icuSources/samples/ustring/ustring.cpp

   1 /*
   2 *******************************************************************************
   3 *
   4 *   © 2016 and later: Unicode, Inc. and others.
   5 *   License & terms of use: http://www.unicode.org/copyright.html#License
   6 *
   7 *******************************************************************************
   8 *******************************************************************************
   9 *
  10 *   Copyright (C) 2000-2014, International Business Machines
  11 *   Corporation and others.  All Rights Reserved.
  12 *
  13 *******************************************************************************
  14 *   file name:  ustring.c
  15 *   encoding:   UTF-8
  16 *   tab size:   8 (not used)
  17 *   indentation:4
  18 *
  19 *   created on: 2000aug15
  20 *   created by: Markus W. Scherer
  21 *
  22 *   This file contains sample code that illustrates the use of Unicode strings
  23 *   with ICU.
  24 */
  25
  26 #include <stdio.h>
  27 #include "unicode/utypes.h"
  28 #include "unicode/uchar.h"
  29 #include "unicode/locid.h"
  30 #include "unicode/ustring.h"
  31 #include "unicode/ucnv.h"
  32 #include "unicode/unistr.h"
  33
  34 using namespace icu;
  35
  36 #ifndef UPRV_LENGTHOF
  37 #define UPRV_LENGTHOF(array) (int32_t)(sizeof(array)/sizeof((array)[0]))
  38 #endif
  39
  40 // helper functions -------------------------------------------------------- ***
  41
  42 // default converter for the platform encoding
  43 static UConverter *cnv=NULL;
  44
  45 static void
  46 printUString(const char *announce, const UChar *s, int32_t length) {
  47     static char out[200];
  48     UChar32 c;
  49     int32_t i;
  50     UErrorCode errorCode=U_ZERO_ERROR;
  51
  52     /*
  53      * Convert to the "platform encoding". See notes in printUnicodeString().
  54      * ucnv_fromUChars(), like most ICU APIs understands length==-1
  55      * to mean that the string is NUL-terminated.
  56      */
  57     ucnv_fromUChars(cnv, out, sizeof(out), s, length, &errorCode);
  58     if(U_FAILURE(errorCode) || errorCode==U_STRING_NOT_TERMINATED_WARNING) {
  59         printf("%sproblem converting string from Unicode: %s\n", announce, u_errorName(errorCode));
  60         return;
  61     }
  62
  63     printf("%s%s {", announce, out);
  64
  65     /* output the code points (not code units) */
  66     if(length>=0) {
  67         /* s is not NUL-terminated */
  68         for(i=0; i<length; /* U16_NEXT post-increments */) {
  69             U16_NEXT(s, i, length, c);
  70             printf(" %04x", c);
  71         }
  72     } else {
  73         /* s is NUL-terminated */
  74         for(i=0; /* condition in loop body */; /* U16_NEXT post-increments */) {
  75             U16_NEXT(s, i, length, c);
  76             if(c==0) {
  77                 break;
  78             }
  79             printf(" %04x", c);
  80         }
  81     }
  82     printf(" }\n");
  83 }
  84
  85 static void
  86 printUnicodeString(const char *announce, const UnicodeString &s) {
  87     static char out[200];
  88     int32_t i, length;
  89
  90     // output the string, converted to the platform encoding
  91
  92     // Note for Windows: The "platform encoding" defaults to the "ANSI codepage",
  93     // which is different from the "OEM codepage" in the console window.
  94     // However, if you pipe the output into a file and look at it with Notepad
  95     // or similar, then "ANSI" characters will show correctly.
  96     // Production code should be aware of what encoding is required,
  97     // and use a UConverter or at least a charset name explicitly.
  98     out[s.extract(0, 99, out)]=0;
  99     printf("%s%s {", announce, out);
 100
 101     // output the code units (not code points)
 102     length=s.length();
 103     for(i=0; i<length; ++i) {
 104         printf(" %04x", s.charAt(i));
 105     }
 106     printf(" }\n");
 107 }
 108
 109 // sample code for utf.h macros -------------------------------------------- ***
 110
 111 static void
 112 demo_utf_h_macros() {
 113     static UChar input[]={ 0x0061, 0xd800, 0xdc00, 0xdbff, 0xdfff, 0x0062 };
 114     UChar32 c;
 115     int32_t i;
 116     UBool isError;
 117
 118     printf("\n* demo_utf_h_macros() -------------- ***\n\n");
 119
 120     printUString("iterate forward through: ", input, UPRV_LENGTHOF(input));
 121     for(i=0; i<UPRV_LENGTHOF(input); /* U16_NEXT post-increments */) {
 122         /* Iterating forwards
 123            Codepoint at offset 0: U+0061
 124            Codepoint at offset 1: U+10000
 125            Codepoint at offset 3: U+10ffff
 126            Codepoint at offset 5: U+0062
 127         */
 128         printf("Codepoint at offset %d: U+", i);
 129         U16_NEXT(input, i, UPRV_LENGTHOF(input), c);
 130         printf("%04x\n", c);
 131     }
 132
 133     puts("");
 134
 135     isError=FALSE;
 136     i=1; /* write position, gets post-incremented so needs to be in an l-value */
 137     U16_APPEND(input, i, UPRV_LENGTHOF(input), 0x0062, isError);
 138
 139     printUString("iterate backward through: ", input, UPRV_LENGTHOF(input));
 140     for(i=UPRV_LENGTHOF(input); i>0; /* U16_PREV pre-decrements */) {
 141         U16_PREV(input, 0, i, c);
 142         /* Iterating backwards
 143            Codepoint at offset 5: U+0062
 144            Codepoint at offset 3: U+10ffff
 145            Codepoint at offset 2: U+dc00 -- unpaired surrogate because lead surr. overwritten
 146            Codepoint at offset 1: U+0062 -- by this BMP code point
 147            Codepoint at offset 0: U+0061
 148         */
 149         printf("Codepoint at offset %d: U+%04x\n", i, c);
 150     }
 151 }
 152
 153 // sample code for Unicode strings in C ------------------------------------ ***
 154
 155 static void demo_C_Unicode_strings() {
 156     printf("\n* demo_C_Unicode_strings() --------- ***\n\n");
 157
 158     static const UChar text[]={ 0x41, 0x42, 0x43, 0 };          /* "ABC" */
 159     static const UChar appendText[]={ 0x61, 0x62, 0x63, 0 };    /* "abc" */
 160     static const UChar cmpText[]={ 0x61, 0x53, 0x73, 0x43, 0 }; /* "aSsC" */
 161     UChar buffer[32];
 162     int32_t compare;
 163     int32_t length=u_strlen(text); /* length=3 */
 164
 165     /* simple ANSI C-style functions */
 166     buffer[0]=0;                    /* empty, NUL-terminated string */
 167     u_strncat(buffer, text, 1);     /* append just n=1 character ('A') */
 168     u_strcat(buffer, appendText);   /* buffer=="Aabc" */
 169     length=u_strlen(buffer);        /* length=4 */
 170     printUString("should be \"Aabc\": ", buffer, -1);
 171
 172     /* bitwise comparing buffer with text */
 173     compare=u_strcmp(buffer, text);
 174     if(compare<=0) {
 175         printf("String comparison error, expected \"Aabc\" > \"ABC\"\n");
 176     }
 177
 178     /* Build "A<sharp s>C" in the buffer... */
 179     u_strcpy(buffer, text);
 180     buffer[1]=0xdf; /* sharp s, case-compares equal to "ss" */
 181     printUString("should be \"A<sharp s>C\": ", buffer, -1);
 182
 183     /* Compare two strings case-insensitively using full case folding */
 184     compare=u_strcasecmp(buffer, cmpText, U_FOLD_CASE_DEFAULT);
 185     if(compare!=0) {
 186         printf("String case insensitive comparison error, expected \"AbC\" to be equal to \"ABC\"\n");
 187     }
 188 }
 189
 190 // sample code for case mappings with C APIs -------------------------------- ***
 191
 192 static void demoCaseMapInC() {
 193     /*
 194      * input=
 195      *   "aB<capital sigma>"
 196      *   "iI<small dotless i><capital dotted I> "
 197      *   "<sharp s> <small lig. ffi>"
 198      *   "<small final sigma><small sigma><capital sigma>"
 199      */
 200     static const UChar input[]={
 201         0x61, 0x42, 0x3a3,
 202         0x69, 0x49, 0x131, 0x130, 0x20,
 203         0xdf, 0x20, 0xfb03,
 204         0x3c2, 0x3c3, 0x3a3, 0
 205     };
 206     UChar buffer[32];
 207
 208     UErrorCode errorCode;
 209     UChar32 c;
 210     int32_t i, j, length;
 211     UBool isError;
 212
 213     printf("\n* demoCaseMapInC() ----------------- ***\n\n");
 214
 215     /*
 216      * First, use simple case mapping functions which provide
 217      * 1:1 code point mappings without context/locale ID.
 218      *
 219      * Note that some mappings will not be "right" because some "real"
 220      * case mappings require context, depend on the locale ID,
 221      * and/or result in a change in the number of code points.
 222      */
 223     printUString("input string: ", input, -1);
 224
 225     /* uppercase */
 226     isError=FALSE;
 227     for(i=j=0; j<UPRV_LENGTHOF(buffer) && !isError; /* U16_NEXT post-increments */) {
 228         U16_NEXT(input, i, INT32_MAX, c); /* without length because NUL-terminated */
 229         if(c==0) {
 230             break; /* stop at terminating NUL, no need to terminate buffer */
 231         }
 232         c=u_toupper(c);
 233         U16_APPEND(buffer, j, UPRV_LENGTHOF(buffer), c, isError);
 234     }
 235     printUString("simple-uppercased: ", buffer, j);
 236     /* lowercase */
 237     isError=FALSE;
 238     for(i=j=0; j<UPRV_LENGTHOF(buffer) && !isError; /* U16_NEXT post-increments */) {
 239         U16_NEXT(input, i, INT32_MAX, c); /* without length because NUL-terminated */
 240         if(c==0) {
 241             break; /* stop at terminating NUL, no need to terminate buffer */
 242         }
 243         c=u_tolower(c);
 244         U16_APPEND(buffer, j, UPRV_LENGTHOF(buffer), c, isError);
 245     }
 246     printUString("simple-lowercased: ", buffer, j);
 247     /* titlecase */
 248     isError=FALSE;
 249     for(i=j=0; j<UPRV_LENGTHOF(buffer) && !isError; /* U16_NEXT post-increments */) {
 250         U16_NEXT(input, i, INT32_MAX, c); /* without length because NUL-terminated */
 251         if(c==0) {
 252             break; /* stop at terminating NUL, no need to terminate buffer */
 253         }
 254         c=u_totitle(c);
 255         U16_APPEND(buffer, j, UPRV_LENGTHOF(buffer), c, isError);
 256     }
 257     printUString("simple-titlecased: ", buffer, j);
 258     /* case-fold/default */
 259     isError=FALSE;
 260     for(i=j=0; j<UPRV_LENGTHOF(buffer) && !isError; /* U16_NEXT post-increments */) {
 261         U16_NEXT(input, i, INT32_MAX, c); /* without length because NUL-terminated */
 262         if(c==0) {
 263             break; /* stop at terminating NUL, no need to terminate buffer */
 264         }
 265         c=u_foldCase(c, U_FOLD_CASE_DEFAULT);
 266         U16_APPEND(buffer, j, UPRV_LENGTHOF(buffer), c, isError);
 267     }
 268     printUString("simple-case-folded/default: ", buffer, j);
 269     /* case-fold/Turkic */
 270     isError=FALSE;
 271     for(i=j=0; j<UPRV_LENGTHOF(buffer) && !isError; /* U16_NEXT post-increments */) {
 272         U16_NEXT(input, i, INT32_MAX, c); /* without length because NUL-terminated */
 273         if(c==0) {
 274             break; /* stop at terminating NUL, no need to terminate buffer */
 275         }
 276         c=u_foldCase(c, U_FOLD_CASE_EXCLUDE_SPECIAL_I);
 277         U16_APPEND(buffer, j, UPRV_LENGTHOF(buffer), c, isError);
 278     }
 279     printUString("simple-case-folded/Turkic: ", buffer, j);
 280
 281     /*
 282      * Second, use full case mapping functions which provide
 283      * 1:n code point mappings (n can be 0!) and are sensitive to context and locale ID.
 284      *
 285      * Note that lower/upper/titlecasing take a locale ID while case-folding
 286      * has bit flag options instead, by design of the Unicode SpecialCasing.txt UCD file.
 287      *
 288      * Also, string titlecasing requires a BreakIterator to find starts of words.
 289      * The sample code here passes in a NULL pointer; u_strToTitle() will open and close a default
 290      * titlecasing BreakIterator automatically.
 291      * For production code where many strings are titlecased it would be more efficient
 292      * to open a BreakIterator externally and pass it in.
 293      */
 294     printUString("\ninput string: ", input, -1);
 295
 296     /* lowercase/English */
 297     errorCode=U_ZERO_ERROR;
 298     length=u_strToLower(buffer, UPRV_LENGTHOF(buffer), input, -1, "en", &errorCode);
 299     if(U_SUCCESS(errorCode)) {
 300         printUString("full-lowercased/en: ", buffer, length);
 301     } else {
 302         printf("error in u_strToLower(en)=%ld error=%s\n", length, u_errorName(errorCode));
 303     }
 304     /* lowercase/Turkish */
 305     errorCode=U_ZERO_ERROR;
 306     length=u_strToLower(buffer, UPRV_LENGTHOF(buffer), input, -1, "tr", &errorCode);
 307     if(U_SUCCESS(errorCode)) {
 308         printUString("full-lowercased/tr: ", buffer, length);
 309     } else {
 310         printf("error in u_strToLower(tr)=%ld error=%s\n", length, u_errorName(errorCode));
 311     }
 312     /* uppercase/English */
 313     errorCode=U_ZERO_ERROR;
 314     length=u_strToUpper(buffer, UPRV_LENGTHOF(buffer), input, -1, "en", &errorCode);
 315     if(U_SUCCESS(errorCode)) {
 316         printUString("full-uppercased/en: ", buffer, length);
 317     } else {
 318         printf("error in u_strToUpper(en)=%ld error=%s\n", length, u_errorName(errorCode));
 319     }
 320     /* uppercase/Turkish */
 321     errorCode=U_ZERO_ERROR;
 322     length=u_strToUpper(buffer, UPRV_LENGTHOF(buffer), input, -1, "tr", &errorCode);
 323     if(U_SUCCESS(errorCode)) {
 324         printUString("full-uppercased/tr: ", buffer, length);
 325     } else {
 326         printf("error in u_strToUpper(tr)=%ld error=%s\n", length, u_errorName(errorCode));
 327     }
 328     /* titlecase/English */
 329     errorCode=U_ZERO_ERROR;
 330     length=u_strToTitle(buffer, UPRV_LENGTHOF(buffer), input, -1, NULL, "en", &errorCode);
 331     if(U_SUCCESS(errorCode)) {
 332         printUString("full-titlecased/en: ", buffer, length);
 333     } else {
 334         printf("error in u_strToTitle(en)=%ld error=%s\n", length, u_errorName(errorCode));
 335     }
 336     /* titlecase/Turkish */
 337     errorCode=U_ZERO_ERROR;
 338     length=u_strToTitle(buffer, UPRV_LENGTHOF(buffer), input, -1, NULL, "tr", &errorCode);
 339     if(U_SUCCESS(errorCode)) {
 340         printUString("full-titlecased/tr: ", buffer, length);
 341     } else {
 342         printf("error in u_strToTitle(tr)=%ld error=%s\n", length, u_errorName(errorCode));
 343     }
 344     /* case-fold/default */
 345     errorCode=U_ZERO_ERROR;
 346     length=u_strFoldCase(buffer, UPRV_LENGTHOF(buffer), input, -1, U_FOLD_CASE_DEFAULT, &errorCode);
 347     if(U_SUCCESS(errorCode)) {
 348         printUString("full-case-folded/default: ", buffer, length);
 349     } else {
 350         printf("error in u_strFoldCase(default)=%ld error=%s\n", length, u_errorName(errorCode));
 351     }
 352     /* case-fold/Turkic */
 353     errorCode=U_ZERO_ERROR;
 354     length=u_strFoldCase(buffer, UPRV_LENGTHOF(buffer), input, -1, U_FOLD_CASE_EXCLUDE_SPECIAL_I, &errorCode);
 355     if(U_SUCCESS(errorCode)) {
 356         printUString("full-case-folded/Turkic: ", buffer, length);
 357     } else {
 358         printf("error in u_strFoldCase(Turkic)=%ld error=%s\n", length, u_errorName(errorCode));
 359     }
 360 }
 361
 362 // sample code for case mappings with C++ APIs ------------------------------ ***
 363
 364 static void demoCaseMapInCPlusPlus() {
 365     /*
 366      * input=
 367      *   "aB<capital sigma>"
 368      *   "iI<small dotless i><capital dotted I> "
 369      *   "<sharp s> <small lig. ffi>"
 370      *   "<small final sigma><small sigma><capital sigma>"
 371      */
 372     static const UChar input[]={
 373         0x61, 0x42, 0x3a3,
 374         0x69, 0x49, 0x131, 0x130, 0x20,
 375         0xdf, 0x20, 0xfb03,
 376         0x3c2, 0x3c3, 0x3a3, 0
 377     };
 378
 379     printf("\n* demoCaseMapInCPlusPlus() --------- ***\n\n");
 380
 381     UnicodeString s(input), t;
 382     const Locale &en=Locale::getEnglish();
 383     Locale tr("tr");
 384
 385     /*
 386      * Full case mappings as in demoCaseMapInC(), using UnicodeString functions.
 387      * These functions modify the string object itself.
 388      * Since we want to keep the input string around, we copy it each time
 389      * and case-map the copy.
 390      */
 391     printUnicodeString("input string: ", s);
 392
 393     /* lowercase/English */
 394     printUnicodeString("full-lowercased/en: ", (t=s).toLower(en));
 395     /* lowercase/Turkish */
 396     printUnicodeString("full-lowercased/tr: ", (t=s).toLower(tr));
 397     /* uppercase/English */
 398     printUnicodeString("full-uppercased/en: ", (t=s).toUpper(en));
 399     /* uppercase/Turkish */
 400     printUnicodeString("full-uppercased/tr: ", (t=s).toUpper(tr));
 401     /* titlecase/English */
 402     printUnicodeString("full-titlecased/en: ", (t=s).toTitle(NULL, en));
 403     /* titlecase/Turkish */
 404     printUnicodeString("full-titlecased/tr: ", (t=s).toTitle(NULL, tr));
 405     /* case-folde/default */
 406     printUnicodeString("full-case-folded/default: ", (t=s).foldCase(U_FOLD_CASE_DEFAULT));
 407     /* case-folde/Turkic */
 408     printUnicodeString("full-case-folded/Turkic: ", (t=s).foldCase(U_FOLD_CASE_EXCLUDE_SPECIAL_I));
 409 }
 410
 411 // sample code for UnicodeString storage models ----------------------------- ***
 412
 413 static const UChar readonly[]={
 414     0x61, 0x31, 0x20ac
 415 };
 416 static UChar writeable[]={
 417     0x62, 0x32, 0xdbc0, 0xdc01 // includes a surrogate pair for a supplementary code point
 418 };
 419 static char out[100];
 420
 421 static void
 422 demoUnicodeStringStorage() {
 423     // These sample code lines illustrate how to use UnicodeString, and the
 424     // comments tell what happens internally. There are no APIs to observe
 425     // most of this programmatically, except for stepping into the code
 426     // with a debugger.
 427     // This is by design to hide such details from the user.
 428     int32_t i;
 429
 430     printf("\n* demoUnicodeStringStorage() ------- ***\n\n");
 431
 432     // * UnicodeString with internally stored contents
 433     // instantiate a UnicodeString from a single code point
 434     // the few (2) UChars will be stored in the object itself
 435     UnicodeString one((UChar32)0x24001);
 436     // this copies the few UChars into the "two" object
 437     UnicodeString two=one;
 438     printf("length of short string copy: %d\n", two.length());
 439     // set "one" to contain the 3 UChars from readonly
 440     // this setTo() variant copies the characters
 441     one.setTo(readonly, UPRV_LENGTHOF(readonly));
 442
 443     // * UnicodeString with allocated contents
 444     // build a longer string that will not fit into the object's buffer
 445     one+=UnicodeString(writeable, UPRV_LENGTHOF(writeable));
 446     one+=one;
 447     one+=one;
 448     printf("length of longer string: %d\n", one.length());
 449     // copying will use the same allocated buffer and increment the reference
 450     // counter
 451     two=one;
 452     printf("length of longer string copy: %d\n", two.length());
 453
 454     // * UnicodeString using readonly-alias to a const UChar array
 455     // construct a string that aliases a readonly buffer
 456     UnicodeString three(FALSE, readonly, UPRV_LENGTHOF(readonly));
 457     printUnicodeString("readonly-alias string: ", three);
 458     // copy-on-write: any modification to the string results in
 459     // a copy to either the internal buffer or to a newly allocated one
 460     three.setCharAt(1, 0x39);
 461     printUnicodeString("readonly-aliasing string after modification: ", three);
 462     // the aliased array is not modified
 463     for(i=0; i<three.length(); ++i) {
 464         printf("readonly buffer[%d] after modifying its string: 0x%lx\n",
 465                i, readonly[i]);
 466     }
 467     // setTo() readonly alias
 468     one.setTo(FALSE, writeable, UPRV_LENGTHOF(writeable));
 469     // copying the readonly-alias object with fastCopyFrom() (new in ICU 2.4)
 470     // will readonly-alias the same buffer
 471     two.fastCopyFrom(one);
 472     printUnicodeString("fastCopyFrom(readonly alias of \"writeable\" array): ", two);
 473     printf("verify that a fastCopyFrom(readonly alias) uses the same buffer pointer: %d (should be 1)\n",
 474         one.getBuffer()==two.getBuffer());
 475     // a normal assignment will clone the contents (new in ICU 2.4)
 476     two=one;
 477     printf("verify that a regular copy of a readonly alias uses a different buffer pointer: %d (should be 0)\n",
 478         one.getBuffer()==two.getBuffer());
 479
 480     // * UnicodeString using writeable-alias to a non-const UChar array
 481     UnicodeString four(writeable, UPRV_LENGTHOF(writeable), UPRV_LENGTHOF(writeable));
 482     printUnicodeString("writeable-alias string: ", four);
 483     // a modification writes through to the buffer
 484     four.setCharAt(1, 0x39);
 485     for(i=0; i<four.length(); ++i) {
 486         printf("writeable-alias backing buffer[%d]=0x%lx "
 487                "after modification\n", i, writeable[i]);
 488     }
 489     // a copy will not alias any more;
 490     // instead, it will get a copy of the contents into allocated memory
 491     two=four;
 492     two.setCharAt(1, 0x21);
 493     for(i=0; i<two.length(); ++i) {
 494         printf("writeable-alias backing buffer[%d]=0x%lx after "
 495                "modification of string copy\n", i, writeable[i]);
 496     }
 497     // setTo() writeable alias, capacity==length
 498     one.setTo(writeable, UPRV_LENGTHOF(writeable), UPRV_LENGTHOF(writeable));
 499     // grow the string - it will not fit into the backing buffer any more
 500     // and will get copied before modification
 501     one.append((UChar)0x40);
 502     // shrink it back so it would fit
 503     one.truncate(one.length()-1);
 504     // we still operate on the copy
 505     one.setCharAt(1, 0x25);
 506     printf("string after growing too much and then shrinking[1]=0x%lx\n"
 507            "                          backing store for this[1]=0x%lx\n",
 508            one.charAt(1), writeable[1]);
 509     // if we need it in the original buffer, then extract() to it
 510     // extract() does not do anything if the string aliases that same buffer
 511     // i=min(one.length(), length of array)
 512     if(one.length()<UPRV_LENGTHOF(writeable)) {
 513         i=one.length();
 514     } else {
 515         i=UPRV_LENGTHOF(writeable);
 516     }
 517     one.extract(0, i, writeable);
 518     for(i=0; i<UPRV_LENGTHOF(writeable); ++i) {
 519         printf("writeable-alias backing buffer[%d]=0x%lx after re-extract\n",
 520                i, writeable[i]);
 521     }
 522 }
 523
 524 // sample code for UnicodeString instantiations ----------------------------- ***
 525
 526 static void
 527 demoUnicodeStringInit() {
 528     // *** Make sure to read about invariant characters in utypes.h! ***
 529     // Initialization of Unicode strings from C literals works _only_ for
 530     // invariant characters!
 531
 532     printf("\n* demoUnicodeStringInit() ---------- ***\n\n");
 533
 534     // the string literal is 32 chars long - this must be counted for the macro
 535     UnicodeString invariantOnly=UNICODE_STRING("such characters are safe 123 %-.", 32);
 536
 537     /*
 538      * In C, we need two macros: one to declare the UChar[] array, and
 539      * one to populate it; the second one is a noop on platforms where
 540      * wchar_t is compatible with UChar and ASCII-based.
 541      * The length of the string literal must be counted for both macros.
 542      */
 543     /* declare the invString array for the string */
 544     U_STRING_DECL(invString, "such characters are safe 123 %-.", 32);
 545     /* populate it with the characters */
 546     U_STRING_INIT(invString, "such characters are safe 123 %-.", 32);
 547
 548     // compare the C and C++ strings
 549     printf("C and C++ Unicode strings are equal: %d\n", invariantOnly==UnicodeString(TRUE, invString, 32));
 550
 551     /*
 552      * convert between char * and UChar * strings that
 553      * contain only invariant characters
 554      */
 555     static const char *cs1="such characters are safe 123 %-.";
 556     static UChar us1[40];
 557     static char cs2[40];
 558     u_charsToUChars(cs1, us1, 33); /* include the terminating NUL */
 559     u_UCharsToChars(us1, cs2, 33);
 560     printf("char * -> UChar * -> char * with only "
 561            "invariant characters: \"%s\"\n",
 562            cs2);
 563
 564     // initialize a UnicodeString from a string literal that contains
 565     // escape sequences written with invariant characters
 566     // do not forget to duplicate the backslashes for ICU to see them
 567     // then, count each double backslash only once!
 568     UnicodeString german=UNICODE_STRING(
 569         "Sch\\u00f6nes Auto: \\u20ac 11240.\\fPrivates Zeichen: \\U00102345\\n", 64).
 570         unescape();
 571     printUnicodeString("german UnicodeString from unescaping:\n    ", german);
 572
 573     /*
 574      * C: convert and unescape a char * string with only invariant
 575      * characters to fill a UChar * string
 576      */
 577     UChar buffer[200];
 578     int32_t length;
 579     length=u_unescape(
 580         "Sch\\u00f6nes Auto: \\u20ac 11240.\\fPrivates Zeichen: \\U00102345\\n",
 581         buffer, UPRV_LENGTHOF(buffer));
 582     printf("german C Unicode string from char * unescaping: (length %d)\n    ", length);
 583     printUnicodeString("", UnicodeString(buffer));
 584 }
 585
 586 extern int
 587 main(int argc, const char *argv[]) {
 588     UErrorCode errorCode=U_ZERO_ERROR;
 589
 590     // Note: Using a global variable for any object is not exactly thread-safe...
 591
 592     // You can change this call to e.g. ucnv_open("UTF-8", &errorCode) if you pipe
 593     // the output to a file and look at it with a Unicode-capable editor.
 594     // This will currently affect only the printUString() function, see the code above.
 595     // printUnicodeString() could use this, too, by changing to an extract() overload
 596     // that takes a UConverter argument.
 597     cnv=ucnv_open(NULL, &errorCode);
 598     if(U_FAILURE(errorCode)) {
 599         fprintf(stderr, "error %s opening the default converter\n", u_errorName(errorCode));
 600         return errorCode;
 601     }
 602
 603     ucnv_setFromUCallBack(cnv, UCNV_FROM_U_CALLBACK_ESCAPE, UCNV_ESCAPE_C, NULL, NULL, &errorCode);
 604     if(U_FAILURE(errorCode)) {
 605         fprintf(stderr, "error %s setting the escape callback in the default converter\n", u_errorName(errorCode));
 606         ucnv_close(cnv);
 607         return errorCode;
 608     }
 609
 610     demo_utf_h_macros();
 611     demo_C_Unicode_strings();
 612     demoCaseMapInC();
 613     demoCaseMapInCPlusPlus();
 614     demoUnicodeStringStorage();
 615     demoUnicodeStringInit();
 616
 617     ucnv_close(cnv);
 618     return 0;
 619 }