]>
git.ipfire.org Git - thirdparty/cups.git/blob - cups/testi18n.c
4 * Internationalization test for CUPS.
6 * Copyright 2007-2014 by Apple Inc.
7 * Copyright 1997-2006 by Easy Software Products.
9 * These coded instructions, statements, and computer programs are the
10 * property of Apple Inc. and are protected by Federal copyright
11 * law. Distribution and use rights are outlined in the file "LICENSE.txt"
12 * which should have been included with this file. If this file is
13 * file is missing or damaged, see the license at "http://www.cups.org/".
15 * This file is subject to the Apple OS-Developed Software exception.
19 * Include necessary headers...
22 #include "string-private.h"
23 #include "language-private.h"
33 static const char * const lang_encodings
[] =
34 { /* Encoding strings */
35 "us-ascii", "iso-8859-1",
36 "iso-8859-2", "iso-8859-3",
37 "iso-8859-4", "iso-8859-5",
38 "iso-8859-6", "iso-8859-7",
39 "iso-8859-8", "iso-8859-9",
40 "iso-8859-10", "utf-8",
41 "iso-8859-13", "iso-8859-14",
42 "iso-8859-15", "windows-874",
43 "windows-1250", "windows-1251",
44 "windows-1252", "windows-1253",
45 "windows-1254", "windows-1255",
46 "windows-1256", "windows-1257",
47 "windows-1258", "koi8-r",
48 "koi8-u", "iso-8859-11",
49 "iso-8859-16", "mac-roman",
67 "windows-932", "windows-936",
68 "windows-949", "windows-950",
69 "windows-1361", "unknown",
109 static void print_utf8(const char *msg
, const cups_utf8_t
*src
);
113 * 'main()' - Main entry for internationalization test module.
116 int /* O - Exit code */
117 main(int argc
, /* I - Argument Count */
118 char *argv
[]) /* I - Arguments */
120 FILE *fp
; /* File pointer */
121 int count
; /* File line counter */
122 int status
, /* Status of current test */
123 errors
; /* Error count */
124 char line
[1024]; /* File line source string */
125 int len
; /* Length (count) of string */
126 char legsrc
[1024], /* Legacy source string */
127 legdest
[1024], /* Legacy destination string */
128 *legptr
; /* Pointer into legacy string */
129 cups_utf8_t utf8latin
[] = /* UTF-8 Latin-1 source */
130 { 0x41, 0x20, 0x21, 0x3D, 0x20, 0xC3, 0x84, 0x2E, 0x00 };
131 /* "A != <A WITH DIAERESIS>." - use ISO 8859-1 */
132 cups_utf8_t utf8repla
[] = /* UTF-8 Latin-1 replacement */
133 { 0x41, 0x20, 0xE2, 0x89, 0xA2, 0x20, 0xC3, 0x84, 0x2E, 0x00 };
134 /* "A <NOT IDENTICAL TO> <A WITH DIAERESIS>." */
135 cups_utf8_t utf8greek
[] = /* UTF-8 Greek source string */
136 { 0x41, 0x20, 0x21, 0x3D, 0x20, 0xCE, 0x91, 0x2E, 0x00 };
137 /* "A != <ALPHA>." - use ISO 8859-7 */
138 cups_utf8_t utf8japan
[] = /* UTF-8 Japanese source */
139 { 0x41, 0x20, 0x21, 0x3D, 0x20, 0xEE, 0x9C, 0x80, 0x2E, 0x00 };
140 /* "A != <PRIVATE U+E700>." - use Windows 932 or EUC-JP */
141 cups_utf8_t utf8taiwan
[] = /* UTF-8 Chinese source */
142 { 0x41, 0x20, 0x21, 0x3D, 0x20, 0xE4, 0xB9, 0x82, 0x2E, 0x00 };
143 /* "A != <CJK U+4E42>." - use Windows 950 (Big5) or EUC-TW */
144 cups_utf8_t utf8dest
[1024]; /* UTF-8 destination string */
145 cups_utf32_t utf32dest
[1024]; /* UTF-32 destination string */
150 int i
; /* Looping var */
151 cups_encoding_t encoding
; /* Source encoding */
156 puts("Usage: ./testi18n [filename charset]");
160 if ((fp
= fopen(argv
[1], "rb")) == NULL
)
166 for (i
= 0, encoding
= CUPS_AUTO_ENCODING
;
167 i
< (int)(sizeof(lang_encodings
) / sizeof(lang_encodings
[0]));
169 if (!_cups_strcasecmp(lang_encodings
[i
], argv
[2]))
171 encoding
= (cups_encoding_t
)i
;
175 if (encoding
== CUPS_AUTO_ENCODING
)
177 fprintf(stderr
, "%s: Unknown character set!\n", argv
[2]);
181 while (fgets(line
, sizeof(line
), fp
))
183 if (cupsCharsetToUTF8(utf8dest
, line
, sizeof(utf8dest
), encoding
) < 0)
185 fprintf(stderr
, "%s: Unable to convert line: %s", argv
[1], line
);
189 fputs((char *)utf8dest
, stdout
);
197 * Start with some conversion tests from a UTF-8 test file.
202 if ((fp
= fopen("utf8demo.txt", "rb")) == NULL
)
204 perror("utf8demo.txt");
212 fputs("cupsUTF8ToUTF32 of utfdemo.txt: ", stdout
);
214 for (count
= 0, status
= 0; fgets(line
, sizeof(line
), fp
);)
218 if (cupsUTF8ToUTF32(utf32dest
, (cups_utf8_t
*)line
, 1024) < 0)
220 printf("FAIL (UTF-8 to UTF-32 on line %d)\n", count
);
231 * cupsUTF8ToCharset(CUPS_EUC_JP)
234 fputs("cupsUTF8ToCharset(CUPS_EUC_JP) of utfdemo.txt: ", stdout
);
238 for (count
= 0, status
= 0; fgets(line
, sizeof(line
), fp
);)
242 len
= cupsUTF8ToCharset(legdest
, (cups_utf8_t
*)line
, 1024, CUPS_EUC_JP
);
245 printf("FAIL (UTF-8 to EUC-JP on line %d)\n", count
);
258 * Test UTF-8 to legacy charset (ISO 8859-1)...
261 fputs("cupsUTF8ToCharset(CUPS_ISO8859_1): ", stdout
);
265 len
= cupsUTF8ToCharset(legdest
, utf8latin
, 1024, CUPS_ISO8859_1
);
268 printf("FAIL (len=%d)\n", len
);
278 fputs("cupsCharsetToUTF8(CUPS_ISO8859_1): ", stdout
);
280 strlcpy(legsrc
, legdest
, sizeof(legsrc
));
282 len
= cupsCharsetToUTF8(utf8dest
, legsrc
, 1024, CUPS_ISO8859_1
);
283 if ((size_t)len
!= strlen((char *)utf8latin
))
285 printf("FAIL (len=%d, expected %d)\n", len
, (int)strlen((char *)utf8latin
));
286 print_utf8(" utf8latin", utf8latin
);
287 print_utf8(" utf8dest", utf8dest
);
290 else if (memcmp(utf8latin
, utf8dest
, (size_t)len
))
292 puts("FAIL (results do not match)");
293 print_utf8(" utf8latin", utf8latin
);
294 print_utf8(" utf8dest", utf8dest
);
297 else if (cupsUTF8ToCharset(legdest
, utf8repla
, 1024, CUPS_ISO8859_1
) < 0)
299 puts("FAIL (replacement characters do not work!)");
306 * Test UTF-8 to/from legacy charset (ISO 8859-7)...
309 fputs("cupsUTF8ToCharset(CUPS_ISO8859_7): ", stdout
);
311 if (cupsUTF8ToCharset(legdest
, utf8greek
, 1024, CUPS_ISO8859_7
) < 0)
318 for (legptr
= legdest
; *legptr
&& *legptr
!= '?'; legptr
++);
322 puts("FAIL (unknown character)");
329 fputs("cupsCharsetToUTF8(CUPS_ISO8859_7): ", stdout
);
331 strlcpy(legsrc
, legdest
, sizeof(legsrc
));
333 len
= cupsCharsetToUTF8(utf8dest
, legsrc
, 1024, CUPS_ISO8859_7
);
334 if ((size_t)len
!= strlen((char *)utf8greek
))
336 printf("FAIL (len=%d, expected %d)\n", len
, (int)strlen((char *)utf8greek
));
337 print_utf8(" utf8greek", utf8greek
);
338 print_utf8(" utf8dest", utf8dest
);
341 else if (memcmp(utf8greek
, utf8dest
, (size_t)len
))
343 puts("FAIL (results do not match)");
344 print_utf8(" utf8greek", utf8greek
);
345 print_utf8(" utf8dest", utf8dest
);
352 * Test UTF-8 to/from legacy charset (Windows 932)...
355 fputs("cupsUTF8ToCharset(CUPS_WINDOWS_932): ", stdout
);
357 if (cupsUTF8ToCharset(legdest
, utf8japan
, 1024, CUPS_WINDOWS_932
) < 0)
364 for (legptr
= legdest
; *legptr
&& *legptr
!= '?'; legptr
++);
368 puts("FAIL (unknown character)");
375 fputs("cupsCharsetToUTF8(CUPS_WINDOWS_932): ", stdout
);
377 strlcpy(legsrc
, legdest
, sizeof(legsrc
));
379 len
= cupsCharsetToUTF8(utf8dest
, legsrc
, 1024, CUPS_WINDOWS_932
);
380 if ((size_t)len
!= strlen((char *)utf8japan
))
382 printf("FAIL (len=%d, expected %d)\n", len
, (int)strlen((char *)utf8japan
));
383 print_utf8(" utf8japan", utf8japan
);
384 print_utf8(" utf8dest", utf8dest
);
387 else if (memcmp(utf8japan
, utf8dest
, (size_t)len
))
389 puts("FAIL (results do not match)");
390 print_utf8(" utf8japan", utf8japan
);
391 print_utf8(" utf8dest", utf8dest
);
398 * Test UTF-8 to/from legacy charset (EUC-JP)...
401 fputs("cupsUTF8ToCharset(CUPS_EUC_JP): ", stdout
);
403 if (cupsUTF8ToCharset(legdest
, utf8japan
, 1024, CUPS_EUC_JP
) < 0)
410 for (legptr
= legdest
; *legptr
&& *legptr
!= '?'; legptr
++);
414 puts("FAIL (unknown character)");
422 fputs("cupsCharsetToUTF8(CUPS_EUC_JP): ", stdout
);
424 strlcpy(legsrc
, legdest
, sizeof(legsrc
));
426 len
= cupsCharsetToUTF8(utf8dest
, legsrc
, 1024, CUPS_EUC_JP
);
427 if ((size_t)len
!= strlen((char *)utf8japan
))
429 printf("FAIL (len=%d, expected %d)\n", len
, (int)strlen((char *)utf8japan
));
430 print_utf8(" utf8japan", utf8japan
);
431 print_utf8(" utf8dest", utf8dest
);
434 else if (memcmp(utf8japan
, utf8dest
, (size_t)len
))
436 puts("FAIL (results do not match)");
437 print_utf8(" utf8japan", utf8japan
);
438 print_utf8(" utf8dest", utf8dest
);
443 #endif /* !__linux */
446 * Test UTF-8 to/from legacy charset (Windows 950)...
449 fputs("cupsUTF8ToCharset(CUPS_WINDOWS_950): ", stdout
);
451 if (cupsUTF8ToCharset(legdest
, utf8taiwan
, 1024, CUPS_WINDOWS_950
) < 0)
458 for (legptr
= legdest
; *legptr
&& *legptr
!= '?'; legptr
++);
462 puts("FAIL (unknown character)");
469 fputs("cupsCharsetToUTF8(CUPS_WINDOWS_950): ", stdout
);
471 strlcpy(legsrc
, legdest
, sizeof(legsrc
));
473 len
= cupsCharsetToUTF8(utf8dest
, legsrc
, 1024, CUPS_WINDOWS_950
);
474 if ((size_t)len
!= strlen((char *)utf8taiwan
))
476 printf("FAIL (len=%d, expected %d)\n", len
, (int)strlen((char *)utf8taiwan
));
477 print_utf8(" utf8taiwan", utf8taiwan
);
478 print_utf8(" utf8dest", utf8dest
);
481 else if (memcmp(utf8taiwan
, utf8dest
, (size_t)len
))
483 puts("FAIL (results do not match)");
484 print_utf8(" utf8taiwan", utf8taiwan
);
485 print_utf8(" utf8dest", utf8dest
);
492 * Test UTF-8 to/from legacy charset (EUC-TW)...
495 fputs("cupsUTF8ToCharset(CUPS_EUC_TW): ", stdout
);
497 if (cupsUTF8ToCharset(legdest
, utf8taiwan
, 1024, CUPS_EUC_TW
) < 0)
504 for (legptr
= legdest
; *legptr
&& *legptr
!= '?'; legptr
++);
508 puts("FAIL (unknown character)");
515 fputs("cupsCharsetToUTF8(CUPS_EUC_TW): ", stdout
);
517 strlcpy(legsrc
, legdest
, sizeof(legsrc
));
519 len
= cupsCharsetToUTF8(utf8dest
, legsrc
, 1024, CUPS_EUC_TW
);
520 if ((size_t)len
!= strlen((char *)utf8taiwan
))
522 printf("FAIL (len=%d, expected %d)\n", len
, (int)strlen((char *)utf8taiwan
));
523 print_utf8(" utf8taiwan", utf8taiwan
);
524 print_utf8(" utf8dest", utf8dest
);
527 else if (memcmp(utf8taiwan
, utf8dest
, (size_t)len
))
529 puts("FAIL (results do not match)");
530 print_utf8(" utf8taiwan", utf8taiwan
);
531 print_utf8(" utf8dest", utf8dest
);
539 * Test UTF-8 (16-bit) to UTF-32 (w/ BOM)...
542 printf("\ntesti18n: Testing UTF-8 to UTF-32 (w/ BOM)...\n");
543 len
= cupsUTF8ToUTF32(utf32dest
, utf8good
, 1024);
548 print_utf8(" utf8good ", utf8good
);
549 print_utf32(" utf32dest", utf32dest
);
551 memcpy(utf32src
, utf32dest
, (len
+ 1) * sizeof(cups_utf32_t
));
552 len
= cupsUTF32ToUTF8(utf8dest
, utf32src
, 1024);
555 if (len
!= strlen ((char *) utf8good
))
557 if (memcmp(utf8good
, utf8dest
, len
) != 0)
561 * Test invalid UTF-8 (16-bit) to UTF-32 (w/ BOM)...
564 printf("\ntesti18n: Testing UTF-8 bad 16-bit source string...\n");
565 len
= cupsUTF8ToUTF32(utf32dest
, utf8bad
, 1024);
569 print_utf8(" utf8bad ", utf8bad
);
572 * Test _cupsCharmapFlush()...
575 printf("\ntesti18n: Testing _cupsCharmapFlush()...\n");
585 * 'print_utf8()' - Print UTF-8 string with (optional) message.
589 print_utf8(const char *msg
, /* I - Message String */
590 const cups_utf8_t
*src
) /* I - UTF-8 Source String */
592 const char *prefix
; /* Prefix string */
598 for (prefix
= " "; *src
; src
++)
600 printf("%s%02x", prefix
, *src
);
602 if ((src
[0] & 0x80) && (src
[1] & 0x80))