dfhack/depends/lua/src/llex.c

569 lines
16 KiB
C

2011-08-08 15:19:44 -06:00
/*
2021-08-05 21:04:43 -06:00
** $Id: llex.c,v 2.96.1.1 2017/04/19 17:20:42 roberto Exp $
2011-08-08 15:19:44 -06:00
** Lexical Analyzer
** See Copyright Notice in lua.h
*/
2016-07-26 21:23:26 -06:00
#define llex_c
#define LUA_CORE
#include "lprefix.h"
2011-08-08 15:19:44 -06:00
#include <locale.h>
#include <string.h>
#include "lua.h"
#include "lctype.h"
2016-07-26 21:23:26 -06:00
#include "ldebug.h"
2011-08-08 15:19:44 -06:00
#include "ldo.h"
2016-07-26 21:23:26 -06:00
#include "lgc.h"
2011-08-08 15:19:44 -06:00
#include "llex.h"
#include "lobject.h"
#include "lparser.h"
#include "lstate.h"
#include "lstring.h"
#include "ltable.h"
#include "lzio.h"
#define next(ls) (ls->current = zgetc(ls->z))
#define currIsNewline(ls) (ls->current == '\n' || ls->current == '\r')
/* ORDER RESERVED */
static const char *const luaX_tokens [] = {
2011-08-08 15:19:44 -06:00
"and", "break", "do", "else", "elseif",
"end", "false", "for", "function", "goto", "if",
2011-08-08 15:19:44 -06:00
"in", "local", "nil", "not", "or", "repeat",
"return", "then", "true", "until", "while",
2016-07-26 21:23:26 -06:00
"//", "..", "...", "==", ">=", "<=", "~=",
"<<", ">>", "::", "<eof>",
"<number>", "<integer>", "<name>", "<string>"
2011-08-08 15:19:44 -06:00
};
#define save_and_next(ls) (save(ls, ls->current), next(ls))
static l_noret lexerror (LexState *ls, const char *msg, int token);
2011-08-08 15:19:44 -06:00
static void save (LexState *ls, int c) {
Mbuffer *b = ls->buff;
if (luaZ_bufflen(b) + 1 > luaZ_sizebuffer(b)) {
2011-08-08 15:19:44 -06:00
size_t newsize;
2016-07-26 21:23:26 -06:00
if (luaZ_sizebuffer(b) >= MAX_SIZE/2)
lexerror(ls, "lexical element too long", 0);
newsize = luaZ_sizebuffer(b) * 2;
2011-08-08 15:19:44 -06:00
luaZ_resizebuffer(ls->L, b, newsize);
}
b->buffer[luaZ_bufflen(b)++] = cast(char, c);
2011-08-08 15:19:44 -06:00
}
void luaX_init (lua_State *L) {
int i;
2016-07-26 21:23:26 -06:00
TString *e = luaS_newliteral(L, LUA_ENV); /* create env name */
luaC_fix(L, obj2gco(e)); /* never collect this name */
2011-08-08 15:19:44 -06:00
for (i=0; i<NUM_RESERVED; i++) {
TString *ts = luaS_new(L, luaX_tokens[i]);
2016-07-26 21:23:26 -06:00
luaC_fix(L, obj2gco(ts)); /* reserved words are never collected */
ts->extra = cast_byte(i+1); /* reserved word */
2011-08-08 15:19:44 -06:00
}
}
const char *luaX_token2str (LexState *ls, int token) {
2014-03-31 04:15:09 -06:00
if (token < FIRST_RESERVED) { /* single-byte symbols? */
2016-07-26 21:23:26 -06:00
lua_assert(token == cast_uchar(token));
return luaO_pushfstring(ls->L, "'%c'", token);
}
else {
const char *s = luaX_tokens[token - FIRST_RESERVED];
2014-03-31 04:15:09 -06:00
if (token < TK_EOS) /* fixed format (symbols and reserved words)? */
2016-07-26 21:23:26 -06:00
return luaO_pushfstring(ls->L, "'%s'", s);
2014-03-31 04:15:09 -06:00
else /* names, strings, and numerals */
return s;
2011-08-08 15:19:44 -06:00
}
}
static const char *txtToken (LexState *ls, int token) {
switch (token) {
2016-07-26 21:23:26 -06:00
case TK_NAME: case TK_STRING:
case TK_FLT: case TK_INT:
2011-08-08 15:19:44 -06:00
save(ls, '\0');
2016-07-26 21:23:26 -06:00
return luaO_pushfstring(ls->L, "'%s'", luaZ_buffer(ls->buff));
2011-08-08 15:19:44 -06:00
default:
return luaX_token2str(ls, token);
}
}
static l_noret lexerror (LexState *ls, const char *msg, int token) {
2016-07-26 21:23:26 -06:00
msg = luaG_addinfo(ls->L, msg, ls->source, ls->linenumber);
2011-08-08 15:19:44 -06:00
if (token)
luaO_pushfstring(ls->L, "%s near %s", msg, txtToken(ls, token));
2011-08-08 15:19:44 -06:00
luaD_throw(ls->L, LUA_ERRSYNTAX);
}
l_noret luaX_syntaxerror (LexState *ls, const char *msg) {
lexerror(ls, msg, ls->t.token);
2011-08-08 15:19:44 -06:00
}
/*
2016-07-26 21:23:26 -06:00
** creates a new string and anchors it in scanner's table so that
** it will not be collected until the end of the compilation
** (by that time it should be anchored somewhere)
*/
2011-08-08 15:19:44 -06:00
TString *luaX_newstring (LexState *ls, const char *str, size_t l) {
lua_State *L = ls->L;
2016-07-26 21:23:26 -06:00
TValue *o; /* entry for 'str' */
TString *ts = luaS_newlstr(L, str, l); /* create new string */
setsvalue2s(L, L->top++, ts); /* temporarily anchor it in stack */
2016-07-26 21:23:26 -06:00
o = luaH_set(L, ls->h, L->top - 1);
if (ttisnil(o)) { /* not in use yet? */
/* boolean value does not need GC barrier;
table has no metatable, so it does not need to invalidate cache */
setbvalue(o, 1); /* t[string] = true */
luaC_checkGC(L);
}
2014-03-31 04:15:09 -06:00
else { /* string already present */
2016-07-26 21:23:26 -06:00
ts = tsvalue(keyfromval(o)); /* re-use value previously stored */
2014-03-31 04:15:09 -06:00
}
L->top--; /* remove string from stack */
2011-08-08 15:19:44 -06:00
return ts;
}
/*
** increment line number and skips newline sequence (any of
** \n, \r, \n\r, or \r\n)
*/
2011-08-08 15:19:44 -06:00
static void inclinenumber (LexState *ls) {
int old = ls->current;
lua_assert(currIsNewline(ls));
2016-07-26 21:23:26 -06:00
next(ls); /* skip '\n' or '\r' */
2011-08-08 15:19:44 -06:00
if (currIsNewline(ls) && ls->current != old)
2016-07-26 21:23:26 -06:00
next(ls); /* skip '\n\r' or '\r\n' */
2011-08-08 15:19:44 -06:00
if (++ls->linenumber >= MAX_INT)
2016-07-26 21:23:26 -06:00
lexerror(ls, "chunk has too many lines", 0);
2011-08-08 15:19:44 -06:00
}
void luaX_setinput (lua_State *L, LexState *ls, ZIO *z, TString *source,
int firstchar) {
2016-07-26 21:23:26 -06:00
ls->t.token = 0;
2011-08-08 15:19:44 -06:00
ls->L = L;
ls->current = firstchar;
2011-08-08 15:19:44 -06:00
ls->lookahead.token = TK_EOS; /* no look-ahead token */
ls->z = z;
ls->fs = NULL;
ls->linenumber = 1;
ls->lastline = 1;
ls->source = source;
2016-07-26 21:23:26 -06:00
ls->envn = luaS_newliteral(L, LUA_ENV); /* get env name */
2011-08-08 15:19:44 -06:00
luaZ_resizebuffer(ls->L, ls->buff, LUA_MINBUFFER); /* initialize buffer */
}
/*
** =======================================================
** LEXICAL ANALYZER
** =======================================================
*/
2016-07-26 21:23:26 -06:00
static int check_next1 (LexState *ls, int c) {
if (ls->current == c) {
next(ls);
return 1;
}
else return 0;
2011-08-08 15:19:44 -06:00
}
/*
2016-07-26 21:23:26 -06:00
** Check whether current char is in set 'set' (with two chars) and
** saves it
*/
2016-07-26 21:23:26 -06:00
static int check_next2 (LexState *ls, const char *set) {
lua_assert(set[2] == '\0');
if (ls->current == set[0] || ls->current == set[1]) {
save_and_next(ls);
return 1;
2011-08-08 15:19:44 -06:00
}
2016-07-26 21:23:26 -06:00
else return 0;
2011-08-08 15:19:44 -06:00
}
/* LUA_NUMBER */
2012-06-17 05:59:03 -06:00
/*
2016-07-26 21:23:26 -06:00
** this function is quite liberal in what it accepts, as 'luaO_str2num'
2012-06-17 05:59:03 -06:00
** will reject ill-formed numerals.
*/
2016-07-26 21:23:26 -06:00
static int read_numeral (LexState *ls, SemInfo *seminfo) {
TValue obj;
2012-06-17 05:59:03 -06:00
const char *expo = "Ee";
int first = ls->current;
lua_assert(lisdigit(ls->current));
2012-06-17 05:59:03 -06:00
save_and_next(ls);
2016-07-26 21:23:26 -06:00
if (first == '0' && check_next2(ls, "xX")) /* hexadecimal? */
2012-06-17 05:59:03 -06:00
expo = "Pp";
for (;;) {
2016-07-26 21:23:26 -06:00
if (check_next2(ls, expo)) /* exponent part? */
check_next2(ls, "-+"); /* optional exponent sign */
if (lisxdigit(ls->current))
save_and_next(ls);
else if (ls->current == '.')
2012-06-17 05:59:03 -06:00
save_and_next(ls);
2016-07-26 21:23:26 -06:00
else break;
2012-06-17 05:59:03 -06:00
}
2011-08-08 15:19:44 -06:00
save(ls, '\0');
2016-07-26 21:23:26 -06:00
if (luaO_str2num(luaZ_buffer(ls->buff), &obj) == 0) /* format error? */
lexerror(ls, "malformed number", TK_FLT);
if (ttisinteger(&obj)) {
seminfo->i = ivalue(&obj);
return TK_INT;
}
else {
lua_assert(ttisfloat(&obj));
seminfo->r = fltvalue(&obj);
return TK_FLT;
}
2011-08-08 15:19:44 -06:00
}
/*
2021-08-05 21:04:43 -06:00
** reads a sequence '[=*[' or ']=*]', leaving the last bracket.
** If sequence is well formed, return its number of '='s + 2; otherwise,
** return 1 if there is no '='s or 0 otherwise (an unfinished '[==...').
*/
2021-08-05 21:04:43 -06:00
static size_t skip_sep (LexState *ls) {
size_t count = 0;
2011-08-08 15:19:44 -06:00
int s = ls->current;
lua_assert(s == '[' || s == ']');
save_and_next(ls);
while (ls->current == '=') {
save_and_next(ls);
count++;
}
2021-08-05 21:04:43 -06:00
return (ls->current == s) ? count + 2
: (count == 0) ? 1
: 0;
2011-08-08 15:19:44 -06:00
}
2021-08-05 21:04:43 -06:00
static void read_long_string (LexState *ls, SemInfo *seminfo, size_t sep) {
2016-07-26 21:23:26 -06:00
int line = ls->linenumber; /* initial line (for error message) */
save_and_next(ls); /* skip 2nd '[' */
2011-08-08 15:19:44 -06:00
if (currIsNewline(ls)) /* string starts with a newline? */
inclinenumber(ls); /* skip it */
for (;;) {
switch (ls->current) {
2016-07-26 21:23:26 -06:00
case EOZ: { /* error */
const char *what = (seminfo ? "string" : "comment");
const char *msg = luaO_pushfstring(ls->L,
"unfinished long %s (starting at line %d)", what, line);
lexerror(ls, msg, TK_EOS);
2011-08-08 15:19:44 -06:00
break; /* to avoid warnings */
2016-07-26 21:23:26 -06:00
}
2011-08-08 15:19:44 -06:00
case ']': {
if (skip_sep(ls) == sep) {
2016-07-26 21:23:26 -06:00
save_and_next(ls); /* skip 2nd ']' */
2011-08-08 15:19:44 -06:00
goto endloop;
}
break;
}
case '\n': case '\r': {
2011-08-08 15:19:44 -06:00
save(ls, '\n');
inclinenumber(ls);
if (!seminfo) luaZ_resetbuffer(ls->buff); /* avoid wasting space */
break;
}
default: {
if (seminfo) save_and_next(ls);
else next(ls);
}
}
} endloop:
if (seminfo)
2021-08-05 21:04:43 -06:00
seminfo->ts = luaX_newstring(ls, luaZ_buffer(ls->buff) + sep,
luaZ_bufflen(ls->buff) - 2 * sep);
2011-08-08 15:19:44 -06:00
}
2016-07-26 21:23:26 -06:00
static void esccheck (LexState *ls, int c, const char *msg) {
if (!c) {
if (ls->current != EOZ)
save_and_next(ls); /* add current to buffer for error message */
lexerror(ls, msg, TK_STRING);
}
}
static int gethexa (LexState *ls) {
save_and_next(ls);
esccheck (ls, lisxdigit(ls->current), "hexadecimal digit expected");
return luaO_hexavalue(ls->current);
}
static int readhexaesc (LexState *ls) {
2016-07-26 21:23:26 -06:00
int r = gethexa(ls);
r = (r << 4) + gethexa(ls);
luaZ_buffremove(ls->buff, 2); /* remove saved chars from buffer */
return r;
}
static unsigned long readutf8esc (LexState *ls) {
unsigned long r;
int i = 4; /* chars to be removed: '\', 'u', '{', and first digit */
save_and_next(ls); /* skip 'u' */
esccheck(ls, ls->current == '{', "missing '{'");
r = gethexa(ls); /* must have at least one digit */
while ((save_and_next(ls), lisxdigit(ls->current))) {
i++;
r = (r << 4) + luaO_hexavalue(ls->current);
esccheck(ls, r <= 0x10FFFF, "UTF-8 value too large");
}
2016-07-26 21:23:26 -06:00
esccheck(ls, ls->current == '}', "missing '}'");
next(ls); /* skip '}' */
luaZ_buffremove(ls->buff, i); /* remove saved chars from buffer */
return r;
}
2016-07-26 21:23:26 -06:00
static void utf8esc (LexState *ls) {
char buff[UTF8BUFFSZ];
int n = luaO_utf8esc(buff, readutf8esc(ls));
for (; n > 0; n--) /* add 'buff' to string */
save(ls, buff[UTF8BUFFSZ - n]);
}
static int readdecesc (LexState *ls) {
2016-07-26 21:23:26 -06:00
int i;
int r = 0; /* result accumulator */
for (i = 0; i < 3 && lisdigit(ls->current); i++) { /* read up to 3 digits */
2016-07-26 21:23:26 -06:00
r = 10*r + ls->current - '0';
save_and_next(ls);
}
2016-07-26 21:23:26 -06:00
esccheck(ls, r <= UCHAR_MAX, "decimal escape too large");
luaZ_buffremove(ls->buff, i); /* remove read digits from buffer */
return r;
}
2011-08-08 15:19:44 -06:00
static void read_string (LexState *ls, int del, SemInfo *seminfo) {
save_and_next(ls); /* keep delimiter (for error messages) */
2011-08-08 15:19:44 -06:00
while (ls->current != del) {
switch (ls->current) {
case EOZ:
lexerror(ls, "unfinished string", TK_EOS);
break; /* to avoid warnings */
2011-08-08 15:19:44 -06:00
case '\n':
case '\r':
lexerror(ls, "unfinished string", TK_STRING);
break; /* to avoid warnings */
case '\\': { /* escape sequences */
int c; /* final character to be saved */
2016-07-26 21:23:26 -06:00
save_and_next(ls); /* keep '\\' for error messages */
2011-08-08 15:19:44 -06:00
switch (ls->current) {
case 'a': c = '\a'; goto read_save;
case 'b': c = '\b'; goto read_save;
case 'f': c = '\f'; goto read_save;
case 'n': c = '\n'; goto read_save;
case 'r': c = '\r'; goto read_save;
case 't': c = '\t'; goto read_save;
case 'v': c = '\v'; goto read_save;
case 'x': c = readhexaesc(ls); goto read_save;
2016-07-26 21:23:26 -06:00
case 'u': utf8esc(ls); goto no_save;
case '\n': case '\r':
inclinenumber(ls); c = '\n'; goto only_save;
case '\\': case '\"': case '\'':
c = ls->current; goto read_save;
case EOZ: goto no_save; /* will raise an error next loop */
case 'z': { /* zap following span of spaces */
2016-07-26 21:23:26 -06:00
luaZ_buffremove(ls->buff, 1); /* remove '\\' */
next(ls); /* skip the 'z' */
while (lisspace(ls->current)) {
if (currIsNewline(ls)) inclinenumber(ls);
else next(ls);
2011-08-08 15:19:44 -06:00
}
goto no_save;
}
default: {
2016-07-26 21:23:26 -06:00
esccheck(ls, lisdigit(ls->current), "invalid escape sequence");
c = readdecesc(ls); /* digital escape '\ddd' */
goto only_save;
2011-08-08 15:19:44 -06:00
}
}
2016-07-26 21:23:26 -06:00
read_save:
next(ls);
/* go through */
only_save:
luaZ_buffremove(ls->buff, 1); /* remove '\\' */
save(ls, c);
/* go through */
no_save: break;
2011-08-08 15:19:44 -06:00
}
default:
save_and_next(ls);
}
}
save_and_next(ls); /* skip delimiter */
seminfo->ts = luaX_newstring(ls, luaZ_buffer(ls->buff) + 1,
luaZ_bufflen(ls->buff) - 2);
}
static int llex (LexState *ls, SemInfo *seminfo) {
luaZ_resetbuffer(ls->buff);
for (;;) {
switch (ls->current) {
case '\n': case '\r': { /* line breaks */
2011-08-08 15:19:44 -06:00
inclinenumber(ls);
break;
2011-08-08 15:19:44 -06:00
}
case ' ': case '\f': case '\t': case '\v': { /* spaces */
next(ls);
break;
}
case '-': { /* '-' or '--' (comment) */
2011-08-08 15:19:44 -06:00
next(ls);
if (ls->current != '-') return '-';
/* else is a comment */
next(ls);
if (ls->current == '[') { /* long comment? */
2021-08-05 21:04:43 -06:00
size_t sep = skip_sep(ls);
2016-07-26 21:23:26 -06:00
luaZ_resetbuffer(ls->buff); /* 'skip_sep' may dirty the buffer */
2021-08-05 21:04:43 -06:00
if (sep >= 2) {
read_long_string(ls, NULL, sep); /* skip long comment */
luaZ_resetbuffer(ls->buff); /* previous call may dirty the buff. */
break;
2011-08-08 15:19:44 -06:00
}
}
/* else short comment */
while (!currIsNewline(ls) && ls->current != EOZ)
next(ls); /* skip until end of line (or end of file) */
break;
2011-08-08 15:19:44 -06:00
}
case '[': { /* long string or simply '[' */
2021-08-05 21:04:43 -06:00
size_t sep = skip_sep(ls);
if (sep >= 2) {
2011-08-08 15:19:44 -06:00
read_long_string(ls, seminfo, sep);
return TK_STRING;
}
2021-08-05 21:04:43 -06:00
else if (sep == 0) /* '[=...' missing second bracket */
2016-07-26 21:23:26 -06:00
lexerror(ls, "invalid long string delimiter", TK_STRING);
return '[';
2011-08-08 15:19:44 -06:00
}
case '=': {
next(ls);
2016-07-26 21:23:26 -06:00
if (check_next1(ls, '=')) return TK_EQ;
else return '=';
2011-08-08 15:19:44 -06:00
}
case '<': {
next(ls);
2016-07-26 21:23:26 -06:00
if (check_next1(ls, '=')) return TK_LE;
else if (check_next1(ls, '<')) return TK_SHL;
else return '<';
2011-08-08 15:19:44 -06:00
}
case '>': {
next(ls);
2016-07-26 21:23:26 -06:00
if (check_next1(ls, '=')) return TK_GE;
else if (check_next1(ls, '>')) return TK_SHR;
else return '>';
}
case '/': {
next(ls);
if (check_next1(ls, '/')) return TK_IDIV;
else return '/';
2011-08-08 15:19:44 -06:00
}
case '~': {
next(ls);
2016-07-26 21:23:26 -06:00
if (check_next1(ls, '=')) return TK_NE;
else return '~';
2011-08-08 15:19:44 -06:00
}
case ':': {
next(ls);
2016-07-26 21:23:26 -06:00
if (check_next1(ls, ':')) return TK_DBCOLON;
else return ':';
}
case '"': case '\'': { /* short literal strings */
2011-08-08 15:19:44 -06:00
read_string(ls, ls->current, seminfo);
return TK_STRING;
}
case '.': { /* '.', '..', '...', or number */
2011-08-08 15:19:44 -06:00
save_and_next(ls);
2016-07-26 21:23:26 -06:00
if (check_next1(ls, '.')) {
if (check_next1(ls, '.'))
return TK_DOTS; /* '...' */
else return TK_CONCAT; /* '..' */
2011-08-08 15:19:44 -06:00
}
else if (!lisdigit(ls->current)) return '.';
2016-07-26 21:23:26 -06:00
else return read_numeral(ls, seminfo);
}
case '0': case '1': case '2': case '3': case '4':
case '5': case '6': case '7': case '8': case '9': {
2016-07-26 21:23:26 -06:00
return read_numeral(ls, seminfo);
2011-08-08 15:19:44 -06:00
}
case EOZ: {
return TK_EOS;
}
default: {
if (lislalpha(ls->current)) { /* identifier or reserved word? */
2011-08-08 15:19:44 -06:00
TString *ts;
do {
save_and_next(ls);
} while (lislalnum(ls->current));
2011-08-08 15:19:44 -06:00
ts = luaX_newstring(ls, luaZ_buffer(ls->buff),
luaZ_bufflen(ls->buff));
seminfo->ts = ts;
2012-06-17 05:59:03 -06:00
if (isreserved(ts)) /* reserved word? */
2016-07-26 21:23:26 -06:00
return ts->extra - 1 + FIRST_RESERVED;
2011-08-08 15:19:44 -06:00
else {
return TK_NAME;
}
}
else { /* single-char tokens (+ - / ...) */
2011-08-08 15:19:44 -06:00
int c = ls->current;
next(ls);
return c;
2011-08-08 15:19:44 -06:00
}
}
}
}
}
void luaX_next (LexState *ls) {
ls->lastline = ls->linenumber;
if (ls->lookahead.token != TK_EOS) { /* is there a look-ahead token? */
ls->t = ls->lookahead; /* use this one */
ls->lookahead.token = TK_EOS; /* and discharge it */
}
else
ls->t.token = llex(ls, &ls->t.seminfo); /* read next token */
}
int luaX_lookahead (LexState *ls) {
2011-08-08 15:19:44 -06:00
lua_assert(ls->lookahead.token == TK_EOS);
ls->lookahead.token = llex(ls, &ls->lookahead.seminfo);
return ls->lookahead.token;
2011-08-08 15:19:44 -06:00
}