test_pax_filename_encoding.c revision 311041
1/*- 2 * Copyright (c) 2003-2007 Tim Kientzle 3 * All rights reserved. 4 * 5 * Redistribution and use in source and binary forms, with or without 6 * modification, are permitted provided that the following conditions 7 * are met: 8 * 1. Redistributions of source code must retain the above copyright 9 * notice, this list of conditions and the following disclaimer. 10 * 2. Redistributions in binary form must reproduce the above copyright 11 * notice, this list of conditions and the following disclaimer in the 12 * documentation and/or other materials provided with the distribution. 13 * 14 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR 15 * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES 16 * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. 17 * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT, 18 * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT 19 * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, 20 * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY 21 * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT 22 * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF 23 * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. 24 */ 25#include "test.h" 26__FBSDID("$FreeBSD: stable/11/contrib/libarchive/libarchive/test/test_pax_filename_encoding.c 311041 2017-01-02 01:41:31Z mm $"); 27 28#include <locale.h> 29 30/* 31 * Pax interchange is supposed to encode filenames into 32 * UTF-8. Of course, that's not always possible. This 33 * test is intended to verify that filenames always get 34 * stored and restored correctly, regardless of the encodings. 35 */ 36 37/* 38 * Read a manually-created archive that has filenames that are 39 * stored in binary instead of UTF-8 and verify that we get 40 * the right filename returned and that we get a warning only 41 * if the header isn't marked as binary. 42 */ 43static void 44test_pax_filename_encoding_1(void) 45{ 46 static const char testname[] = "test_pax_filename_encoding.tar"; 47 /* 48 * \314\214 is a valid 2-byte UTF-8 sequence. 49 * \374 is invalid in UTF-8. 50 */ 51 char filename[] = "abc\314\214mno\374xyz"; 52 struct archive *a; 53 struct archive_entry *entry; 54 55 /* 56 * Read an archive that has non-UTF8 pax filenames in it. 57 */ 58 extract_reference_file(testname); 59 a = archive_read_new(); 60 assertEqualInt(ARCHIVE_OK, archive_read_support_format_tar(a)); 61 assertEqualInt(ARCHIVE_OK, archive_read_support_filter_all(a)); 62 assertEqualInt(ARCHIVE_OK, 63 archive_read_open_filename(a, testname, 10240)); 64 /* 65 * First entry in this test archive has an invalid UTF-8 sequence 66 * in it, but the header is not marked as hdrcharset=BINARY, so that 67 * requires a warning. 68 */ 69 failure("Invalid UTF8 in a pax archive pathname should cause a warning"); 70 assertEqualInt(ARCHIVE_WARN, archive_read_next_header(a, &entry)); 71 assertEqualString(filename, archive_entry_pathname(entry)); 72 /* 73 * Second entry is identical except that it does have 74 * hdrcharset=BINARY, so no warning should be generated. 75 */ 76 failure("A pathname with hdrcharset=BINARY can have invalid UTF8\n" 77 " characters in it without generating a warning"); 78 assertEqualInt(ARCHIVE_OK, archive_read_next_header(a, &entry)); 79 assertEqualString(filename, archive_entry_pathname(entry)); 80 archive_read_free(a); 81} 82 83/* 84 * Set the locale and write a pathname containing invalid characters. 85 * This should work; the underlying implementation should automatically 86 * fall back to storing the pathname in binary. 87 */ 88static void 89test_pax_filename_encoding_2(void) 90{ 91 char filename[] = "abc\314\214mno\374xyz"; 92 struct archive *a; 93 struct archive_entry *entry; 94 char buff[65536]; 95 char longname[] = "abc\314\214mno\374xyz" 96 "/abc\314\214mno\374xyz/abcdefghijklmnopqrstuvwxyz" 97 "/abc\314\214mno\374xyz/abcdefghijklmnopqrstuvwxyz" 98 "/abc\314\214mno\374xyz/abcdefghijklmnopqrstuvwxyz" 99 "/abc\314\214mno\374xyz/abcdefghijklmnopqrstuvwxyz" 100 "/abc\314\214mno\374xyz/abcdefghijklmnopqrstuvwxyz" 101 "/abc\314\214mno\374xyz/abcdefghijklmnopqrstuvwxyz" 102 ; 103 size_t used; 104 105 /* 106 * We need a starting locale which has invalid sequences. 107 * en_US.UTF-8 seems to be commonly supported. 108 */ 109 /* If it doesn't exist, just warn and return. */ 110 if (NULL == setlocale(LC_ALL, "en_US.UTF-8")) { 111 skipping("invalid encoding tests require a suitable locale;" 112 " en_US.UTF-8 not available on this system"); 113 return; 114 } 115 116 assert((a = archive_write_new()) != NULL); 117 assertEqualIntA(a, 0, archive_write_set_format_pax(a)); 118 assertEqualIntA(a, 0, archive_write_add_filter_none(a)); 119 assertEqualIntA(a, 0, archive_write_set_bytes_per_block(a, 0)); 120 assertEqualInt(0, 121 archive_write_open_memory(a, buff, sizeof(buff), &used)); 122 123 assert((entry = archive_entry_new()) != NULL); 124 /* Set pathname, gname, uname, hardlink to nonconvertible values. */ 125 archive_entry_copy_pathname(entry, filename); 126 archive_entry_copy_gname(entry, filename); 127 archive_entry_copy_uname(entry, filename); 128 archive_entry_copy_hardlink(entry, filename); 129 archive_entry_set_filetype(entry, AE_IFREG); 130 failure("This should generate a warning for nonconvertible names."); 131 assertEqualInt(ARCHIVE_WARN, archive_write_header(a, entry)); 132 archive_entry_free(entry); 133 134 assert((entry = archive_entry_new()) != NULL); 135 /* Set path, gname, uname, and symlink to nonconvertible values. */ 136 archive_entry_copy_pathname(entry, filename); 137 archive_entry_copy_gname(entry, filename); 138 archive_entry_copy_uname(entry, filename); 139 archive_entry_copy_symlink(entry, filename); 140 archive_entry_set_filetype(entry, AE_IFLNK); 141 failure("This should generate a warning for nonconvertible names."); 142 assertEqualInt(ARCHIVE_WARN, archive_write_header(a, entry)); 143 archive_entry_free(entry); 144 145 assert((entry = archive_entry_new()) != NULL); 146 /* Set pathname to a very long nonconvertible value. */ 147 archive_entry_copy_pathname(entry, longname); 148 archive_entry_set_filetype(entry, AE_IFREG); 149 failure("This should generate a warning for nonconvertible names."); 150 assertEqualInt(ARCHIVE_WARN, archive_write_header(a, entry)); 151 archive_entry_free(entry); 152 153 assertEqualIntA(a, ARCHIVE_OK, archive_write_close(a)); 154 assertEqualInt(ARCHIVE_OK, archive_write_free(a)); 155 156 /* 157 * Now read the entries back. 158 */ 159 160 assert((a = archive_read_new()) != NULL); 161 assertEqualInt(0, archive_read_support_format_tar(a)); 162 assertEqualInt(0, archive_read_open_memory(a, buff, used)); 163 164 assertEqualInt(0, archive_read_next_header(a, &entry)); 165 assertEqualString(filename, archive_entry_pathname(entry)); 166 assertEqualString(filename, archive_entry_gname(entry)); 167 assertEqualString(filename, archive_entry_uname(entry)); 168 assertEqualString(filename, archive_entry_hardlink(entry)); 169 170 assertEqualInt(0, archive_read_next_header(a, &entry)); 171 assertEqualString(filename, archive_entry_pathname(entry)); 172 assertEqualString(filename, archive_entry_gname(entry)); 173 assertEqualString(filename, archive_entry_uname(entry)); 174 assertEqualString(filename, archive_entry_symlink(entry)); 175 176 assertEqualInt(0, archive_read_next_header(a, &entry)); 177 assertEqualString(longname, archive_entry_pathname(entry)); 178 179 assertEqualIntA(a, ARCHIVE_OK, archive_read_close(a)); 180 assertEqualInt(ARCHIVE_OK, archive_read_free(a)); 181} 182 183#if 0 /* Disable this until Tim check out it. */ 184 185/* 186 * Create an entry starting from a wide-character Unicode pathname, 187 * read it back into "C" locale, which doesn't support the name. 188 * TODO: Figure out the "right" behavior here. 189 */ 190static void 191test_pax_filename_encoding_3(void) 192{ 193 wchar_t badname[] = L"xxxAyyyBzzz"; 194 const char badname_utf8[] = "xxx\xE1\x88\xB4yyy\xE5\x99\xB8zzz"; 195 struct archive *a; 196 struct archive_entry *entry; 197 char buff[65536]; 198 size_t used; 199 200 badname[3] = 0x1234; 201 badname[7] = 0x5678; 202 203 /* If it doesn't exist, just warn and return. */ 204 if (NULL == setlocale(LC_ALL, "C")) { 205 skipping("Can't set \"C\" locale, so can't exercise " 206 "certain character-conversion failures"); 207 return; 208 } 209 210 /* If wctomb is broken, warn and return. */ 211 if (wctomb(buff, 0x1234) > 0) { 212 skipping("Cannot test conversion failures because \"C\" " 213 "locale on this system has no invalid characters."); 214 return; 215 } 216 217 /* If wctomb is broken, warn and return. */ 218 if (wctomb(buff, 0x1234) > 0) { 219 skipping("Cannot test conversion failures because \"C\" " 220 "locale on this system has no invalid characters."); 221 return; 222 } 223 224 /* Skip test if archive_entry_update_pathname_utf8() is broken. */ 225 /* In particular, this is currently broken on Win32 because 226 * setlocale() does not set the default encoding for CP_ACP. */ 227 entry = archive_entry_new(); 228 if (archive_entry_update_pathname_utf8(entry, badname_utf8)) { 229 archive_entry_free(entry); 230 skipping("Cannot test conversion failures."); 231 return; 232 } 233 archive_entry_free(entry); 234 235 assert((a = archive_write_new()) != NULL); 236 assertEqualIntA(a, 0, archive_write_set_format_pax(a)); 237 assertEqualIntA(a, 0, archive_write_add_filter_none(a)); 238 assertEqualIntA(a, 0, archive_write_set_bytes_per_block(a, 0)); 239 assertEqualInt(0, 240 archive_write_open_memory(a, buff, sizeof(buff), &used)); 241 242 assert((entry = archive_entry_new()) != NULL); 243 /* Set pathname to non-convertible wide value. */ 244 archive_entry_copy_pathname_w(entry, badname); 245 archive_entry_set_filetype(entry, AE_IFREG); 246 assertEqualInt(ARCHIVE_OK, archive_write_header(a, entry)); 247 archive_entry_free(entry); 248 249 assert((entry = archive_entry_new()) != NULL); 250 archive_entry_copy_pathname_w(entry, L"abc"); 251 /* Set gname to non-convertible wide value. */ 252 archive_entry_copy_gname_w(entry, badname); 253 archive_entry_set_filetype(entry, AE_IFREG); 254 assertEqualInt(ARCHIVE_OK, archive_write_header(a, entry)); 255 archive_entry_free(entry); 256 257 assert((entry = archive_entry_new()) != NULL); 258 archive_entry_copy_pathname_w(entry, L"abc"); 259 /* Set uname to non-convertible wide value. */ 260 archive_entry_copy_uname_w(entry, badname); 261 archive_entry_set_filetype(entry, AE_IFREG); 262 assertEqualInt(ARCHIVE_OK, archive_write_header(a, entry)); 263 archive_entry_free(entry); 264 265 assert((entry = archive_entry_new()) != NULL); 266 archive_entry_copy_pathname_w(entry, L"abc"); 267 /* Set hardlink to non-convertible wide value. */ 268 archive_entry_copy_hardlink_w(entry, badname); 269 archive_entry_set_filetype(entry, AE_IFREG); 270 assertEqualInt(ARCHIVE_OK, archive_write_header(a, entry)); 271 archive_entry_free(entry); 272 273 assert((entry = archive_entry_new()) != NULL); 274 archive_entry_copy_pathname_w(entry, L"abc"); 275 /* Set symlink to non-convertible wide value. */ 276 archive_entry_copy_symlink_w(entry, badname); 277 archive_entry_set_filetype(entry, AE_IFLNK); 278 assertEqualInt(ARCHIVE_OK, archive_write_header(a, entry)); 279 archive_entry_free(entry); 280 281 assertEqualIntA(a, ARCHIVE_OK, archive_write_close(a)); 282 assertEqualInt(ARCHIVE_OK, archive_write_free(a)); 283 284 /* 285 * Now read the entries back. 286 */ 287 288 assert((a = archive_read_new()) != NULL); 289 assertEqualInt(0, archive_read_support_format_tar(a)); 290 assertEqualInt(0, archive_read_open_memory(a, buff, used)); 291 292 failure("A non-convertible pathname should cause a warning."); 293 assertEqualInt(ARCHIVE_WARN, archive_read_next_header(a, &entry)); 294 assertEqualWString(badname, archive_entry_pathname_w(entry)); 295 failure("If native locale can't convert, we should get UTF-8 back."); 296 assertEqualString(badname_utf8, archive_entry_pathname(entry)); 297 298 failure("A non-convertible gname should cause a warning."); 299 assertEqualInt(ARCHIVE_WARN, archive_read_next_header(a, &entry)); 300 assertEqualWString(badname, archive_entry_gname_w(entry)); 301 failure("If native locale can't convert, we should get UTF-8 back."); 302 assertEqualString(badname_utf8, archive_entry_gname(entry)); 303 304 failure("A non-convertible uname should cause a warning."); 305 assertEqualInt(ARCHIVE_WARN, archive_read_next_header(a, &entry)); 306 assertEqualWString(badname, archive_entry_uname_w(entry)); 307 failure("If native locale can't convert, we should get UTF-8 back."); 308 assertEqualString(badname_utf8, archive_entry_uname(entry)); 309 310 failure("A non-convertible hardlink should cause a warning."); 311 assertEqualInt(ARCHIVE_WARN, archive_read_next_header(a, &entry)); 312 assertEqualWString(badname, archive_entry_hardlink_w(entry)); 313 failure("If native locale can't convert, we should get UTF-8 back."); 314 assertEqualString(badname_utf8, archive_entry_hardlink(entry)); 315 316 failure("A non-convertible symlink should cause a warning."); 317 assertEqualInt(ARCHIVE_WARN, archive_read_next_header(a, &entry)); 318 assertEqualWString(badname, archive_entry_symlink_w(entry)); 319 assertEqualWString(NULL, archive_entry_hardlink_w(entry)); 320 failure("If native locale can't convert, we should get UTF-8 back."); 321 assertEqualString(badname_utf8, archive_entry_symlink(entry)); 322 323 assertEqualInt(ARCHIVE_EOF, archive_read_next_header(a, &entry)); 324 325 assertEqualIntA(a, ARCHIVE_OK, archive_read_close(a)); 326 assertEqualInt(ARCHIVE_OK, archive_read_free(a)); 327} 328#else 329static void 330test_pax_filename_encoding_3(void) 331{ 332} 333#endif 334 335/* 336 * Verify that KOI8-R filenames are correctly translated to Unicode and UTF-8. 337 */ 338DEFINE_TEST(test_pax_filename_encoding_KOI8R) 339{ 340 struct archive *a; 341 struct archive_entry *entry; 342 char buff[4096]; 343 size_t used; 344 345 if (NULL == setlocale(LC_ALL, "ru_RU.KOI8-R")) { 346 skipping("KOI8-R locale not available on this system."); 347 return; 348 } 349 350 /* Check if the platform completely supports the string conversion. */ 351 a = archive_write_new(); 352 assertEqualInt(ARCHIVE_OK, archive_write_set_format_pax(a)); 353 if (archive_write_set_options(a, "hdrcharset=UTF-8") != ARCHIVE_OK) { 354 skipping("This system cannot convert character-set" 355 " from KOI8-R to UTF-8."); 356 archive_write_free(a); 357 return; 358 } 359 archive_write_free(a); 360 361 /* Re-create a write archive object since filenames should be written 362 * in UTF-8 by default. */ 363 a = archive_write_new(); 364 assertEqualInt(ARCHIVE_OK, archive_write_set_format_pax(a)); 365 assertEqualInt(ARCHIVE_OK, 366 archive_write_open_memory(a, buff, sizeof(buff), &used)); 367 368 entry = archive_entry_new2(a); 369 archive_entry_set_pathname(entry, "\xD0\xD2\xC9"); 370 archive_entry_set_filetype(entry, AE_IFREG); 371 archive_entry_set_size(entry, 0); 372 assertEqualInt(ARCHIVE_OK, archive_write_header(a, entry)); 373 archive_entry_free(entry); 374 assertEqualInt(ARCHIVE_OK, archive_write_free(a)); 375 376 /* Above three characters in KOI8-R should translate to the following 377 * three characters (two bytes each) in UTF-8. */ 378 assertEqualMem(buff + 512, "15 path=\xD0\xBF\xD1\x80\xD0\xB8\x0A", 15); 379} 380 381/* 382 * Verify that CP1251 filenames are correctly translated to Unicode and UTF-8. 383 */ 384DEFINE_TEST(test_pax_filename_encoding_CP1251) 385{ 386 struct archive *a; 387 struct archive_entry *entry; 388 char buff[4096]; 389 size_t used; 390 391 if (NULL == setlocale(LC_ALL, "Russian_Russia") && 392 NULL == setlocale(LC_ALL, "ru_RU.CP1251")) { 393 skipping("KOI8-R locale not available on this system."); 394 return; 395 } 396 397 /* Check if the platform completely supports the string conversion. */ 398 a = archive_write_new(); 399 assertEqualInt(ARCHIVE_OK, archive_write_set_format_pax(a)); 400 if (archive_write_set_options(a, "hdrcharset=UTF-8") != ARCHIVE_OK) { 401 skipping("This system cannot convert character-set" 402 " from KOI8-R to UTF-8."); 403 archive_write_free(a); 404 return; 405 } 406 archive_write_free(a); 407 408 /* Re-create a write archive object since filenames should be written 409 * in UTF-8 by default. */ 410 a = archive_write_new(); 411 assertEqualInt(ARCHIVE_OK, archive_write_set_format_pax(a)); 412 assertEqualInt(ARCHIVE_OK, 413 archive_write_open_memory(a, buff, sizeof(buff), &used)); 414 415 entry = archive_entry_new2(a); 416 archive_entry_set_pathname(entry, "\xef\xf0\xe8"); 417 archive_entry_set_filetype(entry, AE_IFREG); 418 archive_entry_set_size(entry, 0); 419 assertEqualInt(ARCHIVE_OK, archive_write_header(a, entry)); 420 archive_entry_free(entry); 421 assertEqualInt(ARCHIVE_OK, archive_write_free(a)); 422 423 /* Above three characters in KOI8-R should translate to the following 424 * three characters (two bytes each) in UTF-8. */ 425 assertEqualMem(buff + 512, "15 path=\xD0\xBF\xD1\x80\xD0\xB8\x0A", 15); 426} 427 428/* 429 * Verify that EUC-JP filenames are correctly translated to Unicode and UTF-8. 430 */ 431DEFINE_TEST(test_pax_filename_encoding_EUCJP) 432{ 433 struct archive *a; 434 struct archive_entry *entry; 435 char buff[4096]; 436 size_t used; 437 438 if (NULL == setlocale(LC_ALL, "ja_JP.eucJP")) { 439 skipping("eucJP locale not available on this system."); 440 return; 441 } 442 443 /* Check if the platform completely supports the string conversion. */ 444 a = archive_write_new(); 445 assertEqualInt(ARCHIVE_OK, archive_write_set_format_pax(a)); 446 if (archive_write_set_options(a, "hdrcharset=UTF-8") != ARCHIVE_OK) { 447 skipping("This system cannot convert character-set" 448 " from eucJP to UTF-8."); 449 archive_write_free(a); 450 return; 451 } 452 archive_write_free(a); 453 454 /* Re-create a write archive object since filenames should be written 455 * in UTF-8 by default. */ 456 a = archive_write_new(); 457 assertEqualInt(ARCHIVE_OK, archive_write_set_format_pax(a)); 458 assertEqualInt(ARCHIVE_OK, 459 archive_write_open_memory(a, buff, sizeof(buff), &used)); 460 461 entry = archive_entry_new2(a); 462 archive_entry_set_pathname(entry, "\xC9\xBD.txt"); 463 /* Check the Unicode version. */ 464 archive_entry_set_filetype(entry, AE_IFREG); 465 archive_entry_set_size(entry, 0); 466 assertEqualInt(ARCHIVE_OK, archive_write_header(a, entry)); 467 archive_entry_free(entry); 468 assertEqualInt(ARCHIVE_OK, archive_write_free(a)); 469 470 /* Check UTF-8 version. */ 471 assertEqualMem(buff + 512, "16 path=\xE8\xA1\xA8.txt\x0A", 16); 472 473} 474 475/* 476 * Verify that CP932/SJIS filenames are correctly translated to Unicode and UTF-8. 477 */ 478DEFINE_TEST(test_pax_filename_encoding_CP932) 479{ 480 struct archive *a; 481 struct archive_entry *entry; 482 char buff[4096]; 483 size_t used; 484 485 if (NULL == setlocale(LC_ALL, "Japanese_Japan") && 486 NULL == setlocale(LC_ALL, "ja_JP.SJIS")) { 487 skipping("eucJP locale not available on this system."); 488 return; 489 } 490 491 /* Check if the platform completely supports the string conversion. */ 492 a = archive_write_new(); 493 assertEqualInt(ARCHIVE_OK, archive_write_set_format_pax(a)); 494 if (archive_write_set_options(a, "hdrcharset=UTF-8") != ARCHIVE_OK) { 495 skipping("This system cannot convert character-set" 496 " from CP932/SJIS to UTF-8."); 497 archive_write_free(a); 498 return; 499 } 500 archive_write_free(a); 501 502 /* Re-create a write archive object since filenames should be written 503 * in UTF-8 by default. */ 504 a = archive_write_new(); 505 assertEqualInt(ARCHIVE_OK, archive_write_set_format_pax(a)); 506 assertEqualInt(ARCHIVE_OK, 507 archive_write_open_memory(a, buff, sizeof(buff), &used)); 508 509 entry = archive_entry_new2(a); 510 archive_entry_set_pathname(entry, "\x95\x5C.txt"); 511 /* Check the Unicode version. */ 512 archive_entry_set_filetype(entry, AE_IFREG); 513 archive_entry_set_size(entry, 0); 514 assertEqualInt(ARCHIVE_OK, archive_write_header(a, entry)); 515 archive_entry_free(entry); 516 assertEqualInt(ARCHIVE_OK, archive_write_free(a)); 517 518 /* Check UTF-8 version. */ 519 assertEqualMem(buff + 512, "16 path=\xE8\xA1\xA8.txt\x0A", 16); 520 521} 522 523/* 524 * Verify that KOI8-R filenames are not translated to Unicode and UTF-8 525 * when using hdrcharset=BINARY option. 526 */ 527DEFINE_TEST(test_pax_filename_encoding_KOI8R_BINARY) 528{ 529 struct archive *a; 530 struct archive_entry *entry; 531 char buff[4096]; 532 size_t used; 533 534 if (NULL == setlocale(LC_ALL, "ru_RU.KOI8-R")) { 535 skipping("KOI8-R locale not available on this system."); 536 return; 537 } 538 539 a = archive_write_new(); 540 assertEqualInt(ARCHIVE_OK, archive_write_set_format_pax(a)); 541 /* BINARY mode should be accepted. */ 542 assertEqualInt(ARCHIVE_OK, 543 archive_write_set_options(a, "hdrcharset=BINARY")); 544 assertEqualInt(ARCHIVE_OK, 545 archive_write_open_memory(a, buff, sizeof(buff), &used)); 546 547 entry = archive_entry_new2(a); 548 archive_entry_set_pathname(entry, "\xD0\xD2\xC9"); 549 archive_entry_set_filetype(entry, AE_IFREG); 550 archive_entry_set_size(entry, 0); 551 assertEqualInt(ARCHIVE_OK, archive_write_header(a, entry)); 552 archive_entry_free(entry); 553 assertEqualInt(ARCHIVE_OK, archive_write_free(a)); 554 555 /* "hdrcharset=BINARY" pax attribute should be written. */ 556 assertEqualMem(buff + 512, "21 hdrcharset=BINARY\x0A", 21); 557 /* Above three characters in KOI8-R should not translate to any 558 * character-set. */ 559 assertEqualMem(buff + 512+21, "12 path=\xD0\xD2\xC9\x0A", 12); 560} 561 562/* 563 * Pax format writer only accepts both BINARY and UTF-8. 564 * If other character-set name is specified, you will get ARCHIVE_FAILED. 565 */ 566DEFINE_TEST(test_pax_filename_encoding_KOI8R_CP1251) 567{ 568 struct archive *a; 569 570 if (NULL == setlocale(LC_ALL, "ru_RU.KOI8-R")) { 571 skipping("KOI8-R locale not available on this system."); 572 return; 573 } 574 575 a = archive_write_new(); 576 assertEqualInt(ARCHIVE_OK, archive_write_set_format_pax(a)); 577 /* pax format writer only accepts both BINARY and UTF-8. */ 578 assertEqualInt(ARCHIVE_FAILED, 579 archive_write_set_options(a, "hdrcharset=CP1251")); 580 assertEqualInt(ARCHIVE_OK, archive_write_free(a)); 581} 582 583 584DEFINE_TEST(test_pax_filename_encoding) 585{ 586 test_pax_filename_encoding_1(); 587 test_pax_filename_encoding_2(); 588 test_pax_filename_encoding_3(); 589} 590