/* {{{ intl_errors_set_code( intl_error* err, UErrorCode err_code )
*/
void intl_errors_set_code( intl_error* err, UErrorCode err_code )
{
if(err) {
intl_error_set_code( err, err_code );
}
intl_error_set_code( NULL, err_code );
}
static inline int convert_cp(UChar32* pcp, zval *zcp) {
zend_long cp = -1;
if (Z_TYPE_P(zcp) == IS_LONG) {
cp = Z_LVAL_P(zcp);
} else if (Z_TYPE_P(zcp) == IS_STRING) {
int32_t i = 0;
size_t zcp_len = Z_STRLEN_P(zcp);
if (ZEND_SIZE_T_INT_OVFL(zcp_len)) {
intl_error_set_code(NULL, U_ILLEGAL_ARGUMENT_ERROR);
intl_error_set_custom_msg(NULL, "Input string is too long.", 0);
return FAILURE;
}
U8_NEXT(Z_STRVAL_P(zcp), i, zcp_len, cp);
if ((size_t)i != zcp_len) {
intl_error_set_code(NULL, U_ILLEGAL_ARGUMENT_ERROR);
intl_error_set_custom_msg(NULL, "Passing a UTF-8 character for codepoint requires a string which is exactly one UTF-8 codepoint long.", 0);
return FAILURE;
}
} else {
intl_error_set_code(NULL, U_ILLEGAL_ARGUMENT_ERROR);
intl_error_set_custom_msg(NULL, "Invalid parameter for unicode point. Must be either integer or UTF-8 sequence.", 0);
return FAILURE;
}
if ((cp < UCHAR_MIN_VALUE) || (cp > UCHAR_MAX_VALUE)) {
intl_error_set_code(NULL, U_ILLEGAL_ARGUMENT_ERROR);
intl_error_set_custom_msg(NULL, "Codepoint out of range", 0);
return FAILURE;
}
*pcp = (UChar32)cp;
return SUCCESS;
}
static void php_intl_idn_to(INTERNAL_FUNCTION_PARAMETERS,
const char *domain, int32_t domain_len, uint32_t option, int mode)
{
UChar* ustring = NULL;
int ustring_len = 0;
UErrorCode status;
char *converted_utf8;
size_t converted_utf8_len;
UChar converted[MAXPATHLEN];
int32_t converted_ret_len;
/* convert the string to UTF-16. */
status = U_ZERO_ERROR;
intl_convert_utf8_to_utf16(&ustring, &ustring_len, domain, domain_len, &status);
if (U_FAILURE(status)) {
intl_error_set_code(NULL, status);
/* Set error messages. */
intl_error_set_custom_msg( NULL, "Error converting input string to UTF-16", 0 );
if (ustring) {
efree(ustring);
}
RETURN_FALSE;
} else {
UParseError parse_error;
status = U_ZERO_ERROR;
if (mode == INTL_IDN_TO_ASCII) {
converted_ret_len = uidna_IDNToASCII(ustring, ustring_len, converted, MAXPATHLEN, (int32_t)option, &parse_error, &status);
} else {
converted_ret_len = uidna_IDNToUnicode(ustring, ustring_len, converted, MAXPATHLEN, (int32_t)option, &parse_error, &status);
}
efree(ustring);
if (U_FAILURE(status)) {
intl_error_set( NULL, status, "idn_to_ascii: cannot convert to ASCII", 0 );
RETURN_FALSE;
}
status = U_ZERO_ERROR;
intl_convert_utf16_to_utf8(&converted_utf8, &converted_utf8_len, converted, converted_ret_len, &status);
if (U_FAILURE(status)) {
/* Set global error code. */
intl_error_set_code(NULL, status);
/* Set error messages. */
intl_error_set_custom_msg( NULL, "Error converting output string to UTF-8", 0 );
efree(converted_utf8);
RETURN_FALSE;
}
}
/* return the allocated string, not a duplicate */
RETVAL_STRINGL(converted_utf8, converted_utf8_len);
//????
efree(converted_utf8);
}
static UBool enumCharNames_callback(enumCharNames_data *context,
UChar32 code, UCharNameChoice nameChoice,
const char *name, int32_t length) {
zval retval;
zval args[3];
ZVAL_NULL(&retval);
ZVAL_LONG(&args[0], code);
ZVAL_LONG(&args[1], nameChoice);
ZVAL_STRINGL(&args[2], name, length);
context->fci.retval = &retval;
context->fci.param_count = 3;
context->fci.params = args;
if (zend_call_function(&context->fci, &context->fci_cache) == FAILURE) {
intl_error_set_code(NULL, U_INTERNAL_PROGRAM_ERROR);
intl_error_set_custom_msg(NULL, "enumCharNames callback failed", 0);
zval_dtor(&retval);
zval_dtor(&args[2]);
return 0;
}
zval_dtor(&retval);
zval_dtor(&args[2]);
return 1;
}
static UBool enumCharType_callback(enumCharType_data *context,
UChar32 start, UChar32 limit,
UCharCategory type) {
zval retval;
zval args[3];
ZVAL_NULL(&retval);
/* Note that $start is INclusive, while $limit is EXclusive
* Therefore (0, 32, 15) means CPs 0..31 are of type 15
*/
ZVAL_LONG(&args[0], start);
ZVAL_LONG(&args[1], limit);
ZVAL_LONG(&args[2], type);
context->fci.retval = &retval;
context->fci.param_count = 3;
context->fci.params = args;
if (zend_call_function(&context->fci, &context->fci_cache) == FAILURE) {
intl_error_set_code(NULL, U_INTERNAL_PROGRAM_ERROR);
intl_errors_set_custom_msg(NULL, "enumCharTypes callback failed", 0);
zval_dtor(&retval);
return 0;
}
zval_dtor(&retval);
return 1;
}
/* {{{ clone handler for Transliterator */
static zend_object *Transliterator_clone_obj( zval *object )
{
Transliterator_object *to_orig,
*to_new;
zend_object *ret_val;
intl_error_reset( NULL );
to_orig = Z_INTL_TRANSLITERATOR_P( object );
intl_error_reset( INTL_DATA_ERROR_P( to_orig ) );
ret_val = Transliterator_ce_ptr->create_object( Z_OBJCE_P( object ) );
to_new = php_intl_transliterator_fetch_object( ret_val );
zend_objects_clone_members( &to_new->zo, &to_orig->zo );
if( to_orig->utrans != NULL )
{
zval tempz; /* dummy zval to pass to transliterator_object_construct */
/* guaranteed to return NULL if it fails */
UTransliterator *utrans = utrans_clone( to_orig->utrans, TRANSLITERATOR_ERROR_CODE_P( to_orig ) );
if( U_FAILURE( TRANSLITERATOR_ERROR_CODE( to_orig ) ) )
goto err;
ZVAL_OBJ(&tempz, ret_val);
transliterator_object_construct( &tempz, utrans,
TRANSLITERATOR_ERROR_CODE_P( to_orig ) );
if( U_FAILURE( TRANSLITERATOR_ERROR_CODE( to_orig ) ) )
{
zend_string *err_msg;
err:
if( utrans != NULL )
transliterator_object_destroy( to_new );
/* set the error anyway, in case in the future we decide not to
* throw an error. It also helps build the error message */
intl_error_set_code( NULL, INTL_DATA_ERROR_CODE( to_orig ) );
intl_errors_set_custom_msg( TRANSLITERATOR_ERROR_P( to_orig ),
"Could not clone transliterator", 0 );
err_msg = intl_error_get_message( TRANSLITERATOR_ERROR_P( to_orig ) );
zend_throw_error( NULL, "%s", ZSTR_VAL(err_msg) );
zend_string_free( err_msg ); /* if it's changed into a warning */
/* do not destroy tempz; we need to return something */
}
}
else
{
/* We shouldn't have unconstructed objects in the first place */
php_error_docref( NULL, E_WARNING,
"Cloning unconstructed transliterator." );
}
return ret_val;
}
/* {{{ resourcebundle_array_fetch */
static void resourcebundle_array_fetch(zval *object, zval *offset, zval *return_value, int fallback)
{
int32_t meindex = 0;
char * mekey = NULL;
zend_bool is_numeric = 0;
char *pbuf;
ResourceBundle_object *rb;
intl_error_reset( NULL );
RESOURCEBUNDLE_METHOD_FETCH_OBJECT;
if(Z_TYPE_P(offset) == IS_LONG) {
is_numeric = 1;
meindex = (int32_t)Z_LVAL_P(offset);
rb->child = ures_getByIndex( rb->me, meindex, rb->child, &INTL_DATA_ERROR_CODE(rb) );
} else if(Z_TYPE_P(offset) == IS_STRING) {
mekey = Z_STRVAL_P(offset);
rb->child = ures_getByKey(rb->me, mekey, rb->child, &INTL_DATA_ERROR_CODE(rb) );
} else {
intl_errors_set(INTL_DATA_ERROR_P(rb), U_ILLEGAL_ARGUMENT_ERROR,
"resourcebundle_get: index should be integer or string", 0);
RETURN_NULL();
}
intl_error_set_code( NULL, INTL_DATA_ERROR_CODE(rb) );
if (U_FAILURE(INTL_DATA_ERROR_CODE(rb))) {
if (is_numeric) {
spprintf( &pbuf, 0, "Cannot load resource element %d", meindex );
} else {
spprintf( &pbuf, 0, "Cannot load resource element '%s'", mekey );
}
intl_errors_set_custom_msg( INTL_DATA_ERROR_P(rb), pbuf, 1 );
efree(pbuf);
RETURN_NULL();
}
if (!fallback && (INTL_DATA_ERROR_CODE(rb) == U_USING_FALLBACK_WARNING || INTL_DATA_ERROR_CODE(rb) == U_USING_DEFAULT_WARNING)) {
UErrorCode icuerror;
const char * locale = ures_getLocaleByType( rb->me, ULOC_ACTUAL_LOCALE, &icuerror );
if (is_numeric) {
spprintf( &pbuf, 0, "Cannot load element %d without fallback from to %s", meindex, locale );
} else {
spprintf( &pbuf, 0, "Cannot load element '%s' without fallback from to %s", mekey, locale );
}
intl_errors_set_custom_msg( INTL_DATA_ERROR_P(rb), pbuf, 1 );
efree(pbuf);
RETURN_NULL();
}
resourcebundle_extract_value( return_value, rb );
}
/* like INTL_CHECK_STATUS, but as a function and varying the name of the func */
static int php_intl_idn_check_status(UErrorCode err, const char *msg)
{
intl_error_set_code(NULL, err);
if (U_FAILURE(err)) {
char *buff;
spprintf(&buff, 0, "%s: %s",
get_active_function_name(),
msg);
intl_error_set_custom_msg(NULL, buff, 1);
efree(buff);
return FAILURE;
}
return SUCCESS;
}
/* like INTL_CHECK_STATUS, but as a function and varying the name of the func */
static int php_intl_idn_check_status(UErrorCode err, const char *msg, int mode)
{
intl_error_set_code(NULL, err);
if (U_FAILURE(err)) {
char *buff;
spprintf(&buff, 0, "%s: %s",
mode == INTL_IDN_TO_ASCII ? "idn_to_ascii" : "idn_to_utf8",
msg);
intl_error_set_custom_msg(NULL, buff, 1);
efree(buff);
return FAILURE;
}
return SUCCESS;
}
static void regexp_ctor(INTERNAL_FUNCTION_PARAMETERS)
{
zval *object;
Regexp_object *ro;
char *pattern;
int32_t pattern_len;
UChar *upattern = NULL;
int32_t upattern_len = 0;
zval *zflags = NULL;
uint32_t flags = 0;
UParseError pe = { -1, -1, {0}, {0} };
intl_error_reset(NULL TSRMLS_CC);
object = return_value;
if (FAILURE == zend_parse_parameters(ZEND_NUM_ARGS() TSRMLS_CC, "s|z", &pattern, &pattern_len, &zflags)) {
intl_error_set(NULL, U_ILLEGAL_ARGUMENT_ERROR, "bad arguments", 0 TSRMLS_CC);
zval_dtor(object);
RETURN_NULL();
}
if (NULL != zflags) {
switch (Z_TYPE_P(zflags)) {
case IS_LONG:
flags = (uint32_t) Z_LVAL_P(zflags);
break;
case IS_STRING:
{
const char *p;
for (p = Z_STRVAL_P(zflags); '\0' != *p; p++) {
switch (*p) {
case 'i': flags |= UREGEX_CASE_INSENSITIVE; break;
case 'm': flags |= UREGEX_MULTILINE; break;
case 's': flags |= UREGEX_DOTALL; break;
case 'x': flags |= UREGEX_COMMENTS; break;
case 'w': flags |= UREGEX_UWORD; break;
default:
intl_error_set(NULL, U_ILLEGAL_ARGUMENT_ERROR, "invalid modifier", 0 TSRMLS_CC);
zval_dtor(object);
RETURN_NULL();
}
}
break;
}
default:
intl_error_set(NULL, U_ILLEGAL_ARGUMENT_ERROR, "bad arguments", 0 TSRMLS_CC);
zval_dtor(object);
RETURN_NULL();
}
}
ro = (Regexp_object *) zend_object_store_get_object(object TSRMLS_CC);
intl_convert_utf8_to_utf16(&upattern, &upattern_len, pattern, pattern_len, REGEXP_ERROR_CODE_P(ro));
INTL_CTOR_CHECK_STATUS(ro, "string conversion of pattern to UTF-16 failed");
ro->uregex = uregex_open(upattern, upattern_len, flags, &pe, REGEXP_ERROR_CODE_P(ro));
efree(upattern);
if (U_FAILURE(REGEXP_ERROR_CODE(ro))) {
intl_error_set_code(NULL, REGEXP_ERROR_CODE(ro) TSRMLS_CC);
if (-1 != pe.line) {
regexp_parse_error_to_string(pe, pattern, pattern_len);
} else {
intl_error_set_custom_msg(NULL, "unable to compile ICU regular expression", 0 TSRMLS_CC);
}
zval_dtor(object);
RETURN_NULL();
}
}
static void php_intl_idn_to(INTERNAL_FUNCTION_PARAMETERS,
const zend_string *domain, uint32_t option, int mode)
{
UChar* ustring = NULL;
int ustring_len = 0;
UErrorCode status;
zend_string *u8str;
/* convert the string to UTF-16. */
status = U_ZERO_ERROR;
intl_convert_utf8_to_utf16(&ustring, &ustring_len, ZSTR_VAL(domain), ZSTR_LEN(domain), &status);
if (U_FAILURE(status)) {
intl_error_set_code(NULL, status);
/* Set error messages. */
intl_error_set_custom_msg( NULL, "Error converting input string to UTF-16", 0 );
if (ustring) {
efree(ustring);
}
RETURN_FALSE;
} else {
UChar converted[MAXPATHLEN];
int32_t converted_ret_len;
status = U_ZERO_ERROR;
#if U_ICU_VERSION_MAJOR_NUM >= 55
UIDNAInfo info = UIDNA_INFO_INITIALIZER;
UIDNA *idna = uidna_openUTS46((int32_t)option, &status);
if (U_FAILURE(status)) {
intl_error_set( NULL, status, "idn_to_ascii: failed to create an UIDNA instance", 0 );
RETURN_FALSE;
}
if (mode == INTL_IDN_TO_ASCII) {
converted_ret_len = uidna_nameToASCII(idna, ustring, ustring_len, converted, MAXPATHLEN, &info, &status);
} else {
converted_ret_len = uidna_nameToUnicode(idna, ustring, ustring_len, converted, MAXPATHLEN, &info, &status);
}
uidna_close(idna);
#else
UParseError parse_error;
if (mode == INTL_IDN_TO_ASCII) {
converted_ret_len = uidna_IDNToASCII(ustring, ustring_len, converted, MAXPATHLEN, (int32_t)option, &parse_error, &status);
} else {
converted_ret_len = uidna_IDNToUnicode(ustring, ustring_len, converted, MAXPATHLEN, (int32_t)option, &parse_error, &status);
}
#endif
efree(ustring);
if (U_FAILURE(status)) {
intl_error_set( NULL, status, "idn_to_ascii: cannot convert to ASCII", 0 );
RETURN_FALSE;
}
status = U_ZERO_ERROR;
u8str = intl_convert_utf16_to_utf8(converted, converted_ret_len, &status);
if (!u8str) {
/* Set global error code. */
intl_error_set_code(NULL, status);
/* Set error messages. */
intl_error_set_custom_msg( NULL, "Error converting output string to UTF-8", 0 );
RETURN_FALSE;
}
}
/* return the allocated string, not a duplicate */
RETVAL_NEW_STR(u8str);
}
static int create_transliterator( char *str_id, int str_id_len, zend_long direction, zval *object )
{
Transliterator_object *to;
UChar *ustr_id = NULL;
int32_t ustr_id_len = 0;
UTransliterator *utrans;
UParseError parse_error = {0, -1};
intl_error_reset( NULL );
if( ( direction != TRANSLITERATOR_FORWARD ) && (direction != TRANSLITERATOR_REVERSE ) )
{
intl_error_set( NULL, U_ILLEGAL_ARGUMENT_ERROR,
"transliterator_create: invalid direction", 0 );
return FAILURE;
}
object_init_ex( object, Transliterator_ce_ptr );
TRANSLITERATOR_METHOD_FETCH_OBJECT_NO_CHECK; /* fetch zend object from zval "object" into "to" */
/* Convert transliterator id to UTF-16 */
intl_convert_utf8_to_utf16( &ustr_id, &ustr_id_len, str_id, str_id_len, TRANSLITERATOR_ERROR_CODE_P( to ) );
if( U_FAILURE( TRANSLITERATOR_ERROR_CODE( to ) ) )
{
intl_error_set_code( NULL, TRANSLITERATOR_ERROR_CODE( to ) );
intl_error_set_custom_msg( NULL, "String conversion of id to UTF-16 failed", 0 );
zval_dtor( object );
return FAILURE;
}
/* Open ICU Transliterator. */
utrans = utrans_openU( ustr_id, ustr_id_len, (UTransDirection ) direction,
NULL, -1, &parse_error, TRANSLITERATOR_ERROR_CODE_P( to ) );
if (ustr_id) {
efree( ustr_id );
}
if( U_FAILURE( TRANSLITERATOR_ERROR_CODE( to ) ) )
{
char *buf = NULL;
intl_error_set_code( NULL, TRANSLITERATOR_ERROR_CODE( to ) );
spprintf( &buf, 0, "transliterator_create: unable to open ICU transliterator"
" with id \"%s\"", str_id );
if( buf == NULL ) {
intl_error_set_custom_msg( NULL,
"transliterator_create: unable to open ICU transliterator", 0 );
}
else
{
intl_error_set_custom_msg( NULL, buf, /* copy message */ 1 );
efree( buf );
}
zval_dtor( object );
return FAILURE;
}
transliterator_object_construct( object, utrans, TRANSLITERATOR_ERROR_CODE_P( to ) );
/* no need to close the transliterator manually on construction error */
if( U_FAILURE( TRANSLITERATOR_ERROR_CODE( to ) ) )
{
intl_error_set_code( NULL, TRANSLITERATOR_ERROR_CODE( to ) );
intl_error_set_custom_msg( NULL,
"transliterator_create: internal constructor call failed", 0 );
zval_dtor( object );
return FAILURE;
}
return SUCCESS;
}
/* {{{ void intl_error_set( intl_error* err, UErrorCode code, char* msg, int copyMsg )
* Set error code and message.
*/
void intl_error_set( intl_error* err, UErrorCode code, char* msg, int copyMsg )
{
intl_error_set_code( err, code );
intl_error_set_custom_msg( err, msg, copyMsg );
}
static void php_intl_idn_to(INTERNAL_FUNCTION_PARAMETERS, int mode)
{
unsigned char* domain;
int domain_len;
long option = 0;
UChar* ustring = NULL;
int ustring_len = 0;
UErrorCode status;
char *converted_utf8;
int32_t converted_utf8_len;
UChar converted[MAXPATHLEN];
int32_t converted_ret_len;
if (zend_parse_parameters(ZEND_NUM_ARGS() TSRMLS_CC, "s|l", (char **)&domain, &domain_len, &option) == FAILURE) {
return;
}
if (domain_len < 1) {
intl_error_set( NULL, U_ILLEGAL_ARGUMENT_ERROR, "idn_to_ascii: empty domain name", 0 TSRMLS_CC );
RETURN_FALSE;
}
/* convert the string to UTF-16. */
status = U_ZERO_ERROR;
intl_convert_utf8_to_utf16(&ustring, &ustring_len, (char*) domain, domain_len, &status );
if (U_FAILURE(status)) {
intl_error_set_code(NULL, status TSRMLS_CC);
/* Set error messages. */
intl_error_set_custom_msg( NULL, "Error converting input string to UTF-16", 0 TSRMLS_CC );
efree(ustring);
RETURN_FALSE;
} else {
UParseError parse_error;
status = U_ZERO_ERROR;
if (mode == INTL_IDN_TO_ASCII) {
converted_ret_len = uidna_IDNToASCII(ustring, ustring_len, converted, MAXPATHLEN, (int32_t)option, &parse_error, &status);
} else {
converted_ret_len = uidna_IDNToUnicode(ustring, ustring_len, converted, MAXPATHLEN, (int32_t)option, &parse_error, &status);
}
efree(ustring);
if (U_FAILURE(status)) {
intl_error_set( NULL, status, "idn_to_ascii: cannot convert to ASCII", 0 TSRMLS_CC );
RETURN_FALSE;
}
status = U_ZERO_ERROR;
intl_convert_utf16_to_utf8(&converted_utf8, &converted_utf8_len, converted, converted_ret_len, &status);
if (U_FAILURE(status)) {
/* Set global error code. */
intl_error_set_code(NULL, status TSRMLS_CC);
/* Set error messages. */
intl_error_set_custom_msg( NULL, "Error converting output string to UTF-8", 0 TSRMLS_CC );
efree(converted_utf8);
RETURN_FALSE;
}
}
/* return the allocated string, not a duplicate */
RETURN_STRINGL(((char *)converted_utf8), converted_utf8_len, 0);
}
/* {{{ grapheme_strpos_utf16 - strrpos using utf16*/
int
grapheme_strpos_utf16(unsigned char *haystack, int32_t haystack_len, unsigned char*needle, int32_t needle_len, int32_t offset, int32_t *puchar_pos, int f_ignore_case TSRMLS_DC)
{
UChar *uhaystack, *puhaystack, *uneedle;
int32_t uhaystack_len, uneedle_len;
int ret_pos;
unsigned char u_break_iterator_buffer[U_BRK_SAFECLONE_BUFFERSIZE];
UBreakIterator* bi;
UErrorCode status;
*puchar_pos = -1;
/* convert the strings to UTF-16. */
uhaystack = NULL;
uhaystack_len = 0;
status = U_ZERO_ERROR;
intl_convert_utf8_to_utf16(&uhaystack, &uhaystack_len, (char *) haystack, haystack_len, &status );
if ( U_FAILURE( status ) ) {
/* Set global error code. */
intl_error_set_code( NULL, status TSRMLS_CC );
/* Set error messages. */
intl_error_set_custom_msg( NULL, "Error converting input string to UTF-16", 0 TSRMLS_CC );
efree( uhaystack );
return -1;
}
/* get a pointer to the haystack taking into account the offset */
bi = NULL;
status = U_ZERO_ERROR;
bi = grapheme_get_break_iterator(u_break_iterator_buffer, &status TSRMLS_CC );
puhaystack = grapheme_get_haystack_offset(bi, uhaystack, uhaystack_len, offset);
uhaystack_len = (uhaystack_len - ( puhaystack - uhaystack));
if ( NULL == puhaystack ) {
intl_error_set( NULL, U_ILLEGAL_ARGUMENT_ERROR, "grapheme_strpos: Offset not contained in string", 1 TSRMLS_CC );
efree( uhaystack );
ubrk_close (bi);
return -1;
}
if ( f_ignore_case ) {
grapheme_intl_case_fold(&uhaystack, &puhaystack, &uhaystack_len, &status );
}
uneedle = NULL;
uneedle_len = 0;
status = U_ZERO_ERROR;
intl_convert_utf8_to_utf16(&uneedle, &uneedle_len, (char *) needle, needle_len, &status );
if ( U_FAILURE( status ) ) {
/* Set global error code. */
intl_error_set_code( NULL, status TSRMLS_CC );
/* Set error messages. */
intl_error_set_custom_msg( NULL, "Error converting input string to UTF-16", 0 TSRMLS_CC );
efree( uhaystack );
efree( uneedle );
ubrk_close (bi);
return -1;
}
if ( f_ignore_case ) {
grapheme_intl_case_fold(&uneedle, &uneedle, &uneedle_len, &status );
}
ret_pos = grapheme_memnstr_grapheme(bi, puhaystack, uneedle, uneedle_len, puhaystack + uhaystack_len );
*puchar_pos = ubrk_current(bi);
efree( uhaystack );
efree( uneedle );
ubrk_close (bi);
return ret_pos;
}
/* {{{ grapheme_strrpos_utf16 - strrpos using utf16 */
int
grapheme_strrpos_utf16(unsigned char *haystack, int32_t haystack_len, unsigned char*needle, int32_t needle_len, int32_t offset, int f_ignore_case TSRMLS_DC)
{
UChar *uhaystack, *puhaystack, *uhaystack_end, *uneedle;
int32_t uhaystack_len, uneedle_len;
UErrorCode status;
unsigned char u_break_iterator_buffer[U_BRK_SAFECLONE_BUFFERSIZE];
UBreakIterator* bi = NULL;
int ret_pos, pos;
/* convert the strings to UTF-16. */
uhaystack = NULL;
uhaystack_len = 0;
status = U_ZERO_ERROR;
intl_convert_utf8_to_utf16(&uhaystack, &uhaystack_len, (char *) haystack, haystack_len, &status );
if ( U_FAILURE( status ) ) {
/* Set global error code. */
intl_error_set_code( NULL, status TSRMLS_CC );
/* Set error messages. */
intl_error_set_custom_msg( NULL, "Error converting input string to UTF-16", 0 TSRMLS_CC );
efree( uhaystack );
return -1;
}
if ( f_ignore_case ) {
grapheme_intl_case_fold(&uhaystack, &uhaystack, &uhaystack_len, &status );
}
/* get a pointer to the haystack taking into account the offset */
bi = NULL;
status = U_ZERO_ERROR;
bi = grapheme_get_break_iterator(u_break_iterator_buffer, &status TSRMLS_CC );
puhaystack = grapheme_get_haystack_offset(bi, uhaystack, uhaystack_len, offset);
if ( NULL == puhaystack ) {
intl_error_set( NULL, U_ILLEGAL_ARGUMENT_ERROR, "grapheme_strpos: Offset not contained in string", 1 TSRMLS_CC );
efree( uhaystack );
ubrk_close (bi);
return -1;
}
uneedle = NULL;
uneedle_len = 0;
status = U_ZERO_ERROR;
intl_convert_utf8_to_utf16(&uneedle, &uneedle_len, (char *) needle, needle_len, &status );
if ( U_FAILURE( status ) ) {
/* Set global error code. */
intl_error_set_code( NULL, status TSRMLS_CC );
/* Set error messages. */
intl_error_set_custom_msg( NULL, "Error converting input string to UTF-16", 0 TSRMLS_CC );
efree( uhaystack );
efree( uneedle );
ubrk_close (bi);
return -1;
}
if ( f_ignore_case ) {
grapheme_intl_case_fold(&uneedle, &uneedle, &uneedle_len, &status );
}
ret_pos = -1; /* -1 represents 'not found' */
/* back up until there's needle_len characters to compare */
uhaystack_end = uhaystack + uhaystack_len;
pos = ubrk_last(bi);
puhaystack = uhaystack + pos;
while ( uhaystack_end - puhaystack < uneedle_len ) {
pos = ubrk_previous(bi);
if ( UBRK_DONE == pos ) {
break;
}
puhaystack = uhaystack + pos;
}
/* is there enough haystack left to hold the needle? */
if ( ( uhaystack_end - puhaystack ) < uneedle_len ) {
/* not enough, not found */
goto exit;
}
while ( UBRK_DONE != pos ) {
if (!u_memcmp(uneedle, puhaystack, uneedle_len)) { /* needle_len - 1 in zend memnstr? */
/* does the grapheme in the haystack end at the same place as the last grapheme in the needle? */
if ( ubrk_isBoundary(bi, pos + uneedle_len) ) {
/* found it, get grapheme count offset */
ret_pos = grapheme_count_graphemes(bi, uhaystack, pos);
break;
}
/* set position back */
ubrk_isBoundary(bi, pos);
}
pos = ubrk_previous(bi);
puhaystack = uhaystack + pos;
}
exit:
efree( uhaystack );
efree( uneedle );
ubrk_close (bi);
return ret_pos;
}
static void _php_intlrbbi_constructor_body(INTERNAL_FUNCTION_PARAMETERS)
{
zval *object = getThis();
char *rules;
int rules_len;
zend_bool compiled = 0;
UErrorCode status = U_ZERO_ERROR;
intl_error_reset(NULL TSRMLS_CC);
if (zend_parse_parameters(ZEND_NUM_ARGS() TSRMLS_CC, "s|b",
&rules, &rules_len, &compiled) == FAILURE) {
intl_error_set(NULL, U_ILLEGAL_ARGUMENT_ERROR,
"rbbi_create_instance: bad arguments", 0 TSRMLS_CC);
RETURN_NULL();
}
// instantiation of ICU object
RuleBasedBreakIterator *rbbi;
if (!compiled) {
UnicodeString rulesStr;
UParseError parseError = UParseError();
if (intl_stringFromChar(rulesStr, rules, rules_len, &status)
== FAILURE) {
intl_error_set(NULL, U_ILLEGAL_ARGUMENT_ERROR,
"rbbi_create_instance: rules were not a valid UTF-8 string",
0 TSRMLS_CC);
RETURN_NULL();
}
rbbi = new RuleBasedBreakIterator(rulesStr, parseError, status);
intl_error_set_code(NULL, status TSRMLS_CC);
if (U_FAILURE(status)) {
char *msg;
smart_str parse_error_str;
parse_error_str = intl_parse_error_to_string(&parseError);
spprintf(&msg, 0, "rbbi_create_instance: unable to create "
"RuleBasedBreakIterator from rules (%s)", parse_error_str.c);
smart_str_free(&parse_error_str);
intl_error_set_custom_msg(NULL, msg, 1 TSRMLS_CC);
efree(msg);
delete rbbi;
RETURN_NULL();
}
} else { // compiled
#if U_ICU_VERSION_MAJOR_NUM * 10 + U_ICU_VERSION_MINOR_NUM >= 48
rbbi = new RuleBasedBreakIterator((uint8_t*)rules, rules_len, status);
if (U_FAILURE(status)) {
intl_error_set(NULL, status, "rbbi_create_instance: unable to "
"create instance from compiled rules", 0 TSRMLS_CC);
delete rbbi;
RETURN_NULL();
}
#else
intl_error_set(NULL, U_UNSUPPORTED_ERROR, "rbbi_create_instance: "
"compiled rules require ICU >= 4.8", 0 TSRMLS_CC);
RETURN_NULL();
#endif
}
breakiterator_object_create(return_value, rbbi TSRMLS_CC);
}