/*
- * Copyright (C) 2012, 2013
+ * Copyright (C) 2012, 2013, 2014
* Wolfgang Bumiller
*
* Permission is hereby granted, free of charge, to any person obtaining a copy of
* OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
* SOFTWARE.
*/
-#include <ctype.h>
#include <string.h>
#include <stdlib.h>
#include "gmqcc.h"
#include "lexer.h"
+
/*
* List of Keywords
*/
"return",
"const"
};
-static size_t num_keywords_qc = sizeof(keywords_qc) / sizeof(keywords_qc[0]);
-
/* For fte/gmgqcc */
static const char *keywords_fg[] = {
"switch", "case", "default",
"__builtin_debug_printtype"
};
-static size_t num_keywords_fg = sizeof(keywords_fg) / sizeof(keywords_fg[0]);
/*
* Lexer code
*/
-
static char* *lex_filenames;
static void lexerror(lex_file *lex, const char *fmt, ...)
static bool lexwarn(lex_file *lex, int warntype, const char *fmt, ...)
{
- bool r;
- lex_ctx ctx;
- va_list ap;
+ bool r;
+ lex_ctx_t ctx;
+ va_list ap;
- ctx.file = lex->name;
- ctx.line = lex->sline;
+ ctx.file = lex->name;
+ ctx.line = lex->sline;
+ ctx.column = lex->column;
va_start(ap, fmt);
r = vcompile_warning(ctx, warntype, fmt, ap);
return r;
}
-
-#if 0
-token* token_new()
-{
- token *tok = (token*)mem_a(sizeof(token));
- if (!tok)
- return NULL;
- memset(tok, 0, sizeof(*tok));
- return tok;
-}
-
-void token_delete(token *self)
-{
- if (self->next && self->next->prev == self)
- self->next->prev = self->prev;
- if (self->prev && self->prev->next == self)
- self->prev->next = self->next;
- MEM_VECTOR_CLEAR(self, value);
- mem_d(self);
-}
-
-token* token_copy(const token *cp)
-{
- token* self = token_new();
- if (!self)
- return NULL;
- /* copy the value */
- self->value_alloc = cp->value_count + 1;
- self->value_count = cp->value_count;
- self->value = (char*)mem_a(self->value_alloc);
- if (!self->value) {
- mem_d(self);
- return NULL;
- }
- memcpy(self->value, cp->value, cp->value_count);
- self->value[self->value_alloc-1] = 0;
-
- /* rest */
- self->ctx = cp->ctx;
- self->ttype = cp->ttype;
- memcpy(&self->constval, &cp->constval, sizeof(self->constval));
- return self;
-}
-
-void token_delete_all(token *t)
-{
- token *n;
-
- do {
- n = t->next;
- token_delete(t);
- t = n;
- } while(t);
-}
-
-token* token_copy_all(const token *cp)
-{
- token *cur;
- token *out;
-
- out = cur = token_copy(cp);
- if (!out)
- return NULL;
-
- while (cp->next) {
- cp = cp->next;
- cur->next = token_copy(cp);
- if (!cur->next) {
- token_delete_all(out);
- return NULL;
- }
- cur->next->prev = cur;
- cur = cur->next;
- }
-
- return out;
-}
-#else
static void lex_token_new(lex_file *lex)
{
-#if 0
- if (lex->tok)
- token_delete(lex->tok);
- lex->tok = token_new();
-#else
if (lex->tok.value)
vec_shrinkto(lex->tok.value, 0);
lex->tok.ctx.line = lex->sline;
lex->tok.ctx.file = lex->name;
lex->tok.ctx.column = lex->column;
-#endif
}
-#endif
+
+static void lex_ungetch(lex_file *lex, int ch);
+static int lex_getch(lex_file *lex);
lex_file* lex_open(const char *file)
{
- lex_file *lex;
- FILE *in = fs_file_open(file, "rb");
+ lex_file *lex;
+ fs_file_t *in = fs_file_open(file, "rb");
+ uint32_t read;
if (!in) {
lexerror(NULL, "open failed: '%s'\n", file);
lex->peekpos = 0;
lex->eof = false;
+ /* handle BOM */
+ if ((read = (lex_getch(lex) << 16) | (lex_getch(lex) << 8) | lex_getch(lex)) != 0xEFBBBF) {
+ lex_ungetch(lex, (read & 0x0000FF));
+ lex_ungetch(lex, (read & 0x00FF00) >> 8);
+ lex_ungetch(lex, (read & 0xFF0000) >> 16);
+ } else {
+ /*
+ * otherwise the lexer has advanced 3 bytes for the BOM, we need
+ * to set the column back to 0
+ */
+ lex->column = 0;
+ }
+
vec_push(lex_filenames, lex->name);
return lex;
}
if (lex->file)
fs_file_close(lex->file);
-#if 0
- if (lex->tok)
- token_delete(lex->tok);
-#else
+
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) {
}
if (lex->open_string) {
if (lex->open_string_pos >= lex->open_string_length)
- return EOF;
+ return FS_FILE_EOF;
lex->column++;
return lex->open_string[lex->open_string_pos++];
}
- return EOF;
+ return FS_FILE_EOF;
}
/* Get or put-back data
* are working on.
* The are merely wrapping get/put in order to count line numbers.
*/
-static void lex_ungetch(lex_file *lex, int ch);
static int lex_try_trigraph(lex_file *lex, int old)
{
int c2, c3;
if (lex->peekpos) {
lex->peekpos--;
- if (!lex->push_line && lex->peek[lex->peekpos] == '\n')
+ if (!lex->push_line && lex->peek[lex->peekpos] == '\n') {
lex->line++;
+ lex->column = 0;
+ }
return lex->peek[lex->peekpos];
}
ch = lex_fgetc(lex);
- if (!lex->push_line && ch == '\n')
+ if (!lex->push_line && ch == '\n') {
lex->line++;
+ lex->column = 0;
+ }
else if (ch == '?')
return lex_try_trigraph(lex, ch);
else if (!lex->flags.nodigraphs && (ch == '<' || ch == ':' || ch == '%'))
/* Idents are alphanumberic, but they start with alpha or _ */
static bool isident_start(int ch)
{
- return isalpha(ch) || ch == '_';
+ return util_isalpha(ch) || ch == '_';
}
static bool isident(int ch)
{
- return isident_start(ch) || isdigit(ch);
+ return isident_start(ch) || util_isdigit(ch);
}
/* isxdigit_only is used when we already know it's not a digit
goto unroll;
lex->line = line;
- while (ch != '\n' && ch != EOF)
+ while (ch != '\n' && ch != FS_FILE_EOF)
ch = lex_getch(lex);
vec_free(command);
vec_free(param);
do
{
ch = lex_getch(lex);
- while (ch != EOF && isspace(ch)) {
+ while (ch != FS_FILE_EOF && util_isspace(ch)) {
if (ch == '\n') {
if (lex_try_pragma(lex))
continue;
if (lex->flags.preprocessing) {
haswhite = true;
- /*
- lex_tokench(lex, '/');
- lex_tokench(lex, '/');
- */
lex_tokench(lex, ' ');
lex_tokench(lex, ' ');
}
- while (ch != EOF && ch != '\n') {
+ while (ch != FS_FILE_EOF && ch != '\n') {
if (lex->flags.preprocessing)
lex_tokench(lex, ' '); /* ch); */
ch = lex_getch(lex);
/* multiline comment */
if (lex->flags.preprocessing) {
haswhite = true;
- /*
- lex_tokench(lex, '/');
- lex_tokench(lex, '*');
- */
lex_tokench(lex, ' ');
lex_tokench(lex, ' ');
}
- while (ch != EOF)
+ while (ch != FS_FILE_EOF)
{
ch = lex_getch(lex);
if (ch == '*') {
ch = lex_getch(lex);
if (ch == '/') {
if (lex->flags.preprocessing) {
- /*
- lex_tokench(lex, '*');
- lex_tokench(lex, '/');
- */
lex_tokench(lex, ' ');
lex_tokench(lex, ' ');
}
if (ch == '\n')
lex_tokench(lex, '\n');
else
- lex_tokench(lex, ' '); /* ch); */
+ lex_tokench(lex, ' ');
}
}
ch = ' '; /* cause TRUE in the isspace check */
ch = '/';
break;
}
- } while (ch != EOF && isspace(ch));
+ } while (ch != FS_FILE_EOF && util_isspace(ch));
if (haswhite) {
lex_endtoken(lex);
int ch;
ch = lex_getch(lex);
- while (ch != EOF && isident(ch))
+ while (ch != FS_FILE_EOF && isident(ch))
{
lex_tokench(lex, ch);
ch = lex_getch(lex);
lex_token_new(lex);
ch = lex_getch(lex);
- while (ch != EOF && ch != '\n' && isspace(ch))
+ while (ch != FS_FILE_EOF && ch != '\n' && util_isspace(ch))
ch = lex_getch(lex);
if (ch == '\n')
static int GMQCC_WARN lex_finish_string(lex_file *lex, int quote)
{
- uchar_t chr;
+ utf8ch_t chr = 0;
int ch = 0;
int nextch;
bool hex;
+ bool oct;
char u8buf[8]; /* way more than enough */
int u8len, uc;
- while (ch != EOF)
+ while (ch != FS_FILE_EOF)
{
ch = lex_getch(lex);
if (ch == quote)
if (lex->flags.preprocessing && ch == '\\') {
lex_tokench(lex, ch);
ch = lex_getch(lex);
- if (ch == EOF) {
+ if (ch == FS_FILE_EOF) {
lexerror(lex, "unexpected end of file");
- lex_ungetch(lex, EOF); /* next token to be TOKEN_EOF */
+ lex_ungetch(lex, FS_FILE_EOF); /* next token to be TOKEN_EOF */
return (lex->tok.ttype = TOKEN_ERROR);
}
lex_tokench(lex, ch);
}
else if (ch == '\\') {
ch = lex_getch(lex);
- if (ch == EOF) {
+ if (ch == FS_FILE_EOF) {
lexerror(lex, "unexpected end of file");
- lex_ungetch(lex, EOF); /* next token to be TOKEN_EOF */
+ lex_ungetch(lex, FS_FILE_EOF); /* next token to be TOKEN_EOF */
return (lex->tok.ttype = TOKEN_ERROR);
}
chr = 0;
nextch = lex_getch(lex);
hex = (nextch == 'x');
- if (!hex)
+ oct = (nextch == '0');
+ if (!hex && !oct)
lex_ungetch(lex, nextch);
for (nextch = lex_getch(lex); nextch != '}'; nextch = lex_getch(lex)) {
- if (!hex) {
+ if (!hex && !oct) {
if (nextch >= '0' && nextch <= '9')
chr = chr * 10 + nextch - '0';
else {
lexerror(lex, "bad character code");
return (lex->tok.ttype = TOKEN_ERROR);
}
- } else {
+ } else if (!oct) {
if (nextch >= '0' && nextch <= '9')
chr = chr * 0x10 + nextch - '0';
else if (nextch >= 'a' && nextch <= 'f')
lexerror(lex, "bad character code");
return (lex->tok.ttype = TOKEN_ERROR);
}
+ } else {
+ if (nextch >= '0' && nextch <= '9')
+ chr = chr * 8 + chr - '0';
+ else {
+ lexerror(lex, "bad character code");
+ return (lex->tok.ttype = TOKEN_ERROR);
+ }
}
if (chr > 0x10FFFF || (!OPTS_FLAG(UTF8) && chr > 255))
{
}
}
if (OPTS_FLAG(UTF8) && chr >= 128) {
- u8len = u8_fromchar(chr, u8buf, sizeof(u8buf));
+ u8len = utf8_from(u8buf, chr);
if (!u8len)
ch = 0;
else {
lex->column += u8len;
for (uc = 0; uc < u8len; ++uc)
lex_tokench(lex, u8buf[uc]);
- /* the last character will be inserted with the tokench() call
+ /*
+ * the last character will be inserted with the tokench() call
* below the switch
*/
ch = u8buf[uc];
lex_tokench(lex, ch);
}
lexerror(lex, "unexpected end of file within string constant");
- lex_ungetch(lex, EOF); /* next token to be TOKEN_EOF */
+ lex_ungetch(lex, FS_FILE_EOF); /* next token to be TOKEN_EOF */
return (lex->tok.ttype = TOKEN_ERROR);
}
lex_tokench(lex, ch);
ch = lex_getch(lex);
- if (ch != '.' && !isdigit(ch))
+ if (ch != '.' && !util_isdigit(ch))
{
if (lastch != '0' || ch != 'x')
{
{
lex_tokench(lex, ch);
ch = lex_getch(lex);
- while (isdigit(ch) || (ishex && isxdigit_only(ch)))
+ while (util_isdigit(ch) || (ishex && isxdigit_only(ch)))
{
lex_tokench(lex, ch);
ch = lex_getch(lex);
/* continue digits-only */
ch = lex_getch(lex);
- while (isdigit(ch))
+ while (util_isdigit(ch))
{
lex_tokench(lex, ch);
ch = lex_getch(lex);
bool hadwhite = false;
lex_token_new(lex);
-#if 0
- if (!lex->tok)
- return TOKEN_FATAL;
-#endif
while (true) {
ch = lex_skipwhite(lex, hadwhite);
if (lex->eof)
return (lex->tok.ttype = TOKEN_FATAL);
- if (ch == EOF) {
+ if (ch == FS_FILE_EOF) {
lex->eof = true;
return (lex->tok.ttype = TOKEN_EOF);
}
if (!strcmp(v, "framevalue"))
{
ch = lex_getch(lex);
- while (ch != EOF && isspace(ch) && ch != '\n')
+ while (ch != FS_FILE_EOF && util_isspace(ch) && ch != '\n')
ch = lex_getch(lex);
- if (!isdigit(ch)) {
+ if (!util_isdigit(ch)) {
lexerror(lex, "$framevalue requires an integer parameter");
return lex_do(lex);
}
vec_free(lex->frames);
/* skip line (fteqcc does it too) */
ch = lex_getch(lex);
- while (ch != EOF && ch != '\n')
+ while (ch != FS_FILE_EOF && ch != '\n')
ch = lex_getch(lex);
return lex_do(lex);
}
{
/* skip line */
ch = lex_getch(lex);
- while (ch != EOF && ch != '\n')
+ while (ch != FS_FILE_EOF && ch != '\n')
ch = lex_getch(lex);
return lex_do(lex);
}
if (ch == '.') {
nextch = lex_getch(lex);
/* digits starting with a dot */
- if (isdigit(nextch)) {
+ if (util_isdigit(nextch)) {
lex_ungetch(lex, nextch);
lex->tok.ttype = lex_finish_digit(lex, ch);
lex_endtoken(lex);
*/
switch (ch)
{
- /*
- case '+':
- case '-':
- */
case '*':
case '/':
case '<':
}
if (ch == '+' || ch == '-' || /* ++, --, +=, -= and -> as well! */
- ch == '>' || ch == '<' || /* <<, >>, <=, >= */
+ ch == '>' || ch == '<' || /* <<, >>, <=, >= and >< as well! */
ch == '=' || ch == '!' || /* <=>, ==, != */
ch == '&' || ch == '|' || /* &&, ||, &=, |= */
ch == '~' || ch == '^' /* ~=, ~, ^ */
) {
lex_tokench(lex, ch);
-
nextch = lex_getch(lex);
- if ((nextch == '=' && ch != '<') || (nextch == ch && ch != '!')) {
+
+ if ((nextch == '=' && ch != '<') || (nextch == '<' && ch == '>'))
+ lex_tokench(lex, nextch);
+ else if (nextch == ch && ch != '!') {
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);
if ((thirdch = lex_getch(lex)) == '>')
}
}
else if (lex->flags.preprocessing &&
- ch == '-' && isdigit(nextch))
+ ch == '-' && util_isdigit(nextch))
{
lex->tok.ttype = lex_finish_digit(lex, nextch);
if (lex->tok.ttype == TOKEN_INTCONST)
return (lex->tok.ttype = TOKEN_OPERATOR);
}
- /*
- if (ch == '^' || ch == '~' || ch == '!')
- {
- lex_tokench(lex, ch);
- lex_endtoken(lex);
- return (lex->tok.ttype = TOKEN_OPERATOR);
- }
- */
-
if (ch == '*' || ch == '/') /* *=, /= */
{
lex_tokench(lex, ch);
return (lex->tok.ttype = TOKEN_OPERATOR);
}
+ /* length operator */
+ if (ch == 'l') {
+ if ((nextch = lex_getch(lex)) == 'e') {
+ if ((nextch = lex_getch(lex)) == 'n') {
+ if ((nextch = lex_getch(lex)) == 'g') {
+ if ((nextch = lex_getch(lex)) == 't') {
+ if ((nextch = lex_getch(lex)) == 'h') {
+ lex_tokench(lex, 'l');
+ lex_tokench(lex, 'e');
+ lex_tokench(lex, 'n');
+ lex_tokench(lex, 'g');
+ lex_tokench(lex, 't');
+ lex_tokench(lex, 'h');
+ lex_endtoken(lex);
+ return (lex->tok.ttype = TOKEN_OPERATOR);
+ } else lex_ungetch(lex, nextch);
+ } else lex_ungetch(lex, nextch);
+ } else lex_ungetch(lex, nextch);
+ } else lex_ungetch(lex, nextch);
+ } else lex_ungetch(lex, nextch);
+ }
+
if (isident_start(ch))
{
const char *v;
lex->tok.constval.t = TYPE_VECTOR;
} else {
size_t kw;
- for (kw = 0; kw < num_keywords_qc; ++kw) {
+ for (kw = 0; kw < GMQCC_ARRAY_COUNT(keywords_qc); ++kw) {
if (!strcmp(v, keywords_qc[kw]))
return (lex->tok.ttype = TOKEN_KEYWORD);
}
if (OPTS_OPTION_U32(OPTION_STANDARD) != COMPILER_QCC) {
- for (kw = 0; kw < num_keywords_fg; ++kw) {
+ for (kw = 0; kw < GMQCC_ARRAY_COUNT(keywords_fg); ++kw) {
if (!strcmp(v, keywords_fg[kw]))
return (lex->tok.ttype = TOKEN_KEYWORD);
}
lex_endtoken(lex);
lex->tok.ttype = TOKEN_CHARCONST;
- /* It's a vector if we can successfully scan 3 floats */
-#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
- if (sscanf(lex->tok.value, " %f %f %f ",
+
+ /* It's a vector if we can successfully scan 3 floats */
+ if (util_sscanf(lex->tok.value, " %f %f %f ",
&lex->tok.constval.v.x, &lex->tok.constval.v.y, &lex->tok.constval.v.z) == 3)
-#endif
{
lex->tok.ttype = TOKEN_VECTORCONST;
else
{
if (!lex->flags.preprocessing && strlen(lex->tok.value) > 1) {
- uchar_t u8char;
+ utf8ch_t u8char;
/* check for a valid utf8 character */
- if (!OPTS_FLAG(UTF8) || !u8_analyze(lex->tok.value, NULL, NULL, &u8char, 8)) {
+ if (!OPTS_FLAG(UTF8) || !utf8_to(&u8char, (const unsigned char *)lex->tok.value, 8)) {
if (lexwarn(lex, WARN_MULTIBYTE_CHARACTER,
( OPTS_FLAG(UTF8) ? "invalid multibyte character sequence `%s`"
: "multibyte character: `%s`" ),
return lex->tok.ttype;
}
- if (isdigit(ch))
+ if (util_isdigit(ch))
{
lex->tok.ttype = lex_finish_digit(lex, ch);
lex_endtoken(lex);
return (lex->tok.ttype = ch);
}
- lexerror(lex, "unknown token: `%s`", lex->tok.value);
+ lexerror(lex, "unknown token: `%c`", ch);
return (lex->tok.ttype = TOKEN_ERROR);
}