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 #define __STDC_FORMAT_MACROS 1
  27 #include <inttypes.h>
  28
  29 #include <stdio.h>
  30 #include "unicode/utypes.h"
  31 #include "unicode/uchar.h"
  32 #include "unicode/locid.h"
  33 #include "unicode/ustring.h"
  34 #include "unicode/ucnv.h"
  35 #include "unicode/unistr.h"
  36
  37 using namespace icu;
  38
  39 #ifndef UPRV_LENGTHOF
  40 #define UPRV_LENGTHOF(array) (int32_t)(sizeof(array)/sizeof((array)[0]))
  41 #endif
  42
  43 // helper functions -------------------------------------------------------- ***
  44
  45 // default converter for the platform encoding
  46 static UConverter *cnv=NULL;
  47
  48 static void
  49 printUString(const char *announce, const UChar *s, int32_t length) {
  50     static char out[200];
  51     UChar32 c;
  52     int32_t i;
  53     UErrorCode errorCode=U_ZERO_ERROR;
  54
  55     /*
  56      * Convert to the "platform encoding". See notes in printUnicodeString().
  57      * ucnv_fromUChars(), like most ICU APIs understands length==-1
  58      * to mean that the string is NUL-terminated.
  59      */
  60     ucnv_fromUChars(cnv, out, sizeof(out), s, length, &errorCode);
  61     if(U_FAILURE(errorCode) || errorCode==U_STRING_NOT_TERMINATED_WARNING) {
  62         printf("%sproblem converting string from Unicode: %s\n", announce, u_errorName(errorCode));
  63         return;
  64     }
  65
  66     printf("%s%s {", announce, out);
  67
  68     /* output the code points (not code units) */
  69     if(length>=0) {
  70         /* s is not NUL-terminated */
  71         for(i=0; i<length; /* U16_NEXT post-increments */) {
  72             U16_NEXT(s, i, length, c);
  73             printf(" %04x", c);
  74         }
  75     } else {
  76         /* s is NUL-terminated */
  77         for(i=0; /* condition in loop body */; /* U16_NEXT post-increments */) {
  78             U16_NEXT(s, i, length, c);
  79             if(c==0) {
  80                 break;
  81             }
  82             printf(" %04x", c);
  83         }
  84     }
  85     printf(" }\n");
  86 }
  87
  88 static void
  89 printUnicodeString(const char *announce, const UnicodeString &s) {
  90     static char out[200];
  91     int32_t i, length;
  92
  93     // output the string, converted to the platform encoding
  94
  95     // Note for Windows: The "platform encoding" defaults to the "ANSI codepage",
  96     // which is different from the "OEM codepage" in the console window.
  97     // However, if you pipe the output into a file and look at it with Notepad
  98     // or similar, then "ANSI" characters will show correctly.
  99     // Production code should be aware of what encoding is required,
 100     // and use a UConverter or at least a charset name explicitly.
 101     out[s.extract(0, 99, out)]=0;
 102     printf("%s%s {", announce, out);
 103
 104     // output the code units (not code points)
 105     length=s.length();
 106     for(i=0; i<length; ++i) {
 107         printf(" %04x", s.charAt(i));
 108     }
 109     printf(" }\n");
 110 }
 111
 112 // sample code for utf.h macros -------------------------------------------- ***
 113
 114 static void
 115 demo_utf_h_macros() {
 116     static UChar input[]={ 0x0061, 0xd800, 0xdc00, 0xdbff, 0xdfff, 0x0062 };
 117     UChar32 c;
 118     int32_t i;
 119     UBool isError;
 120
 121     printf("\n* demo_utf_h_macros() -------------- ***\n\n");
 122
 123     printUString("iterate forward through: ", input, UPRV_LENGTHOF(input));
 124     for(i=0; i<UPRV_LENGTHOF(input); /* U16_NEXT post-increments */) {
 125         /* Iterating forwards
 126            Codepoint at offset 0: U+0061
 127            Codepoint at offset 1: U+10000
 128            Codepoint at offset 3: U+10ffff
 129            Codepoint at offset 5: U+0062
 130         */
 131         printf("Codepoint at offset %d: U+", i);
 132         U16_NEXT(input, i, UPRV_LENGTHOF(input), c);
 133         printf("%04x\n", c);
 134     }
 135
 136     puts("");
 137
 138     isError=FALSE;
 139     i=1; /* write position, gets post-incremented so needs to be in an l-value */
 140     U16_APPEND(input, i, UPRV_LENGTHOF(input), 0x0062, isError);
 141
 142     printUString("iterate backward through: ", input, UPRV_LENGTHOF(input));
 143     for(i=UPRV_LENGTHOF(input); i>0; /* U16_PREV pre-decrements */) {
 144         U16_PREV(input, 0, i, c);
 145         /* Iterating backwards
 146            Codepoint at offset 5: U+0062
 147            Codepoint at offset 3: U+10ffff
 148            Codepoint at offset 2: U+dc00 -- unpaired surrogate because lead surr. overwritten
 149            Codepoint at offset 1: U+0062 -- by this BMP code point
 150            Codepoint at offset 0: U+0061
 151         */
 152         printf("Codepoint at offset %d: U+%04x\n", i, c);
 153     }
 154 }
 155
 156 // sample code for Unicode strings in C ------------------------------------ ***
 157
 158 static void demo_C_Unicode_strings() {
 159     printf("\n* demo_C_Unicode_strings() --------- ***\n\n");
 160
 161     static const UChar text[]={ 0x41, 0x42, 0x43, 0 };          /* "ABC" */
 162     static const UChar appendText[]={ 0x61, 0x62, 0x63, 0 };    /* "abc" */
 163     static const UChar cmpText[]={ 0x61, 0x53, 0x73, 0x43, 0 }; /* "aSsC" */
 164     UChar buffer[32];
 165     int32_t compare;
 166     int32_t length=u_strlen(text); /* length=3 */
 167
 168     /* simple ANSI C-style functions */
 169     buffer[0]=0;                    /* empty, NUL-terminated string */
 170     u_strncat(buffer, text, 1);     /* append just n=1 character ('A') */
 171     u_strcat(buffer, appendText);   /* buffer=="Aabc" */
 172     length=u_strlen(buffer);        /* length=4 */
 173     printUString("should be \"Aabc\": ", buffer, -1);
 174
 175     /* bitwise comparing buffer with text */
 176     compare=u_strcmp(buffer, text);
 177     if(compare<=0) {
 178         printf("String comparison error, expected \"Aabc\" > \"ABC\"\n");
 179     }
 180
 181     /* Build "A<sharp s>C" in the buffer... */
 182     u_strcpy(buffer, text);
 183     buffer[1]=0xdf; /* sharp s, case-compares equal to "ss" */
 184     printUString("should be \"A<sharp s>C\": ", buffer, -1);
 185
 186     /* Compare two strings case-insensitively using full case folding */
 187     compare=u_strcasecmp(buffer, cmpText, U_FOLD_CASE_DEFAULT);
 188     if(compare!=0) {
 189         printf("String case insensitive comparison error, expected \"AbC\" to be equal to \"ABC\"\n");
 190     }
 191 }
 192
 193 // sample code for case mappings with C APIs -------------------------------- ***
 194
 195 static void demoCaseMapInC() {
 196     /*
 197      * input=
 198      *   "aB<capital sigma>"
 199      *   "iI<small dotless i><capital dotted I> "
 200      *   "<sharp s> <small lig. ffi>"
 201      *   "<small final sigma><small sigma><capital sigma>"
 202      */
 203     static const UChar input[]={
 204         0x61, 0x42, 0x3a3,
 205         0x69, 0x49, 0x131, 0x130, 0x20,
 206         0xdf, 0x20, 0xfb03,
 207         0x3c2, 0x3c3, 0x3a3, 0
 208     };
 209     UChar buffer[32];
 210
 211     UErrorCode errorCode;
 212     UChar32 c;
 213     int32_t i, j, length;
 214     UBool isError;
 215
 216     printf("\n* demoCaseMapInC() ----------------- ***\n\n");
 217
 218     /*
 219      * First, use simple case mapping functions which provide
 220      * 1:1 code point mappings without context/locale ID.
 221      *
 222      * Note that some mappings will not be "right" because some "real"
 223      * case mappings require context, depend on the locale ID,
 224      * and/or result in a change in the number of code points.
 225      */
 226     printUString("input string: ", input, -1);
 227
 228     /* uppercase */
 229     isError=FALSE;
 230     for(i=j=0; j<UPRV_LENGTHOF(buffer) && !isError; /* U16_NEXT post-increments */) {
 231         U16_NEXT(input, i, INT32_MAX, c); /* without length because NUL-terminated */
 232         if(c==0) {
 233             break; /* stop at terminating NUL, no need to terminate buffer */
 234         }
 235         c=u_toupper(c);
 236         U16_APPEND(buffer, j, UPRV_LENGTHOF(buffer), c, isError);
 237     }
 238     printUString("simple-uppercased: ", buffer, j);
 239     /* lowercase */
 240     isError=FALSE;
 241     for(i=j=0; j<UPRV_LENGTHOF(buffer) && !isError; /* U16_NEXT post-increments */) {
 242         U16_NEXT(input, i, INT32_MAX, c); /* without length because NUL-terminated */
 243         if(c==0) {
 244             break; /* stop at terminating NUL, no need to terminate buffer */
 245         }
 246         c=u_tolower(c);
 247         U16_APPEND(buffer, j, UPRV_LENGTHOF(buffer), c, isError);
 248     }
 249     printUString("simple-lowercased: ", buffer, j);
 250     /* titlecase */
 251     isError=FALSE;
 252     for(i=j=0; j<UPRV_LENGTHOF(buffer) && !isError; /* U16_NEXT post-increments */) {
 253         U16_NEXT(input, i, INT32_MAX, c); /* without length because NUL-terminated */
 254         if(c==0) {
 255             break; /* stop at terminating NUL, no need to terminate buffer */
 256         }
 257         c=u_totitle(c);
 258         U16_APPEND(buffer, j, UPRV_LENGTHOF(buffer), c, isError);
 259     }
 260     printUString("simple-titlecased: ", buffer, j);
 261     /* case-fold/default */
 262     isError=FALSE;
 263     for(i=j=0; j<UPRV_LENGTHOF(buffer) && !isError; /* U16_NEXT post-increments */) {
 264         U16_NEXT(input, i, INT32_MAX, c); /* without length because NUL-terminated */
 265         if(c==0) {
 266             break; /* stop at terminating NUL, no need to terminate buffer */
 267         }
 268         c=u_foldCase(c, U_FOLD_CASE_DEFAULT);
 269         U16_APPEND(buffer, j, UPRV_LENGTHOF(buffer), c, isError);
 270     }
 271     printUString("simple-case-folded/default: ", buffer, j);
 272     /* case-fold/Turkic */
 273     isError=FALSE;
 274     for(i=j=0; j<UPRV_LENGTHOF(buffer) && !isError; /* U16_NEXT post-increments */) {
 275         U16_NEXT(input, i, INT32_MAX, c); /* without length because NUL-terminated */
 276         if(c==0) {
 277             break; /* stop at terminating NUL, no need to terminate buffer */
 278         }
 279         c=u_foldCase(c, U_FOLD_CASE_EXCLUDE_SPECIAL_I);
 280         U16_APPEND(buffer, j, UPRV_LENGTHOF(buffer), c, isError);
 281     }
 282     printUString("simple-case-folded/Turkic: ", buffer, j);
 283
 284     /*
 285      * Second, use full case mapping functions which provide
 286      * 1:n code point mappings (n can be 0!) and are sensitive to context and locale ID.
 287      *
 288      * Note that lower/upper/titlecasing take a locale ID while case-folding
 289      * has bit flag options instead, by design of the Unicode SpecialCasing.txt UCD file.
 290      *
 291      * Also, string titlecasing requires a BreakIterator to find starts of words.
 292      * The sample code here passes in a NULL pointer; u_strToTitle() will open and close a default
 293      * titlecasing BreakIterator automatically.
 294      * For production code where many strings are titlecased it would be more efficient
 295      * to open a BreakIterator externally and pass it in.
 296      */
 297     printUString("\ninput string: ", input, -1);
 298
 299     /* lowercase/English */
 300     errorCode=U_ZERO_ERROR;
 301     length=u_strToLower(buffer, UPRV_LENGTHOF(buffer), input, -1, "en", &errorCode);
 302     if(U_SUCCESS(errorCode)) {
 303         printUString("full-lowercased/en: ", buffer, length);
 304     } else {
 305         printf("error in u_strToLower(en)=%" PRId32 " error=%s\n", length, u_errorName(errorCode));
 306     }
 307     /* lowercase/Turkish */
 308     errorCode=U_ZERO_ERROR;
 309     length=u_strToLower(buffer, UPRV_LENGTHOF(buffer), input, -1, "tr", &errorCode);
 310     if(U_SUCCESS(errorCode)) {
 311         printUString("full-lowercased/tr: ", buffer, length);
 312     } else {
 313         printf("error in u_strToLower(tr)=%" PRId32 " error=%s\n", length, u_errorName(errorCode));
 314     }
 315     /* uppercase/English */
 316     errorCode=U_ZERO_ERROR;
 317     length=u_strToUpper(buffer, UPRV_LENGTHOF(buffer), input, -1, "en", &errorCode);
 318     if(U_SUCCESS(errorCode)) {
 319         printUString("full-uppercased/en: ", buffer, length);
 320     } else {
 321         printf("error in u_strToUpper(en)=%" PRId32 " error=%s\n", length, u_errorName(errorCode));
 322     }
 323     /* uppercase/Turkish */
 324     errorCode=U_ZERO_ERROR;
 325     length=u_strToUpper(buffer, UPRV_LENGTHOF(buffer), input, -1, "tr", &errorCode);
 326     if(U_SUCCESS(errorCode)) {
 327         printUString("full-uppercased/tr: ", buffer, length);
 328     } else {
 329         printf("error in u_strToUpper(tr)=%" PRId32 " error=%s\n", length, u_errorName(errorCode));
 330     }
 331     /* titlecase/English */
 332     errorCode=U_ZERO_ERROR;
 333     length=u_strToTitle(buffer, UPRV_LENGTHOF(buffer), input, -1, NULL, "en", &errorCode);
 334     if(U_SUCCESS(errorCode)) {
 335         printUString("full-titlecased/en: ", buffer, length);
 336     } else {
 337         printf("error in u_strToTitle(en)=%" PRId32 " error=%s\n", length, u_errorName(errorCode));
 338     }
 339     /* titlecase/Turkish */
 340     errorCode=U_ZERO_ERROR;
 341     length=u_strToTitle(buffer, UPRV_LENGTHOF(buffer), input, -1, NULL, "tr", &errorCode);
 342     if(U_SUCCESS(errorCode)) {
 343         printUString("full-titlecased/tr: ", buffer, length);
 344     } else {
 345         printf("error in u_strToTitle(tr)=%" PRId32 " error=%s\n", length, u_errorName(errorCode));
 346     }
 347     /* case-fold/default */
 348     errorCode=U_ZERO_ERROR;
 349     length=u_strFoldCase(buffer, UPRV_LENGTHOF(buffer), input, -1, U_FOLD_CASE_DEFAULT, &errorCode);
 350     if(U_SUCCESS(errorCode)) {
 351         printUString("full-case-folded/default: ", buffer, length);
 352     } else {
 353         printf("error in u_strFoldCase(default)=%" PRId32 " error=%s\n", length, u_errorName(errorCode));
 354     }
 355     /* case-fold/Turkic */
 356     errorCode=U_ZERO_ERROR;
 357     length=u_strFoldCase(buffer, UPRV_LENGTHOF(buffer), input, -1, U_FOLD_CASE_EXCLUDE_SPECIAL_I, &errorCode);
 358     if(U_SUCCESS(errorCode)) {
 359         printUString("full-case-folded/Turkic: ", buffer, length);
 360     } else {
 361         printf("error in u_strFoldCase(Turkic)=%" PRId32 " error=%s\n", length, u_errorName(errorCode));
 362     }
 363 }
 364
 365 // sample code for case mappings with C++ APIs ------------------------------ ***
 366
 367 static void demoCaseMapInCPlusPlus() {
 368     /*
 369      * input=
 370      *   "aB<capital sigma>"
 371      *   "iI<small dotless i><capital dotted I> "
 372      *   "<sharp s> <small lig. ffi>"
 373      *   "<small final sigma><small sigma><capital sigma>"
 374      */
 375     static const UChar input[]={
 376         0x61, 0x42, 0x3a3,
 377         0x69, 0x49, 0x131, 0x130, 0x20,
 378         0xdf, 0x20, 0xfb03,
 379         0x3c2, 0x3c3, 0x3a3, 0
 380     };
 381
 382     printf("\n* demoCaseMapInCPlusPlus() --------- ***\n\n");
 383
 384     UnicodeString s(input), t;
 385     const Locale &en=Locale::getEnglish();
 386     Locale tr("tr");
 387
 388     /*
 389      * Full case mappings as in demoCaseMapInC(), using UnicodeString functions.
 390      * These functions modify the string object itself.
 391      * Since we want to keep the input string around, we copy it each time
 392      * and case-map the copy.
 393      */
 394     printUnicodeString("input string: ", s);
 395
 396     /* lowercase/English */
 397     printUnicodeString("full-lowercased/en: ", (t=s).toLower(en));
 398     /* lowercase/Turkish */
 399     printUnicodeString("full-lowercased/tr: ", (t=s).toLower(tr));
 400     /* uppercase/English */
 401     printUnicodeString("full-uppercased/en: ", (t=s).toUpper(en));
 402     /* uppercase/Turkish */
 403     printUnicodeString("full-uppercased/tr: ", (t=s).toUpper(tr));
 404     /* titlecase/English */
 405     printUnicodeString("full-titlecased/en: ", (t=s).toTitle(NULL, en));
 406     /* titlecase/Turkish */
 407     printUnicodeString("full-titlecased/tr: ", (t=s).toTitle(NULL, tr));
 408     /* case-folde/default */
 409     printUnicodeString("full-case-folded/default: ", (t=s).foldCase(U_FOLD_CASE_DEFAULT));
 410     /* case-folde/Turkic */
 411     printUnicodeString("full-case-folded/Turkic: ", (t=s).foldCase(U_FOLD_CASE_EXCLUDE_SPECIAL_I));
 412 }
 413
 414 // sample code for UnicodeString storage models ----------------------------- ***
 415
 416 static const UChar readonly[]={
 417     0x61, 0x31, 0x20ac
 418 };
 419 static UChar writeable[]={
 420     0x62, 0x32, 0xdbc0, 0xdc01 // includes a surrogate pair for a supplementary code point
 421 };
 422 static char out[100];
 423
 424 static void
 425 demoUnicodeStringStorage() {
 426     // These sample code lines illustrate how to use UnicodeString, and the
 427     // comments tell what happens internally. There are no APIs to observe
 428     // most of this programmatically, except for stepping into the code
 429     // with a debugger.
 430     // This is by design to hide such details from the user.
 431     int32_t i;
 432
 433     printf("\n* demoUnicodeStringStorage() ------- ***\n\n");
 434
 435     // * UnicodeString with internally stored contents
 436     // instantiate a UnicodeString from a single code point
 437     // the few (2) UChars will be stored in the object itself
 438     UnicodeString one((UChar32)0x24001);
 439     // this copies the few UChars into the "two" object
 440     UnicodeString two=one;
 441     printf("length of short string copy: %d\n", two.length());
 442     // set "one" to contain the 3 UChars from readonly
 443     // this setTo() variant copies the characters
 444     one.setTo(readonly, UPRV_LENGTHOF(readonly));
 445
 446     // * UnicodeString with allocated contents
 447     // build a longer string that will not fit into the object's buffer
 448     one+=UnicodeString(writeable, UPRV_LENGTHOF(writeable));
 449     one+=one;
 450     one+=one;
 451     printf("length of longer string: %d\n", one.length());
 452     // copying will use the same allocated buffer and increment the reference
 453     // counter
 454     two=one;
 455     printf("length of longer string copy: %d\n", two.length());
 456
 457     // * UnicodeString using readonly-alias to a const UChar array
 458     // construct a string that aliases a readonly buffer
 459     UnicodeString three(FALSE, readonly, UPRV_LENGTHOF(readonly));
 460     printUnicodeString("readonly-alias string: ", three);
 461     // copy-on-write: any modification to the string results in
 462     // a copy to either the internal buffer or to a newly allocated one
 463     three.setCharAt(1, 0x39);
 464     printUnicodeString("readonly-aliasing string after modification: ", three);
 465     // the aliased array is not modified
 466     for(i=0; i<three.length(); ++i) {
 467         printf("readonly buffer[%d] after modifying its string: 0x%" PRId32 "\n",
 468                i, readonly[i]);
 469     }
 470     // setTo() readonly alias
 471     one.setTo(FALSE, writeable, UPRV_LENGTHOF(writeable));
 472     // copying the readonly-alias object with fastCopyFrom() (new in ICU 2.4)
 473     // will readonly-alias the same buffer
 474     two.fastCopyFrom(one);
 475     printUnicodeString("fastCopyFrom(readonly alias of \"writeable\" array): ", two);
 476     printf("verify that a fastCopyFrom(readonly alias) uses the same buffer pointer: %d (should be 1)\n",
 477         one.getBuffer()==two.getBuffer());
 478     // a normal assignment will clone the contents (new in ICU 2.4)
 479     two=one;
 480     printf("verify that a regular copy of a readonly alias uses a different buffer pointer: %d (should be 0)\n",
 481         one.getBuffer()==two.getBuffer());
 482
 483     // * UnicodeString using writeable-alias to a non-const UChar array
 484     UnicodeString four(writeable, UPRV_LENGTHOF(writeable), UPRV_LENGTHOF(writeable));
 485     printUnicodeString("writeable-alias string: ", four);
 486     // a modification writes through to the buffer
 487     four.setCharAt(1, 0x39);
 488     for(i=0; i<four.length(); ++i) {
 489         printf("writeable-alias backing buffer[%d]=0x%" PRId32 " "
 490                "after modification\n", i, writeable[i]);
 491     }
 492     // a copy will not alias any more;
 493     // instead, it will get a copy of the contents into allocated memory
 494     two=four;
 495     two.setCharAt(1, 0x21);
 496     for(i=0; i<two.length(); ++i) {
 497         printf("writeable-alias backing buffer[%d]=0x%" PRId32 " after "
 498                "modification of string copy\n", i, writeable[i]);
 499     }
 500     // setTo() writeable alias, capacity==length
 501     one.setTo(writeable, UPRV_LENGTHOF(writeable), UPRV_LENGTHOF(writeable));
 502     // grow the string - it will not fit into the backing buffer any more
 503     // and will get copied before modification
 504     one.append((UChar)0x40);
 505     // shrink it back so it would fit
 506     one.truncate(one.length()-1);
 507     // we still operate on the copy
 508     one.setCharAt(1, 0x25);
 509     printf("string after growing too much and then shrinking[1]=0x%" PRId32 "\n"
 510            "                          backing store for this[1]=0x%" PRId32 "\n",
 511            one.charAt(1), writeable[1]);
 512     // if we need it in the original buffer, then extract() to it
 513     // extract() does not do anything if the string aliases that same buffer
 514     // i=min(one.length(), length of array)
 515     if(one.length()<UPRV_LENGTHOF(writeable)) {
 516         i=one.length();
 517     } else {
 518         i=UPRV_LENGTHOF(writeable);
 519     }
 520     one.extract(0, i, writeable);
 521     for(i=0; i<UPRV_LENGTHOF(writeable); ++i) {
 522         printf("writeable-alias backing buffer[%d]=0x%" PRId32 " after re-extract\n",
 523                i, writeable[i]);
 524     }
 525 }
 526
 527 // sample code for UnicodeString instantiations ----------------------------- ***
 528
 529 static void
 530 demoUnicodeStringInit() {
 531     // *** Make sure to read about invariant characters in utypes.h! ***
 532     // Initialization of Unicode strings from C literals works _only_ for
 533     // invariant characters!
 534
 535     printf("\n* demoUnicodeStringInit() ---------- ***\n\n");
 536
 537     // the string literal is 32 chars long - this must be counted for the macro
 538     UnicodeString invariantOnly=UNICODE_STRING("such characters are safe 123 %-.", 32);
 539
 540     /*
 541      * In C, we need two macros: one to declare the UChar[] array, and
 542      * one to populate it; the second one is a noop on platforms where
 543      * wchar_t is compatible with UChar and ASCII-based.
 544      * The length of the string literal must be counted for both macros.
 545      */
 546     /* declare the invString array for the string */
 547     U_STRING_DECL(invString, "such characters are safe 123 %-.", 32);
 548     /* populate it with the characters */
 549     U_STRING_INIT(invString, "such characters are safe 123 %-.", 32);
 550
 551     // compare the C and C++ strings
 552     printf("C and C++ Unicode strings are equal: %d\n", invariantOnly==UnicodeString(TRUE, invString, 32));
 553
 554     /*
 555      * convert between char * and UChar * strings that
 556      * contain only invariant characters
 557      */
 558     static const char *cs1="such characters are safe 123 %-.";
 559     static UChar us1[40];
 560     static char cs2[40];
 561     u_charsToUChars(cs1, us1, 33); /* include the terminating NUL */
 562     u_UCharsToChars(us1, cs2, 33);
 563     printf("char * -> UChar * -> char * with only "
 564            "invariant characters: \"%s\"\n",
 565            cs2);
 566
 567     // initialize a UnicodeString from a string literal that contains
 568     // escape sequences written with invariant characters
 569     // do not forget to duplicate the backslashes for ICU to see them
 570     // then, count each double backslash only once!
 571     UnicodeString german=UNICODE_STRING(
 572         "Sch\\u00f6nes Auto: \\u20ac 11240.\\fPrivates Zeichen: \\U00102345\\n", 64).
 573         unescape();
 574     printUnicodeString("german UnicodeString from unescaping:\n    ", german);
 575
 576     /*
 577      * C: convert and unescape a char * string with only invariant
 578      * characters to fill a UChar * string
 579      */
 580     UChar buffer[200];
 581     int32_t length;
 582     length=u_unescape(
 583         "Sch\\u00f6nes Auto: \\u20ac 11240.\\fPrivates Zeichen: \\U00102345\\n",
 584         buffer, UPRV_LENGTHOF(buffer));
 585     printf("german C Unicode string from char * unescaping: (length %d)\n    ", length);
 586     printUnicodeString("", UnicodeString(buffer));
 587 }
 588
 589 extern int
 590 main(int argc, const char *argv[]) {
 591     UErrorCode errorCode=U_ZERO_ERROR;
 592
 593     // Note: Using a global variable for any object is not exactly thread-safe...
 594
 595     // You can change this call to e.g. ucnv_open("UTF-8", &errorCode) if you pipe
 596     // the output to a file and look at it with a Unicode-capable editor.
 597     // This will currently affect only the printUString() function, see the code above.
 598     // printUnicodeString() could use this, too, by changing to an extract() overload
 599     // that takes a UConverter argument.
 600     cnv=ucnv_open(NULL, &errorCode);
 601     if(U_FAILURE(errorCode)) {
 602         fprintf(stderr, "error %s opening the default converter\n", u_errorName(errorCode));
 603         return errorCode;
 604     }
 605
 606     ucnv_setFromUCallBack(cnv, UCNV_FROM_U_CALLBACK_ESCAPE, UCNV_ESCAPE_C, NULL, NULL, &errorCode);
 607     if(U_FAILURE(errorCode)) {
 608         fprintf(stderr, "error %s setting the escape callback in the default converter\n", u_errorName(errorCode));
 609         ucnv_close(cnv);
 610         return errorCode;
 611     }
 612
 613     demo_utf_h_macros();
 614     demo_C_Unicode_strings();
 615     demoCaseMapInC();
 616     demoCaseMapInCPlusPlus();
 617     demoUnicodeStringStorage();
 618     demoUnicodeStringInit();
 619
 620     ucnv_close(cnv);
 621     return 0;
 622 }