static token charLit(int n, int indLine, int indCol)
{
if (ch == '\'')
error(indLine, indCol, "empty literal");
if (ch == '\\')
{
next_ch();
if (ch == 'n')
n = '\n';
else
if (ch == '\\')
n = '\\';
else
error(indLine, indCol, "unknow esacape char");
}
next_ch();
if (ch != '\'')
error(indLine, indCol, "it is not character literal");
next_ch();
return (token){tk_Integer, indLine, indCol, n, NULL};
}
static token identOrInt(int indLine, int indCol)
{
int n;
int position = 0;
int isNumber = TRUE;
static char id[MAXIDENT];
while (isalnum(ch) || ch == '_')
{
id[position++] = ch;
if (!isdigit(ch))
isNumber = FALSE;
next_ch();
}
if (position == 0)
error(indLine, indCol, "unrecognized character (%d) '%c'\n", ch, ch);
id[position++] = '\0';
if (isdigit(id[0]))
{
if (!isNumber)
error(indLine, indCol, "invalid number: %s\n", id);
n = strtol(id, NULL, 0);
if (n == LONG_MAX && errno == ERANGE)
error(indLine, indCol, "Number out of limits");
return (token){tk_Integer, indLine, indCol, n, NULL};
}
return (token){getIdentType(id, indLine, indCol), indLine, indCol, 0, id};
}
static token stringLit(char start, int indLine, int indCol)
{
static char buf[MAXSTRING];
int position = 0;
next_ch();
while (ch != start)
{
if (ch == '\n') error(indLine, indCol, "EOL in string");
if (ch == EOF) error(indLine, indCol, "EOF in string");
buf[position++] = ch;
next_ch();
}
buf[position++] = '\0';
next_ch();
return (token){tk_String, indLine, indCol, 0, buf};
}
static token follow(int expect, TokenType ifyes, TokenType ifno, int indLine, int indCol)
{
if (ch == expect)
{
next_ch();
return (token){ifyes, indLine, indCol, 0, NULL};
}
if (ifno == tk_EOI)
error(indLine, indCol, "follow: unrecognized character '%c' (%d)\n", ch, ch);
return (token){ifno, indLine, indCol, 0, NULL};
}
static token divOrCmt(int indLine, int indCol)
{
if (ch == '/')
{
next_ch();
while (ch != EOF && ch != '\n')
next_ch();
if (ch == EOF)
return (token){tk_EOI, indLine, indCol, 0, NULL};
if (ch == '\n')
{
next_ch();
return getTok();
}
}
if (ch != '*')
return (token){tk_Div, indLine, indCol, 0, NULL};
/* comment detected */
while (TRUE)
{
switch (ch)
{
case '*':
next_ch();
if (ch == '/')
{
next_ch();
return getTok();
}
break;
case EOF:
error(indLine, indCol, "EOF in comment");
}
next_ch();
}
}
static void next_sym(prolog_obj_t *o)
{ again: switch(o->ch)
{ case ' ': case '\t': next_ch(o); goto again;
case '\n': o->lc++; next_ch(o); goto again;
case '\0': case EOF: o->sym = EOI; break;
case '#': comment(o); next_ch(o); goto again;
case '(': o->sym = LPAR; next_ch(o); break;
case ')': o->sym = RPAR; next_ch(o); break;
case '[': o->sym = LSPAR; next_ch(o); break;
case ']': o->sym = RSPAR; next_ch(o); break;
case '.': o->sym = PERIOD; next_ch(o); break;
case ',': o->sym = COMMA; next_ch(o); break;
case '?': next_ch(o);
if(o->ch != '-') RECOVER(o,"expected '-'");
o->sym = QUERY;
next_ch(o);
break;
case ':': next_ch(o);
if(o->ch != '-') RECOVER(o, "expected '-'");
o->sym = ASSIGN;
next_ch(o);
break;
default:
if(isdigit(o->ch)) { /*integer*/
o->ival = 0;
while(isdigit(o->ch)) {
o->ival = o->ival*10 + (o->ch - '0');
next_ch(o);
}
o->sym = INT;
} else if(isalpha(o->ch)) { /*variable, atom or keyword*/
size_t i = 0;
while(isalnum(o->ch)) {
if(i > ID_MAX) RECOVER(o, "identifier too longer");
o->id[i++] = o->ch;
next_ch(o);
}
o->id[i] = '\0';
if(isupper(o->id[0])) { /*variable*/
o->sym = VARLEX;
break;
} else { /*id or keyword*/
o->sym = 0;
while(words[o->sym] && strcmp(words[o->sym], o->id))
o->sym++;
if(!words[o->sym]) /*id*/
o->sym = ID;
break;
}
} else {
RECOVER(o, "invalid char");
}
break;
}
}
token getTok()
{
/* skip white space */
while (isspace(ch))
next_ch();
int indLine = line, indCol = col;
switch (ch)
{
case '{': next_ch(); return (token){tk_Lbrace, indLine, indCol, 0, NULL};
case '}': next_ch(); return (token){tk_Rbrace, indLine, indCol, 0, NULL};
case '(': next_ch(); return (token){tk_Lparen, indLine, indCol, 0, NULL};
case ')': next_ch(); return (token){tk_Rparen, indLine, indCol, 0, NULL};
case '+': next_ch(); return (token){tk_Add, indLine, indCol, 0, NULL};
case '-': next_ch(); return (token){tk_Sub, indLine, indCol, 0, NULL};
case '*': next_ch(); return (token){tk_Mul, indLine, indCol, 0, NULL};
case '%': next_ch(); return (token){tk_Mod, indLine, indCol, 0, NULL};
case ';': next_ch(); return (token){tk_Semi, indLine, indCol, 0, NULL};
case ',': next_ch(); return (token){tk_Comma, indLine, indCol, 0, NULL};
case '/': next_ch(); return divOrCmt(indLine, indCol);
case '\'': next_ch(); return charLit(ch, indLine, indCol);
case '<': next_ch(); return follow('=', tk_Leq, tk_Lss, indLine, indCol);
case '>': next_ch(); return follow('=', tk_Geq, tk_Gtr, indLine, indCol);
case '=': next_ch(); return follow('=', tk_Eq, tk_Assign, indLine, indCol);
case '!': next_ch(); return follow('=', tk_Neq, tk_Not, indLine, indCol);
case '&': next_ch(); return follow('&', tk_And, tk_EOI, indLine, indCol);
case '|': next_ch(); return follow('|', tk_Or, tk_EOI, indLine, indCol);
case '"' : return stringLit(ch, indLine, indCol);
case EOF: return (token){tk_EOI, indLine, indCol, 0, NULL};
default: return identOrInt(indLine, indCol);
}
}
