static REBSER *make_string(REBVAL *arg, REBOOL make)
{
REBSER *ser = 0;
// MAKE <type> 123
if (make && (IS_INTEGER(arg) || IS_DECIMAL(arg))) {
ser = Make_Binary(Int32s(arg, 0));
}
// MAKE/TO <type> <binary!>
else if (IS_BINARY(arg)) {
REBYTE *bp = VAL_BIN_DATA(arg);
REBCNT len = VAL_LEN(arg);
switch (What_UTF(bp, len)) {
case 0:
break;
case 8: // UTF-8 encoded
bp += 3;
len -= 3;
break;
default:
Trap0(RE_BAD_DECODE);
}
ser = Decode_UTF_String(bp, len, 8); // UTF-8
}
// MAKE/TO <type> <any-string>
else if (ANY_BINSTR(arg)) {
ser = Copy_String(VAL_SERIES(arg), VAL_INDEX(arg), VAL_LEN(arg));
}
// MAKE/TO <type> <any-word>
else if (ANY_WORD(arg)) {
ser = Copy_Mold_Value(arg, TRUE);
//ser = Append_UTF8(0, Get_Word_Name(arg), -1);
}
// MAKE/TO <type> #"A"
else if (IS_CHAR(arg)) {
ser = (VAL_CHAR(arg) > 0xff) ? Make_Unicode(2) : Make_Binary(2);
Append_Byte(ser, VAL_CHAR(arg));
}
// MAKE/TO <type> <any-value>
// else if (IS_NONE(arg)) {
// ser = Make_Binary(0);
// }
else
ser = Copy_Form_Value(arg, 1<<MOPT_TIGHT);
return ser;
}
static REBSER *MAKE_TO_String_Common(const REBVAL *arg)
{
REBSER *ser = 0;
// MAKE/TO <type> <binary!>
if (IS_BINARY(arg)) {
REBYTE *bp = VAL_BIN_AT(arg);
REBCNT len = VAL_LEN_AT(arg);
switch (What_UTF(bp, len)) {
case 0:
break;
case 8: // UTF-8 encoded
bp += 3;
len -= 3;
break;
default:
fail (Error(RE_BAD_UTF8));
}
ser = Decode_UTF_String(bp, len, 8); // UTF-8
}
// MAKE/TO <type> <any-string>
else if (ANY_BINSTR(arg)) {
ser = Copy_String_Slimming(VAL_SERIES(arg), VAL_INDEX(arg), VAL_LEN_AT(arg));
}
// MAKE/TO <type> <any-word>
else if (ANY_WORD(arg)) {
ser = Copy_Mold_Value(arg, 0 /* opts... MOPT_0? */);
}
// MAKE/TO <type> #"A"
else if (IS_CHAR(arg)) {
ser = (VAL_CHAR(arg) > 0xff) ? Make_Unicode(2) : Make_Binary(2);
Append_Codepoint_Raw(ser, VAL_CHAR(arg));
}
else
ser = Copy_Form_Value(arg, 1 << MOPT_TIGHT);
return ser;
}
x*/ void Modify_StringX(REBCNT action, REBVAL *string, REBVAL *arg)
/*
** Actions: INSERT, APPEND, CHANGE
**
** string [string!] {Series at point to insert}
** value [any-type!] {The value to insert}
** /part {Limits to a given length or position.}
** length [number! series! pair!]
** /only {Inserts a series as a series.}
** /dup {Duplicates the insert a specified number of times.}
** count [number! pair!]
**
***********************************************************************/
{
REBSER *series = VAL_SERIES(string);
REBCNT index = VAL_INDEX(string);
REBCNT tail = VAL_TAIL(string);
REBINT rlen; // length to be removed
REBINT ilen = 1; // length to be inserted
REBINT cnt = 1; // DUP count
REBINT size;
REBVAL *val;
REBSER *arg_ser = 0; // argument series
// Length of target (may modify index): (arg can be anything)
rlen = Partial1((action == A_CHANGE) ? string : arg, DS_ARG(AN_LENGTH));
index = VAL_INDEX(string);
if (action == A_APPEND || index > tail) index = tail;
// If the arg is not a string, then we need to create a string:
if (IS_BINARY(string)) {
if (IS_INTEGER(arg)) {
if (VAL_INT64(arg) > 255 || VAL_INT64(arg) < 0)
Trap_Range(arg);
arg_ser = Make_Binary(1);
Append_Byte(arg_ser, VAL_CHAR(arg)); // check for size!!!
}
else if (!ANY_BINSTR(arg)) Trap_Arg(arg);
}
else if (IS_BLOCK(arg)) {
// MOVE!
REB_MOLD mo = {0};
arg_ser = mo.series = Make_Unicode(VAL_BLK_LEN(arg) * 10); // GC!?
for (val = VAL_BLK_DATA(arg); NOT_END(val); val++)
Mold_Value(&mo, val, 0);
}
else if (IS_CHAR(arg)) {
// Optimize this case !!!
arg_ser = Make_Unicode(1);
Append_Byte(arg_ser, VAL_CHAR(arg));
}
else if (!ANY_STR(arg) || IS_TAG(arg)) {
arg_ser = Copy_Form_Value(arg, 0);
}
if (arg_ser) Set_String(arg, arg_ser);
else arg_ser = VAL_SERIES(arg);
// Length of insertion:
ilen = (action != A_CHANGE && DS_REF(AN_PART)) ? rlen : VAL_LEN(arg);
// If Source == Destination we need to prevent possible conflicts.
// Clone the argument just to be safe.
// (Note: It may be possible to optimize special cases like append !!)
if (series == VAL_SERIES(arg)) {
arg_ser = Copy_Series_Part(arg_ser, VAL_INDEX(arg), ilen); // GC!?
}
// Get /DUP count:
if (DS_REF(AN_DUP)) {
cnt = Int32(DS_ARG(AN_COUNT));
if (cnt <= 0) return; // no changes
}
// Total to insert:
size = cnt * ilen;
if (action != A_CHANGE) {
// Always expand series for INSERT and APPEND actions:
Expand_Series(series, index, size);
} else {
if (size > rlen)
Expand_Series(series, index, size-rlen);
else if (size < rlen && DS_REF(AN_PART))
Remove_Series(series, index, rlen-size);
else if (size + index > tail) {
EXPAND_SERIES_TAIL(series, size - (tail - index));
}
}
// For dup count:
for (; cnt > 0; cnt--) {
Insert_String(series, index, arg_ser, VAL_INDEX(arg), ilen, TRUE);
index += ilen;
}
TERM_SERIES(series);
VAL_INDEX(string) = (action == A_APPEND) ? 0 : index;
}
//
// Modify_String: C
//
// Returns new dst_idx.
//
REBCNT Modify_String(
REBCNT action, // INSERT, APPEND, CHANGE
REBSER *dst_ser, // target
REBCNT dst_idx, // position
const REBVAL *src_val, // source
REBFLGS flags, // AN_PART
REBINT dst_len, // length to remove
REBINT dups // dup count
) {
REBSER *src_ser = 0;
REBCNT src_idx = 0;
REBCNT src_len;
REBCNT tail = SER_LEN(dst_ser);
REBINT size; // total to insert
REBOOL needs_free;
REBINT limit;
// For INSERT/PART and APPEND/PART
if (action != SYM_CHANGE && GET_FLAG(flags, AN_PART))
limit = dst_len; // should be non-negative
else
limit = -1;
if (limit == 0 || dups < 0) return (action == SYM_APPEND) ? 0 : dst_idx;
if (action == SYM_APPEND || dst_idx > tail) dst_idx = tail;
// If the src_val is not a string, then we need to create a string:
if (GET_FLAG(flags, AN_SERIES)) { // used to indicate a BINARY series
if (IS_INTEGER(src_val)) {
src_ser = Make_Series_Codepoint(Int8u(src_val));
needs_free = TRUE;
limit = -1;
}
else if (IS_BLOCK(src_val)) {
src_ser = Join_Binary(src_val, limit); // NOTE: it's the shared FORM buffer!
needs_free = FALSE;
limit = -1;
}
else if (IS_CHAR(src_val)) {
//
// "UTF-8 was originally specified to allow codepoints with up to
// 31 bits (or 6 bytes). But with RFC3629, this was reduced to 4
// bytes max. to be more compatible to UTF-16." So depending on
// which RFC you consider "the UTF-8", max size is either 4 or 6.
//
src_ser = Make_Binary(6);
SET_SERIES_LEN(
src_ser,
Encode_UTF8_Char(BIN_HEAD(src_ser), VAL_CHAR(src_val))
);
needs_free = TRUE;
limit = -1;
}
else if (ANY_STRING(src_val)) {
src_len = VAL_LEN_AT(src_val);
if (limit >= 0 && src_len > cast(REBCNT, limit))
src_len = limit;
src_ser = Make_UTF8_From_Any_String(src_val, src_len, 0);
needs_free = TRUE;
limit = -1;
}
else if (!IS_BINARY(src_val))
fail (Error_Invalid_Arg(src_val));
}
else if (IS_CHAR(src_val)) {
src_ser = Make_Series_Codepoint(VAL_CHAR(src_val));
needs_free = TRUE;
}
else if (IS_BLOCK(src_val)) {
src_ser = Form_Tight_Block(src_val);
needs_free = TRUE;
}
else if (!ANY_STRING(src_val) || IS_TAG(src_val)) {
src_ser = Copy_Form_Value(src_val, 0);
needs_free = TRUE;
}
// Use either new src or the one that was passed:
if (src_ser) {
src_len = SER_LEN(src_ser);
}
else {
src_ser = VAL_SERIES(src_val);
src_idx = VAL_INDEX(src_val);
src_len = VAL_LEN_AT(src_val);
needs_free = FALSE;
}
if (limit >= 0) src_len = limit;
// If Source == Destination we need to prevent possible conflicts.
// Clone the argument just to be safe.
// (Note: It may be possible to optimize special cases like append !!)
if (dst_ser == src_ser) {
assert(!needs_free);
src_ser = Copy_Sequence_At_Len(src_ser, src_idx, src_len);
needs_free = TRUE;
src_idx = 0;
}
// Total to insert:
size = dups * src_len;
if (action != SYM_CHANGE) {
// Always expand dst_ser for INSERT and APPEND actions:
Expand_Series(dst_ser, dst_idx, size);
} else {
if (size > dst_len)
Expand_Series(dst_ser, dst_idx, size - dst_len);
else if (size < dst_len && GET_FLAG(flags, AN_PART))
Remove_Series(dst_ser, dst_idx, dst_len - size);
else if (size + dst_idx > tail) {
EXPAND_SERIES_TAIL(dst_ser, size - (tail - dst_idx));
}
}
// For dup count:
for (; dups > 0; dups--) {
Insert_String(dst_ser, dst_idx, src_ser, src_idx, src_len, TRUE);
dst_idx += src_len;
}
TERM_SEQUENCE(dst_ser);
if (needs_free) {
// If we did not use the series that was passed in, but rather
// created an internal temporary one, we need to free it.
Free_Series(src_ser);
}
return (action == SYM_APPEND) ? 0 : dst_idx;
}
*/ REBCNT Modify_String(REBCNT action, REBSER *dst_ser, REBCNT dst_idx, const REBVAL *src_val, REBCNT flags, REBINT dst_len, REBINT dups)
/*
** action: INSERT, APPEND, CHANGE
**
** dst_ser: target
** dst_idx: position
** src_val: source
** flags: AN_PART
** dst_len: length to remove
** dups: dup count
**
** return: new dst_idx
**
***********************************************************************/
{
REBSER *src_ser = 0;
REBCNT src_idx = 0;
REBCNT src_len;
REBCNT tail = SERIES_TAIL(dst_ser);
REBINT size; // total to insert
if (dups < 0) return (action == A_APPEND) ? 0 : dst_idx;
if (action == A_APPEND || dst_idx > tail) dst_idx = tail;
// If the src_val is not a string, then we need to create a string:
if (GET_FLAG(flags, AN_SERIES)) { // used to indicate a BINARY series
if (IS_INTEGER(src_val)) {
src_ser = Append_Byte(0, Int8u(src_val)); // creates a binary
}
else if (IS_BLOCK(src_val)) {
src_ser = Join_Binary(src_val); // NOTE: it's the shared FORM buffer!
}
else if (IS_CHAR(src_val)) {
src_ser = Make_Binary(6); // (I hate unicode)
src_ser->tail = Encode_UTF8_Char(BIN_HEAD(src_ser), VAL_CHAR(src_val));
}
else if (!ANY_BINSTR(src_val)) Trap_Arg_DEAD_END(src_val);
}
else if (IS_CHAR(src_val)) {
src_ser = Append_Byte(0, VAL_CHAR(src_val)); // unicode ok too
}
else if (IS_BLOCK(src_val)) {
src_ser = Form_Tight_Block(src_val);
}
else if (!ANY_STR(src_val) || IS_TAG(src_val)) {
src_ser = Copy_Form_Value(src_val, 0);
}
// Use either new src or the one that was passed:
if (src_ser) {
src_len = SERIES_TAIL(src_ser);
}
else {
src_ser = VAL_SERIES(src_val);
src_idx = VAL_INDEX(src_val);
src_len = VAL_LEN(src_val);
}
// For INSERT or APPEND with /PART use the dst_len not src_len:
if (action != A_CHANGE && GET_FLAG(flags, AN_PART)) src_len = dst_len;
// If Source == Destination we need to prevent possible conflicts.
// Clone the argument just to be safe.
// (Note: It may be possible to optimize special cases like append !!)
if (dst_ser == src_ser) {
src_ser = Copy_Series_Part(src_ser, src_idx, src_len);
src_idx = 0;
}
// Total to insert:
size = dups * src_len;
if (action != A_CHANGE) {
// Always expand dst_ser for INSERT and APPEND actions:
Expand_Series(dst_ser, dst_idx, size);
} else {
if (size > dst_len)
Expand_Series(dst_ser, dst_idx, size - dst_len);
else if (size < dst_len && GET_FLAG(flags, AN_PART))
Remove_Series(dst_ser, dst_idx, dst_len - size);
else if (size + dst_idx > tail) {
EXPAND_SERIES_TAIL(dst_ser, size - (tail - dst_idx));
}
}
// For dup count:
for (; dups > 0; dups--) {
Insert_String(dst_ser, dst_idx, src_ser, src_idx, src_len, TRUE);
dst_idx += src_len;
}
TERM_SERIES(dst_ser);
return (action == A_APPEND) ? 0 : dst_idx;
}
if (Form_Value_Throws(
D_OUT,
&mo,
&pending_delimiter,
value,
(REF(quote) || !IS_BLOCK(value) ? 0 : FORM_FLAG_REDUCE),
delimiter,
0 // depth
)) {
return R_OUT_IS_THROWN;
}
Val_Init_String(D_OUT, Pop_Molded_String(&mo));
}
else {
Val_Init_String(D_OUT, Copy_Form_Value(value, 0));
}
return R_OUT;
}
//
// mold: native [
//
// "Converts a value to a REBOL-readable string."
//
// value [any-value!] "The value to mold"
// /only {For a block value, mold only its contents, no outer []}
// /all "Use construction syntax"
// /flat "No indentation"
// ]
*/ static REBCNT Parse_To(REBPARSE *parse, REBCNT index, REBVAL *item, REBFLG is_thru)
/*
** Parse TO a specific:
** 1. integer - index position
** 2. END - end of input
** 3. value - according to datatype
** 4. block of values - the first one we hit
**
***********************************************************************/
{
REBSER *series = parse->series;
REBCNT i;
REBSER *ser;
// TO a specific index position.
if (IS_INTEGER(item)) {
i = (REBCNT)Int32(item) - (is_thru ? 0 : 1);
if (i > series->tail) i = series->tail;
}
// END
else if (IS_WORD(item) && VAL_WORD_CANON(item) == SYM_END) {
i = series->tail;
}
else if (IS_BLOCK(item)) {
i = To_Thru(parse, index, item, is_thru);
}
else {
if (IS_BLOCK_INPUT(parse)) {
REBVAL word; /// !!!Temp, but where can we put it?
if (IS_LIT_WORD(item)) { // patch to search for word, not lit.
word = *item;
VAL_SET(&word, REB_WORD);
item = &word;
}
///i = Find_Value(series, index, tail-index, item, 1, (REBOOL)(PF_CASE & flags), FALSE, 1);
i = Find_Block(series, index, series->tail, item, 1, HAS_CASE(parse)?AM_FIND_CASE:0, 1);
if (i != NOT_FOUND && is_thru) i++;
}
else {
// "str"
if (ANY_BINSTR(item)) {
if (!IS_STRING(item) && !IS_BINARY(item)) {
// !!! Can this be optimized not to use COPY?
ser = Copy_Form_Value(item, 0);
i = Find_Str_Str(series, 0, index, series->tail, 1, ser, 0, ser->tail, HAS_CASE(parse));
if (i != NOT_FOUND && is_thru) i += ser->tail;
}
else {
i = Find_Str_Str(series, 0, index, series->tail, 1, VAL_SERIES(item), VAL_INDEX(item), VAL_LEN(item), HAS_CASE(parse));
if (i != NOT_FOUND && is_thru) i += VAL_LEN(item);
}
}
// #"A"
else if (IS_CHAR(item)) {
i = Find_Str_Char(series, 0, index, series->tail, 1, VAL_CHAR(item), HAS_CASE(parse));
if (i != NOT_FOUND && is_thru) i++;
}
}
}
return i;
}
*/ static REBCNT Parse_Next_String(REBPARSE *parse, REBCNT index, REBVAL *item, REBCNT depth)
/*
** Match the next item in the string ruleset.
**
** If it matches, return the index just past it.
** Otherwise return NOT_FOUND.
**
***********************************************************************/
{
// !!! THIS CODE NEEDS CLEANUP AND REWRITE BASED ON OTHER CHANGES
REBSER *series = parse->series;
REBSER *ser;
REBCNT flags = parse->flags | AM_FIND_MATCH | AM_FIND_TAIL;
int rewrite_needed;
if (Trace_Level) {
Trace_Value(7, item);
Trace_String(8, STR_SKIP(series, index), series->tail - index);
}
if (IS_NONE(item)) return index;
if (index >= series->tail) return NOT_FOUND;
switch (VAL_TYPE(item)) {
// Do we match a single character?
case REB_CHAR:
if (HAS_CASE(parse))
index = (VAL_CHAR(item) == GET_ANY_CHAR(series, index)) ? index+1 : NOT_FOUND;
else
index = (UP_CASE(VAL_CHAR(item)) == UP_CASE(GET_ANY_CHAR(series, index))) ? index+1 : NOT_FOUND;
break;
case REB_EMAIL:
case REB_STRING:
case REB_BINARY:
index = Find_Str_Str(series, 0, index, SERIES_TAIL(series), 1, VAL_SERIES(item), VAL_INDEX(item), VAL_LEN(item), flags);
break;
// Do we match to a char set?
case REB_BITSET:
flags = Check_Bit(VAL_SERIES(item), GET_ANY_CHAR(series, index), !HAS_CASE(parse));
index = flags ? index + 1 : NOT_FOUND;
break;
/*
case REB_DATATYPE: // Currently: integer!
if (VAL_DATATYPE(item) == REB_INTEGER) {
REBCNT begin = index;
while (IS_LEX_NUMBER(*str)) str++, index++;
if (begin == index) index = NOT_FOUND;
}
break;
*/
case REB_TAG:
case REB_FILE:
// case REB_ISSUE:
// !! Can be optimized (w/o COPY)
ser = Copy_Form_Value(item, 0);
index = Find_Str_Str(series, 0, index, SERIES_TAIL(series), 1, ser, 0, ser->tail, flags);
break;
case REB_NONE:
break;
// Parse a sub-rule block:
case REB_BLOCK:
index = Parse_Rules_Loop(parse, index, VAL_BLK_DATA(item), depth);
break;
// Do an expression:
case REB_PAREN:
item = Do_Block_Value_Throw(item); // might GC
// old: if (IS_ERROR(item)) Throw_Error(VAL_ERR_OBJECT(item));
index = MIN(index, series->tail); // may affect tail
break;
default:
Trap1(RE_PARSE_RULE, item);
}
return index;
}
