DrSocalkwe3n/luajit
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Cleanup lexer source code.

Browse Source
This commit is contained in:
Mike Pall
2013-02-28 01:11:49 +01:00
parent 116cdd7e9a
commit 87c51e7f57
5 changed files with 218 additions and 204 deletions
Download Patch File Download Diff File Expand all files Collapse all files

301
src/lj_lex.c
View File

@@ -38,37 +38,48 @@ TKDEF(TKSTR1, TKSTR2)
/* -- Buffer handling ----------------------------------------------------- */
#define char2int(c) ((int)(uint8_t)(c))
#define next(ls) \
(ls->current = (ls->n--) > 0 ? char2int(*ls->p++) : fillbuf(ls))
#define save_and_next(ls) (save(ls, ls->current), next(ls))
#define currIsNewline(ls) (ls->current == '\n' || ls->current == '\r')
#define END_OF_STREAM (-1)
#define LEX_EOF (-1)
#define lex_iseol(ls) (ls->c == '\n' || ls->c == '\r')
static int fillbuf(LexState *ls)
/* Get more input from reader. */
static LJ_NOINLINE LexChar lex_more(LexState *ls)
{
size_t sz;
const char *buf = ls->rfunc(ls->L, ls->rdata, &sz);
if (buf == NULL || sz == 0) return END_OF_STREAM;
if (buf == NULL || sz == 0) return LEX_EOF;
ls->n = (MSize)sz - 1;
ls->p = buf;
return char2int(*(ls->p++));
return (LexChar)(uint8_t)*ls->p++;
}
static LJ_AINLINE void save(LexState *ls, int c)
/* Get next character. */
static LJ_AINLINE LexChar lex_next(LexState *ls)
{
return (ls->c = ls->n ? (ls->n--,(LexChar)(uint8_t)*ls->p++) : lex_more(ls));
}
/* Save character. */
static LJ_AINLINE void lex_save(LexState *ls, LexChar c)
{
lj_buf_putb(ls->L, &ls->sb, c);
}
static void inclinenumber(LexState *ls)
/* Save previous character and get next character. */
static LJ_AINLINE LexChar lex_savenext(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' */
lex_save(ls, ls->c);
return lex_next(ls);
}
/* 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->token, LJ_ERR_XLINES);
lj_lex_error(ls, ls->tok, LJ_ERR_XLINES);
}
/* -- Scanner for terminals ----------------------------------------------- */
@@ -77,18 +88,16 @@ static void inclinenumber(LexState *ls)
static void lex_number(LexState *ls, TValue *tv)
{
StrScanFmt fmt;
int c, xp = 'e';
lua_assert(lj_char_isdigit(ls->current));
if ((c = ls->current) == '0') {
save_and_next(ls);
if ((ls->current | 0x20) == 'x') xp = 'p';
LexChar c, xp = 'e';
lua_assert(lj_char_isdigit(ls->c));
if ((c = ls->c) == '0' && (lex_savenext(ls) | 0x20) == 'x')
xp = 'p';
while (lj_char_isident(ls->c) || ls->c == '.' ||
((ls->c == '-' || ls->c == '+') && (c | 0x20) == xp)) {
c = ls->c;
lex_savenext(ls);
}
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');
lex_save(ls, '\0');
fmt = lj_strscan_scan((const uint8_t *)sbufB(&ls->sb), tv,
(LJ_DUALNUM ? STRSCAN_OPT_TOINT : STRSCAN_OPT_TONUM) |
(LJ_HASFFI ? (STRSCAN_OPT_LL|STRSCAN_OPT_IMAG) : 0));
@@ -122,44 +131,42 @@ static void lex_number(LexState *ls, TValue *tv)
}
}
static int skip_sep(LexState *ls)
/* Skip equal signs for "[=...=[" and "]=...=]" and return their count. */
static int lex_skipeq(LexState *ls)
{
int count = 0;
int s = ls->current;
LexChar s = ls->c;
lua_assert(s == '[' || s == ']');
save_and_next(ls);
while (ls->current == '=') {
save_and_next(ls);
while (lex_savenext(ls) == '=')
count++;
}
return (ls->current == s) ? count : (-count) - 1;
return (ls->c == s) ? count : (-count) - 1;
}
static void read_long_string(LexState *ls, TValue *tv, int sep)
/* Parse a long string or long comment (tv set to NULL). */
static void lex_longstring(LexState *ls, TValue *tv, int sep)
{
save_and_next(ls); /* skip 2nd `[' */
if (currIsNewline(ls)) /* string starts with a newline? */
inclinenumber(ls); /* skip it */
lex_savenext(ls); /* Skip second '['. */
if (lex_iseol(ls)) /* Skip initial newline. */
lex_newline(ls);
for (;;) {
switch (ls->current) {
case END_OF_STREAM:
switch (ls->c) {
case LEX_EOF:
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 `]' */
if (lex_skipeq(ls) == sep) {
lex_savenext(ls); /* Skip second ']'. */
goto endloop;
}
break;
case '\n':
case '\r':
save(ls, '\n');
inclinenumber(ls);
lex_save(ls, '\n');
lex_newline(ls);
if (!tv) lj_buf_reset(&ls->sb); /* Don't waste space for comments. */
break;
default:
if (tv) save_and_next(ls);
else next(ls);
lex_savenext(ls);
break;
}
} endloop:
@@ -170,12 +177,14 @@ static void read_long_string(LexState *ls, TValue *tv, int sep)
}
}
static void read_string(LexState *ls, int delim, TValue *tv)
/* Parse a string. */
static void lex_string(LexState *ls, TValue *tv)
{
save_and_next(ls);
while (ls->current != delim) {
switch (ls->current) {
case END_OF_STREAM:
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':
@@ -183,7 +192,7 @@ static void read_string(LexState *ls, int delim, TValue *tv)
lj_lex_error(ls, TK_string, LJ_ERR_XSTR);
continue;
case '\\': {
int c = next(ls); /* Skip the '\\'. */
LexChar c = lex_next(ls); /* Skip the '\\'. */
switch (c) {
case 'a': c = '\a'; break;
case 'b': c = '\b'; break;
@@ -193,112 +202,112 @@ static void read_string(LexState *ls, int delim, TValue *tv)
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 = (lex_next(ls) & 15u) << 4;
if (!lj_char_isdigit(ls->c)) {
if (!lj_char_isxdigit(ls->c)) 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 += (lex_next(ls) & 15u);
if (!lj_char_isdigit(ls->c)) {
if (!lj_char_isxdigit(ls->c)) 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);
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': save(ls, '\n'); inclinenumber(ls); continue;
case '\n': case '\r': lex_save(ls, '\n'); lex_newline(ls); continue;
case '\\': case '\"': case '\'': break;
case END_OF_STREAM: continue;
case LEX_EOF: 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 (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);
}
next(ls);
lex_next(ls);
}
}
save(ls, c);
lex_save(ls, c);
continue;
}
save(ls, c);
next(ls);
lex_save(ls, c);
lex_next(ls);
continue;
}
default:
save_and_next(ls);
lex_savenext(ls);
break;
}
}
save_and_next(ls); /* skip delimiter */
lex_savenext(ls); /* Skip trailing delimiter. */
setstrV(ls->L, tv,
lj_parse_keepstr(ls, sbufB(&ls->sb)+1, sbuflen(&ls->sb)-2));
}
/* -- Main lexical scanner ------------------------------------------------ */
static int llex(LexState *ls, TValue *tv)
/* Get next lexical token. */
static LexToken lex_scan(LexState *ls, TValue *tv)
{
lj_buf_reset(&ls->sb);
for (;;) {
if (lj_char_isident(ls->current)) {
if (lj_char_isident(ls->c)) {
GCstr *s;
if (lj_char_isdigit(ls->current)) { /* Numeric literal. */
if (lj_char_isdigit(ls->c)) { /* Numeric literal. */
lex_number(ls, tv);
return TK_number;
}
/* Identifier or reserved word. */
do {
save_and_next(ls);
} while (lj_char_isident(ls->current));
lex_savenext(ls);
} while (lj_char_isident(ls->c));
s = lj_parse_keepstr(ls, sbufB(&ls->sb), sbuflen(&ls->sb));
setstrV(ls->L, tv, s);
if (s->reserved > 0) /* Reserved word? */
return TK_OFS + s->reserved;
return TK_name;
}
switch (ls->current) {
switch (ls->c) {
case '\n':
case '\r':
inclinenumber(ls);
lex_newline(ls);
continue;
case ' ':
case '\t':
case '\v':
case '\f':
next(ls);
lex_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_buf_reset(&ls->sb); /* `skip_sep' may dirty the buffer */
lex_next(ls);
if (ls->c != '-') return '-';
lex_next(ls);
if (ls->c == '[') { /* Long comment "--[=*[...]=*]". */
int sep = lex_skipeq(ls);
lj_buf_reset(&ls->sb); /* `lex_skipeq' may dirty the buffer */
if (sep >= 0) {
read_long_string(ls, NULL, sep); /* long comment */
lex_longstring(ls, NULL, sep);
lj_buf_reset(&ls->sb);
continue;
}
}
/* else short comment */
while (!currIsNewline(ls) && ls->current != END_OF_STREAM)
next(ls);
/* Short comment "--.*\n". */
while (!lex_iseol(ls) && ls->c != LEX_EOF)
lex_next(ls);
continue;
case '[': {
int sep = skip_sep(ls);
int sep = lex_skipeq(ls);
if (sep >= 0) {
read_long_string(ls, tv, sep);
lex_longstring(ls, tv, sep);
return TK_string;
} else if (sep == -1) {
return '[';
@@ -308,44 +317,43 @@ static int llex(LexState *ls, TValue *tv)
}
}
case '=':
next(ls);
if (ls->current != '=') return '='; else { next(ls); return TK_eq; }
lex_next(ls);
if (ls->c != '=') return '='; else { lex_next(ls); return TK_eq; }
case '<':
next(ls);
if (ls->current != '=') return '<'; else { next(ls); return TK_le; }
lex_next(ls);
if (ls->c != '=') return '<'; else { lex_next(ls); return TK_le; }
case '>':
next(ls);
if (ls->current != '=') return '>'; else { next(ls); return TK_ge; }
lex_next(ls);
if (ls->c != '=') return '>'; else { lex_next(ls); return TK_ge; }
case '~':
next(ls);
if (ls->current != '=') return '~'; else { next(ls); return TK_ne; }
lex_next(ls);
if (ls->c != '=') return '~'; else { lex_next(ls); return TK_ne; }
case ':':
next(ls);
if (ls->current != ':') return ':'; else { next(ls); return TK_label; }
lex_next(ls);
if (ls->c != ':') return ':'; else { lex_next(ls); return TK_label; }
case '"':
case '\'':
read_string(ls, ls->current, tv);
lex_string(ls, tv);
return TK_string;
case '.':
save_and_next(ls);
if (ls->current == '.') {
next(ls);
if (ls->current == '.') {
next(ls);
if (lex_savenext(ls) == '.') {
lex_next(ls);
if (ls->c == '.') {
lex_next(ls);
return TK_dots; /* ... */
}
return TK_concat; /* .. */
} else if (!lj_char_isdigit(ls->current)) {
} else if (!lj_char_isdigit(ls->c)) {
return '.';
} else {
lex_number(ls, tv);
return TK_number;
}
case END_OF_STREAM:
case LEX_EOF:
return TK_eof;
default: {
int c = ls->current;
next(ls);
LexChar c = ls->c;
lex_next(ls);
return c; /* Single-char tokens (+ - / ...). */
}
}
@@ -370,23 +378,23 @@ int lj_lex_setup(lua_State *L, LexState *ls)
ls->lookahead = TK_eof; /* No look-ahead token. */
ls->linenumber = 1;
ls->lastline = 1;
next(ls); /* Read-ahead first char. */
if (ls->current == 0xef && ls->n >= 2 && char2int(ls->p[0]) == 0xbb &&
char2int(ls->p[1]) == 0xbf) { /* Skip UTF-8 BOM (if buffered). */
lex_next(ls); /* Read-ahead first char. */
if (ls->c == 0xef && ls->n >= 2 && (uint8_t)ls->p[0] == 0xbb &&
(uint8_t)ls->p[1] == 0xbf) { /* Skip UTF-8 BOM (if buffered). */
ls->n -= 2;
ls->p += 2;
next(ls);
lex_next(ls);
header = 1;
}
if (ls->current == '#') { /* Skip POSIX #! header line. */
if (ls->c == '#') { /* Skip POSIX #! header line. */
do {
next(ls);
if (ls->current == END_OF_STREAM) return 0;
} while (!currIsNewline(ls));
inclinenumber(ls);
lex_next(ls);
if (ls->c == LEX_EOF) return 0;
} while (!lex_iseol(ls));
lex_newline(ls);
header = 1;
}
if (ls->current == LUA_SIGNATURE[0]) { /* Bytecode dump. */
if (ls->c == LUA_SIGNATURE[0]) { /* Bytecode dump. */
if (header) {
/*
** Loading bytecode with an extra header is disabled for security
@@ -411,52 +419,57 @@ void lj_lex_cleanup(lua_State *L, LexState *ls)
lj_buf_free(g, &ls->sb);
}
/* Return next lexical token. */
void lj_lex_next(LexState *ls)
{
ls->lastline = ls->linenumber;
if (LJ_LIKELY(ls->lookahead == TK_eof)) { /* No lookahead token? */
ls->token = llex(ls, &ls->tokenval); /* Get next token. */
ls->tok = lex_scan(ls, &ls->tokval); /* Get next token. */
} else { /* Otherwise return lookahead token. */
ls->token = ls->lookahead;
ls->tok = ls->lookahead;
ls->lookahead = TK_eof;
ls->tokenval = ls->lookaheadval;
ls->tokval = ls->lookaheadval;
}
}
/* Look ahead for the next token. */
LexToken lj_lex_lookahead(LexState *ls)
{
lua_assert(ls->lookahead == TK_eof);
ls->lookahead = llex(ls, &ls->lookaheadval);
ls->lookahead = lex_scan(ls, &ls->lookaheadval);
return ls->lookahead;
}
const char *lj_lex_token2str(LexState *ls, LexToken token)
/* Convert token to string. */
const char *lj_lex_token2str(LexState *ls, LexToken tok)
{
if (token > TK_OFS)
return tokennames[token-TK_OFS-1];
else if (!lj_char_iscntrl(token))
return lj_str_pushf(ls->L, "%c", token);
if (tok > TK_OFS)
return tokennames[tok-TK_OFS-1];
else if (!lj_char_iscntrl(tok))
return lj_str_pushf(ls->L, "%c", tok);
else
return lj_str_pushf(ls->L, "char(%d)", token);
return lj_str_pushf(ls->L, "char(%d)", tok);
}
void lj_lex_error(LexState *ls, LexToken token, ErrMsg em, ...)
/* Lexer error. */
void lj_lex_error(LexState *ls, LexToken tok, ErrMsg em, ...)
{
const char *tok;
const char *tokstr;
va_list argp;
if (token == 0) {
tok = NULL;
} else if (token == TK_name || token == TK_string || token == TK_number) {
save(ls, '\0');
tok = sbufB(&ls->sb);
if (tok == 0) {
tokstr = NULL;
} else if (tok == TK_name || tok == TK_string || tok == TK_number) {
lex_save(ls, '\0');
tokstr = sbufB(&ls->sb);
} else {
tok = lj_lex_token2str(ls, token);
tokstr = lj_lex_token2str(ls, tok);
}
va_start(argp, em);
lj_err_lex(ls->L, ls->chunkname, tok, ls->linenumber, em, argp);
lj_err_lex(ls->L, ls->chunkname, tokstr, ls->linenumber, em, argp);
va_end(argp);
}
/* Initialize strings for reserved words. */
void lj_lex_init(lua_State *L)
{
uint32_t i;