inline void do_to_title(
ErlNifBinary in,
ErlNifBinary& out,
int32_t& ulen,
UBreakIterator* iter,
const char* locale,
UErrorCode& status)
{
status = U_ZERO_ERROR;
if (!enif_alloc_binary(FROM_ULEN(ulen), &out)) {
status = U_MEMORY_ALLOCATION_ERROR;
return;
}
ulen = u_strToTitle(
(UChar*) out.data, /* src */
ulen, /* len of src */
(UChar*) in.data,
TO_ULEN(in.size),
iter, /* Iterator */
locale,
&status);
if (status == U_BUFFER_OVERFLOW_ERROR) {
/* enlarge buffer if it was too small */
enif_release_binary(&out);
return;
}
if (FROM_ULEN(ulen) != out.size) {
/* shrink binary if it was too large */
enif_realloc_binary(&out, FROM_ULEN(ulen));
}
}
int32_t gtm_strToTitle(UChar *dst, int32_t dstlen, const UChar *src, int32_t srclen,
const char *locale, UErrorCode *status)
{
# ifdef DEBUG
if (gtm_white_box_test_case_enabled && (WBTEST_OPFNZCONVERT_FILE_ACCESS_ERROR == gtm_white_box_test_case_number))
{
*status = U_FILE_ACCESS_ERROR;
return -1;
}
# endif
return u_strToTitle(dst, dstlen, src, srclen, NULL, locale, status);
}
// title {{{
static PyObject *
icu_title(PyObject *self, PyObject *args) {
char *input, *ans, *buf3 = NULL;
const char *loc;
int32_t sz;
UChar *buf, *buf2;
PyObject *ret;
UErrorCode status = U_ZERO_ERROR;
if (!PyArg_ParseTuple(args, "ses", &loc, "UTF-8", &input)) return NULL;
sz = (int32_t)strlen(input);
buf = (UChar*)calloc(sz*4 + 1, sizeof(UChar));
buf2 = (UChar*)calloc(sz*8 + 1, sizeof(UChar));
if (buf == NULL || buf2 == NULL) return PyErr_NoMemory();
u_strFromUTF8(buf, sz*4, NULL, input, sz, &status);
u_strToTitle(buf2, sz*8, buf, -1, NULL, loc, &status);
ans = input;
sz = u_strlen(buf2);
free(buf);
if (U_SUCCESS(status) && sz > 0) {
buf3 = (char*)calloc(sz*5+1, sizeof(char));
if (buf3 == NULL) return PyErr_NoMemory();
u_strToUTF8(buf3, sz*5, NULL, buf2, -1, &status);
if (U_SUCCESS(status)) ans = buf3;
}
ret = PyUnicode_DecodeUTF8(ans, strlen(ans), "replace");
if (ret == NULL) return PyErr_NoMemory();
free(buf2);
if (buf3 != NULL) free(buf3);
PyMem_Free(input);
return ret;
} // }}}
static PyObject* icu_change_case(PyObject *self, PyObject *args) {
char *locale = NULL;
PyObject *input = NULL, *result = NULL;
int which = UPPER_CASE;
UErrorCode status = U_ZERO_ERROR;
UChar *input_buf = NULL, *output_buf = NULL;
int32_t sz = 0;
if (!PyArg_ParseTuple(args, "Oiz", &input, &which, &locale)) return NULL;
if (locale == NULL) {
PyErr_SetString(PyExc_NotImplementedError, "You must specify a locale"); // We deliberately use NotImplementedError so that this error can be unambiguously identified
return NULL;
}
input_buf = python_to_icu(input, &sz, 1);
if (input_buf == NULL) goto end;
output_buf = (UChar*) calloc(3 * sz, sizeof(UChar));
if (output_buf == NULL) { PyErr_NoMemory(); goto end; }
switch (which) {
case TITLE_CASE:
sz = u_strToTitle(output_buf, 3 * sz, input_buf, sz, NULL, locale, &status);
break;
case UPPER_CASE:
sz = u_strToUpper(output_buf, 3 * sz, input_buf, sz, locale, &status);
break;
default:
sz = u_strToLower(output_buf, 3 * sz, input_buf, sz, locale, &status);
}
if (U_FAILURE(status)) { PyErr_SetString(PyExc_ValueError, u_errorName(status)); goto end; }
result = icu_to_python(output_buf, sz);
end:
if (input_buf != NULL) free(input_buf);
if (output_buf != NULL) free(output_buf);
return result;
} // }}}
static uint32_t icu_utf16_sub(struct icu_buf_utf16 *dest16,
struct icu_buf_utf16 *src16,
const char *locale, char action,
UErrorCode *status)
{
switch (action)
{
case 'l':
case 'L':
return u_strToLower(dest16->utf16, dest16->utf16_cap,
src16->utf16, src16->utf16_len,
locale, status);
case 'u':
case 'U':
return u_strToUpper(dest16->utf16, dest16->utf16_cap,
src16->utf16, src16->utf16_len,
locale, status);
break;
case 't':
case 'T':
return u_strToTitle(dest16->utf16, dest16->utf16_cap,
src16->utf16, src16->utf16_len,
0, locale, status);
break;
case 'f':
case 'F':
return u_strFoldCase(dest16->utf16, dest16->utf16_cap,
src16->utf16, src16->utf16_len,
U_FOLD_CASE_DEFAULT, status);
break;
default:
*status = U_UNSUPPORTED_ERROR;
break;
}
return 0;
}
static void
TestCaseDutchTitle(void) {
static const UChar
beforeTitle[]= { 0x69, 0x6A, 0x73, 0x73, 0x45, 0x6c, 0x20, 0x69, 0x67, 0x6c, 0x4f, 0x6f , 0x20 , 0x49, 0x4A, 0x53, 0x53, 0x45, 0x4C },
titleRoot[]= { 0x49, 0x6A, 0x73, 0x73, 0x65, 0x6c, 0x20, 0x49, 0x67, 0x6c, 0x6f, 0x6f , 0x20 , 0x49, 0x6A, 0x73, 0x73, 0x65, 0x6C },
titleDutch[]= { 0x49, 0x4A, 0x73, 0x73, 0x65, 0x6c, 0x20, 0x49, 0x67, 0x6c, 0x6f, 0x6f , 0x20 , 0x49, 0x4A, 0x73, 0x73, 0x65, 0x6C };
UChar buffer[32];
UBreakIterator *titleIterWord;
int32_t length;
UErrorCode errorCode;
errorCode=U_ZERO_ERROR;
titleIterWord=ubrk_open(UBRK_WORD, "", beforeTitle, sizeof(beforeTitle)/U_SIZEOF_UCHAR, &errorCode);
if(U_FAILURE(errorCode)) {
log_err_status(errorCode, "error: ubrk_open(UBRK_WORD)->%s\n", u_errorName(errorCode));
return;
}
/* titlecase with default locale */
buffer[0]=0xabcd;
errorCode=U_ZERO_ERROR;
length=u_strToTitle(buffer, sizeof(buffer)/U_SIZEOF_UCHAR,
beforeTitle, sizeof(beforeTitle)/U_SIZEOF_UCHAR,
titleIterWord, "",
&errorCode);
if( U_FAILURE(errorCode) ||
length!=(sizeof(titleRoot)/U_SIZEOF_UCHAR) ||
uprv_memcmp(titleRoot, buffer, length*U_SIZEOF_UCHAR)!=0 ||
buffer[length]!=0
) {
char charsOut[21];
u_UCharsToChars(buffer,charsOut,sizeof(charsOut));
log_err("error in u_strToTitle(UBRK_CHARACTERS)=%ld error=%s root locale string matches: %s\noutput buffer is {%s}\n",
length,
u_errorName(errorCode),
uprv_memcmp(titleRoot, buffer, length*U_SIZEOF_UCHAR)==0 && buffer[length]==0 ? "yes" : "no", charsOut);
}
/* titlecase with Dutch locale */
buffer[0]=0xabcd;
errorCode=U_ZERO_ERROR;
length=u_strToTitle(buffer, sizeof(buffer)/U_SIZEOF_UCHAR,
beforeTitle, sizeof(beforeTitle)/U_SIZEOF_UCHAR,
titleIterWord, "nl",
&errorCode);
if( U_FAILURE(errorCode) ||
length!=(sizeof(titleDutch)/U_SIZEOF_UCHAR) ||
uprv_memcmp(titleDutch, buffer, length*U_SIZEOF_UCHAR)!=0 ||
buffer[length]!=0
) {
char charsOut[21];
u_UCharsToChars(buffer,charsOut,sizeof(charsOut));
log_err("error in u_strToTitle(UBRK_CHARACTERS)=%ld error=%s dutch locale string matches: %s\noutput buffer is {%s}\n",
length,
u_errorName(errorCode),
uprv_memcmp(titleDutch, buffer, length*U_SIZEOF_UCHAR)==0 && buffer[length]==0 ? "yes" : "no", charsOut);
}
ubrk_close(titleIterWord);
}
static void
TestCaseTitle(void) {
static const UChar
beforeTitle[]= { 0x61, 0x42, 0x20, 0x69, 0x3c2, 0x20, 0xdf, 0x3c3, 0x2f, 0xfb03, 0xd93f, 0xdfff },
titleWord[]= { 0x41, 0x62, 0x20, 0x49, 0x3c2, 0x20, 0x53, 0x73, 0x3c3, 0x2f, 0x46, 0x66, 0x69, 0xd93f, 0xdfff },
titleChar[]= { 0x41, 0x42, 0x20, 0x49, 0x3a3, 0x20, 0x53, 0x73, 0x3a3, 0x2f, 0x46, 0x66, 0x69, 0xd93f, 0xdfff };
UChar buffer[32];
UBreakIterator *titleIterChars;
int32_t length;
UErrorCode errorCode;
errorCode=U_ZERO_ERROR;
titleIterChars=ubrk_open(UBRK_CHARACTER, "", beforeTitle, sizeof(beforeTitle)/U_SIZEOF_UCHAR, &errorCode);
if(U_FAILURE(errorCode)) {
log_err_status(errorCode, "error: ubrk_open(UBRK_CHARACTER)->%s\n", u_errorName(errorCode));
return;
}
/* titlecase with standard break iterator and in the same buffer */
uprv_memcpy(buffer, beforeTitle, sizeof(beforeTitle));
errorCode=U_ZERO_ERROR;
length=u_strToTitle(buffer, sizeof(buffer)/U_SIZEOF_UCHAR,
buffer, sizeof(beforeTitle)/U_SIZEOF_UCHAR,
NULL, "",
&errorCode);
if( U_FAILURE(errorCode) ||
length!=(sizeof(titleWord)/U_SIZEOF_UCHAR) ||
uprv_memcmp(titleWord, buffer, length*U_SIZEOF_UCHAR)!=0 ||
buffer[length]!=0
) {
log_err("error in u_strToTitle(standard iterator)=%ld error=%s string matches: %s\n",
length,
u_errorName(errorCode),
uprv_memcmp(titleWord, buffer, length*U_SIZEOF_UCHAR)==0 && buffer[length]==0 ? "yes" : "no");
}
/* titlecase with UBRK_CHARACTERS and separate buffers */
buffer[0]=0xabcd;
errorCode=U_ZERO_ERROR;
length=u_strToTitle(buffer, sizeof(buffer)/U_SIZEOF_UCHAR,
beforeTitle, sizeof(beforeTitle)/U_SIZEOF_UCHAR,
titleIterChars, "",
&errorCode);
if( U_FAILURE(errorCode) ||
length!=(sizeof(titleChar)/U_SIZEOF_UCHAR) ||
uprv_memcmp(titleChar, buffer, length*U_SIZEOF_UCHAR)!=0 ||
buffer[length]!=0
) {
log_err("error in u_strToTitle(UBRK_CHARACTERS)=%ld error=%s string matches: %s\n",
length,
u_errorName(errorCode),
uprv_memcmp(titleChar, buffer, length*U_SIZEOF_UCHAR)==0 && buffer[length]==0 ? "yes" : "no");
}
/* test preflighting */
errorCode=U_ZERO_ERROR;
length=u_strToTitle(NULL, 0,
beforeTitle, sizeof(beforeTitle)/U_SIZEOF_UCHAR,
titleIterChars, "",
&errorCode);
if( errorCode!=U_BUFFER_OVERFLOW_ERROR ||
length!=(sizeof(titleChar)/U_SIZEOF_UCHAR)
) {
log_err("error in u_strToTitle(UBRK_CHARACTERS pure preflighting)=%ld error=%s\n",
length,
u_errorName(errorCode));
}
/* test error handling */
buffer[0]=0xabcd;
errorCode=U_ZERO_ERROR;
length=u_strToTitle(buffer, sizeof(buffer)/U_SIZEOF_UCHAR,
NULL, sizeof(beforeTitle)/U_SIZEOF_UCHAR,
titleIterChars, "",
&errorCode);
if( errorCode!=U_ILLEGAL_ARGUMENT_ERROR ||
buffer[0]!=0xabcd
) {
log_err("error in u_strToTitle(UBRK_CHARACTERS src=NULL)=%ld error=%s buffer[0]==0x%lx\n",
length,
u_errorName(errorCode),
buffer[0]);
}
buffer[0]=0xabcd;
errorCode=U_ZERO_ERROR;
length=u_strToTitle(buffer, sizeof(buffer)/U_SIZEOF_UCHAR,
beforeTitle, -2,
titleIterChars, "",
&errorCode);
if( errorCode!=U_ILLEGAL_ARGUMENT_ERROR ||
buffer[0]!=0xabcd
) {
log_err("error in u_strToTitle(UBRK_CHARACTERS srcLength=-2)=%ld error=%s buffer[0]==0x%lx\n",
length,
u_errorName(errorCode),
buffer[0]);
}
ubrk_close(titleIterChars);
}
static void demoCaseMapInC() {
/*
* input=
* "aB<capital sigma>"
* "iI<small dotless i><capital dotted I> "
* "<sharp s> <small lig. ffi>"
* "<small final sigma><small sigma><capital sigma>"
*/
static const UChar input[]={
0x61, 0x42, 0x3a3,
0x69, 0x49, 0x131, 0x130, 0x20,
0xdf, 0x20, 0xfb03,
0x3c2, 0x3c3, 0x3a3, 0
};
UChar buffer[32];
UErrorCode errorCode;
UChar32 c;
int32_t i, j, length;
UBool isError;
printf("\n* demoCaseMapInC() ----------------- ***\n\n");
/*
* First, use simple case mapping functions which provide
* 1:1 code point mappings without context/locale ID.
*
* Note that some mappings will not be "right" because some "real"
* case mappings require context, depend on the locale ID,
* and/or result in a change in the number of code points.
*/
printUString("input string: ", input, -1);
/* uppercase */
isError=FALSE;
for(i=j=0; j<UPRV_LENGTHOF(buffer) && !isError; /* U16_NEXT post-increments */) {
U16_NEXT(input, i, INT32_MAX, c); /* without length because NUL-terminated */
if(c==0) {
break; /* stop at terminating NUL, no need to terminate buffer */
}
c=u_toupper(c);
U16_APPEND(buffer, j, UPRV_LENGTHOF(buffer), c, isError);
}
printUString("simple-uppercased: ", buffer, j);
/* lowercase */
isError=FALSE;
for(i=j=0; j<UPRV_LENGTHOF(buffer) && !isError; /* U16_NEXT post-increments */) {
U16_NEXT(input, i, INT32_MAX, c); /* without length because NUL-terminated */
if(c==0) {
break; /* stop at terminating NUL, no need to terminate buffer */
}
c=u_tolower(c);
U16_APPEND(buffer, j, UPRV_LENGTHOF(buffer), c, isError);
}
printUString("simple-lowercased: ", buffer, j);
/* titlecase */
isError=FALSE;
for(i=j=0; j<UPRV_LENGTHOF(buffer) && !isError; /* U16_NEXT post-increments */) {
U16_NEXT(input, i, INT32_MAX, c); /* without length because NUL-terminated */
if(c==0) {
break; /* stop at terminating NUL, no need to terminate buffer */
}
c=u_totitle(c);
U16_APPEND(buffer, j, UPRV_LENGTHOF(buffer), c, isError);
}
printUString("simple-titlecased: ", buffer, j);
/* case-fold/default */
isError=FALSE;
for(i=j=0; j<UPRV_LENGTHOF(buffer) && !isError; /* U16_NEXT post-increments */) {
U16_NEXT(input, i, INT32_MAX, c); /* without length because NUL-terminated */
if(c==0) {
break; /* stop at terminating NUL, no need to terminate buffer */
}
c=u_foldCase(c, U_FOLD_CASE_DEFAULT);
U16_APPEND(buffer, j, UPRV_LENGTHOF(buffer), c, isError);
}
printUString("simple-case-folded/default: ", buffer, j);
/* case-fold/Turkic */
isError=FALSE;
for(i=j=0; j<UPRV_LENGTHOF(buffer) && !isError; /* U16_NEXT post-increments */) {
U16_NEXT(input, i, INT32_MAX, c); /* without length because NUL-terminated */
if(c==0) {
break; /* stop at terminating NUL, no need to terminate buffer */
}
c=u_foldCase(c, U_FOLD_CASE_EXCLUDE_SPECIAL_I);
U16_APPEND(buffer, j, UPRV_LENGTHOF(buffer), c, isError);
}
printUString("simple-case-folded/Turkic: ", buffer, j);
/*
* Second, use full case mapping functions which provide
* 1:n code point mappings (n can be 0!) and are sensitive to context and locale ID.
*
* Note that lower/upper/titlecasing take a locale ID while case-folding
* has bit flag options instead, by design of the Unicode SpecialCasing.txt UCD file.
*
* Also, string titlecasing requires a BreakIterator to find starts of words.
* The sample code here passes in a NULL pointer; u_strToTitle() will open and close a default
* titlecasing BreakIterator automatically.
* For production code where many strings are titlecased it would be more efficient
* to open a BreakIterator externally and pass it in.
*/
printUString("\ninput string: ", input, -1);
/* lowercase/English */
errorCode=U_ZERO_ERROR;
length=u_strToLower(buffer, UPRV_LENGTHOF(buffer), input, -1, "en", &errorCode);
if(U_SUCCESS(errorCode)) {
printUString("full-lowercased/en: ", buffer, length);
} else {
printf("error in u_strToLower(en)=%ld error=%s\n", length, u_errorName(errorCode));
}
/* lowercase/Turkish */
errorCode=U_ZERO_ERROR;
length=u_strToLower(buffer, UPRV_LENGTHOF(buffer), input, -1, "tr", &errorCode);
if(U_SUCCESS(errorCode)) {
printUString("full-lowercased/tr: ", buffer, length);
} else {
printf("error in u_strToLower(tr)=%ld error=%s\n", length, u_errorName(errorCode));
}
/* uppercase/English */
errorCode=U_ZERO_ERROR;
length=u_strToUpper(buffer, UPRV_LENGTHOF(buffer), input, -1, "en", &errorCode);
if(U_SUCCESS(errorCode)) {
printUString("full-uppercased/en: ", buffer, length);
} else {
printf("error in u_strToUpper(en)=%ld error=%s\n", length, u_errorName(errorCode));
}
/* uppercase/Turkish */
errorCode=U_ZERO_ERROR;
length=u_strToUpper(buffer, UPRV_LENGTHOF(buffer), input, -1, "tr", &errorCode);
if(U_SUCCESS(errorCode)) {
printUString("full-uppercased/tr: ", buffer, length);
} else {
printf("error in u_strToUpper(tr)=%ld error=%s\n", length, u_errorName(errorCode));
}
/* titlecase/English */
errorCode=U_ZERO_ERROR;
length=u_strToTitle(buffer, UPRV_LENGTHOF(buffer), input, -1, NULL, "en", &errorCode);
if(U_SUCCESS(errorCode)) {
printUString("full-titlecased/en: ", buffer, length);
} else {
printf("error in u_strToTitle(en)=%ld error=%s\n", length, u_errorName(errorCode));
}
/* titlecase/Turkish */
errorCode=U_ZERO_ERROR;
length=u_strToTitle(buffer, UPRV_LENGTHOF(buffer), input, -1, NULL, "tr", &errorCode);
if(U_SUCCESS(errorCode)) {
printUString("full-titlecased/tr: ", buffer, length);
} else {
printf("error in u_strToTitle(tr)=%ld error=%s\n", length, u_errorName(errorCode));
}
/* case-fold/default */
errorCode=U_ZERO_ERROR;
length=u_strFoldCase(buffer, UPRV_LENGTHOF(buffer), input, -1, U_FOLD_CASE_DEFAULT, &errorCode);
if(U_SUCCESS(errorCode)) {
printUString("full-case-folded/default: ", buffer, length);
} else {
printf("error in u_strFoldCase(default)=%ld error=%s\n", length, u_errorName(errorCode));
}
/* case-fold/Turkic */
errorCode=U_ZERO_ERROR;
length=u_strFoldCase(buffer, UPRV_LENGTHOF(buffer), input, -1, U_FOLD_CASE_EXCLUDE_SPECIAL_I, &errorCode);
if(U_SUCCESS(errorCode)) {
printUString("full-case-folded/Turkic: ", buffer, length);
} else {
printf("error in u_strFoldCase(Turkic)=%ld error=%s\n", length, u_errorName(errorCode));
}
}
int icu_case(lua_State *L) {
size_t input_l;
const char* input = luaL_checklstring(L, 1, &input_l);
const char* locale = luaL_checkstring(L, 2);
const char* recase = luaL_checkstring(L, 3);
/* Convert input to ICU-friendly UChars */
UChar *input_as_uchar;
int32_t l;
utf8_to_uchar(input, input_l, input_as_uchar, l);
/* Now do the conversion */
UChar *output;
int32_t l2 = 0;
UErrorCode err = U_ZERO_ERROR;
if (strcmp(recase, "title") == 0) {
l2 = u_strToTitle(NULL, 0, input_as_uchar, l, NULL, locale, &err);
err = U_ZERO_ERROR;
output = malloc(l2 * sizeof(UChar));
u_strToTitle(output, l2, input_as_uchar, l, NULL, locale, &err);
} else {
conversion_function_t conversion;
if (strcmp(recase, "upper") == 0) {
conversion = u_strToUpper;
} else if (strcmp(recase, "lower") == 0) {
conversion = u_strToLower;
} else {
free(input_as_uchar);
return luaL_error(L, "Unknown case conversion type %s", recase);
}
l2 = conversion(NULL, 0, input_as_uchar, l, locale, &err);
err = U_ZERO_ERROR;
output = malloc(l2 * sizeof(UChar));
conversion(output, l2, input_as_uchar, l, locale, &err);
}
if (!U_SUCCESS(err)) {
free(input_as_uchar);
free(output);
return luaL_error(L, "Error in case conversion %s", u_errorName(err));
}
int32_t l3 = 0;
char possibleOutbuf[4096];
u_strToUTF8(possibleOutbuf, 4096, &l3, output, l2, &err);
if (U_SUCCESS(err)) {
lua_pushstring(L, possibleOutbuf);
free(input_as_uchar);
free(output);
return 1;
}
char *utf8output;
if (err == U_BUFFER_OVERFLOW_ERROR) {
utf8output = malloc(l3);
u_strToUTF8(utf8output, l3, NULL, output, l2, &err);
if (!U_SUCCESS(err)) goto fail;
utf8output[l3] = '\0';
lua_pushstring(L, utf8output);
free(input_as_uchar);
free(output);
free(utf8output);
return 1;
}
fail:
return luaL_error(L, "Error in UTF8 conversion %s", u_errorName(err));
}
int32_t gtm_strToTitle(UChar *dst, int32_t dstlen, const UChar *src, int32_t srclen,
const char *locale, UErrorCode *status)
{
return u_strToTitle(dst, dstlen, src, srclen, NULL, locale, status);
}
