icuSources/samples/ustring/ustring.cpp

   1 /*
   2 *******************************************************************************
   3 *
   4 *   Copyright (C) 2000-2002, 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 #define LENGTHOF(array) (sizeof(array)/sizeof((array)[0]))
  29
  30 // helper functions -------------------------------------------------------- ***
  31
  32 // default converter for the platform encoding
  33 static UConverter *cnv=NULL;
  34
  35 static void
  36 printUString(const char *announce, const UChar *s, int32_t length) {
  37     static char out[200];
  38     UChar32 c;
  39     int32_t i;
  40     UErrorCode errorCode=U_ZERO_ERROR;
  41
  42     /*
  43      * Convert to the "platform encoding". See notes in printUnicodeString().
  44      * ucnv_fromUChars(), like most ICU APIs understands length==-1
  45      * to mean that the string is NUL-terminated.
  46      */
  47     ucnv_fromUChars(cnv, out, sizeof(out), s, length, &errorCode);
  48     if(U_FAILURE(errorCode) || errorCode==U_STRING_NOT_TERMINATED_WARNING) {
  49         printf("%sproblem converting string from Unicode: %s\n", announce, u_errorName(errorCode));
  50         return;
  51     }
  52
  53     printf("%s%s {", announce, out);
  54
  55     /* output the code points (not code units) */
  56     if(length>=0) {
  57         /* s is not NUL-terminated */
  58         for(i=0; i<length; /* U16_NEXT post-increments */) {
  59             U16_NEXT(s, i, length, c);
  60             printf(" %04x", c);
  61         }
  62     } else {
  63         /* s is NUL-terminated */
  64         for(i=0; /* condition in loop body */; /* U16_NEXT post-increments */) {
  65             U16_NEXT(s, i, length, c);
  66             if(c==0) {
  67                 break;
  68             }
  69             printf(" %04x", c);
  70         }
  71     }
  72     printf(" }\n");
  73 }
  74
  75 static void
  76 printUnicodeString(const char *announce, const UnicodeString &s) {
  77     static char out[200];
  78     int32_t i, length;
  79
  80     // output the string, converted to the platform encoding
  81
  82     // Note for Windows: The "platform encoding" defaults to the "ANSI codepage",
  83     // which is different from the "OEM codepage" in the console window.
  84     // However, if you pipe the output into a file and look at it with Notepad
  85     // or similar, then "ANSI" characters will show correctly.
  86     // Production code should be aware of what encoding is required,
  87     // and use a UConverter or at least a charset name explicitly.
  88     out[s.extract(0, 99, out)]=0;
  89     printf("%s%s {", announce, out);
  90
  91     // output the code units (not code points)
  92     length=s.length();
  93     for(i=0; i<length; ++i) {
  94         printf(" %04x", s.charAt(i));
  95     }
  96     printf(" }\n");
  97 }
  98
  99 // sample code for utf.h macros -------------------------------------------- ***
 100
 101 static void
 102 demo_utf_h_macros() {
 103     static UChar input[]={ 0x0061, 0xd800, 0xdc00, 0xdbff, 0xdfff, 0x0062 };
 104     UChar32 c;
 105     int32_t i;
 106     UBool isError;
 107
 108     printf("\n* demo_utf_h_macros() -------------- ***\n\n");
 109
 110     printUString("iterate forward through: ", input, LENGTHOF(input));
 111     for(i=0; i<LENGTHOF(input); /* U16_NEXT post-increments */) {
 112         /* Iterating forwards
 113            Codepoint at offset 0: U+0061
 114            Codepoint at offset 1: U+10000
 115            Codepoint at offset 3: U+10ffff
 116            Codepoint at offset 5: U+0062
 117         */
 118         printf("Codepoint at offset %d: U+", i);
 119         U16_NEXT(input, i, LENGTHOF(input), c);
 120         printf("%04x\n", c);
 121     }
 122
 123     puts("");
 124
 125     isError=FALSE;
 126     i=1; /* write position, gets post-incremented so needs to be in an l-value */
 127     U16_APPEND(input, i, LENGTHOF(input), 0x0062, isError);
 128
 129     printUString("iterate backward through: ", input, LENGTHOF(input));
 130     for(i=LENGTHOF(input); i>0; /* U16_PREV pre-decrements */) {
 131         U16_PREV(input, 0, i, c);
 132         /* Iterating backwards
 133            Codepoint at offset 5: U+0062
 134            Codepoint at offset 3: U+10ffff
 135            Codepoint at offset 2: U+dc00 -- unpaired surrogate because lead surr. overwritten
 136            Codepoint at offset 1: U+0062 -- by this BMP code point
 137            Codepoint at offset 0: U+0061
 138         */
 139         printf("Codepoint at offset %d: U+%04x\n", i, c);
 140     }
 141 }
 142
 143 // sample code for Unicode strings in C ------------------------------------ ***
 144
 145 static void demo_C_Unicode_strings() {
 146     printf("\n* demo_C_Unicode_strings() --------- ***\n\n");
 147
 148     static const UChar text[]={ 0x41, 0x42, 0x43, 0 };          /* "ABC" */
 149     static const UChar appendText[]={ 0x61, 0x62, 0x63, 0 };    /* "abc" */
 150     static const UChar cmpText[]={ 0x61, 0x53, 0x73, 0x43, 0 }; /* "aSsC" */
 151     UChar buffer[32];
 152     int32_t compare;
 153     int32_t length=u_strlen(text); /* length=3 */
 154
 155     /* simple ANSI C-style functions */
 156     buffer[0]=0;                    /* empty, NUL-terminated string */
 157     u_strncat(buffer, text, 1);     /* append just n=1 character ('A') */
 158     u_strcat(buffer, appendText);   /* buffer=="Aabc" */
 159     length=u_strlen(buffer);        /* length=4 */
 160     printUString("should be \"Aabc\": ", buffer, -1);
 161
 162     /* bitwise comparing buffer with text */
 163     compare=u_strcmp(buffer, text);
 164     if(compare<=0) {
 165         printf("String comparison error, expected \"Aabc\" > \"ABC\"\n");
 166     }
 167
 168     /* Build "A<sharp s>C" in the buffer... */
 169     u_strcpy(buffer, text);
 170     buffer[1]=0xdf; /* sharp s, case-compares equal to "ss" */
 171     printUString("should be \"A<sharp s>C\": ", buffer, -1);
 172
 173     /* Compare two strings case-insensitively using full case folding */
 174     compare=u_strcasecmp(buffer, cmpText, U_FOLD_CASE_DEFAULT);
 175     if(compare!=0) {
 176         printf("String case insensitive comparison error, expected \"AbC\" to be equal to \"ABC\"\n");
 177     }
 178 }
 179
 180 // sample code for case mappings with C APIs -------------------------------- ***
 181
 182 static void demoCaseMapInC() {
 183     /*
 184      * input=
 185      *   "aB<capital sigma>"
 186      *   "iI<small dotless i><capital dotted I> "
 187      *   "<sharp s> <small lig. ffi>"
 188      *   "<small final sigma><small sigma><capital sigma>"
 189      */
 190     static const UChar input[]={
 191         0x61, 0x42, 0x3a3,
 192         0x69, 0x49, 0x131, 0x130, 0x20,
 193         0xdf, 0x20, 0xfb03,
 194         0x3c2, 0x3c3, 0x3a3, 0
 195     };
 196     UChar buffer[32];
 197
 198     UErrorCode errorCode;
 199     UChar32 c;
 200     int32_t i, j, length;
 201     UBool isError;
 202
 203     printf("\n* demoCaseMapInC() ----------------- ***\n\n");
 204
 205     /*
 206      * First, use simple case mapping functions which provide
 207      * 1:1 code point mappings without context/locale ID.
 208      *
 209      * Note that some mappings will not be "right" because some "real"
 210      * case mappings require context, depend on the locale ID,
 211      * and/or result in a change in the number of code points.
 212      */
 213     printUString("input string: ", input, -1);
 214
 215     /* uppercase */
 216     isError=FALSE;
 217     for(i=j=0; j<LENGTHOF(buffer) && !isError; /* U16_NEXT post-increments */) {
 218         U16_NEXT(input, i, INT32_MAX, c); /* without length because NUL-terminated */
 219         if(c==0) {
 220             break; /* stop at terminating NUL, no need to terminate buffer */
 221         }
 222         c=u_toupper(c);
 223         U16_APPEND(buffer, j, LENGTHOF(buffer), c, isError);
 224     }
 225     printUString("simple-uppercased: ", buffer, j);
 226     /* lowercase */
 227     isError=FALSE;
 228     for(i=j=0; j<LENGTHOF(buffer) && !isError; /* U16_NEXT post-increments */) {
 229         U16_NEXT(input, i, INT32_MAX, c); /* without length because NUL-terminated */
 230         if(c==0) {
 231             break; /* stop at terminating NUL, no need to terminate buffer */
 232         }
 233         c=u_tolower(c);
 234         U16_APPEND(buffer, j, LENGTHOF(buffer), c, isError);
 235     }
 236     printUString("simple-lowercased: ", buffer, j);
 237     /* titlecase */
 238     isError=FALSE;
 239     for(i=j=0; j<LENGTHOF(buffer) && !isError; /* U16_NEXT post-increments */) {
 240         U16_NEXT(input, i, INT32_MAX, c); /* without length because NUL-terminated */
 241         if(c==0) {
 242             break; /* stop at terminating NUL, no need to terminate buffer */
 243         }
 244         c=u_totitle(c);
 245         U16_APPEND(buffer, j, LENGTHOF(buffer), c, isError);
 246     }
 247     printUString("simple-titlecased: ", buffer, j);
 248     /* case-fold/default */
 249     isError=FALSE;
 250     for(i=j=0; j<LENGTHOF(buffer) && !isError; /* U16_NEXT post-increments */) {
 251         U16_NEXT(input, i, INT32_MAX, c); /* without length because NUL-terminated */
 252         if(c==0) {
 253             break; /* stop at terminating NUL, no need to terminate buffer */
 254         }
 255         c=u_foldCase(c, U_FOLD_CASE_DEFAULT);
 256         U16_APPEND(buffer, j, LENGTHOF(buffer), c, isError);
 257     }
 258     printUString("simple-case-folded/default: ", buffer, j);
 259     /* case-fold/Turkic */
 260     isError=FALSE;
 261     for(i=j=0; j<LENGTHOF(buffer) && !isError; /* U16_NEXT post-increments */) {
 262         U16_NEXT(input, i, INT32_MAX, c); /* without length because NUL-terminated */
 263         if(c==0) {
 264             break; /* stop at terminating NUL, no need to terminate buffer */
 265         }
 266         c=u_foldCase(c, U_FOLD_CASE_EXCLUDE_SPECIAL_I);
 267         U16_APPEND(buffer, j, LENGTHOF(buffer), c, isError);
 268     }
 269     printUString("simple-case-folded/Turkic: ", buffer, j);
 270
 271     /*
 272      * Second, use full case mapping functions which provide
 273      * 1:n code point mappings (n can be 0!) and are sensitive to context and locale ID.
 274      *
 275      * Note that lower/upper/titlecasing take a locale ID while case-folding
 276      * has bit flag options instead, by design of the Unicode SpecialCasing.txt UCD file.
 277      *
 278      * Also, string titlecasing requires a BreakIterator to find starts of words.
 279      * The sample code here passes in a NULL pointer; u_strToTitle() will open and close a default
 280      * titlecasing BreakIterator automatically.
 281      * For production code where many strings are titlecased it would be more efficient
 282      * to open a BreakIterator externally and pass it in.
 283      */
 284     printUString("\ninput string: ", input, -1);
 285
 286     /* lowercase/English */
 287     errorCode=U_ZERO_ERROR;
 288     length=u_strToLower(buffer, LENGTHOF(buffer), input, -1, "en", &errorCode);
 289     if(U_SUCCESS(errorCode)) {
 290         printUString("full-lowercased/en: ", buffer, length);
 291     } else {
 292         printf("error in u_strToLower(en)=%ld error=%s\n", length, u_errorName(errorCode));
 293     }
 294     /* lowercase/Turkish */
 295     errorCode=U_ZERO_ERROR;
 296     length=u_strToLower(buffer, LENGTHOF(buffer), input, -1, "tr", &errorCode);
 297     if(U_SUCCESS(errorCode)) {
 298         printUString("full-lowercased/tr: ", buffer, length);
 299     } else {
 300         printf("error in u_strToLower(tr)=%ld error=%s\n", length, u_errorName(errorCode));
 301     }
 302     /* uppercase/English */
 303     errorCode=U_ZERO_ERROR;
 304     length=u_strToUpper(buffer, LENGTHOF(buffer), input, -1, "en", &errorCode);
 305     if(U_SUCCESS(errorCode)) {
 306         printUString("full-uppercased/en: ", buffer, length);
 307     } else {
 308         printf("error in u_strToUpper(en)=%ld error=%s\n", length, u_errorName(errorCode));
 309     }
 310     /* uppercase/Turkish */
 311     errorCode=U_ZERO_ERROR;
 312     length=u_strToUpper(buffer, LENGTHOF(buffer), input, -1, "tr", &errorCode);
 313     if(U_SUCCESS(errorCode)) {
 314         printUString("full-uppercased/tr: ", buffer, length);
 315     } else {
 316         printf("error in u_strToUpper(tr)=%ld error=%s\n", length, u_errorName(errorCode));
 317     }
 318     /* titlecase/English */
 319     errorCode=U_ZERO_ERROR;
 320     length=u_strToTitle(buffer, LENGTHOF(buffer), input, -1, NULL, "en", &errorCode);
 321     if(U_SUCCESS(errorCode)) {
 322         printUString("full-titlecased/en: ", buffer, length);
 323     } else {
 324         printf("error in u_strToTitle(en)=%ld error=%s\n", length, u_errorName(errorCode));
 325     }
 326     /* titlecase/Turkish */
 327     errorCode=U_ZERO_ERROR;
 328     length=u_strToTitle(buffer, LENGTHOF(buffer), input, -1, NULL, "tr", &errorCode);
 329     if(U_SUCCESS(errorCode)) {
 330         printUString("full-titlecased/tr: ", buffer, length);
 331     } else {
 332         printf("error in u_strToTitle(tr)=%ld error=%s\n", length, u_errorName(errorCode));
 333     }
 334     /* case-fold/default */
 335     errorCode=U_ZERO_ERROR;
 336     length=u_strFoldCase(buffer, LENGTHOF(buffer), input, -1, U_FOLD_CASE_DEFAULT, &errorCode);
 337     if(U_SUCCESS(errorCode)) {
 338         printUString("full-case-folded/default: ", buffer, length);
 339     } else {
 340         printf("error in u_strFoldCase(default)=%ld error=%s\n", length, u_errorName(errorCode));
 341     }
 342     /* case-fold/Turkic */
 343     errorCode=U_ZERO_ERROR;
 344     length=u_strFoldCase(buffer, LENGTHOF(buffer), input, -1, U_FOLD_CASE_EXCLUDE_SPECIAL_I, &errorCode);
 345     if(U_SUCCESS(errorCode)) {
 346         printUString("full-case-folded/Turkic: ", buffer, length);
 347     } else {
 348         printf("error in u_strFoldCase(Turkic)=%ld error=%s\n", length, u_errorName(errorCode));
 349     }
 350 }
 351
 352 // sample code for case mappings with C++ APIs ------------------------------ ***
 353
 354 static void demoCaseMapInCPlusPlus() {
 355     /*
 356      * input=
 357      *   "aB<capital sigma>"
 358      *   "iI<small dotless i><capital dotted I> "
 359      *   "<sharp s> <small lig. ffi>"
 360      *   "<small final sigma><small sigma><capital sigma>"
 361      */
 362     static const UChar input[]={
 363         0x61, 0x42, 0x3a3,
 364         0x69, 0x49, 0x131, 0x130, 0x20,
 365         0xdf, 0x20, 0xfb03,
 366         0x3c2, 0x3c3, 0x3a3, 0
 367     };
 368
 369     printf("\n* demoCaseMapInCPlusPlus() --------- ***\n\n");
 370
 371     UnicodeString s(input), t;
 372     const Locale &en=Locale::getEnglish();
 373     Locale tr("tr");
 374
 375     /*
 376      * Full case mappings as in demoCaseMapInC(), using UnicodeString functions.
 377      * These functions modify the string object itself.
 378      * Since we want to keep the input string around, we copy it each time
 379      * and case-map the copy.
 380      */
 381     printUnicodeString("input string: ", s);
 382
 383     /* lowercase/English */
 384     printUnicodeString("full-lowercased/en: ", (t=s).toLower(en));
 385     /* lowercase/Turkish */
 386     printUnicodeString("full-lowercased/tr: ", (t=s).toLower(tr));
 387     /* uppercase/English */
 388     printUnicodeString("full-uppercased/en: ", (t=s).toUpper(en));
 389     /* uppercase/Turkish */
 390     printUnicodeString("full-uppercased/tr: ", (t=s).toUpper(tr));
 391     /* titlecase/English */
 392     printUnicodeString("full-titlecased/en: ", (t=s).toTitle(NULL, en));
 393     /* titlecase/Turkish */
 394     printUnicodeString("full-titlecased/tr: ", (t=s).toTitle(NULL, tr));
 395     /* case-folde/default */
 396     printUnicodeString("full-case-folded/default: ", (t=s).foldCase(U_FOLD_CASE_DEFAULT));
 397     /* case-folde/Turkic */
 398     printUnicodeString("full-case-folded/Turkic: ", (t=s).foldCase(U_FOLD_CASE_EXCLUDE_SPECIAL_I));
 399 }
 400
 401 // sample code for UnicodeString storage models ----------------------------- ***
 402
 403 static const UChar readonly[]={
 404     0x61, 0x31, 0x20ac
 405 };
 406 static UChar writeable[]={
 407     0x62, 0x32, 0xdbc0, 0xdc01 // includes a surrogate pair for a supplementary code point
 408 };
 409 static char out[100];
 410
 411 static void
 412 demoUnicodeStringStorage() {
 413     // These sample code lines illustrate how to use UnicodeString, and the
 414     // comments tell what happens internally. There are no APIs to observe
 415     // most of this programmatically, except for stepping into the code
 416     // with a debugger.
 417     // This is by design to hide such details from the user.
 418     int32_t i;
 419
 420     printf("\n* demoUnicodeStringStorage() ------- ***\n\n");
 421
 422     // * UnicodeString with internally stored contents
 423     // instantiate a UnicodeString from a single code point
 424     // the few (2) UChars will be stored in the object itself
 425     UnicodeString one((UChar32)0x24001);
 426     // this copies the few UChars into the "two" object
 427     UnicodeString two=one;
 428     printf("length of short string copy: %d\n", two.length());
 429     // set "one" to contain the 3 UChars from readonly
 430     // this setTo() variant copies the characters
 431     one.setTo(readonly, LENGTHOF(readonly));
 432
 433     // * UnicodeString with allocated contents
 434     // build a longer string that will not fit into the object's buffer
 435     one+=UnicodeString(writeable, LENGTHOF(writeable));
 436     one+=one;
 437     one+=one;
 438     printf("length of longer string: %d\n", one.length());
 439     // copying will use the same allocated buffer and increment the reference
 440     // counter
 441     two=one;
 442     printf("length of longer string copy: %d\n", two.length());
 443
 444     // * UnicodeString using readonly-alias to a const UChar array
 445     // construct a string that aliases a readonly buffer
 446     UnicodeString three(FALSE, readonly, LENGTHOF(readonly));
 447     printUnicodeString("readonly-alias string: ", three);
 448     // copy-on-write: any modification to the string results in
 449     // a copy to either the internal buffer or to a newly allocated one
 450     three.setCharAt(1, 0x39);
 451     printUnicodeString("readonly-aliasing string after modification: ", three);
 452     // the aliased array is not modified
 453     for(i=0; i<three.length(); ++i) {
 454         printf("readonly buffer[%d] after modifying its string: 0x%lx\n",
 455                i, readonly[i]);
 456     }
 457     // setTo() readonly alias
 458     one.setTo(FALSE, writeable, LENGTHOF(writeable));
 459     // copying the readonly-alias object with fastCopyFrom() (new in ICU 2.4)
 460     // will readonly-alias the same buffer
 461     two.fastCopyFrom(one);
 462     printUnicodeString("fastCopyFrom(readonly alias of \"writeable\" array): ", two);
 463     printf("verify that a fastCopyFrom(readonly alias) uses the same buffer pointer: %d (should be 1)\n",
 464         one.getBuffer()==two.getBuffer());
 465     // a normal assignment will clone the contents (new in ICU 2.4)
 466     two=one;
 467     printf("verify that a regular copy of a readonly alias uses a different buffer pointer: %d (should be 0)\n",
 468         one.getBuffer()==two.getBuffer());
 469
 470     // * UnicodeString using writeable-alias to a non-const UChar array
 471     UnicodeString four(writeable, LENGTHOF(writeable), LENGTHOF(writeable));
 472     printUnicodeString("writeable-alias string: ", four);
 473     // a modification writes through to the buffer
 474     four.setCharAt(1, 0x39);
 475     for(i=0; i<four.length(); ++i) {
 476         printf("writeable-alias backing buffer[%d]=0x%lx "
 477                "after modification\n", i, writeable[i]);
 478     }
 479     // a copy will not alias any more;
 480     // instead, it will get a copy of the contents into allocated memory
 481     two=four;
 482     two.setCharAt(1, 0x21);
 483     for(i=0; i<two.length(); ++i) {
 484         printf("writeable-alias backing buffer[%d]=0x%lx after "
 485                "modification of string copy\n", i, writeable[i]);
 486     }
 487     // setTo() writeable alias, capacity==length
 488     one.setTo(writeable, LENGTHOF(writeable), LENGTHOF(writeable));
 489     // grow the string - it will not fit into the backing buffer any more
 490     // and will get copied before modification
 491     one.append((UChar)0x40);
 492     // shrink it back so it would fit
 493     one.truncate(one.length()-1);
 494     // we still operate on the copy
 495     one.setCharAt(1, 0x25);
 496     printf("string after growing too much and then shrinking[1]=0x%lx\n"
 497            "                          backing store for this[1]=0x%lx\n",
 498            one.charAt(1), writeable[1]);
 499     // if we need it in the original buffer, then extract() to it
 500     // extract() does not do anything if the string aliases that same buffer
 501     // i=min(one.length(), length of array)
 502     if(one.length()<LENGTHOF(writeable)) {
 503         i=one.length();
 504     } else {
 505         i=LENGTHOF(writeable);
 506     }
 507     one.extract(0, i, writeable);
 508     for(i=0; i<LENGTHOF(writeable); ++i) {
 509         printf("writeable-alias backing buffer[%d]=0x%lx after re-extract\n",
 510                i, writeable[i]);
 511     }
 512 }
 513
 514 // sample code for UnicodeString instantiations ----------------------------- ***
 515
 516 static void
 517 demoUnicodeStringInit() {
 518     // *** Make sure to read about invariant characters in utypes.h! ***
 519     // Initialization of Unicode strings from C literals works _only_ for
 520     // invariant characters!
 521
 522     printf("\n* demoUnicodeStringInit() ---------- ***\n\n");
 523
 524     // the string literal is 32 chars long - this must be counted for the macro
 525     UnicodeString invariantOnly=UNICODE_STRING("such characters are safe 123 %-.", 32);
 526
 527     /*
 528      * In C, we need two macros: one to declare the UChar[] array, and
 529      * one to populate it; the second one is a noop on platforms where
 530      * wchar_t is compatible with UChar and ASCII-based.
 531      * The length of the string literal must be counted for both macros.
 532      */
 533     /* declare the invString array for the string */
 534     U_STRING_DECL(invString, "such characters are safe 123 %-.", 32);
 535     /* populate it with the characters */
 536     U_STRING_INIT(invString, "such characters are safe 123 %-.", 32);
 537
 538     // compare the C and C++ strings
 539     printf("C and C++ Unicode strings are equal: %d\n", invariantOnly==UnicodeString(TRUE, invString, 32));
 540
 541     /*
 542      * convert between char * and UChar * strings that
 543      * contain only invariant characters
 544      */
 545     static const char *cs1="such characters are safe 123 %-.";
 546     static UChar us1[40];
 547     static char cs2[40];
 548     u_charsToUChars(cs1, us1, 33); /* include the terminating NUL */
 549     u_UCharsToChars(us1, cs2, 33);
 550     printf("char * -> UChar * -> char * with only "
 551            "invariant characters: \"%s\"\n",
 552            cs2);
 553
 554     // initialize a UnicodeString from a string literal that contains
 555     // escape sequences written with invariant characters
 556     // do not forget to duplicate the backslashes for ICU to see them
 557     // then, count each double backslash only once!
 558     UnicodeString german=UNICODE_STRING(
 559         "Sch\\u00f6nes Auto: \\u20ac 11240.\\fPrivates Zeichen: \\U00102345\\n", 64).
 560         unescape();
 561     printUnicodeString("german UnicodeString from unescaping:\n    ", german);
 562
 563     /*
 564      * C: convert and unescape a char * string with only invariant
 565      * characters to fill a UChar * string
 566      */
 567     UChar buffer[200];
 568     int32_t length;
 569     length=u_unescape(
 570         "Sch\\u00f6nes Auto: \\u20ac 11240.\\fPrivates Zeichen: \\U00102345\\n",
 571         buffer, LENGTHOF(buffer));
 572     printf("german C Unicode string from char * unescaping: (length %d)\n    ", length);
 573     printUnicodeString("", UnicodeString(buffer));
 574 }
 575
 576 extern int
 577 main(int argc, const char *argv[]) {
 578     UErrorCode errorCode=U_ZERO_ERROR;
 579
 580     // Note: Using a global variable for any object is not exactly thread-safe...
 581
 582     // You can change this call to e.g. ucnv_open("UTF-8", &errorCode) if you pipe
 583     // the output to a file and look at it with a Unicode-capable editor.
 584     // This will currently affect only the printUString() function, see the code above.
 585     // printUnicodeString() could use this, too, by changing to an extract() overload
 586     // that takes a UConverter argument.
 587     cnv=ucnv_open(NULL, &errorCode);
 588     if(U_FAILURE(errorCode)) {
 589         fprintf(stderr, "error %s opening the default converter\n", u_errorName(errorCode));
 590         return errorCode;
 591     }
 592
 593     ucnv_setFromUCallBack(cnv, UCNV_FROM_U_CALLBACK_ESCAPE, UCNV_ESCAPE_C, NULL, NULL, &errorCode);
 594     if(U_FAILURE(errorCode)) {
 595         fprintf(stderr, "error %s setting the escape callback in the default converter\n", u_errorName(errorCode));
 596         ucnv_close(cnv);
 597         return errorCode;
 598     }
 599
 600     demo_utf_h_macros();
 601     demo_C_Unicode_strings();
 602     demoCaseMapInC();
 603     demoCaseMapInCPlusPlus();
 604     demoUnicodeStringStorage();
 605     demoUnicodeStringInit();
 606
 607     ucnv_close(cnv);
 608     return 0;
 609 }