static void read_string(LexState *ls, int delim, TValue *tv)
{
save_and_next(ls);
while (ls->current != delim) {
switch (ls->current) {
case END_OF_STREAM:
lj_lex_error(ls, TK_eof, LJ_ERR_XSTR);
continue;
case '\n':
case '\r':
lj_lex_error(ls, TK_string, LJ_ERR_XSTR);
continue;
case '\\': {
int c;
next(ls); /* do not save the `\' */
switch (ls->current) {
case 'a': c = '\a'; break;
case 'b': c = '\b'; break;
case 'f': c = '\f'; break;
case 'n': c = '\n'; break;
case 'r': c = '\r'; break;
case 't': c = '\t'; break;
case 'v': c = '\v'; break;
case '\n': case '\r': save(ls, '\n'); inclinenumber(ls); continue;
case END_OF_STREAM: continue; /* will raise an error next loop */
default:
if (!lj_ctype_isdigit(ls->current)) {
save_and_next(ls); /* handles \\, \", \', and \? */
} else { /* \xxx */
int i = 0;
c = 0;
do {
c = 10*c + (ls->current-'0');
next(ls);
} while (++i<3 && lj_ctype_isdigit(ls->current));
if (c > UCHAR_MAX)
lj_lex_error(ls, TK_string, LJ_ERR_XESC);
save(ls, c);
}
continue;
}
save(ls, c);
next(ls);
continue;
}
default:
save_and_next(ls);
break;
}
}
save_and_next(ls); /* skip delimiter */
setstrV(ls->L, tv, lj_parse_keepstr(ls, ls->sb.buf + 1, ls->sb.n - 2));
}
static void read_long_string(LexState *ls, TValue *tv, int sep)
{
save_and_next(ls); /* skip 2nd `[' */
if (currIsNewline(ls)) /* string starts with a newline? */
inclinenumber(ls); /* skip it */
for (;;) {
switch (ls->current) {
case END_OF_STREAM:
lj_lex_error(ls, TK_eof, tv ? LJ_ERR_XLSTR : LJ_ERR_XLCOM);
break;
case ']':
if (skip_sep(ls) == sep) {
save_and_next(ls); /* skip 2nd `]' */
goto endloop;
}
break;
case '\n':
case '\r':
save(ls, '\n');
inclinenumber(ls);
if (!tv) lj_str_resetbuf(&ls->sb); /* avoid wasting space */
break;
default:
if (tv) save_and_next(ls);
else next(ls);
break;
}
} endloop:
if (tv) {
GCstr *str = lj_parse_keepstr(ls, ls->sb.buf + (2 + (MSize)sep),
ls->sb.n - 2*(2 + (MSize)sep));
setstrV(ls->L, tv, str);
}
}
/* Parse a long string or long comment (tv set to NULL). */
static void lex_longstring(LexState *ls, TValue *tv, int sep)
{
lex_savenext(ls); /* Skip second '['. */
if (lex_iseol(ls)) /* Skip initial newline. */
lex_newline(ls);
for (;;) {
switch (ls->c) {
case LEX_EOF:
lj_lex_error(ls, TK_eof, tv ? LJ_ERR_XLSTR : LJ_ERR_XLCOM);
break;
case ']':
if (lex_skipeq(ls) == sep) {
lex_savenext(ls); /* Skip second ']'. */
goto endloop;
}
break;
case '\n':
case '\r':
lex_save(ls, '\n');
lex_newline(ls);
if (!tv) lj_buf_reset(&ls->sb); /* Don't waste space for comments. */
break;
default:
lex_savenext(ls);
break;
}
} endloop:
if (tv) {
GCstr *str = lj_parse_keepstr(ls, sbufB(&ls->sb) + (2 + (MSize)sep),
sbuflen(&ls->sb) - 2*(2 + (MSize)sep));
setstrV(ls->L, tv, str);
}
}
static LJ_NOINLINE void save_grow(LexState *ls, int c)
{
MSize newsize;
if (ls->sb.sz >= LJ_MAX_STR/2)
lj_lex_error(ls, 0, LJ_ERR_XELEM);
newsize = ls->sb.sz * 2;
lj_str_resizebuf(ls->L, &ls->sb, newsize);
ls->sb.buf[ls->sb.n++] = (char)c;
}
/* Skip line break. Handles "\n", "\r", "\r\n" or "\n\r". */
static void lex_newline(LexState *ls)
{
LexChar old = ls->c;
lua_assert(lex_iseol(ls));
lex_next(ls); /* Skip "\n" or "\r". */
if (lex_iseol(ls) && ls->c != old) lex_next(ls); /* Skip "\n\r" or "\r\n". */
if (++ls->linenumber >= LJ_MAX_LINE)
lj_lex_error(ls, ls->tok, LJ_ERR_XLINES);
}
static void inclinenumber(LexState *ls)
{
int old = ls->current;
lua_assert(currIsNewline(ls));
next(ls); /* skip `\n' or `\r' */
if (currIsNewline(ls) && ls->current != old)
next(ls); /* skip `\n\r' or `\r\n' */
if (++ls->linenumber >= LJ_MAX_LINE)
lj_lex_error(ls, ls->token, LJ_ERR_XLINES);
}
static void save(LexState *ls, int c)
{
if (ls->sb.n + 1 > ls->sb.sz) {
MSize newsize;
if (ls->sb.sz >= LJ_MAX_STR/2)
lj_lex_error(ls, 0, LJ_ERR_XELEM);
newsize = ls->sb.sz * 2;
lj_str_resizebuf(ls->L, &ls->sb, newsize);
}
ls->sb.buf[ls->sb.n++] = cast(char, c);
}
/* Parse a number literal. */
static void lex_number(LexState *ls)
{
StrScanFmt fmt;
TValue *tv = &ls->tokenval;
int c, xp = 'e';
if (ls->current == '-' || ls->current == '+') {
save_and_next(ls);
}
if ((c = ls->current) == '0') {
save_and_next(ls);
if ((ls->current | 0x20) == 'x') xp = 'p';
}
while (lj_char_isident(ls->current) || ls->current == '.' ||
((ls->current == '-' || ls->current == '+') && (c | 0x20) == xp)) {
c = ls->current;
save_and_next(ls);
}
save(ls, '\0');
fmt = lj_strscan_scan((const uint8_t *)ls->sb.buf, tv,
(LJ_DUALNUM ? STRSCAN_OPT_TOINT : STRSCAN_OPT_TONUM) |
(LJ_HASFFI ? (STRSCAN_OPT_LL|STRSCAN_OPT_IMAG) : 0));
ls->token = TK_number;
if (LJ_DUALNUM && fmt == STRSCAN_INT) {
setitype(tv, LJ_TISNUM);
} else if (fmt == STRSCAN_NUM) {
/* Already in correct format. */
#if LJ_HASFFI
} else if (fmt != STRSCAN_ERROR) {
lua_State *L = ls->L;
GCcdata *cd;
lua_assert(fmt == STRSCAN_I64 || fmt == STRSCAN_U64 || fmt == STRSCAN_IMAG);
if (!ctype_ctsG(G(L))) {
ptrdiff_t oldtop = savestack(L, L->top);
luaopen_ffi(L); /* Load FFI library on-demand. */
L->top = restorestack(L, oldtop);
}
if (fmt == STRSCAN_IMAG) {
cd = lj_cdata_new_(L, CTID_COMPLEX_DOUBLE, 2*sizeof(double));
((double *)cdataptr(cd))[0] = 0;
((double *)cdataptr(cd))[1] = numV(tv);
} else {
cd = lj_cdata_new_(L, fmt==STRSCAN_I64 ? CTID_INT64 : CTID_UINT64, 8);
*(uint64_t *)cdataptr(cd) = tv->u64;
}
lj_parse_keepcdata(ls, tv, cd);
#endif
} else {
lua_assert(fmt == STRSCAN_ERROR);
lj_lex_error(ls, TK_number, LJ_ERR_XNUMBER);
}
}
static void read_numeral(LexState *ls, TValue *tv)
{
lua_assert(lj_ctype_isdigit(ls->current));
do {
save_and_next(ls);
} while (lj_ctype_isdigit(ls->current) || ls->current == '.');
if (check_next(ls, "Ee")) /* `E'? */
check_next(ls, "+-"); /* optional exponent sign */
while (lj_ctype_isident(ls->current))
save_and_next(ls);
save(ls, '\0');
if (!lj_str_numconv(ls->sb.buf, tv))
lj_lex_error(ls, TK_number, LJ_ERR_XNUMBER);
}
/* Parse a number literal. */
static void lex_number(LexState *ls, TValue *tv)
{
int c;
lua_assert(lj_char_isdigit(ls->current));
do {
c = ls->current;
save_and_next(ls);
} while (lj_char_isident(ls->current) || ls->current == '.' ||
((ls->current == '-' || ls->current == '+') &&
((c & ~0x20) == 'E' || (c & ~0x20) == 'P')));
#if LJ_HASFFI
c &= ~0x20;
if ((c == 'I' || c == 'L' || c == 'U') && !ctype_ctsG(G(ls->L)))
lex_loadffi(ls->L);
if (c == 'I') /* Parse imaginary part of complex number. */
ls->sb.n--;
#endif
save(ls, '\0');
#if LJ_HASFFI
if ((c == 'L' || c == 'U') && lex_number64(ls, tv)) { /* Parse 64 bit int. */
return;
} else
#endif
if (lj_str_numconv(ls->sb.buf, tv)) {
#if LJ_HASFFI
if (c == 'I') { /* Return cdata holding a complex number. */
GCcdata *cd = lj_cdata_new_(ls->L, CTID_COMPLEX_DOUBLE, 2*sizeof(double));
((double *)cdataptr(cd))[0] = 0;
((double *)cdataptr(cd))[1] = numberVnum(tv);
lj_parse_keepcdata(ls, tv, cd);
}
#endif
if (LJ_DUALNUM && tvisnum(tv)) {
int32_t k = lj_num2int(numV(tv));
if ((lua_Number)k == numV(tv)) /* -0 cannot end up here. */
setintV(tv, k);
}
return;
}
lj_lex_error(ls, TK_number, LJ_ERR_XNUMBER);
}
static int llex(LexState *ls, TValue *tv)
{
lj_str_resetbuf(&ls->sb);
for (;;) {
if (lj_char_isident(ls->current)) {
GCstr *s;
if (lj_char_isdigit(ls->current)) { /* Numeric literal. */
lex_number(ls, tv);
return TK_number;
}
/* Identifier or reserved word. */
do {
save_and_next(ls);
} while (lj_char_isident(ls->current));
s = lj_parse_keepstr(ls, ls->sb.buf, ls->sb.n);
setstrV(ls->L, tv, s);
if (s->reserved > 0) /* Reserved word? */
return TK_OFS + s->reserved;
return TK_name;
}
switch (ls->current) {
case '\n':
case '\r':
inclinenumber(ls);
continue;
case ' ':
case '\t':
case '\v':
case '\f':
next(ls);
continue;
case '-':
next(ls);
if (ls->current != '-') return '-';
/* else is a comment */
next(ls);
if (ls->current == '[') {
int sep = skip_sep(ls);
lj_str_resetbuf(&ls->sb); /* `skip_sep' may dirty the buffer */
if (sep >= 0) {
read_long_string(ls, NULL, sep); /* long comment */
lj_str_resetbuf(&ls->sb);
continue;
}
}
/* else short comment */
while (!currIsNewline(ls) && ls->current != END_OF_STREAM)
next(ls);
continue;
case '[': {
int sep = skip_sep(ls);
if (sep >= 0) {
read_long_string(ls, tv, sep);
return TK_string;
} else if (sep == -1) {
return '[';
} else {
lj_lex_error(ls, TK_string, LJ_ERR_XLDELIM);
continue;
}
}
case '=':
next(ls);
if (ls->current != '=') return '='; else { next(ls); return TK_eq; }
case '<':
next(ls);
if (ls->current != '=') return '<'; else { next(ls); return TK_le; }
case '>':
next(ls);
if (ls->current != '=') return '>'; else { next(ls); return TK_ge; }
case '~':
next(ls);
if (ls->current != '=') return '~'; else { next(ls); return TK_ne; }
case ':':
next(ls);
if (ls->current != ':') return ':'; else { next(ls); return TK_label; }
case '"':
case '\'':
read_string(ls, ls->current, tv);
return TK_string;
case '.':
save_and_next(ls);
if (ls->current == '.') {
next(ls);
if (ls->current == '.') {
next(ls);
return TK_dots; /* ... */
}
return TK_concat; /* .. */
} else if (!lj_char_isdigit(ls->current)) {
return '.';
} else {
lex_number(ls, tv);
return TK_number;
}
case END_OF_STREAM:
return TK_eof;
default: {
int c = ls->current;
next(ls);
return c; /* Single-char tokens (+ - / ...). */
}
}
}
}
static void read_string(LexState *ls, int delim, TValue *tv)
{
save_and_next(ls);
while (ls->current != delim) {
switch (ls->current) {
case END_OF_STREAM:
lj_lex_error(ls, TK_eof, LJ_ERR_XSTR);
continue;
case '\n':
case '\r':
lj_lex_error(ls, TK_string, LJ_ERR_XSTR);
continue;
case '\\': {
int c = next(ls); /* Skip the '\\'. */
switch (c) {
case 'a': c = '\a'; break;
case 'b': c = '\b'; break;
case 'f': c = '\f'; break;
case 'n': c = '\n'; break;
case 'r': c = '\r'; break;
case 't': c = '\t'; break;
case 'v': c = '\v'; break;
case 'x': /* Hexadecimal escape '\xXX'. */
c = (next(ls) & 15u) << 4;
if (!lj_char_isdigit(ls->current)) {
if (!lj_char_isxdigit(ls->current)) goto err_xesc;
c += 9 << 4;
}
c += (next(ls) & 15u);
if (!lj_char_isdigit(ls->current)) {
if (!lj_char_isxdigit(ls->current)) goto err_xesc;
c += 9;
}
break;
case 'z': /* Skip whitespace. */
next(ls);
while (lj_char_isspace(ls->current))
if (currIsNewline(ls)) inclinenumber(ls); else next(ls);
continue;
case '\n': case '\r': save(ls, '\n'); inclinenumber(ls); continue;
case '\\': case '\"': case '\'': break;
case END_OF_STREAM: continue;
default:
if (!lj_char_isdigit(c))
goto err_xesc;
c -= '0'; /* Decimal escape '\ddd'. */
if (lj_char_isdigit(next(ls))) {
c = c*10 + (ls->current - '0');
if (lj_char_isdigit(next(ls))) {
c = c*10 + (ls->current - '0');
if (c > 255) {
err_xesc:
lj_lex_error(ls, TK_string, LJ_ERR_XESC);
}
next(ls);
}
}
save(ls, c);
continue;
}
save(ls, c);
next(ls);
continue;
}
default:
save_and_next(ls);
break;
}
}
save_and_next(ls); /* skip delimiter */
setstrV(ls->L, tv, lj_parse_keepstr(ls, ls->sb.buf + 1, ls->sb.n - 2));
}
static void read_string(LexState *ls)
{
save_and_next(ls);
while (ls->current != '"') {
switch (ls->current) {
case END_OF_STREAM:
lj_lex_error(ls, TK_eof, LJ_ERR_XSTR);
continue;
case '\\': {
int c = next(ls); /* Skip the '\\'. */
switch (c) {
case 'a': c = '\a'; break;
case 'b': c = '\b'; break;
case 'f': c = '\f'; break;
case 'n': c = '\n'; break;
case 'r': c = '\r'; break;
case 't': c = '\t'; break;
case 'v': c = '\v'; break;
case 'x': /* Hexadecimal escape '\xXX'. */
c = (next(ls) & 15u) << 4;
if (!lj_char_isdigit(ls->current)) {
if (!lj_char_isxdigit(ls->current)) goto err_xesc;
c += 9 << 4;
}
c += (next(ls) & 15u);
if (!lj_char_isdigit(ls->current)) {
if (!lj_char_isxdigit(ls->current)) goto err_xesc;
c += 9;
}
break;
case 'u': /* Unicode escape '\uXXXX'. */
c = (next(ls) & 15u) << 12;
if (!lj_char_isdigit(ls->current)) {
if (!lj_char_isxdigit(ls->current)) goto err_xesc;
c += 9 << 12;
}
c += (next(ls) & 15u) << 8;
if (!lj_char_isdigit(ls->current)) {
if (!lj_char_isxdigit(ls->current)) goto err_xesc;
c += 9 << 8;
}
c += (next(ls) & 15u) << 4;
if (!lj_char_isdigit(ls->current)) {
if (!lj_char_isxdigit(ls->current)) goto err_xesc;
c += 9 << 4;
}
c += (next(ls) & 15u);
if (!lj_char_isdigit(ls->current)) {
if (!lj_char_isxdigit(ls->current)) goto err_xesc;
c += 9;
}
if (c >= 0x0800) {
save(ls, 0xE0 | (c >> 12));
save(ls, 0x80 | ((c >> 6) & 0x3f));
c = 0x80 | (c & 0x3f);
}
else if (c >= 0x0080) {
save(ls, 0xC0 | (c >> 6));
c = 0x80 | (c & 0x3f);
}
break;
case 'U': /* Unicode escape '\UXXXXXXXX'. */
c = (next(ls) & 15u) << 28;
if (!lj_char_isdigit(ls->current)) {
if (!lj_char_isxdigit(ls->current)) goto err_xesc;
c += 9 << 28;
}
c += (next(ls) & 15u) << 24;
if (!lj_char_isdigit(ls->current)) {
if (!lj_char_isxdigit(ls->current)) goto err_xesc;
c += 9 << 24;
}
c += (next(ls) & 15u) << 20;
if (!lj_char_isdigit(ls->current)) {
if (!lj_char_isxdigit(ls->current)) goto err_xesc;
c += 9 << 20;
}
c += (next(ls) & 15u) << 16;
if (!lj_char_isdigit(ls->current)) {
if (!lj_char_isxdigit(ls->current)) goto err_xesc;
c += 9 << 16;
}
c += (next(ls) & 15u) << 12;
if (!lj_char_isdigit(ls->current)) {
if (!lj_char_isxdigit(ls->current)) goto err_xesc;
c += 9 << 12;
}
c += (next(ls) & 15u) << 8;
if (!lj_char_isdigit(ls->current)) {
if (!lj_char_isxdigit(ls->current)) goto err_xesc;
c += 9 << 8;
}
c += (next(ls) & 15u) << 4;
if (!lj_char_isdigit(ls->current)) {
if (!lj_char_isxdigit(ls->current)) goto err_xesc;
c += 9 << 4;
}
c += (next(ls) & 15u);
if (!lj_char_isdigit(ls->current)) {
if (!lj_char_isxdigit(ls->current)) goto err_xesc;
c += 9;
}
if (c >= 0x4000000) {
save(ls, 0xFC | (c >> 30));
save(ls, 0x80 | ((c >> 24) & 0x3f));
save(ls, 0x80 | ((c >> 18) & 0x3f));
save(ls, 0x80 | ((c >> 12) & 0x3f));
save(ls, 0x80 | ((c >> 6) & 0x3f));
c = 0x80 | (c & 0x3f);
}
else if (c >= 0x200000) {
/* Parse a string. */
static void lex_string(LexState *ls, TValue *tv)
{
LexChar delim = ls->c; /* Delimiter is '\'' or '"'. */
lex_savenext(ls);
while (ls->c != delim) {
switch (ls->c) {
case LEX_EOF:
lj_lex_error(ls, TK_eof, LJ_ERR_XSTR);
continue;
case '\n':
case '\r':
lj_lex_error(ls, TK_string, LJ_ERR_XSTR);
continue;
case '\\': {
LexChar c = lex_next(ls); /* Skip the '\\'. */
switch (c) {
case 'a': c = '\a'; break;
case 'b': c = '\b'; break;
case 'f': c = '\f'; break;
case 'n': c = '\n'; break;
case 'r': c = '\r'; break;
case 't': c = '\t'; break;
case 'v': c = '\v'; break;
case 'x': /* Hexadecimal escape '\xXX'. */
c = (lex_next(ls) & 15u) << 4;
if (!lj_char_isdigit(ls->c)) {
if (!lj_char_isxdigit(ls->c))
lj_lex_error(ls, TK_string, LJ_ERR_XHEX);
c += 9 << 4;
}
c += (lex_next(ls) & 15u);
if (!lj_char_isdigit(ls->c)) {
if (!lj_char_isxdigit(ls->c))
lj_lex_error(ls, TK_string, LJ_ERR_XHEX);
c += 9;
}
break;
case 'u': /* Unicode escape '\u{XX...}'. */
if (lex_next(ls) != '{') goto err_xesc;
lex_next(ls);
c = 0;
do {
c = (c << 4) | (ls->c & 15u);
if (!lj_char_isdigit(ls->c)) {
if (!lj_char_isxdigit(ls->c)) goto err_xesc;
c += 9;
}
if (c >= 0x110000) goto err_xesc; /* Out of Unicode range. */
} while (lex_next(ls) != '}');
if (c < 0x800) {
if (c < 0x80) break;
lex_save(ls, 0xc0 | (c >> 6));
} else {
if (c >= 0x10000) {
lex_save(ls, 0xf0 | (c >> 18));
lex_save(ls, 0x80 | ((c >> 12) & 0x3f));
} else {
if (c >= 0xd800 && c < 0xe000) goto err_xesc; /* No surrogates. */
lex_save(ls, 0xe0 | (c >> 12));
}
lex_save(ls, 0x80 | ((c >> 6) & 0x3f));
}
c = 0x80 | (c & 0x3f);
break;
case 'z': /* Skip whitespace. */
lex_next(ls);
while (lj_char_isspace(ls->c))
if (lex_iseol(ls)) lex_newline(ls); else lex_next(ls);
continue;
case '\n': case '\r': lex_save(ls, '\n'); lex_newline(ls); continue;
case '\\': case '\"': case '\'': break;
case LEX_EOF: continue;
default:
if (!lj_char_isdigit(c))
goto err_xesc;
c -= '0'; /* Decimal escape '\ddd'. */
if (lj_char_isdigit(lex_next(ls))) {
c = c*10 + (ls->c - '0');
if (lj_char_isdigit(lex_next(ls))) {
c = c*10 + (ls->c - '0');
if (c > 255) {
err_xesc:
lj_lex_error(ls, TK_string, LJ_ERR_XESC);
}
lex_next(ls);
}
}
lex_save(ls, c);
continue;
}
lex_save(ls, c);
lex_next(ls);
continue;
}
