#include "gmqcc.h"
#include "lexer.h"
-MEM_VEC_FUNCTIONS(token, char, value)
-MEM_VEC_FUNCTIONS(lex_file, frame_macro, frames)
-
-VECTOR_MAKE(char*, lex_filenames);
+char* *lex_filenames;
void lexerror(lex_file *lex, const char *fmt, ...)
{
va_list ap;
va_start(ap, fmt);
- vprintmsg(LVL_ERROR, lex->name, lex->sline, "parse error", fmt, ap);
+ if (lex)
+ con_vprintmsg(LVL_ERROR, lex->name, lex->sline, "parse error", fmt, ap);
+ else
+ con_vprintmsg(LVL_ERROR, "", 0, "parse error", fmt, ap);
va_end(ap);
}
lvl = LVL_ERROR;
va_start(ap, fmt);
- vprintmsg(lvl, lex->name, lex->sline, "warning", fmt, ap);
+ con_vprintmsg(lvl, lex->name, lex->sline, "warning", fmt, ap);
va_end(ap);
return opts_werror;
token_delete(lex->tok);
lex->tok = token_new();
#else
- lex->tok.value_count = 0;
+ if (lex->tok.value)
+ vec_shrinkto(lex->tok.value, 0);
lex->tok.constval.t = 0;
lex->tok.ctx.line = lex->sline;
lex->tok.ctx.file = lex->name;
lex->peekpos = 0;
lex->eof = false;
- lex_filenames_add(lex->name);
+ vec_push(lex_filenames, lex->name);
+ return lex;
+}
+
+lex_file* lex_open_string(const char *str, size_t len, const char *name)
+{
+ lex_file *lex;
+
+ lex = (lex_file*)mem_a(sizeof(*lex));
+ if (!lex) {
+ lexerror(NULL, "out of memory\n");
+ return NULL;
+ }
+
+ memset(lex, 0, sizeof(*lex));
+
+ lex->file = NULL;
+ lex->open_string = str;
+ lex->open_string_length = len;
+ lex->open_string_pos = 0;
+
+ lex->name = util_strdup(name ? name : "<string-source>");
+ lex->line = 1; /* we start counting at 1 */
+
+ lex->peekpos = 0;
+ lex->eof = false;
+
+ vec_push(lex_filenames, lex->name);
return lex;
}
void lex_cleanup(void)
{
size_t i;
- for (i = 0; i < lex_filenames_elements; ++i)
- mem_d(lex_filenames_data[i]);
- mem_d(lex_filenames_data);
+ for (i = 0; i < vec_size(lex_filenames); ++i)
+ mem_d(lex_filenames[i]);
+ vec_free(lex_filenames);
}
void lex_close(lex_file *lex)
{
size_t i;
- for (i = 0; i < lex->frames_count; ++i)
+ for (i = 0; i < vec_size(lex->frames); ++i)
mem_d(lex->frames[i].name);
- MEM_VECTOR_CLEAR(lex, frames);
+ vec_free(lex->frames);
if (lex->modelname)
- mem_d(lex->modelname);
+ vec_free(lex->modelname);
if (lex->file)
fclose(lex->file);
if (lex->tok)
token_delete(lex->tok);
#else
- MEM_VECTOR_CLEAR(&(lex->tok), value);
+ vec_free(lex->tok.value);
#endif
/* mem_d(lex->name); collected in lex_filenames */
mem_d(lex);
}
+static int lex_fgetc(lex_file *lex)
+{
+ if (lex->file)
+ return fgetc(lex->file);
+ if (lex->open_string) {
+ if (lex->open_string_pos >= lex->open_string_length)
+ return EOF;
+ return lex->open_string[lex->open_string_pos++];
+ }
+ return EOF;
+}
+
/* Get or put-back data
* The following to functions do NOT understand what kind of data they
* are working on.
static int lex_try_trigraph(lex_file *lex, int old)
{
int c2, c3;
- c2 = fgetc(lex->file);
+ c2 = lex_fgetc(lex);
if (c2 != '?') {
lex_ungetch(lex, c2);
return old;
}
- c3 = fgetc(lex->file);
+ c3 = lex_fgetc(lex);
switch (c3) {
case '=': return '#';
case '/': return '\\';
static int lex_try_digraph(lex_file *lex, int ch)
{
int c2;
- c2 = fgetc(lex->file);
+ c2 = lex_fgetc(lex);
if (ch == '<' && c2 == ':')
return '[';
else if (ch == ':' && c2 == '>')
return lex->peek[lex->peekpos];
}
- ch = fgetc(lex->file);
+ ch = lex_fgetc(lex);
if (ch == '\n')
lex->line++;
else if (ch == '?')
}
/* Append a character to the token buffer */
-static bool GMQCC_WARN lex_tokench(lex_file *lex, int ch)
+static void lex_tokench(lex_file *lex, int ch)
{
- if (!token_value_add(&lex->tok, ch)) {
- lexerror(lex, "out of memory");
- return false;
- }
- return true;
+ vec_push(lex->tok.value, ch);
}
/* Append a trailing null-byte */
-static bool GMQCC_WARN lex_endtoken(lex_file *lex)
+static void lex_endtoken(lex_file *lex)
{
- if (!token_value_add(&lex->tok, 0)) {
- lexerror(lex, "out of memory");
- return false;
- }
- lex->tok.value_count--;
- return true;
+ vec_push(lex->tok.value, 0);
+ vec_shrinkby(lex->tok.value, 1);
}
/* Skip whitespace and comments and return the first
if (ch == '\n') {
/* end-of-line */
/* see if there was whitespace first */
- if (haswhite) { /* (lex->tok.value_count) { */
+ if (haswhite) { /* (vec_size(lex->tok.value)) { */
lex_ungetch(lex, ch);
- if (!lex_endtoken(lex))
- return TOKEN_FATAL;
+ lex_endtoken(lex);
return TOKEN_WHITE;
}
/* otherwise return EOL */
return TOKEN_EOL;
}
haswhite = true;
- if (!lex_tokench(lex, ch))
- return TOKEN_FATAL;
+ lex_tokench(lex, ch);
}
ch = lex_getch(lex);
}
if (lex->flags.preprocessing) {
haswhite = true;
- if (!lex_tokench(lex, '/') ||
- !lex_tokench(lex, '/'))
- {
- return TOKEN_FATAL;
- }
+ /*
+ lex_tokench(lex, '/');
+ lex_tokench(lex, '/');
+ */
+ lex_tokench(lex, ' ');
+ lex_tokench(lex, ' ');
}
while (ch != EOF && ch != '\n') {
- if (lex->flags.preprocessing && !lex_tokench(lex, ch))
- return TOKEN_FATAL;
+ if (lex->flags.preprocessing)
+ lex_tokench(lex, ' '); /* ch); */
ch = lex_getch(lex);
}
if (lex->flags.preprocessing) {
lex_ungetch(lex, '\n');
- if (!lex_endtoken(lex))
- return TOKEN_FATAL;
+ lex_endtoken(lex);
return TOKEN_WHITE;
}
continue;
/* multiline comment */
if (lex->flags.preprocessing) {
haswhite = true;
- if (!lex_tokench(lex, '/') ||
- !lex_tokench(lex, '*'))
- {
- return TOKEN_FATAL;
- }
+ /*
+ lex_tokench(lex, '/');
+ lex_tokench(lex, '*');
+ */
+ lex_tokench(lex, ' ');
+ lex_tokench(lex, ' ');
}
while (ch != EOF)
ch = lex_getch(lex);
if (ch == '/') {
if (lex->flags.preprocessing) {
- if (!lex_tokench(lex, '*') ||
- !lex_tokench(lex, '/'))
- {
- return TOKEN_FATAL;
- }
+ /*
+ lex_tokench(lex, '*');
+ lex_tokench(lex, '/');
+ */
+ lex_tokench(lex, ' ');
+ lex_tokench(lex, ' ');
}
break;
}
}
if (lex->flags.preprocessing) {
- if (!lex_tokench(lex, ch))
- return TOKEN_FATAL;
+ lex_tokench(lex, ' '); /* ch); */
}
}
ch = ' '; /* cause TRUE in the isspace check */
} while (ch != EOF && isspace(ch));
if (haswhite) {
- if (!lex_endtoken(lex))
- return TOKEN_FATAL;
+ lex_endtoken(lex);
lex_ungetch(lex, ch);
return TOKEN_WHITE;
}
ch = lex_getch(lex);
while (ch != EOF && isident(ch))
{
- if (!lex_tokench(lex, ch))
- return (lex->tok.ttype = TOKEN_FATAL);
+ lex_tokench(lex, ch);
ch = lex_getch(lex);
}
return -1;
}
- if (!lex_tokench(lex, ch))
- return -1;
+ lex_tokench(lex, ch);
if (!lex_finish_ident(lex))
return -1;
- if (!lex_endtoken(lex))
- return -1;
+ lex_endtoken(lex);
return 0;
}
if (rc < 0) /* error */
return false;
- for (i = 0; i < lex->frames_count; ++i) {
+ for (i = 0; i < vec_size(lex->frames); ++i) {
if (!strcmp(lex->tok.value, lex->frames[i].name)) {
lex->frames[i].value = lex->framevalue++;
if (lexwarn(lex, WARN_FRAME_MACROS, "duplicate frame macro defined: `%s`", lex->tok.value))
break;
}
}
- if (i < lex->frames_count)
+ if (i < vec_size(lex->frames))
continue;
m.value = lex->framevalue++;
- m.name = lex->tok.value;
- lex->tok.value = NULL;
- lex->tok.value_alloc = lex->tok.value_count = 0;
- if (!lex_file_frames_add(lex, m)) {
- lexerror(lex, "out of memory");
- return false;
- }
+ m.name = util_strdup(lex->tok.value);
+ vec_shrinkto(lex->tok.value, 0);
+ vec_push(lex->frames, m);
} while (true);
}
default:
lexwarn(lex, WARN_UNKNOWN_CONTROL_SEQUENCE, "unrecognized control sequence: \\%c", ch);
/* so we just add the character plus backslash no matter what it actually is */
- if (!lex_tokench(lex, '\\'))
- return (lex->tok.ttype = TOKEN_FATAL);
+ lex_tokench(lex, '\\');
}
/* add the character finally */
- if (!lex_tokench(lex, ch))
- return (lex->tok.ttype = TOKEN_FATAL);
+ lex_tokench(lex, ch);
}
- else if (!lex_tokench(lex, ch))
- return (lex->tok.ttype = TOKEN_FATAL);
+ else
+ lex_tokench(lex, ch);
}
lexerror(lex, "unexpected end of file within string constant");
lex_ungetch(lex, EOF); /* next token to be TOKEN_EOF */
/* parse a number... */
lex->tok.ttype = TOKEN_INTCONST;
- if (!lex_tokench(lex, ch))
- return (lex->tok.ttype = TOKEN_FATAL);
+ lex_tokench(lex, ch);
ch = lex_getch(lex);
if (ch != '.' && !isdigit(ch))
{
/* end of the number or EOF */
lex_ungetch(lex, ch);
- if (!lex_endtoken(lex))
- return (lex->tok.ttype = TOKEN_FATAL);
+ lex_endtoken(lex);
lex->tok.constval.i = lastch - '0';
return lex->tok.ttype;
if (ch != '.')
{
- if (!lex_tokench(lex, ch))
- return (lex->tok.ttype = TOKEN_FATAL);
+ lex_tokench(lex, ch);
ch = lex_getch(lex);
while (isdigit(ch) || (ishex && isxdigit_only(ch)))
{
- if (!lex_tokench(lex, ch))
- return (lex->tok.ttype = TOKEN_FATAL);
+ lex_tokench(lex, ch);
ch = lex_getch(lex);
}
}
{
/* Allow floating comma in non-hex mode */
lex->tok.ttype = TOKEN_FLOATCONST;
- if (!lex_tokench(lex, ch))
- return (lex->tok.ttype = TOKEN_FATAL);
+ lex_tokench(lex, ch);
/* continue digits-only */
ch = lex_getch(lex);
while (isdigit(ch))
{
- if (!lex_tokench(lex, ch))
- return (lex->tok.ttype = TOKEN_FATAL);
+ lex_tokench(lex, ch);
ch = lex_getch(lex);
}
}
}
lex_ungetch(lex, ch);
- if (!lex_endtoken(lex))
- return (lex->tok.ttype = TOKEN_FATAL);
+ lex_endtoken(lex);
if (lex->tok.ttype == TOKEN_FLOATCONST)
lex->tok.constval.f = strtod(lex->tok.value, NULL);
else
return TOKEN_FATAL;
#endif
- ch = lex_skipwhite(lex);
+ while (true) {
+ ch = lex_skipwhite(lex);
+ if (!lex->flags.mergelines || ch != '\\')
+ break;
+ ch = lex_getch(lex);
+ if (ch != '\n') {
+ lex_ungetch(lex, ch);
+ ch = '\\';
+ break;
+ }
+ /* we reached a linemerge */
+ lex_tokench(lex, '\n');
+ continue;
+ }
+
lex->sline = lex->line;
lex->tok.ctx.line = lex->sline;
lex->tok.ctx.file = lex->name;
lexerror(lex, "hanging '$' modelgen/spritegen command line");
return lex_do(lex);
}
- if (!lex_tokench(lex, ch))
- return (lex->tok.ttype = TOKEN_FATAL);
+ lex_tokench(lex, ch);
if (!lex_finish_ident(lex))
return (lex->tok.ttype = TOKEN_ERROR);
- if (!lex_endtoken(lex))
- return (lex->tok.ttype = TOKEN_FATAL);
+ lex_endtoken(lex);
/* skip the known commands */
v = lex->tok.value;
lex_token_new(lex);
lex->tok.ttype = lex_finish_digit(lex, ch);
- if (!lex_endtoken(lex))
- return (lex->tok.ttype = TOKEN_FATAL);
+ lex_endtoken(lex);
if (lex->tok.ttype != TOKEN_INTCONST) {
lexerror(lex, "$framevalue requires an integer parameter");
return lex_do(lex);
return (lex->tok.ttype = TOKEN_FATAL);
v = lex->tok.value;
- for (frame = 0; frame < lex->frames_count; ++frame) {
+ for (frame = 0; frame < vec_size(lex->frames); ++frame) {
if (!strcmp(v, lex->frames[frame].name)) {
lex->framevalue = lex->frames[frame].value;
return lex_do(lex);
rc = lex_parse_frame(lex);
if (rc > 0) {
- lexerror(lex, "$framerestore requires a framename parameter");
+ lexerror(lex, "$modelname requires a parameter");
return lex_do(lex);
}
if (rc < 0)
m.value = lex->framevalue;
m.name = lex->modelname;
lex->modelname = NULL;
- if (!lex_file_frames_add(lex, m)) {
- lexerror(lex, "out of memory");
- return (lex->tok.ttype = TOKEN_FATAL);
- }
+ vec_push(lex->frames, m);
}
lex->modelname = lex->tok.value;
lex->tok.value = NULL;
- lex->tok.value_alloc = lex->tok.value_count = 0;
- for (frame = 0; frame < lex->frames_count; ++frame) {
- if (!strcmp(v, lex->frames[frame].name)) {
- lex->framevalue = lex->frames[frame].value;
- break;
- }
- }
return lex_do(lex);
}
if (!strcmp(v, "flush"))
{
size_t frame;
- for (frame = 0; frame < lex->frames_count; ++frame)
+ for (frame = 0; frame < vec_size(lex->frames); ++frame)
mem_d(lex->frames[frame].name);
- MEM_VECTOR_CLEAR(lex, frames);
+ vec_free(lex->frames);
/* skip line (fteqcc does it too) */
ch = lex_getch(lex);
while (ch != EOF && ch != '\n')
return lex_do(lex);
}
- for (frame = 0; frame < lex->frames_count; ++frame) {
+ for (frame = 0; frame < vec_size(lex->frames); ++frame) {
if (!strcmp(v, lex->frames[frame].name)) {
lex->tok.constval.i = lex->frames[frame].value;
return (lex->tok.ttype = TOKEN_INTCONST);
/* single-character tokens */
switch (ch)
{
+ case '[':
case '(':
- if (!lex_tokench(lex, ch) ||
- !lex_endtoken(lex))
- {
- return (lex->tok.ttype = TOKEN_FATAL);
- }
+ lex_tokench(lex, ch);
+ lex_endtoken(lex);
if (lex->flags.noops)
return (lex->tok.ttype = ch);
else
return (lex->tok.ttype = TOKEN_OPERATOR);
case ')':
case ';':
+ case ':':
case '{':
case '}':
- case '[':
case ']':
case '#':
- if (!lex_tokench(lex, ch) ||
- !lex_endtoken(lex))
- {
- return (lex->tok.ttype = TOKEN_FATAL);
- }
+ lex_tokench(lex, ch);
+ lex_endtoken(lex);
return (lex->tok.ttype = ch);
default:
break;
case '~':
case ',':
case '!':
- if (!lex_tokench(lex, ch) ||
- !lex_endtoken(lex))
- {
- return (lex->tok.ttype = TOKEN_FATAL);
- }
+ lex_tokench(lex, ch);
+ lex_endtoken(lex);
return (lex->tok.ttype = ch);
default:
break;
if (ch == '.')
{
- if (!lex_tokench(lex, ch))
- return (lex->tok.ttype = TOKEN_FATAL);
+ lex_tokench(lex, ch);
/* peak ahead once */
nextch = lex_getch(lex);
if (nextch != '.') {
lex_ungetch(lex, nextch);
- if (!lex_endtoken(lex))
- return (lex->tok.ttype = TOKEN_FATAL);
+ lex_endtoken(lex);
return (lex->tok.ttype = ch);
}
/* peak ahead again */
if (nextch != '.') {
lex_ungetch(lex, nextch);
lex_ungetch(lex, nextch);
- if (!lex_endtoken(lex))
- return (lex->tok.ttype = TOKEN_FATAL);
+ lex_endtoken(lex);
return (lex->tok.ttype = ch);
}
/* fill the token to be "..." */
- if (!lex_tokench(lex, ch) ||
- !lex_tokench(lex, ch) ||
- !lex_endtoken(lex))
- {
- return (lex->tok.ttype = TOKEN_FATAL);
- }
+ lex_tokench(lex, ch);
+ lex_tokench(lex, ch);
+ lex_endtoken(lex);
return (lex->tok.ttype = TOKEN_DOTS);
}
}
if (ch == ',' || ch == '.') {
- if (!lex_tokench(lex, ch) ||
- !lex_endtoken(lex))
- {
- return (lex->tok.ttype = TOKEN_FATAL);
- }
+ lex_tokench(lex, ch);
+ lex_endtoken(lex);
return (lex->tok.ttype = TOKEN_OPERATOR);
}
ch == '=' || ch == '!' || /* ==, != */
ch == '&' || ch == '|') /* &&, ||, &=, |= */
{
- if (!lex_tokench(lex, ch))
- return (lex->tok.ttype = TOKEN_FATAL);
+ lex_tokench(lex, ch);
nextch = lex_getch(lex);
if (nextch == ch || nextch == '=') {
- if (!lex_tokench(lex, nextch))
- return (lex->tok.ttype = TOKEN_FATAL);
+ lex_tokench(lex, nextch);
} else if (ch == '-' && nextch == '>') {
- if (!lex_tokench(lex, nextch))
- return (lex->tok.ttype = TOKEN_FATAL);
+ lex_tokench(lex, nextch);
} else
lex_ungetch(lex, nextch);
- if (!lex_endtoken(lex))
- return (lex->tok.ttype = TOKEN_FATAL);
+ lex_endtoken(lex);
return (lex->tok.ttype = TOKEN_OPERATOR);
}
/*
if (ch == '^' || ch == '~' || ch == '!')
{
- if (!lex_tokench(lex, ch) ||
- !lex_endtoken(lex))
- {
- return (lex->tok.ttype = TOKEN_FATAL);
- }
+ lex_tokench(lex, ch);
+ lex_endtoken(lex);
return (lex->tok.ttype = TOKEN_OPERATOR);
}
*/
if (ch == '*' || ch == '/') /* *=, /= */
{
- if (!lex_tokench(lex, ch))
- return (lex->tok.ttype = TOKEN_FATAL);
+ lex_tokench(lex, ch);
nextch = lex_getch(lex);
if (nextch == '=') {
- if (!lex_tokench(lex, nextch))
- return (lex->tok.ttype = TOKEN_FATAL);
+ lex_tokench(lex, nextch);
} else
lex_ungetch(lex, nextch);
- if (!lex_endtoken(lex))
- return (lex->tok.ttype = TOKEN_FATAL);
+ lex_endtoken(lex);
return (lex->tok.ttype = TOKEN_OPERATOR);
}
{
const char *v;
- if (!lex_tokench(lex, ch))
- return (lex->tok.ttype = TOKEN_FATAL);
+ lex_tokench(lex, ch);
if (!lex_finish_ident(lex)) {
/* error? */
return (lex->tok.ttype = TOKEN_ERROR);
}
- if (!lex_endtoken(lex))
- return (lex->tok.ttype = TOKEN_FATAL);
+ lex_endtoken(lex);
lex->tok.ttype = TOKEN_IDENT;
v = lex->tok.value;
!strcmp(v, "else") ||
!strcmp(v, "local") ||
!strcmp(v, "return") ||
+ !strcmp(v, "not") ||
!strcmp(v, "const"))
{
lex->tok.ttype = TOKEN_KEYWORD;
!strcmp(v, "struct") ||
!strcmp(v, "union") ||
!strcmp(v, "break") ||
- !strcmp(v, "continue"))
+ !strcmp(v, "continue") ||
+ !strcmp(v, "var"))
{
lex->tok.ttype = TOKEN_KEYWORD;
}
if (ch == '"')
{
lex->flags.nodigraphs = true;
- if (lex->flags.preprocessing && !lex_tokench(lex, ch))
- return TOKEN_FATAL;
+ if (lex->flags.preprocessing)
+ lex_tokench(lex, ch);
lex->tok.ttype = lex_finish_string(lex, '"');
- if (lex->flags.preprocessing && !lex_tokench(lex, ch))
- return TOKEN_FATAL;
+ if (lex->flags.preprocessing)
+ lex_tokench(lex, ch);
while (!lex->flags.preprocessing && lex->tok.ttype == TOKEN_STRINGCONST)
{
/* Allow c style "string" "continuation" */
lex->tok.ttype = lex_finish_string(lex, '"');
}
lex->flags.nodigraphs = false;
- if (!lex_endtoken(lex))
- return (lex->tok.ttype = TOKEN_FATAL);
+ lex_endtoken(lex);
return lex->tok.ttype;
}
* Likewise actual unescaping has to be done by the parser.
* The difference is we don't allow 'char' 'continuation'.
*/
- if (lex->flags.preprocessing && !lex_tokench(lex, ch))
- return TOKEN_FATAL;
+ if (lex->flags.preprocessing)
+ lex_tokench(lex, ch);
lex->tok.ttype = lex_finish_string(lex, '\'');
- if (lex->flags.preprocessing && !lex_tokench(lex, ch))
- return TOKEN_FATAL;
- if (!lex_endtoken(lex))
- return (lex->tok.ttype = TOKEN_FATAL);
+ if (lex->flags.preprocessing)
+ lex_tokench(lex, ch);
+ lex_endtoken(lex);
/* It's a vector if we can successfully scan 3 floats */
#ifdef WIN32
if (isdigit(ch))
{
lex->tok.ttype = lex_finish_digit(lex, ch);
- if (!lex_endtoken(lex))
- return (lex->tok.ttype = TOKEN_FATAL);
+ lex_endtoken(lex);
return lex->tok.ttype;
}