|
| 283 |
|
|---|
| 284 | /* find line after begin marker */
|
|---|
| 285 | const char *pcszLineAfterBegin = strchr(pcszBegin + cchBeginMarker, '\n');
|
|---|
| 286 | if (!pcszLineAfterBegin)
|
|---|
| 287 | return printErr("No newline after begin-file marker found.\n");
|
|---|
| 288 | ++pcszLineAfterBegin;
|
|---|
| 289 |
|
|---|
| 290 | /* find filename end quote in begin marker line */
|
|---|
| 291 | const char *pcszStartFilename = pcszBegin + cchBeginMarker;
|
|---|
| 292 | const char *pcszEndQuote = (const char *)memchr(pcszStartFilename, '\"', pcszLineAfterBegin - pcszStartFilename);
|
|---|
| 293 | if (!pcszEndQuote)
|
|---|
| 294 | return printErr("Can't parse filename after begin-file marker (line %lu).\n",
|
|---|
| 295 | lineNumber(pcszContent, s_szBeginMarker));
|
|---|
| 296 |
|
|---|
| 297 | /* find end marker */
|
|---|
| 298 | const char *pcszEnd = strstr(pcszLineAfterBegin, s_szEndMarker);
|
|---|
| 299 | if (!pcszEnd)
|
|---|
| 300 | return printErr("No matching end-line marker for begin-file marker found (line %lu).\n",
|
|---|
| 301 | lineNumber(pcszContent, s_szBeginMarker));
|
|---|
| 302 |
|
|---|
| 303 | /* construct output filename */
|
|---|
| 304 | size_t cchFilename = pcszEndQuote - pcszStartFilename;
|
|---|
| 305 | char *pszFilename = (char *)malloc(cchOutDir + 1 + cchFilename + 1);
|
|---|
| 306 | if (!pszFilename)
|
|---|
| 307 | return printErr("Can't allocate memory for filename.\n");
|
|---|
| 308 |
|
|---|
| 309 | memcpy(pszFilename, pcszOutDir, cchOutDir);
|
|---|
| 310 | pszFilename[cchOutDir] = '/';
|
|---|
| 311 | memcpy(pszFilename + cchOutDir + 1, pcszStartFilename, cchFilename);
|
|---|
| 312 | pszFilename[cchFilename + 1 + cchOutDir] = '\0';
|
|---|
| 313 |
|
|---|
| 314 | /* Write the file only if necessary. */
|
|---|
| 315 | if (compareSubFile(pszFilename, pcszLineAfterBegin, pcszEnd - pcszLineAfterBegin))
|
|---|
| 316 | cFilesUnchanged++;
|
|---|
| 317 | else
|
|---|
| 318 | {
|
|---|
| 319 | rc = writeSubFile(pszFilename, pcszLineAfterBegin, pcszEnd - pcszLineAfterBegin);
|
|---|
| 320 | cFilesWritten++;
|
|---|
| 321 | }
|
|---|
| 322 |
|
|---|
| 323 | if (!rc)
|
|---|
| 324 | rc = addFileToMakefileList(pFileList, pszFilename);
|
|---|
| 325 |
|
|---|
| 326 | free(pszFilename);
|
|---|
| 327 |
|
|---|
| 328 | pcszSearch = pcszEnd;
|
|---|
| 329 | } while (rc == 0 && pcszSearch);
|
|---|
| 330 |
|
|---|
| 331 | printf("filesplitter: Out of %lu files: %lu rewritten, %lu unchanged. (%s)\n",
|
|---|
| 332 | cFilesWritten + cFilesUnchanged, cFilesWritten, cFilesUnchanged, pcszOutDir);
|
|---|
| 333 | return rc;
|
|---|
| 334 | }
|
|---|
| 335 |
|
|---|
| 336 |
|
|---|
| 337 | int main(int argc, char *argv[])
|
|---|
| 338 | {
|
|---|
| 339 | int rc = 0;
|
|---|
| 340 |
|
|---|
| 341 | if (argc == 3 || argc == 5)
|
|---|
| 342 | {
|
|---|
| 343 | struct stat DirStat;
|
|---|
| 344 | if ( stat(argv[2], &DirStat) == 0
|
|---|
| 345 | && S_ISDIR(DirStat.st_mode))
|
|---|
| 346 | {
|
|---|
| 347 | char *pszContent;
|
|---|
| 348 | rc = readFile(argv[1], &pszContent, NULL);
|
|---|
| 349 | if (!rc)
|
|---|
| 350 | {
|
|---|
| 351 | FILE *pFileList = NULL;
|
|---|
| 352 | if (argc == 5)
|
|---|
| 353 | rc = openMakefileList(argv[3], argv[4], &pFileList);
|
|---|
| 354 |
|
|---|
| 355 | if (argc < 4 || pFileList)
|
|---|
| 356 | rc = splitFile(argv[2], pszContent, pFileList);
|
|---|
| 357 |
|
|---|
| 358 | if (pFileList)
|
|---|
| 359 | rc = closeMakefileList(pFileList, rc);
|
|---|
| 360 | free(pszContent);
|
|---|
| 361 | }
|
|---|
| 362 | }
|
|---|
| 363 | else
|
|---|
| 364 | rc = printErr("Given argument \"%s\" is not a valid directory.\n", argv[2]);
|
|---|
| 365 | }
|
|---|
| 366 | else
|
|---|
| 367 | rc = printErr("Syntax error: usage: filesplitter <infile> <outdir> [<list.kmk> <kmkvar>]\n");
|
|---|
| 368 | return rc;
|
|---|
| 369 | }
|
|---|
