| /******************************************************************** | |
| * COPYRIGHT: | |
| * Copyright (c) 1999-2010, International Business Machines Corporation and | |
| * others. All Rights Reserved. | |
| ********************************************************************/ | |
| #include "unicode/unistr.h" | |
| #include <stdio.h> | |
| #include <stdlib.h> | |
| // Verify that a UErrorCode is successful; exit(1) if not | |
| void check(UErrorCode& status, const char* msg) { | |
| if (U_FAILURE(status)) { | |
| printf("ERROR: %s (%s)\n", u_errorName(status), msg); | |
| exit(1); | |
| } | |
| // printf("Ok: %s\n", msg); | |
| } | |
| // Append a hex string to the target | |
| static UnicodeString& appendHex(uint32_t number, | |
| int8_t digits, | |
| UnicodeString& target) { | |
| static const UnicodeString DIGIT_STRING("0123456789ABCDEF"); | |
| while (digits > 0) { | |
| target += DIGIT_STRING[(number >> ((--digits) * 4)) & 0xF]; | |
| } | |
| return target; | |
| } | |
| // Replace nonprintable characters with unicode escapes | |
| UnicodeString escape(const UnicodeString &source) { | |
| int32_t i; | |
| UnicodeString target; | |
| target += "\""; | |
| for (i=0; i<source.length(); ++i) { | |
| UChar ch = source[i]; | |
| if (ch < 0x09 || (ch > 0x0A && ch < 0x20) || ch > 0x7E) { | |
| target += "\\u"; | |
| appendHex(ch, 4, target); | |
| } else { | |
| target += ch; | |
| } | |
| } | |
| target += "\""; | |
| return target; | |
| } | |
| // Print the given string to stdout | |
| void uprintf(const UnicodeString &str) { | |
| char *buf = 0; | |
| int32_t len = str.length(); | |
| // int32_t bufLen = str.extract(0, len, buf); // Preflight | |
| /* Preflighting seems to be broken now, so assume 1-1 conversion, | |
| plus some slop. */ | |
| int32_t bufLen = len + 16; | |
| int32_t actualLen; | |
| buf = new char[bufLen + 1]; | |
| actualLen = str.extract(0, len, buf/*, bufLen*/); // Default codepage conversion | |
| buf[actualLen] = 0; | |
| printf("%s", buf); | |
| delete [] buf; | |
| } |