*/ REBSER *Decode_UTF_String(REBYTE *bp, REBCNT len, REBINT utf, REBFLG ccr)
/*
** Do all the details to decode a string.
** Input is a byte series. Len is len of input.
** The utf is 0, 8, +/-16, +/-32.
** A special -1 means use the BOM.
**
***********************************************************************/
{
REBSER *ser = BUF_UTF8; // buffer is Unicode width
REBSER *dst;
REBINT size;
//REBFLG ccr = FALSE; // in original R3-alpha if was TRUE
//@@ https://github.com/rebol/rebol-issues/issues/2336
if (utf == -1) {
utf = What_UTF(bp, len);
if (utf) {
if (utf == 8) bp += 3, len -= 3;
else if (utf == -16 || utf == 16) bp += 2, len -= 2;
else if (utf == -32 || utf == 32) bp += 4, len -= 4;
}
}
if (utf == 0 || utf == 8) {
size = Decode_UTF8((REBUNI*)Reset_Buffer(ser, len), bp, len, ccr);
}
else if (utf == -16 || utf == 16) {
size = Decode_UTF16((REBUNI*)Reset_Buffer(ser, len/2 + 1), bp, len, utf < 0, ccr);
}
else if (utf == -32 || utf == 32) {
size = Decode_UTF32((REBUNI*)Reset_Buffer(ser, len/4 + 1), bp, len, utf < 0, ccr);
}
else {
return NULL;
}
if (size < 0) {
size = -size;
dst = Make_Binary(size);
Append_Uni_Bytes(dst, UNI_HEAD(ser), size);
}
else {
dst = Make_Unicode(size);
Append_Uni_Uni(dst, UNI_HEAD(ser), size);
}
return dst;
}
*/ void Debug_Uni(const REBSER *ser)
/*
** Print debug unicode string followed by a newline.
**
***********************************************************************/
{
REBCNT ul;
REBCNT bl;
REBYTE buf[1024];
REBUNI *up = UNI_HEAD(ser);
REBINT size = Length_As_UTF8(up, SERIES_TAIL(ser), TRUE, OS_CRLF);
REBINT disabled = GC_Disabled;
GC_Disabled = 1;
while (size > 0) {
ul = Encode_UTF8(buf, MIN(size, 1020), up, &bl, TRUE, OS_CRLF);
Debug_String(buf, bl, 0, 0);
size -= ul;
up += ul;
}
Debug_Line();
assert(GC_Disabled == 1);
GC_Disabled = disabled;
}
//
// Enline_Uni: C
//
void Enline_Uni(REBSER *ser, REBCNT idx, REBCNT len)
{
REBCNT cnt = 0;
REBUNI *bp;
REBUNI c = 0;
REBCNT tail;
// Calculate the size difference by counting the number of LF's
// that have no CR's in front of them.
bp = UNI_AT(ser, idx);
for (; len > 0; len--) {
if (*bp == LF && c != CR) cnt++;
c = *bp++;
}
if (cnt == 0) return;
// Extend series:
len = SER_LEN(ser); // before expansion
EXPAND_SERIES_TAIL(ser, cnt);
tail = SER_LEN(ser); // after expansion
bp = UNI_HEAD(ser); // expand may change it
// Add missing CRs:
while (cnt > 0) {
bp[tail--] = bp[len]; // Copy src to dst.
if (bp[len] == LF && (len == 0 || bp[len - 1] != CR)) {
bp[tail--] = CR;
cnt--;
}
len--;
}
}
*/ REBSER *Copy_Wide_Str(void *src, REBINT len)
/*
** Create a REBOL string series from a wide char string.
** Minimize to bytes if possible
*/
{
REBSER *dst;
REBUNI *str = (REBUNI*)src;
if (Is_Wide(str, len)) {
REBUNI *up;
dst = Make_Unicode(len);
SERIES_TAIL(dst) = len;
up = UNI_HEAD(dst);
while (len-- > 0) *up++ = *str++;
*up = 0;
}
else {
REBYTE *bp;
dst = Make_Binary(len);
SERIES_TAIL(dst) = len;
bp = BIN_HEAD(dst);
while (len-- > 0) *bp++ = (REBYTE)*str++;
*bp = 0;
}
return dst;
}
*/ REBSER *Make_Unicode(REBCNT length)
/*
** Make a unicode string series. Used for internal strings.
** Add 1 extra for terminator.
**
***********************************************************************/
{
REBSER *series = Make_Series(length + 1, sizeof(REBUNI), MKS_NONE);
LABEL_SERIES(series, "make unicode");
// !!! Clients seem to have different expectations of if `length` is
// total capacity (and the binary should be empty) or actually is
// specifically being preallocated at a fixed length. Until this
// is straightened out, terminate for both possibilities.
UNI_HEAD(series)[length] = 0;
TERM_SEQUENCE(series);
return series;
}
*/ REBSER *Parse_Lines(REBSER *src)
/*
** Convert a string buffer to a block of strings.
** Note that the string must already be converted
** to REBOL LF format (no CRs).
**
***********************************************************************/
{
REBSER *blk;
REBUNI c;
REBCNT i;
REBCNT s;
REBVAL *val;
REBOOL uni = !BYTE_SIZE(src);
REBYTE *bp = BIN_HEAD(src);
REBUNI *up = UNI_HEAD(src);
blk = BUF_EMIT;
RESET_SERIES(blk);
// Scan string, looking for LF and CR terminators:
for (i = s = 0; i < SERIES_TAIL(src); i++) {
c = uni ? up[i] : bp[i];
if (c == LF || c == CR) {
val = Append_Value(blk);
Set_String(val, Copy_String(src, s, i - s));
VAL_SET_LINE(val);
// Skip CRLF if found:
if (c == CR && LF == uni ? up[i] : bp[i]) i++;
s = i;
}
}
// Partial line (no linefeed):
if (s + 1 != i) {
val = Append_Value(blk);
Set_String(val, Copy_String(src, s, i - s));
VAL_SET_LINE(val);
}
return Copy_Block(blk, 0);
}
*/ REBCHR *Val_Str_To_OS_Managed(REBSER **out, REBVAL *val)
/*
** This is used to pass a REBOL value string to an OS API.
**
** The REBOL (input) string can be byte or wide sized.
** The OS (output) string is in the native OS format.
** On Windows, its a wide-char, but on Linux, its UTF-8.
**
** If we know that the string can be used directly as-is,
** (because it's in the OS size format), we can used it
** like that.
**
** !!! The series is created but just let up to the garbage
** collector to free. This is a "leaky" approach. You may
** optionally request to have the series returned if it is
** important for you to protect it from GC, but you cannot
** currently get a "freeable" series out of this.
**
***********************************************************************/
{
#ifdef OS_WIDE_CHAR
if (VAL_BYTE_SIZE(val)) {
// On windows, we need to convert byte to wide:
REBINT n = VAL_LEN(val);
REBSER *up = Make_Unicode(n);
// !!!"Leaks" in the sense that the GC has to take care of this
MANAGE_SERIES(up);
n = Decode_UTF8(UNI_HEAD(up), VAL_BIN_DATA(val), n, FALSE);
SERIES_TAIL(up) = abs(n);
UNI_TERM(up);
if (out) *out = up;
return cast(REBCHR*, UNI_HEAD(up));
}
else {
// Already wide, we can use it as-is:
// !Assumes the OS uses same wide format!
if (out) *out = VAL_SERIES(val);
STOID Mold_Uni_Char(REBSER *dst, REBUNI chr, REBOOL molded, REBOOL parened)
{
REBCNT tail = SERIES_TAIL(dst);
REBUNI *up;
if (!molded) {
EXPAND_SERIES_TAIL(dst, 1);
*UNI_SKIP(dst, tail) = chr;
}
else {
EXPAND_SERIES_TAIL(dst, 10); // worst case: #"^(1234)"
up = UNI_SKIP(dst, tail);
*up++ = '#';
*up++ = '"';
up = Emit_Uni_Char(up, chr, parened);
*up++ = '"';
dst->tail = up - UNI_HEAD(dst);
}
UNI_TERM(dst);
}
*/ REBSER *Copy_Bytes_To_Unicode(REBYTE *src, REBINT len)
/*
** Convert a byte string to a unicode string. This can
** be used for ASCII or LATIN-8 strings.
**
***********************************************************************/
{
REBSER *series;
REBUNI *dst;
series = Make_Unicode(len);
dst = UNI_HEAD(series);
SERIES_TAIL(series) = len;
for (; len > 0; len--) {
*dst++ = (REBUNI)(*src++);
}
UNI_TERM(series);
return series;
}
*/ REBSER *Copy_Buffer(REBSER *buf, void *end)
/*
** Copy a shared buffer. Set tail and termination.
**
***********************************************************************/
{
REBSER *ser;
REBCNT len;
len = BYTE_SIZE(buf) ? ((REBYTE *)end) - BIN_HEAD(buf)
: ((REBUNI *)end) - UNI_HEAD(buf);
ser = Make_Series(
len + 1,
SERIES_WIDE(buf),
Is_Array_Series(buf) ? MKS_ARRAY : MKS_NONE
);
memcpy(ser->data, buf->data, SERIES_WIDE(buf) * len);
ser->tail = len;
TERM_SERIES(ser);
return ser;
}
*/ int Encode_UTF8_Line(REBSER *dst, REBSER *src, REBCNT idx)
/*
** Encode a unicode source buffer into a binary line of UTF8.
** Include the LF terminator in the result.
** Return the length of the line buffer.
**
***********************************************************************/
{
REBUNI *up = UNI_HEAD(src);
REBCNT len = SERIES_TAIL(src);
REBCNT tail;
REBUNI c;
REBINT n;
REBYTE buf[8];
tail = RESET_TAIL(dst);
while (idx < len) {
if ((c = up[idx]) < 0x80) {
EXPAND_SERIES_TAIL(dst, 1);
BIN_HEAD(dst)[tail++] = (REBYTE)c;
}
else {
n = Encode_UTF8_Char(buf, c);
EXPAND_SERIES_TAIL(dst, n);
memcpy(BIN_SKIP(dst, tail), buf, n);
tail += n;
}
idx++;
if (c == LF) break;
}
BIN_HEAD(dst)[tail] = 0;
SERIES_TAIL(dst) = tail;
return idx;
}
STOID Mold_String_Series(REBVAL *value, REB_MOLD *mold)
{
REBCNT len = VAL_LEN(value);
REBSER *ser = VAL_SERIES(value);
REBCNT idx = VAL_INDEX(value);
REB_STRF sf = {0};
REBYTE *bp;
REBUNI *up;
REBUNI *dp;
REBOOL uni = !BYTE_SIZE(ser);
REBCNT n;
REBUNI c;
// Empty string:
if (idx >= VAL_TAIL(value)) {
Append_Bytes(mold->series, "\"\""); //Trap0(RE_PAST_END);
return;
}
Sniff_String(ser, idx, &sf);
if (!GET_MOPT(mold, MOPT_ANSI_ONLY)) sf.paren = 0;
// Source can be 8 or 16 bits:
if (uni) up = UNI_HEAD(ser);
else bp = STR_HEAD(ser);
// If it is a short quoted string, emit it as "string":
if (len <= MAX_QUOTED_STR && sf.quote == 0 && sf.newline < 3) {
dp = Prep_Uni_Series(mold, len + sf.newline + sf.escape + sf.paren + sf.chr1e + 2);
*dp++ = '"';
for (n = idx; n < VAL_TAIL(value); n++) {
c = uni ? up[n] : (REBUNI)(bp[n]);
dp = Emit_Uni_Char(dp, c, (REBOOL)GET_MOPT(mold, MOPT_ANSI_ONLY)); // parened
}
*dp++ = '"';
*dp = 0;
return;
}
// It is a braced string, emit it as {string}:
if (!sf.malign) sf.brace_in = sf.brace_out = 0;
dp = Prep_Uni_Series(mold, len + sf.brace_in + sf.brace_out + sf.escape + sf.paren + sf.chr1e + 2);
*dp++ = '{';
for (n = idx; n < VAL_TAIL(value); n++) {
c = uni ? up[n] : (REBUNI)(bp[n]);
switch (c) {
case '{':
case '}':
if (sf.malign) {
*dp++ = '^';
*dp++ = c;
break;
}
case '\n':
case '"':
*dp++ = c;
break;
default:
dp = Emit_Uni_Char(dp, c, (REBOOL)GET_MOPT(mold, MOPT_ANSI_ONLY)); // parened
}
}
*dp++ = '}';
*dp = 0;
}
//
// Clipboard_Actor: C
//
static REB_R Clipboard_Actor(struct Reb_Call *call_, REBSER *port, REBCNT action)
{
REBREQ *req;
REBINT result;
REBVAL *arg;
REBCNT refs; // refinement argument flags
REBINT len;
REBSER *ser;
Validate_Port(port, action);
arg = DS_ARGC > 1 ? D_ARG(2) : NULL;
req = cast(REBREQ*, Use_Port_State(port, RDI_CLIPBOARD, sizeof(REBREQ)));
switch (action) {
case A_UPDATE:
// Update the port object after a READ or WRITE operation.
// This is normally called by the WAKE-UP function.
arg = OFV(port, STD_PORT_DATA);
if (req->command == RDC_READ) {
// this could be executed twice:
// once for an event READ, once for the CLOSE following the READ
if (!req->common.data) return R_NONE;
len = req->actual;
if (GET_FLAG(req->flags, RRF_WIDE)) {
// convert to UTF8, so that it can be converted back to string!
Val_Init_Binary(arg, Make_UTF8_Binary(
req->common.data,
len / sizeof(REBUNI),
0,
OPT_ENC_UNISRC
));
}
else {
REBSER *ser = Make_Binary(len);
memcpy(BIN_HEAD(ser), req->common.data, len);
SERIES_TAIL(ser) = len;
Val_Init_Binary(arg, ser);
}
OS_FREE(req->common.data); // release the copy buffer
req->common.data = 0;
}
else if (req->command == RDC_WRITE) {
SET_NONE(arg); // Write is done.
}
return R_NONE;
case A_READ:
// This device is opened on the READ:
if (!IS_OPEN(req)) {
if (OS_DO_DEVICE(req, RDC_OPEN))
fail (Error_On_Port(RE_CANNOT_OPEN, port, req->error));
}
// Issue the read request:
CLR_FLAG(req->flags, RRF_WIDE); // allow byte or wide chars
result = OS_DO_DEVICE(req, RDC_READ);
if (result < 0) fail (Error_On_Port(RE_READ_ERROR, port, req->error));
if (result > 0) return R_NONE; /* pending */
// Copy and set the string result:
arg = OFV(port, STD_PORT_DATA);
len = req->actual;
if (GET_FLAG(req->flags, RRF_WIDE)) {
// convert to UTF8, so that it can be converted back to string!
Val_Init_Binary(arg, Make_UTF8_Binary(
req->common.data,
len / sizeof(REBUNI),
0,
OPT_ENC_UNISRC
));
}
else {
REBSER *ser = Make_Binary(len);
memcpy(BIN_HEAD(ser), req->common.data, len);
SERIES_TAIL(ser) = len;
Val_Init_Binary(arg, ser);
}
*D_OUT = *arg;
return R_OUT;
case A_WRITE:
if (!IS_STRING(arg) && !IS_BINARY(arg))
fail (Error(RE_INVALID_PORT_ARG, arg));
// This device is opened on the WRITE:
if (!IS_OPEN(req)) {
if (OS_DO_DEVICE(req, RDC_OPEN))
fail (Error_On_Port(RE_CANNOT_OPEN, port, req->error));
}
refs = Find_Refines(call_, ALL_WRITE_REFS);
// Handle /part refinement:
len = VAL_LEN(arg);
if (refs & AM_WRITE_PART && VAL_INT32(D_ARG(ARG_WRITE_LIMIT)) < len)
len = VAL_INT32(D_ARG(ARG_WRITE_LIMIT));
// If bytes, see if we can fit it:
if (SERIES_WIDE(VAL_SERIES(arg)) == 1) {
#ifdef ARG_STRINGS_ALLOWED
if (!All_Bytes_ASCII(VAL_BIN_DATA(arg), len)) {
Val_Init_String(
arg, Copy_Bytes_To_Unicode(VAL_BIN_DATA(arg), len)
);
} else
req->common.data = VAL_BIN_DATA(arg);
#endif
// Temp conversion:!!!
ser = Make_Unicode(len);
len = Decode_UTF8(UNI_HEAD(ser), VAL_BIN_DATA(arg), len, FALSE);
SERIES_TAIL(ser) = len = abs(len);
UNI_TERM(ser);
Val_Init_String(arg, ser);
req->common.data = cast(REBYTE*, UNI_HEAD(ser));
SET_FLAG(req->flags, RRF_WIDE);
}
else
// If unicode (may be from above conversion), handle it:
if (SERIES_WIDE(VAL_SERIES(arg)) == sizeof(REBUNI)) {
req->common.data = cast(REBYTE *, VAL_UNI_DATA(arg));
SET_FLAG(req->flags, RRF_WIDE);
}