/**
* This is the internal function which does processing on behalf of
* both sscanf() and fscanf()
*
* parameters :
* string literal string to be processed
* format format string
* return_value set with the results of the scan
*/
int string_sscanf(const char *string, const char *format, int numVars,
Variant &return_value) {
int nconversions;
int totalVars = -1;
int64_t value;
char *end;
const char *baseString;
char op = 0;
int base = 0;
int underflow = 0;
size_t width;
long (*fn)(const char *, char **, int) = nullptr;
const char *ch;
char sch;
int flags;
char buf[64]; /* Temporary buffer to hold scanned number
* strings before they are passed to strtoul() */
Array returnArray;
/*
* Check for errors in the format string.
*/
if (ValidateFormat(format, numVars, &totalVars) != SCAN_SUCCESS) {
scan_set_error_return(numVars, return_value);
return SCAN_ERROR_INVALID_FORMAT;
}
baseString = string;
/*
* Iterate over the format string filling in the result objects until
* we reach the end of input, the end of the format string, or there
* is a mismatch.
*/
nconversions = 0;
while (*format != '\0') {
ch = format++;
flags = 0;
/*
* If we see whitespace in the format, skip whitespace in the string.
*/
if ( isspace( (int)*ch ) ) {
sch = *string;
while ( isspace( (int)sch ) ) {
if (*string == '\0') {
goto done;
}
string++;
sch = *string;
}
continue;
}
if (*ch != '%') {
literal:
if (*string == '\0') {
underflow = 1;
goto done;
}
sch = *string;
string++;
if (*ch != sch) {
goto done;
}
continue;
}
ch = format++;
if (*ch == '%') {
goto literal;
}
/*
* Check for assignment suppression ('*') or an XPG3-style
* assignment ('%n$').
*/
if (*ch == '*') {
flags |= SCAN_SUPPRESS;
ch = format++;
} else if ( isdigit(UCHAR(*ch))) {
value = strtoul(format-1, &end, 10);
if (*end == '$') {
format = end+1;
ch = format++;
}
}
/*
* Parse any width specifier.
*/
if ( isdigit(UCHAR(*ch))) {
char *endptr;
width = strtoul(format-1, &endptr, 10);
format = endptr;
ch = format++;
} else {
width = 0;
}
/*
* Ignore size specifier.
*/
if ((*ch == 'l') || (*ch == 'L') || (*ch == 'h')) {
ch = format++;
}
/*
* Handle the various field types.
*/
switch (*ch) {
case 'n':
if (!(flags & SCAN_SUPPRESS)) {
returnArray.append((int)(string - baseString));
}
nconversions++;
continue;
case 'd':
case 'D':
op = 'i';
base = 10;
fn = (long (*)(const char *, char **, int))strtol;
break;
case 'i':
op = 'i';
base = 0;
fn = (long (*)(const char *, char **, int))strtol;
break;
case 'o':
op = 'i';
base = 8;
fn = (long (*)(const char *, char **, int))strtol;
break;
case 'x':
case 'X':
op = 'i';
base = 16;
fn = (long (*)(const char *, char **, int))strtol;
break;
case 'u':
op = 'i';
base = 10;
flags |= SCAN_UNSIGNED;
fn = (long (*)(const char *, char **, int))strtoul;
break;
case 'f':
case 'e':
case 'E':
case 'g':
op = 'f';
break;
case 's':
op = 's';
break;
case 'c':
op = 's';
flags |= SCAN_NOSKIP;
/*-cc-*/
if (0 == width) {
width = 1;
}
/*-cc-*/
break;
case '[':
op = '[';
flags |= SCAN_NOSKIP;
break;
} /* switch */
/*
* At this point, we will need additional characters from the
* string to proceed.
*/
if (*string == '\0') {
underflow = 1;
goto done;
}
/*
* Skip any leading whitespace at the beginning of a field unless
* the format suppresses this behavior.
*/
if (!(flags & SCAN_NOSKIP)) {
while (*string != '\0') {
sch = *string;
if (! isspace((int)sch) ) {
break;
}
string++;
}
if (*string == '\0') {
underflow = 1;
goto done;
}
}
/*
* Perform the requested scanning operation.
*/
switch (op) {
case 'c':
case 's':
/*
* Scan a string up to width characters or whitespace.
*/
if (width == 0) {
width = (size_t) ~0;
}
end = (char*)string;
while (*end != '\0') {
sch = *end;
if ( isspace( (int)sch ) ) {
break;
}
end++;
if (--width == 0) {
break;
}
}
if (!(flags & SCAN_SUPPRESS)) {
returnArray.append(String(string, end-string, CopyString));
}
string = end;
break;
case '[': {
CharSet cset;
if (width == 0) {
width = (size_t) ~0;
}
end = (char*)string;
format = BuildCharSet(&cset, format);
while (*end != '\0') {
sch = *end;
if (!CharInSet(&cset, (int)sch)) {
break;
}
end++;
if (--width == 0) {
break;
}
}
ReleaseCharSet(&cset);
if (string == end) {
/*
* Nothing matched the range, stop processing
*/
goto done;
}
if (!(flags & SCAN_SUPPRESS)) {
returnArray.append(String(string, end-string, CopyString));
}
string = end;
break;
}
case 'i':
/*
* Scan an unsigned or signed integer.
*/
/*-cc-*/
buf[0] = '\0';
/*-cc-*/
if ((width == 0) || (width > sizeof(buf) - 1)) {
width = sizeof(buf) - 1;
}
flags |= SCAN_SIGNOK | SCAN_NODIGITS | SCAN_NOZERO;
for (end = buf; width > 0; width--) {
switch (*string) {
/*
* The 0 digit has special meaning at the beginning of
* a number. If we are unsure of the base, it
* indicates that we are in base 8 or base 16 (if it is
* followed by an 'x').
*/
case '0':
/*-cc-*/
if (base == 16) {
flags |= SCAN_XOK;
}
/*-cc-*/
if (base == 0) {
base = 8;
flags |= SCAN_XOK;
}
if (flags & SCAN_NOZERO) {
flags &= ~(SCAN_SIGNOK | SCAN_NODIGITS | SCAN_NOZERO);
} else {
flags &= ~(SCAN_SIGNOK | SCAN_XOK | SCAN_NODIGITS);
}
goto addToInt;
case '1': case '2': case '3': case '4':
case '5': case '6': case '7':
if (base == 0) {
base = 10;
}
flags &= ~(SCAN_SIGNOK | SCAN_XOK | SCAN_NODIGITS);
goto addToInt;
case '8': case '9':
if (base == 0) {
base = 10;
}
if (base <= 8) {
break;
}
flags &= ~(SCAN_SIGNOK | SCAN_XOK | SCAN_NODIGITS);
goto addToInt;
case 'A': case 'B': case 'C':
case 'D': case 'E': case 'F':
case 'a': case 'b': case 'c':
case 'd': case 'e': case 'f':
if (base <= 10) {
break;
}
flags &= ~(SCAN_SIGNOK | SCAN_XOK | SCAN_NODIGITS);
goto addToInt;
case '+': case '-':
if (flags & SCAN_SIGNOK) {
flags &= ~SCAN_SIGNOK;
goto addToInt;
}
break;
case 'x': case 'X':
if ((flags & SCAN_XOK) && (end == buf+1)) {
base = 16;
flags &= ~SCAN_XOK;
goto addToInt;
}
break;
}
/*
* We got an illegal character so we are done accumulating.
*/
break;
addToInt:
/*
* Add the character to the temporary buffer.
*/
*end++ = *string++;
if (*string == '\0') {
break;
}
}
/*
* Check to see if we need to back up because we only got a
* sign or a trailing x after a 0.
*/
if (flags & SCAN_NODIGITS) {
if (*string == '\0') {
underflow = 1;
}
goto done;
} else if (end[-1] == 'x' || end[-1] == 'X') {
end--;
string--;
}
/*
* Scan the value from the temporary buffer. If we are
* returning a large unsigned value, we have to convert it back
* to a string since PHP only supports signed values.
*/
if (!(flags & SCAN_SUPPRESS)) {
*end = '\0';
value = (int64_t) (*fn)(buf, nullptr, base);
if ((flags & SCAN_UNSIGNED) && (value < 0)) {
snprintf(buf, sizeof(buf), "%lu", (long)value); /* INTL: ISO digit */
returnArray.append(String(buf, CopyString));
} else {
returnArray.append(value);
}
}
break;
case 'f':
/*
* Scan a floating point number
*/
buf[0] = '\0'; /* call me pedantic */
if ((width == 0) || (width > sizeof(buf) - 1)) {
width = sizeof(buf) - 1;
}
flags |= SCAN_SIGNOK | SCAN_NODIGITS | SCAN_PTOK | SCAN_EXPOK;
for (end = buf; width > 0; width--) {
switch (*string) {
case '0': case '1': case '2': case '3':
case '4': case '5': case '6': case '7':
case '8': case '9':
flags &= ~(SCAN_SIGNOK | SCAN_NODIGITS);
goto addToFloat;
case '+':
case '-':
if (flags & SCAN_SIGNOK) {
flags &= ~SCAN_SIGNOK;
goto addToFloat;
}
break;
case '.':
if (flags & SCAN_PTOK) {
flags &= ~(SCAN_SIGNOK | SCAN_PTOK);
goto addToFloat;
}
break;
case 'e':
case 'E':
/*
* An exponent is not allowed until there has
* been at least one digit.
*/
if ((flags & (SCAN_NODIGITS | SCAN_EXPOK)) == SCAN_EXPOK) {
flags = (flags & ~(SCAN_EXPOK|SCAN_PTOK))
| SCAN_SIGNOK | SCAN_NODIGITS;
goto addToFloat;
}
break;
}
/*
* We got an illegal character so we are done accumulating.
*/
break;
addToFloat:
/*
* Add the character to the temporary buffer.
*/
*end++ = *string++;
if (*string == '\0') {
break;
}
}
/*
* Check to see if we need to back up because we saw a
* trailing 'e' or sign.
*/
if (flags & SCAN_NODIGITS) {
if (flags & SCAN_EXPOK) {
/*
* There were no digits at all so scanning has
* failed and we are done.
*/
if (*string == '\0') {
underflow = 1;
}
goto done;
}
/*
* We got a bad exponent ('e' and maybe a sign).
*/
end--;
string--;
if (*end != 'e' && *end != 'E') {
end--;
string--;
}
}
/*
* Scan the value from the temporary buffer.
*/
if (!(flags & SCAN_SUPPRESS)) {
double dvalue;
*end = '\0';
dvalue = strtod(buf, nullptr);
returnArray.append(dvalue);
}
break;
} /* switch (op) */
nconversions++;
} /* while (*format != '\0') */
done:
if (underflow && (0==nconversions)) {
scan_set_error_return(numVars, return_value);
return SCAN_ERROR_EOF;
} else if (nconversions < totalVars) {
/* TODO: not all elements converted. we need to prune the list - cc */
}
return_value = returnArray;
return SCAN_SUCCESS;
}
PHPAPI int php_sscanf_internal( char *string, char *format,
int argCount, zval *args,
int varStart, zval *return_value)
{
int numVars, nconversions, totalVars = -1;
int i, result;
zend_long value;
int objIndex;
char *end, *baseString;
zval *current;
char op = 0;
int base = 0;
int underflow = 0;
size_t width;
zend_long (*fn)() = NULL;
char *ch, sch;
int flags;
char buf[64]; /* Temporary buffer to hold scanned number
* strings before they are passed to strtoul() */
/* do some sanity checking */
if ((varStart > argCount) || (varStart < 0)){
varStart = SCAN_MAX_ARGS + 1;
}
numVars = argCount - varStart;
if (numVars < 0) {
numVars = 0;
}
#if 0
zend_printf("<br>in sscanf_internal : <br> string is \"%s\", format = \"%s\"<br> NumVars = %d. VarStart = %d<br>-------------------------<br>",
string, format, numVars, varStart);
#endif
/*
* Check for errors in the format string.
*/
if (ValidateFormat(format, numVars, &totalVars) != SCAN_SUCCESS) {
scan_set_error_return( numVars, return_value );
return SCAN_ERROR_INVALID_FORMAT;
}
objIndex = numVars ? varStart : 0;
/*
* If any variables are passed, make sure they are all passed by reference
*/
if (numVars) {
for (i = varStart;i < argCount;i++){
if ( ! Z_ISREF(args[ i ] ) ) {
php_error_docref(NULL, E_WARNING, "Parameter %d must be passed by reference", i);
scan_set_error_return(numVars, return_value);
return SCAN_ERROR_VAR_PASSED_BYVAL;
}
}
}
/*
* Allocate space for the result objects. Only happens when no variables
* are specified
*/
if (!numVars) {
zval tmp;
/* allocate an array for return */
array_init(return_value);
for (i = 0; i < totalVars; i++) {
ZVAL_NULL(&tmp);
if (add_next_index_zval(return_value, &tmp) == FAILURE) {
scan_set_error_return(0, return_value);
return FAILURE;
}
}
varStart = 0; /* Array index starts from 0 */
}
baseString = string;
/*
* Iterate over the format string filling in the result objects until
* we reach the end of input, the end of the format string, or there
* is a mismatch.
*/
nconversions = 0;
/* note ! - we need to limit the loop for objIndex to keep it in bounds */
while (*format != '\0') {
ch = format++;
flags = 0;
/*
* If we see whitespace in the format, skip whitespace in the string.
*/
if ( isspace( (int)*ch ) ) {
sch = *string;
while ( isspace( (int)sch ) ) {
if (*string == '\0') {
goto done;
}
string++;
sch = *string;
}
continue;
}
if (*ch != '%') {
literal:
if (*string == '\0') {
underflow = 1;
goto done;
}
sch = *string;
string++;
if (*ch != sch) {
goto done;
}
continue;
}
ch = format++;
if (*ch == '%') {
goto literal;
}
/*
* Check for assignment suppression ('*') or an XPG3-style
* assignment ('%n$').
*/
if (*ch == '*') {
flags |= SCAN_SUPPRESS;
ch = format++;
} else if ( isdigit(UCHAR(*ch))) {
value = ZEND_STRTOUL(format-1, &end, 10);
if (*end == '$') {
format = end+1;
ch = format++;
objIndex = varStart + value - 1;
}
}
/*
* Parse any width specifier.
*/
if ( isdigit(UCHAR(*ch))) {
width = ZEND_STRTOUL(format-1, &format, 10);
ch = format++;
} else {
width = 0;
}
/*
* Ignore size specifier.
*/
if ((*ch == 'l') || (*ch == 'L') || (*ch == 'h')) {
ch = format++;
}
/*
* Handle the various field types.
*/
switch (*ch) {
case 'n':
if (!(flags & SCAN_SUPPRESS)) {
if (numVars && objIndex >= argCount) {
break;
} else if (numVars) {
current = Z_REFVAL(args[objIndex++]);
zval_ptr_dtor(current);
ZVAL_LONG(current, (zend_long)(string - baseString) );
} else {
add_index_long(return_value, objIndex++, string - baseString);
}
}
nconversions++;
continue;
case 'd':
case 'D':
op = 'i';
base = 10;
fn = (zend_long (*)())ZEND_STRTOL_PTR;
break;
case 'i':
op = 'i';
base = 0;
fn = (zend_long (*)())ZEND_STRTOL_PTR;
break;
case 'o':
op = 'i';
base = 8;
fn = (zend_long (*)())ZEND_STRTOL_PTR;
break;
case 'x':
case 'X':
op = 'i';
base = 16;
fn = (zend_long (*)())ZEND_STRTOL_PTR;
break;
case 'u':
op = 'i';
base = 10;
flags |= SCAN_UNSIGNED;
fn = (zend_long (*)())ZEND_STRTOUL_PTR;
break;
case 'f':
case 'e':
case 'E':
case 'g':
op = 'f';
break;
case 's':
op = 's';
break;
case 'c':
op = 's';
flags |= SCAN_NOSKIP;
/*-cc-*/
if (0 == width) {
width = 1;
}
/*-cc-*/
break;
case '[':
op = '[';
flags |= SCAN_NOSKIP;
break;
} /* switch */
/*
* At this point, we will need additional characters from the
* string to proceed.
*/
if (*string == '\0') {
underflow = 1;
goto done;
}
/*
* Skip any leading whitespace at the beginning of a field unless
* the format suppresses this behavior.
*/
if (!(flags & SCAN_NOSKIP)) {
while (*string != '\0') {
sch = *string;
if (! isspace((int)sch) ) {
break;
}
string++;
}
if (*string == '\0') {
underflow = 1;
goto done;
}
}
/*
* Perform the requested scanning operation.
*/
switch (op) {
case 'c':
case 's':
/*
* Scan a string up to width characters or whitespace.
*/
if (width == 0) {
width = (size_t) ~0;
}
end = string;
while (*end != '\0') {
sch = *end;
if ( isspace( (int)sch ) ) {
break;
}
end++;
if (--width == 0) {
break;
}
}
if (!(flags & SCAN_SUPPRESS)) {
if (numVars && objIndex >= argCount) {
break;
} else if (numVars) {
current = Z_REFVAL(args[objIndex++]);
zval_ptr_dtor(current);
ZVAL_STRINGL(current, string, end-string);
} else {
add_index_stringl(return_value, objIndex++, string, end-string);
}
}
string = end;
break;
case '[': {
CharSet cset;
if (width == 0) {
width = (size_t) ~0;
}
end = string;
format = BuildCharSet(&cset, format);
while (*end != '\0') {
sch = *end;
if (!CharInSet(&cset, (int)sch)) {
break;
}
end++;
if (--width == 0) {
break;
}
}
ReleaseCharSet(&cset);
if (string == end) {
/*
* Nothing matched the range, stop processing
*/
goto done;
}
if (!(flags & SCAN_SUPPRESS)) {
if (numVars && objIndex >= argCount) {
break;
} else if (numVars) {
current = Z_REFVAL(args[objIndex++]);
zval_ptr_dtor(current);
ZVAL_STRINGL(current, string, end-string);
} else {
add_index_stringl(return_value, objIndex++, string, end-string);
}
}
string = end;
break;
}
/*
case 'c':
/ Scan a single character./
sch = *string;
string++;
if (!(flags & SCAN_SUPPRESS)) {
if (numVars) {
char __buf[2];
__buf[0] = sch;
__buf[1] = '\0';;
current = args[objIndex++];
zval_dtor(*current);
ZVAL_STRINGL( *current, __buf, 1);
} else {
add_index_stringl(return_value, objIndex++, &sch, 1);
}
}
break;
*/
case 'i':
/*
* Scan an unsigned or signed integer.
*/
/*-cc-*/
buf[0] = '\0';
/*-cc-*/
if ((width == 0) || (width > sizeof(buf) - 1)) {
width = sizeof(buf) - 1;
}
flags |= SCAN_SIGNOK | SCAN_NODIGITS | SCAN_NOZERO;
for (end = buf; width > 0; width--) {
switch (*string) {
/*
* The 0 digit has special meaning at the beginning of
* a number. If we are unsure of the base, it
* indicates that we are in base 8 or base 16 (if it is
* followed by an 'x').
*/
case '0':
/*-cc-*/
if (base == 16) {
flags |= SCAN_XOK;
}
/*-cc-*/
if (base == 0) {
base = 8;
flags |= SCAN_XOK;
}
if (flags & SCAN_NOZERO) {
flags &= ~(SCAN_SIGNOK | SCAN_NODIGITS | SCAN_NOZERO);
} else {
flags &= ~(SCAN_SIGNOK | SCAN_XOK | SCAN_NODIGITS);
}
goto addToInt;
case '1': case '2': case '3': case '4':
case '5': case '6': case '7':
if (base == 0) {
base = 10;
}
flags &= ~(SCAN_SIGNOK | SCAN_XOK | SCAN_NODIGITS);
goto addToInt;
case '8': case '9':
if (base == 0) {
base = 10;
}
if (base <= 8) {
break;
}
flags &= ~(SCAN_SIGNOK | SCAN_XOK | SCAN_NODIGITS);
goto addToInt;
case 'A': case 'B': case 'C':
case 'D': case 'E': case 'F':
case 'a': case 'b': case 'c':
case 'd': case 'e': case 'f':
if (base <= 10) {
break;
}
flags &= ~(SCAN_SIGNOK | SCAN_XOK | SCAN_NODIGITS);
goto addToInt;
case '+': case '-':
if (flags & SCAN_SIGNOK) {
flags &= ~SCAN_SIGNOK;
goto addToInt;
}
break;
case 'x': case 'X':
if ((flags & SCAN_XOK) && (end == buf+1)) {
base = 16;
flags &= ~SCAN_XOK;
goto addToInt;
}
break;
}
/*
* We got an illegal character so we are done accumulating.
*/
break;
addToInt:
/*
* Add the character to the temporary buffer.
*/
*end++ = *string++;
if (*string == '\0') {
break;
}
}
/*
* Check to see if we need to back up because we only got a
* sign or a trailing x after a 0.
*/
if (flags & SCAN_NODIGITS) {
if (*string == '\0') {
underflow = 1;
}
goto done;
} else if (end[-1] == 'x' || end[-1] == 'X') {
end--;
string--;
}
/*
* Scan the value from the temporary buffer. If we are
* returning a large unsigned value, we have to convert it back
* to a string since PHP only supports signed values.
*/
if (!(flags & SCAN_SUPPRESS)) {
*end = '\0';
value = (zend_long) (*fn)(buf, NULL, base);
if ((flags & SCAN_UNSIGNED) && (value < 0)) {
snprintf(buf, sizeof(buf), ZEND_ULONG_FMT, value); /* INTL: ISO digit */
if (numVars && objIndex >= argCount) {
break;
} else if (numVars) {
/* change passed value type to string */
current = Z_REFVAL(args[objIndex++]);
zval_ptr_dtor(current);
ZVAL_STRING(current, buf);
} else {
add_index_string(return_value, objIndex++, buf);
}
} else {
if (numVars && objIndex >= argCount) {
break;
} else if (numVars) {
current = Z_REFVAL(args[objIndex++]);
zval_ptr_dtor(current);
ZVAL_LONG(current, value);
} else {
add_index_long(return_value, objIndex++, value);
}
}
}
break;
case 'f':
/*
* Scan a floating point number
*/
buf[0] = '\0'; /* call me pedantic */
if ((width == 0) || (width > sizeof(buf) - 1)) {
width = sizeof(buf) - 1;
}
flags |= SCAN_SIGNOK | SCAN_NODIGITS | SCAN_PTOK | SCAN_EXPOK;
for (end = buf; width > 0; width--) {
switch (*string) {
case '0': case '1': case '2': case '3':
case '4': case '5': case '6': case '7':
case '8': case '9':
flags &= ~(SCAN_SIGNOK | SCAN_NODIGITS);
goto addToFloat;
case '+':
case '-':
if (flags & SCAN_SIGNOK) {
flags &= ~SCAN_SIGNOK;
goto addToFloat;
}
break;
case '.':
if (flags & SCAN_PTOK) {
flags &= ~(SCAN_SIGNOK | SCAN_PTOK);
goto addToFloat;
}
break;
case 'e':
case 'E':
/*
* An exponent is not allowed until there has
* been at least one digit.
*/
if ((flags & (SCAN_NODIGITS | SCAN_EXPOK)) == SCAN_EXPOK) {
flags = (flags & ~(SCAN_EXPOK|SCAN_PTOK))
| SCAN_SIGNOK | SCAN_NODIGITS;
goto addToFloat;
}
break;
}
/*
* We got an illegal character so we are done accumulating.
*/
break;
addToFloat:
/*
* Add the character to the temporary buffer.
*/
*end++ = *string++;
if (*string == '\0') {
break;
}
}
/*
* Check to see if we need to back up because we saw a
* trailing 'e' or sign.
*/
if (flags & SCAN_NODIGITS) {
if (flags & SCAN_EXPOK) {
/*
* There were no digits at all so scanning has
* failed and we are done.
*/
if (*string == '\0') {
underflow = 1;
}
goto done;
}
/*
* We got a bad exponent ('e' and maybe a sign).
*/
end--;
string--;
if (*end != 'e' && *end != 'E') {
end--;
string--;
}
}
/*
* Scan the value from the temporary buffer.
*/
if (!(flags & SCAN_SUPPRESS)) {
double dvalue;
*end = '\0';
dvalue = zend_strtod(buf, NULL);
if (numVars && objIndex >= argCount) {
break;
} else if (numVars) {
current = Z_REFVAL(args[objIndex++]);
zval_ptr_dtor(current);
ZVAL_DOUBLE(current, dvalue);
} else {
add_index_double(return_value, objIndex++, dvalue );
}
}
break;
} /* switch (op) */
nconversions++;
} /* while (*format != '\0') */
done:
result = SCAN_SUCCESS;
if (underflow && (0==nconversions)) {
scan_set_error_return( numVars, return_value );
result = SCAN_ERROR_EOF;
} else if (numVars) {
convert_to_long(return_value );
Z_LVAL_P(return_value) = nconversions;
} else if (nconversions < totalVars) {
/* TODO: not all elements converted. we need to prune the list - cc */
}
return result;
}
/* ARGSUSED */
int
Tcl_ScanObjCmd(
ClientData dummy, /* Not used. */
Tcl_Interp *interp, /* Current interpreter. */
int objc, /* Number of arguments. */
Tcl_Obj *const objv[]) /* Argument objects. */
{
const char *format;
int numVars, nconversions, totalVars = -1;
int objIndex, offset, i, result, code;
long value;
const char *string, *end, *baseString;
char op = 0;
int width, underflow = 0;
Tcl_WideInt wideValue;
Tcl_UniChar ch, sch;
Tcl_Obj **objs = NULL, *objPtr = NULL;
int flags;
char buf[513]; /* Temporary buffer to hold scanned number
* strings before they are passed to
* strtoul. */
if (objc < 3) {
Tcl_WrongNumArgs(interp, 1, objv,
"string format ?varName ...?");
return TCL_ERROR;
}
format = Tcl_GetStringFromObj(objv[2], NULL);
numVars = objc-3;
/*
* Check for errors in the format string.
*/
if (ValidateFormat(interp, format, numVars, &totalVars) == TCL_ERROR) {
return TCL_ERROR;
}
/*
* Allocate space for the result objects.
*/
if (totalVars > 0) {
objs = ckalloc(sizeof(Tcl_Obj *) * totalVars);
for (i = 0; i < totalVars; i++) {
objs[i] = NULL;
}
}
string = Tcl_GetStringFromObj(objv[1], NULL);
baseString = string;
/*
* Iterate over the format string filling in the result objects until we
* reach the end of input, the end of the format string, or there is a
* mismatch.
*/
objIndex = 0;
nconversions = 0;
while (*format != '\0') {
int parseFlag = TCL_PARSE_NO_WHITESPACE;
format += Tcl_UtfToUniChar(format, &ch);
flags = 0;
/*
* If we see whitespace in the format, skip whitespace in the string.
*/
if (Tcl_UniCharIsSpace(ch)) {
offset = Tcl_UtfToUniChar(string, &sch);
while (Tcl_UniCharIsSpace(sch)) {
if (*string == '\0') {
goto done;
}
string += offset;
offset = Tcl_UtfToUniChar(string, &sch);
}
continue;
}
if (ch != '%') {
literal:
if (*string == '\0') {
underflow = 1;
goto done;
}
string += Tcl_UtfToUniChar(string, &sch);
if (ch != sch) {
goto done;
}
continue;
}
format += Tcl_UtfToUniChar(format, &ch);
if (ch == '%') {
goto literal;
}
/*
* Check for assignment suppression ('*') or an XPG3-style assignment
* ('%n$').
*/
if (ch == '*') {
flags |= SCAN_SUPPRESS;
format += Tcl_UtfToUniChar(format, &ch);
} else if ((ch < 0x80) && isdigit(UCHAR(ch))) { /* INTL: "C" locale. */
char *formatEnd;
value = strtoul(format-1, &formatEnd, 10);/* INTL: "C" locale. */
if (*formatEnd == '$') {
format = formatEnd+1;
format += Tcl_UtfToUniChar(format, &ch);
objIndex = (int) value - 1;
}
}
/*
* Parse any width specifier.
*/
if ((ch < 0x80) && isdigit(UCHAR(ch))) { /* INTL: "C" locale. */
width = (int) strtoul(format-1, (char **) &format, 10);/* INTL: "C" locale. */
format += Tcl_UtfToUniChar(format, &ch);
} else {
width = 0;
}
/*
* Handle any size specifier.
*/
switch (ch) {
case 'l':
if (*format == 'l') {
flags |= SCAN_BIG;
format += 1;
format += Tcl_UtfToUniChar(format, &ch);
break;
}
case 'L':
flags |= SCAN_LONGER;
/*
* Fall through so we skip to the next character.
*/
case 'h':
format += Tcl_UtfToUniChar(format, &ch);
}
/*
* Handle the various field types.
*/
switch (ch) {
case 'n':
if (!(flags & SCAN_SUPPRESS)) {
objPtr = Tcl_NewIntObj(string - baseString);
Tcl_IncrRefCount(objPtr);
CLANG_ASSERT(objs);
objs[objIndex++] = objPtr;
}
nconversions++;
continue;
case 'd':
op = 'i';
parseFlag |= TCL_PARSE_DECIMAL_ONLY;
break;
case 'i':
op = 'i';
parseFlag |= TCL_PARSE_SCAN_PREFIXES;
break;
case 'o':
op = 'i';
parseFlag |= TCL_PARSE_OCTAL_ONLY | TCL_PARSE_SCAN_PREFIXES;
break;
case 'x':
case 'X':
op = 'i';
parseFlag |= TCL_PARSE_HEXADECIMAL_ONLY;
break;
case 'b':
op = 'i';
parseFlag |= TCL_PARSE_BINARY_ONLY;
break;
case 'u':
op = 'i';
parseFlag |= TCL_PARSE_DECIMAL_ONLY;
flags |= SCAN_UNSIGNED;
break;
case 'f':
case 'e':
case 'E':
case 'g':
case 'G':
op = 'f';
break;
case 's':
op = 's';
break;
case 'c':
op = 'c';
flags |= SCAN_NOSKIP;
break;
case '[':
op = '[';
flags |= SCAN_NOSKIP;
break;
}
/*
* At this point, we will need additional characters from the string
* to proceed.
*/
if (*string == '\0') {
underflow = 1;
goto done;
}
/*
* Skip any leading whitespace at the beginning of a field unless the
* format suppresses this behavior.
*/
if (!(flags & SCAN_NOSKIP)) {
while (*string != '\0') {
offset = Tcl_UtfToUniChar(string, &sch);
if (!Tcl_UniCharIsSpace(sch)) {
break;
}
string += offset;
}
if (*string == '\0') {
underflow = 1;
goto done;
}
}
/*
* Perform the requested scanning operation.
*/
switch (op) {
case 's':
/*
* Scan a string up to width characters or whitespace.
*/
if (width == 0) {
width = ~0;
}
end = string;
while (*end != '\0') {
offset = Tcl_UtfToUniChar(end, &sch);
if (Tcl_UniCharIsSpace(sch)) {
break;
}
end += offset;
if (--width == 0) {
break;
}
}
if (!(flags & SCAN_SUPPRESS)) {
objPtr = Tcl_NewStringObj(string, end-string);
Tcl_IncrRefCount(objPtr);
CLANG_ASSERT(objs);
objs[objIndex++] = objPtr;
}
string = end;
break;
case '[': {
CharSet cset;
if (width == 0) {
width = ~0;
}
end = string;
format = BuildCharSet(&cset, format);
while (*end != '\0') {
offset = Tcl_UtfToUniChar(end, &sch);
if (!CharInSet(&cset, (int)sch)) {
break;
}
end += offset;
if (--width == 0) {
break;
}
}
ReleaseCharSet(&cset);
if (string == end) {
/*
* Nothing matched the range, stop processing.
*/
goto done;
}
if (!(flags & SCAN_SUPPRESS)) {
objPtr = Tcl_NewStringObj(string, end-string);
Tcl_IncrRefCount(objPtr);
objs[objIndex++] = objPtr;
}
string = end;
break;
}
case 'c':
/*
* Scan a single Unicode character.
*/
string += Tcl_UtfToUniChar(string, &sch);
if (!(flags & SCAN_SUPPRESS)) {
objPtr = Tcl_NewIntObj((int)sch);
Tcl_IncrRefCount(objPtr);
CLANG_ASSERT(objs);
objs[objIndex++] = objPtr;
}
break;
case 'i':
/*
* Scan an unsigned or signed integer.
*/
objPtr = Tcl_NewLongObj(0);
Tcl_IncrRefCount(objPtr);
if (width == 0) {
width = ~0;
}
if (TCL_OK != TclParseNumber(NULL, objPtr, NULL, string, width,
&end, TCL_PARSE_INTEGER_ONLY | parseFlag)) {
Tcl_DecrRefCount(objPtr);
if (width < 0) {
if (*end == '\0') {
underflow = 1;
}
} else {
if (end == string + width) {
underflow = 1;
}
}
goto done;
}
string = end;
if (flags & SCAN_SUPPRESS) {
Tcl_DecrRefCount(objPtr);
break;
}
if (flags & SCAN_LONGER) {
if (Tcl_GetWideIntFromObj(NULL, objPtr, &wideValue) != TCL_OK) {
wideValue = ~(Tcl_WideUInt)0 >> 1; /* WIDE_MAX */
if (TclGetString(objPtr)[0] == '-') {
wideValue++; /* WIDE_MAX + 1 = WIDE_MIN */
}
}
if ((flags & SCAN_UNSIGNED) && (wideValue < 0)) {
sprintf(buf, "%" TCL_LL_MODIFIER "u",
(Tcl_WideUInt)wideValue);
Tcl_SetStringObj(objPtr, buf, -1);
} else {
Tcl_SetWideIntObj(objPtr, wideValue);
}
} else if (!(flags & SCAN_BIG)) {
if (TclGetLongFromObj(NULL, objPtr, &value) != TCL_OK) {
if (TclGetString(objPtr)[0] == '-') {
value = LONG_MIN;
} else {
value = LONG_MAX;
}
}
if ((flags & SCAN_UNSIGNED) && (value < 0)) {
sprintf(buf, "%lu", value); /* INTL: ISO digit */
Tcl_SetStringObj(objPtr, buf, -1);
} else {
Tcl_SetLongObj(objPtr, value);
}
}
objs[objIndex++] = objPtr;
break;
case 'f':
/*
* Scan a floating point number
*/
objPtr = Tcl_NewDoubleObj(0.0);
Tcl_IncrRefCount(objPtr);
if (width == 0) {
width = ~0;
}
if (TCL_OK != TclParseNumber(NULL, objPtr, NULL, string, width,
&end, TCL_PARSE_DECIMAL_ONLY | TCL_PARSE_NO_WHITESPACE)) {
Tcl_DecrRefCount(objPtr);
if (width < 0) {
if (*end == '\0') {
underflow = 1;
}
} else {
if (end == string + width) {
underflow = 1;
}
}
goto done;
} else if (flags & SCAN_SUPPRESS) {
Tcl_DecrRefCount(objPtr);
string = end;
} else {
double dvalue;
if (Tcl_GetDoubleFromObj(NULL, objPtr, &dvalue) != TCL_OK) {
#ifdef ACCEPT_NAN
if (objPtr->typePtr == &tclDoubleType) {
dvalue = objPtr->internalRep.doubleValue;
} else
#endif
{
Tcl_DecrRefCount(objPtr);
goto done;
}
}
Tcl_SetDoubleObj(objPtr, dvalue);
CLANG_ASSERT(objs);
objs[objIndex++] = objPtr;
string = end;
}
}