/*
- * Copyright (C) 2012
+ * Copyright (C) 2012, 2013
* Wolfgang Bumiller
*
* Permission is hereby granted, free of charge, to any person obtaining a copy of
* Lexer code
*/
-char* *lex_filenames;
+static char* *lex_filenames;
-void lexerror(lex_file *lex, const char *fmt, ...)
+static void lexerror(lex_file *lex, const char *fmt, ...)
{
va_list ap;
va_end(ap);
}
-bool lexwarn(lex_file *lex, int warntype, const char *fmt, ...)
+static bool lexwarn(lex_file *lex, int warntype, const char *fmt, ...)
{
bool r;
lex_ctx ctx;
lex_file* lex_open(const char *file)
{
lex_file *lex;
- FILE *in = util_fopen(file, "rb");
+ FILE *in = fs_file_open(file, "rb");
if (!in) {
lexerror(NULL, "open failed: '%s'\n", file);
lex = (lex_file*)mem_a(sizeof(*lex));
if (!lex) {
- fclose(in);
+ fs_file_close(in);
lexerror(NULL, "out of memory\n");
return NULL;
}
vec_free(lex->modelname);
if (lex->file)
- fclose(lex->file);
+ fs_file_close(lex->file);
#if 0
if (lex->tok)
token_delete(lex->tok);
static int lex_fgetc(lex_file *lex)
{
if (lex->file)
- return fgetc(lex->file);
+ return fs_file_getc(lex->file);
if (lex->open_string) {
if (lex->open_string_pos >= lex->open_string_length)
return EOF;
{
int c2, c3;
c2 = lex_fgetc(lex);
+ if (!lex->push_line && c2 == '\n')
+ lex->line++;
if (c2 != '?') {
lex_ungetch(lex, c2);
return old;
}
c3 = lex_fgetc(lex);
+ if (!lex->push_line && c3 == '\n')
+ lex->line++;
switch (c3) {
case '=': return '#';
case '/': return '\\';
lex->line = line;
while (ch != '\n' && ch != EOF)
ch = lex_getch(lex);
+ vec_free(command);
+ vec_free(param);
+ vec_free(pragma);
return true;
unroll:
if (command) {
vec_pop(command);
while (vec_size(command)) {
- lex_ungetch(lex, vec_last(command));
+ lex_ungetch(lex, (unsigned char)vec_last(command));
vec_pop(command);
}
vec_free(command);
lex_ungetch(lex, ' ');
}
- if (command) {
- vec_pop(command);
- while (vec_size(command)) {
- lex_ungetch(lex, vec_last(command));
- vec_pop(command);
+ if (param) {
+ vec_pop(param);
+ while (vec_size(param)) {
+ lex_ungetch(lex, (unsigned char)vec_last(param));
+ vec_pop(param);
}
- vec_free(command);
+ vec_free(param);
lex_ungetch(lex, ' ');
}
if (pragma) {
vec_pop(pragma);
while (vec_size(pragma)) {
- lex_ungetch(lex, vec_last(pragma));
+ lex_ungetch(lex, (unsigned char)vec_last(pragma));
vec_pop(pragma);
}
vec_free(pragma);
vec_shrinkto(lex->tok.value, 0);
vec_push(lex->frames, m);
} while (true);
+
+ return false;
}
static int GMQCC_WARN lex_finish_string(lex_file *lex, int quote)
{
+ uchar_t chr;
int ch = 0;
int nextch;
+ bool hex;
+ char u8buf[8]; /* way more than enough */
+ int u8len, uc;
while (ch != EOF)
{
case '[': ch = 16; break;
case ']': ch = 17; break;
case '{':
- ch = 0;
+ chr = 0;
+ nextch = lex_getch(lex);
+ hex = (nextch == 'x');
+ if (!hex)
+ lex_ungetch(lex, nextch);
for (nextch = lex_getch(lex); nextch != '}'; nextch = lex_getch(lex)) {
- ch = ch * 10 + nextch - '0';
- if (nextch < '0' || nextch > '9' || ch > 255) {
- lexerror(lex, "bad character code");
+ if (!hex) {
+ if (nextch >= '0' && nextch <= '9')
+ chr = chr * 10 + nextch - '0';
+ else {
+ lexerror(lex, "bad character code");
+ return (lex->tok.ttype = TOKEN_ERROR);
+ }
+ } else {
+ if (nextch >= '0' && nextch <= '9')
+ chr = chr * 0x10 + nextch - '0';
+ else if (nextch >= 'a' && nextch <= 'f')
+ chr = chr * 0x10 + nextch - 'a' + 10;
+ else if (nextch >= 'A' && nextch <= 'F')
+ chr = chr * 0x10 + nextch - 'A' + 10;
+ else {
+ lexerror(lex, "bad character code");
+ return (lex->tok.ttype = TOKEN_ERROR);
+ }
+ }
+ if (chr > 0x10FFFF || (!OPTS_FLAG(UTF8) && chr > 255))
+ {
+ lexerror(lex, "character code out of range");
return (lex->tok.ttype = TOKEN_ERROR);
}
}
+ if (OPTS_FLAG(UTF8) && chr >= 128) {
+ u8len = u8_fromchar(chr, u8buf, sizeof(u8buf));
+ if (!u8len)
+ ch = 0;
+ else {
+ --u8len;
+ for (uc = 0; uc < u8len; ++uc)
+ lex_tokench(lex, u8buf[uc]);
+ /* the last character will be inserted with the tokench() call
+ * below the switch
+ */
+ ch = u8buf[uc];
+ }
+ }
+ else
+ ch = chr;
break;
case '\n': ch = '\n'; break;
if (!lex->flags.mergelines || ch != '\\')
break;
ch = lex_getch(lex);
+ if (ch == '\r')
+ ch = lex_getch(lex);
if (ch != '\n') {
lex_ungetch(lex, ch);
ch = '\\';
switch (ch)
{
case '[':
+ nextch = lex_getch(lex);
+ if (nextch == '[') {
+ lex_tokench(lex, ch);
+ lex_tokench(lex, nextch);
+ lex_endtoken(lex);
+ return (lex->tok.ttype = TOKEN_ATTRIBUTE_OPEN);
+ }
+ lex_ungetch(lex, nextch);
+ /* FALL THROUGH */
case '(':
case ':':
case '?':
return (lex->tok.ttype = ch);
else
return (lex->tok.ttype = TOKEN_OPERATOR);
+
+ case ']':
+ if (lex->flags.noops) {
+ nextch = lex_getch(lex);
+ if (nextch == ']') {
+ lex_tokench(lex, ch);
+ lex_tokench(lex, nextch);
+ lex_endtoken(lex);
+ return (lex->tok.ttype = TOKEN_ATTRIBUTE_CLOSE);
+ }
+ lex_ungetch(lex, nextch);
+ }
+ /* FALL THROUGH */
case ')':
case ';':
case '{':
case '}':
- case ']':
case '#':
lex_tokench(lex, ch);
default:
break;
}
+ }
- if (ch == '.')
- {
- lex_tokench(lex, ch);
- /* peak ahead once */
- nextch = lex_getch(lex);
- if (nextch != '.') {
- lex_ungetch(lex, nextch);
- lex_endtoken(lex);
- return (lex->tok.ttype = ch);
- }
- /* peak ahead again */
- nextch = lex_getch(lex);
- if (nextch != '.') {
- lex_ungetch(lex, nextch);
- lex_ungetch(lex, '.');
- lex_endtoken(lex);
+ if (ch == '.')
+ {
+ lex_tokench(lex, ch);
+ /* peak ahead once */
+ nextch = lex_getch(lex);
+ if (nextch != '.') {
+ lex_ungetch(lex, nextch);
+ lex_endtoken(lex);
+ if (lex->flags.noops)
return (lex->tok.ttype = ch);
- }
- /* fill the token to be "..." */
- lex_tokench(lex, ch);
- lex_tokench(lex, ch);
+ else
+ return (lex->tok.ttype = TOKEN_OPERATOR);
+ }
+ /* peak ahead again */
+ nextch = lex_getch(lex);
+ if (nextch != '.') {
+ lex_ungetch(lex, nextch);
+ lex_ungetch(lex, '.');
lex_endtoken(lex);
- return (lex->tok.ttype = TOKEN_DOTS);
+ if (lex->flags.noops)
+ return (lex->tok.ttype = ch);
+ else
+ return (lex->tok.ttype = TOKEN_OPERATOR);
}
+ /* fill the token to be "..." */
+ lex_tokench(lex, ch);
+ lex_tokench(lex, ch);
+ lex_endtoken(lex);
+ return (lex->tok.ttype = TOKEN_DOTS);
}
if (ch == ',' || ch == '.') {
}
if (ch == '+' || ch == '-' || /* ++, --, +=, -= and -> as well! */
- ch == '>' || ch == '<' || /* <<, >>, <=, >= */
- ch == '=' || ch == '!' || /* ==, != */
- ch == '&' || ch == '|') /* &&, ||, &=, |= */
- {
+ ch == '>' || ch == '<' || /* <<, >>, <=, >= */
+ ch == '=' || ch == '!' || /* <=>, ==, != */
+ ch == '&' || ch == '|' || /* &&, ||, &=, |= */
+ ch == '~' /* ~=, ~ */
+ ) {
lex_tokench(lex, ch);
nextch = lex_getch(lex);
- if (nextch == '=' || (nextch == ch && ch != '!')) {
+ if ((nextch == '=' && ch != '<') || (nextch == ch && ch != '!')) {
lex_tokench(lex, nextch);
+ } else if (ch == '<' && nextch == '=') {
+ lex_tokench(lex, nextch);
+ if ((thirdch = lex_getch(lex)) == '>')
+ lex_tokench(lex, thirdch);
+ else
+ lex_ungetch(lex, thirdch);
+
} else if (ch == '-' && nextch == '>') {
lex_tokench(lex, nextch);
} else if (ch == '&' && nextch == '~') {
lex_tokench(lex, nextch);
lex_tokench(lex, thirdch);
}
- } else
+ }
+ else if (lex->flags.preprocessing &&
+ ch == '-' && isdigit(nextch))
+ {
+ lex->tok.ttype = lex_finish_digit(lex, nextch);
+ if (lex->tok.ttype == TOKEN_INTCONST)
+ lex->tok.constval.i = -lex->tok.constval.i;
+ else
+ lex->tok.constval.f = -lex->tok.constval.f;
+ lex_endtoken(lex);
+ return lex->tok.ttype;
+ } else {
lex_ungetch(lex, nextch);
+ }
lex_endtoken(lex);
return (lex->tok.ttype = TOKEN_OPERATOR);
lex_tokench(lex, ch);
nextch = lex_getch(lex);
- if (nextch == '=') {
+ if (nextch == '=' || nextch == '*') {
lex_tokench(lex, nextch);
} else
lex_ungetch(lex, nextch);
return (lex->tok.ttype = TOKEN_OPERATOR);
}
+ if (ch == '%') {
+ lex_tokench(lex, ch);
+ lex_endtoken(lex);
+ return (lex->tok.ttype = TOKEN_OPERATOR);
+ }
+
if (isident_start(ch))
{
const char *v;
if (!strcmp(v, keywords_qc[kw]))
return (lex->tok.ttype = TOKEN_KEYWORD);
}
- if (opts.standard != COMPILER_QCC) {
+ if (OPTS_OPTION_U32(OPTION_STANDARD) != COMPILER_QCC) {
for (kw = 0; kw < num_keywords_fg; ++kw) {
if (!strcmp(v, keywords_fg[kw]))
return (lex->tok.ttype = TOKEN_KEYWORD);
lex->tok.ttype = TOKEN_CHARCONST;
/* It's a vector if we can successfully scan 3 floats */
-#ifdef WIN32
+#ifdef _MSC_VER
if (sscanf_s(lex->tok.value, " %f %f %f ",
&lex->tok.constval.v.x, &lex->tok.constval.v.y, &lex->tok.constval.v.z) == 3)
#else
else
{
if (!lex->flags.preprocessing && strlen(lex->tok.value) > 1) {
- if (lexwarn(lex, WARN_MULTIBYTE_CHARACTER, "multibyte character: `%s`", lex->tok.value))
- return (lex->tok.ttype = TOKEN_ERROR);
+ uchar_t u8char;
+ /* check for a valid utf8 character */
+ if (!OPTS_FLAG(UTF8) || !u8_analyze(lex->tok.value, NULL, NULL, &u8char, 8)) {
+ if (lexwarn(lex, WARN_MULTIBYTE_CHARACTER,
+ ( OPTS_FLAG(UTF8) ? "invalid multibyte character sequence `%s`"
+ : "multibyte character: `%s`" ),
+ lex->tok.value))
+ return (lex->tok.ttype = TOKEN_ERROR);
+ }
+ else
+ lex->tok.constval.i = u8char;
}
- lex->tok.constval.i = lex->tok.value[0];
+ else
+ lex->tok.constval.i = lex->tok.value[0];
}
return lex->tok.ttype;
return (lex->tok.ttype = ch);
}
- lexerror(lex, "unknown token");
+ lexerror(lex, "unknown token: `%s`", lex->tok.value);
return (lex->tok.ttype = TOKEN_ERROR);
}