icuSources/samples/ustring/ustring.cpp

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