/*
* deal with comments in the fast scanner.
*/
static void
fastcmnt2(int ch)
{
int lastline = ifiles->lineno;
incmnt = 1;
if (ch == '/') { /* C++ comment */
while ((ch = qcchar()) != '\n')
;
unch(ch);
} else if (ch == '*') {
for (;;) {
if ((ch = qcchar()) == 0)
break;
if (ch == '*') {
if ((ch = qcchar()) == '/') {
break;
} else
unch(ch);
} else if (ch == '\n') {
putch('\n');
ifiles->lineno++;
}
}
} else
error("fastcmnt2");
if (ch == 0)
error("comment at line %d never ends", lastline);
incmnt = 0;
}
/*
* Check for (and convert) trigraphs.
*/
int
chktg()
{
int c;
if ((c = inpch()) != '?') {
unch(c);
return 0;
}
switch (c = inpch()) {
case '=': c = '#'; break;
case '(': c = '['; break;
case ')': c = ']'; break;
case '<': c = '{'; break;
case '>': c = '}'; break;
case '/': c = '\\'; break;
case '\'': c = '^'; break;
case '!': c = '|'; break;
case '-': c = '~'; break;
default:
unch(c);
unch('?');
c = 0;
}
return c;
}
/*
* Check for (and convert) trigraphs.
*/
static int
chktg(void)
{
int ch;
if ((ch = inpch()) != '?') {
unch(ch);
return 0;
}
switch (ch = inpch()) {
case '=': return '#';
case '(': return '[';
case ')': return ']';
case '<': return '{';
case '>': return '}';
case '/': return '\\';
case '\'': return '^';
case '!': return '|';
case '-': return '~';
}
unch(ch);
unch('?');
return 0;
}
/*
* deal with comments when -C is active.
* Save comments in expanded macros???
*/
void
Ccmnt2(struct iobuf *ob, int ch)
{
if (skpows)
cntline();
if (ch == '/') { /* C++ comment */
putob(ob, ch);
do {
putob(ob, ch);
} while ((ch = qcchar()) && ch != '\n');
unch(ch);
} else if (ch == '*') {
strtobuf((usch *)"/*", ob);
for (;;) {
ch = qcchar();
putob(ob, ch);
if (ch == '*') {
if ((ch = qcchar()) == '/') {
putob(ob, ch);
break;
} else
unch(ch);
} else if (ch == '\n') {
ifiles->lineno++;
}
}
}
}
/*
* Handle a preprocessor directive.
*/
void
ppdir(void)
{
char bp[20];
int ch, i;
redo: while ((ch = inch()) == ' ' || ch == '\t')
;
if (ch == '/') {
if ((ch = inch()) == '/') {
skpln();
return;
}
if (ch == '*') {
while ((ch = inch()) != -1) {
if (ch == '*') {
if ((ch = inch()) == '/')
goto redo;
unch(ch);
} else if (ch == '\n') {
putch('\n');
ifiles->lineno++;
}
}
}
}
if (ch == '\n') { /* empty directive */
unch(ch);
return;
}
if (ch < 'a' || ch > 'z')
goto out; /* something else, ignore */
i = 0;
do {
bp[i++] = (usch)ch;
if (i == sizeof(bp)-1)
goto out; /* too long */
ch = inch();
} while ((ch >= 'a' && ch <= 'z') || (ch == '_'));
unch(ch);
bp[i++] = 0;
/* got keyword */
for (i = 0; i < NPPD; i++) {
if (bp[0] == ppd[i].name[0] && strcmp(bp, ppd[i].name) == 0) {
(*ppd[i].fun)();
return;
}
}
out:
if (flslvl == 0 && Aflag == 0)
error("invalid preprocessor directive");
unch(ch);
skpln();
}
/*
* Handle a preprocessor directive.
* # is already found.
*/
void
ppdir(void)
{
int ch, i, oldC;
usch *bp;
oldC = Cflag;
redo: Cflag = 0;
if ((ch = fastspc()) == '\n') { /* empty directive */
unch(ch);
Cflag = oldC;
return;
}
Cflag = oldC;
if ((spechr[ch] & C_ID0) == 0)
goto out;
bp = readid(ch);
/* got some keyword */
for (i = 0; i < NPPD; i++) {
if (bp[0] == ppd[i].name[0] &&
strcmp((char *)bp, ppd[i].name) == 0) {
if (flslvl == 0) {
(*ppd[i].fun)();
if (flslvl == 0)
return;
} else {
if (ppd[i].flags & DIR_FLSLVL) {
(*ppd[i].fun)();
if (flslvl == 0)
return;
} else if (ppd[i].flags & DIR_FLSINC)
flslvl++;
}
flscan();
goto redo;
}
}
if (flslvl == 0) {
if (Aflag)
skpln();
return;
}
flscan();
goto redo;
out:
if (flslvl == 0 && Aflag == 0)
error("invalid preprocessor directive");
unch(ch);
skpln();
}
static int
eatcmnt(void)
{
int ch;
if (Cflag) { PUTCH('/'); PUTCH('*'); }
for (;;) {
ch = inch();
if (ch == '\n') {
ifiles->lineno++;
PUTCH('\n');
}
if (ch == -1)
return -1;
if (ch == '*') {
ch = inch();
if (ch == '/') {
if (Cflag) {
PUTCH('*');
PUTCH('/');
} else
PUTCH(' ');
break;
}
unch(ch);
ch = '*';
}
if (Cflag) PUTCH(ch);
}
return 0;
}
/*
* get a preprocessing number and save it as given by ob.
* returns first non-pp-number char.
* We know that this is a valid number already.
*
* pp-number: digit
* . digit
* pp-number digit
* pp-number identifier-nondigit
* pp-number e sign
* pp-number E sign
* pp-number p sign
* pp-number P sign
* pp-number .
*/
int
fastnum(int ch, struct iobuf *ob)
{
int c2;
if ((spechr[ch] & C_DIGIT) == 0) {
/* not digit, dot */
putob(ob, ch);
ch = qcchar();
}
for (;;) {
putob(ob, ch);
if ((ch = qcchar()) == 0)
return 0;
if ((c2 = (ch & 0337)) == 'E' || c2 == 'P') {
if ((c2 = qcchar()) != '-' && c2 != '+') {
if (c2 > 0)
unch(c2);
break;
}
putob(ob, ch);
ch = c2;
} else if (ch == '.' || (spechr[ch] & C_ID)) {
continue;
} else
break;
}
return ch;
}
static void
cppwarning(void)
{
usch *cp;
int c;
if (flslvl)
return;
c = sloscan();
if (c != WSPACE && c != '\n')
error("bad warning");
/* svinp() add an unwanted \n */
cp = stringbuf;
while ((c = inch()) && c != '\n')
savch(c);
savch(0);
if (flslvl)
stringbuf = cp;
else
warning("#warning %s", cp);
unch('\n');
}
static void
eatcmnt(void)
{
int ch;
if (Cflag) { PUTCH('/'); PUTCH('*'); }
for (;;) {
ch = inch();
if (ch == '\n') {
ifiles->lineno++;
putch('\n');
continue;
}
if (ch == -1)
break;
if (ch == '*') {
ch = inch();
if (ch == '/') {
if (Cflag) {
PUTCH('*');
PUTCH('/');
} else
PUTCH(' ');
break;
}
unch(ch);
ch = '*';
}
if (Cflag) PUTCH(ch);
}
}
void
cunput(int c)
{
#ifdef PCC_DEBUG
// if (dflag)printf(": '%c'(%d)\n", c > 31 ? c : ' ', c);
#endif
unch(c);
}
/*
* check for universal-character-name on input, and
* unput to the pushback buffer encoded as UTF-8.
*/
static int
chkucn(void)
{
unsigned long cp, m;
int ch, n;
if ((ch = inch()) == -1)
return 0;
if (ch == 'u')
n = 4;
else if (ch == 'U')
n = 8;
else {
unch(ch);
return 0;
}
cp = 0;
while (n-- > 0) {
if ((ch = inch()) == -1 || (spechr[ch] & C_HEX) == 0) {
warning("invalid universal character name");
// XXX should actually unput the chars and return 0
unch(ch); // XXX eof
break;
}
cp = cp * 16 + dig2num(ch);
}
if ((cp < 0xa0 && cp != 0x24 && cp != 0x40 && cp != 0x60)
|| (cp >= 0xd800 && cp <= 0xdfff)) /* 6.4.3.2 */
error("universal character name cannot be used");
if (cp > 0x7fffffff)
error("universal character name out of range");
n = 0;
m = 0x7f;
while (cp > m) {
unch(0x80 | (cp & 0x3f));
cp >>= 6;
m >>= (n++ ? 1 : 2);
}
unch(((m << 1) ^ 0xfe) | cp);
return 1;
}
static void
skpln(void)
{
/* just ignore the rest of the line */
while (inch() != '\n')
;
unch('\n');
flslvl++;
}
/*
* do an even faster scan than fastscan while at flslvl.
* just search for a new directive.
*/
static void
flscan(void)
{
int ch;
for (;;) {
ch = qcchar();
again: switch (ch) {
case 0:
return;
case '\n':
putch('\n');
ifiles->lineno++;
while ((ch = qcchar()) == ' ' || ch == '\t')
;
if (ch == '#')
return;
if (ch == '%' && (ch = qcchar()) == ':')
return;
goto again;
case '\'':
while ((ch = qcchar()) != '\'') {
if (ch == '\\')
qcchar();
if (ch == '\n')
return;
}
break;
case '\"':
instr = 1;
while ((ch = qcchar()) != '\"') {
switch (ch) {
case '\\':
incmnt = 1;
qcchar();
incmnt = 0;
break;
case '\n':
unch(ch);
/* FALLTHROUGH */
case 0:
instr = 0;
return;
}
}
instr = 0;
break;
case '/':
ch = qcchar();
if (ch == '/' || ch == '*')
fastcmnt2(ch);
goto again;
}
}
}
static void
pragmastmt(void)
{
int ch;
putstr((const usch *)"\n#pragma");
while ((ch = qcchar()) != '\n' && ch > 0)
putch(ch);
unch(ch);
prtline(1);
}
/*
* check for (and eat) end-of-line
*/
static int
chkeol(void)
{
int ch;
ch = inpch();
if (ch == '\r') {
ch = inpch();
if (ch == '\n')
return '\n';
unch(ch);
unch('\r');
return 0;
}
if (ch == '\n')
return '\n';
unch(ch);
return 0;
}
/*
* check for universal-character-name on input, and
* unput to the pushback buffer encoded as UTF-8.
*/
static void
ucn(int n)
{
unsigned long cp, m;
int ch;
if (incmnt) {
struct iobuf *ib = ifiles->ib;
ib->cptr--; /* [uU] */
ib->buf[--ib->cptr] = '\\';
return;
}
cp = 0;
while (n-- > 0) {
if ((ch = qcchar()) == 0 || (spechr[ch] & C_HEX) == 0) {
warning("invalid universal character name");
// XXX should actually unput the chars and return 0
unch(ch); // XXX eof
break;
}
cp = cp * 16 + dig2num(ch);
}
if ((cp < 0xa0 && cp != 0x24 && cp != 0x40 && cp != 0x60)
|| (cp >= 0xd800 && cp <= 0xdfff)) /* 6.4.3.2 */
error("universal character name cannot be used");
if (cp > 0x7fffffff)
error("universal character name out of range");
n = 0;
m = 0x7f;
while (cp > m) {
unch(0x80 | (cp & 0x3f));
cp >>= 6;
m >>= (n++ ? 1 : 2);
}
unch(((m << 1) ^ 0xfe) | cp);
}
/*
* Handle a preprocessor directive.
*/
void
ppdir(void)
{
char bp[20];
int ch, i;
while ((ch = inch()) == ' ' || ch == '\t')
;
if (ch == '\n') { /* empty directive */
unch(ch);
return;
}
if (ch < 'a' || ch > 'z')
goto out; /* something else, ignore */
i = 0;
do {
bp[i++] = (usch)ch;
if (i == sizeof(bp)-1)
goto out; /* too long */
ch = inch();
} while ((ch >= 'a' && ch <= 'z') || (ch == '_'));
unch(ch);
bp[i++] = 0;
/* got keyword */
#define SZ (int)(sizeof(ppd)/sizeof(ppd[0]))
for (i = 0; i < SZ; i++)
if (bp[0] == ppd[i].name[0] && strcmp(bp, ppd[i].name) == 0)
break;
if (i == SZ)
goto out;
/* Found matching keyword */
(*ppd[i].fun)();
return;
out: while ((ch = inch()) != '\n' && ch != -1)
;
unch('\n');
}
static void
skpln(void)
{
int ch;
/* just ignore the rest of the line */
while ((ch = inch()) != -1) {
if (ch == '\n') {
unch('\n');
break;
}
}
}
void
cunput(int c)
{
#ifdef CPP_DEBUG
extern int dflag;
if (dflag)printf(": '%c'(%d)", c > 31 ? c : ' ', c);
#endif
#if 0
if (c == 10) {
printf("c == 10!!!\n");
}
#endif
unch(c);
}
static int
inch(void)
{
int c;
again: switch (c = inpch()) {
case '\\': /* continued lines */
msdos: if ((c = inpch()) == '\n') {
ifiles->lineno++;
goto again;
} else if (c == '\r')
goto msdos;
unch(c);
return '\\';
case '?': /* trigraphs */
if ((c = chktg())) {
unch(c);
goto again;
}
return '?';
default:
return c;
}
}
/*
* readin chars and store in buf. Warn about too long names.
*/
usch *
readid(int ch)
{
int p = 0;
do {
if (p == MAXIDSZ)
warning("identifier exceeds C99 5.2.4.1, truncating");
if (p < MAXIDSZ)
idbuf[p] = ch;
p++;
} while (spechr[ch = qcchar()] & C_ID);
idbuf[p] = 0;
unch(ch);
return idbuf;
}
/*
* readin chars and store in buf. Warn about too long names.
*/
usch *
bufid(int ch, struct iobuf *ob)
{
int n = ob->cptr;
do {
if (ob->cptr - n == MAXIDSZ)
warning("identifier exceeds C99 5.2.4.1");
if (ob->cptr < ob->bsz)
ob->buf[ob->cptr++] = ch;
else
putob(ob, ch);
} while (spechr[ch = qcchar()] & C_ID);
ob->buf[ob->cptr] = 0; /* legal */
unch(ch);
return ob->buf+n;
}
/* save line into stringbuf */
static usch *
savln(void)
{
int c;
usch *cp = stringbuf;
while ((c = inch()) != -1) {
if (c == '\n') {
unch(c);
break;
}
savch(c);
}
savch(0);
return cp;
}
static void
pragmastmt(void)
{
int c;
if (sloscan() != WSPACE)
error("bad pragma");
if (!flslvl)
putstr((const usch *)"#pragma ");
do {
c = inch();
if (!flslvl)
putch(c); /* Do arg expansion instead? */
} while (c && c != '\n');
if (c == '\n')
unch(c);
prtline();
}
static void
chknl(int ignore)
{
void (*f)(const char *, ...);
int t;
f = ignore ? warning : error;
if ((t = fastspc()) != '\n') {
if (t) {
f("newline expected");
/* ignore rest of line */
while ((t = qcchar()) > 0 && t != '\n')
;
} else
f("no newline at end of file");
}
unch(t);
}
/* save line into iobuf */
struct iobuf *
savln(void)
{
struct iobuf *ob = getobuf(BNORMAL);
int c;
while ((c = qcchar()) != 0) {
if (c == '\n') {
unch(c);
break;
}
if (c == '\'' || c == '\"')
faststr(c, ob);
else
putob(ob, c);
}
ob->buf[ob->cptr] = 0;
return ob;
}
/*
* get a string or character constant and save it as given by d.
*/
struct iobuf *
faststr(int bc, struct iobuf *ob)
{
struct iobuf *ib = ifiles->ib;
int ch;
if (ob == NULL)
ob = getobuf(BNORMAL);
instr = 1;
putob(ob, bc);
for (;;) {
if (ib->bsz == ib->cptr)
ch = qcchar();
else if (ISCQ(ch = ib->buf[ib->cptr]))
ch = qcchar();
else
ib->cptr++;
switch (ch) {
case '\\':
putob(ob, ch);
if (ib->cptr == ib->bsz)
inpbuf(0);
incmnt = 1;
putob(ob, qcchar());
incmnt = 0;
continue;
case '\n':
warning("unterminated literal");
instr = 0;
unch(ch);
return ob;
}
putob(ob, ch);
if (ch == bc)
break;
}
putob(ob, 0);
ob->cptr--;
instr = 0;
return ob;
}
int
yylex(void)
{
static int ifdef, noex;
struct symtab *nl;
int ch, c2;
while ((ch = sloscan()) == WSPACE)
;
if (ch < 128 && (spechr[ch] & C_2))
c2 = inch();
else
c2 = 0;
switch (ch) {
case '=':
if (c2 == '=') return EQ;
break;
case '!':
if (c2 == '=') return NE;
break;
case '|':
if (c2 == '|') return OROR;
break;
case '&':
if (c2 == '&') return ANDAND;
break;
case '<':
if (c2 == '<') return LS;
if (c2 == '=') return LE;
break;
case '>':
if (c2 == '>') return RS;
if (c2 == '=') return GE;
break;
case '+':
case '-':
if (ch == c2)
error("invalid preprocessor operator %c%c", ch, c2);
break;
case '/':
if (Cflag == 0 || c2 != '*')
break;
/* Found comment that need to be skipped */
for (;;) {
ch = inch();
c1: if (ch != '*')
continue;
if ((ch = inch()) == '/')
break;
goto c1;
}
return yylex();
case NUMBER:
if (yytext[0] == '\'') {
yylval.node.op = NUMBER;
yylval.node.nd_val = charcon(yytext);
} else
cvtdig(yytext[0] != '0' ? 10 :
yytext[1] == 'x' || yytext[1] == 'X' ? 16 : 8);
return NUMBER;
case IDENT:
if (strcmp((char *)yytext, "defined") == 0) {
ifdef = 1;
return DEFINED;
}
nl = lookup(yytext, FIND);
if (ifdef) {
yylval.node.nd_val = nl != NULL;
ifdef = 0;
} else if (nl && noex == 0) {
usch *och = stringbuf;
int i;
i = kfind(nl);
unch(WARN);
if (i)
unpstr(stringbuf);
else
unpstr(nl->namep);
stringbuf = och;
noex = 1;
return yylex();
} else {
yylval.node.nd_val = 0;
}
yylval.node.op = NUMBER;
return NUMBER;
case WARN:
noex = 0;
/* FALLTHROUGH */
case PHOLD:
return yylex();
default:
return ch;
}
unch(c2);
return ch;
}
int
sloscan(void)
{
int ch;
int yyp;
zagain:
yyp = 0;
ch = inch();
yytext[yyp++] = (usch)ch;
switch (ch) {
case -1:
return 0;
case '\n':
/* sloscan() never passes \n, that's up to fastscan() */
unch(ch);
yytext[yyp] = 0;
return ch;
case '\r': /* Ignore CR's */
yyp = 0;
break;
case '0': case '1': case '2': case '3': case '4': case '5':
case '6': case '7': case '8': case '9':
/* readin a "pp-number" */
ppnum: for (;;) {
ch = inch();
if (ch == -1)
break;
if (spechr[ch] & C_EP) {
yytext[yyp++] = (usch)ch;
ch = inch();
if (ch == '-' || ch == '+') {
yytext[yyp++] = (usch)ch;
} else
unch(ch);
continue;
}
if ((spechr[ch] & C_ID) || ch == '.') {
yytext[yyp++] = (usch)ch;
continue;
}
break;
}
unch(ch);
yytext[yyp] = 0;
return NUMBER;
case '\'':
chlit:
for (;;) {
if ((ch = inch()) == '\\') {
yytext[yyp++] = (usch)ch;
yytext[yyp++] = (usch)inch();
continue;
} else if (ch == -1 || ch == '\n') {
/* not a constant */
while (yyp > 1)
unch(yytext[--yyp]);
ch = '\'';
goto any;
} else
yytext[yyp++] = (usch)ch;
if (ch == '\'')
break;
}
yytext[yyp] = 0;
return NUMBER;
case ' ':
case '\t':
while ((ch = inch()) == ' ' || ch == '\t')
yytext[yyp++] = (usch)ch;
unch(ch);
yytext[yyp] = 0;
return WSPACE;
case '/':
if ((ch = inch()) == '/') {
do {
yytext[yyp++] = (usch)ch;
ch = inch();
} while (ch != -1 && ch != '\n');
yytext[yyp] = 0;
unch(ch);
goto zagain;
} else if (ch == '*') {
int c, wrn;
extern int readmac;
if (Cflag && !flslvl && readmac) {
unch(ch);
yytext[yyp] = 0;
return CMNT;
}
wrn = 0;
more: while ((c = inch()) != '*') {
if (c == -1)
return 0;
if (c == '\n')
putch(c), ifiles->lineno++;
else if (c == EBLOCK) {
(void)inch();
(void)inch();
} else if (c == WARN)
wrn = 1;
}
if ((c = inch()) == -1)
return 0;
if (c != '/') {
unch(c);
goto more;
}
if (!tflag && !Cflag && !flslvl)
unch(' ');
if (wrn)
unch(WARN);
goto zagain;
}
unch(ch);
ch = '/';
goto any;
case '.':
if ((ch = inch()) == -1)
return 0;
if ((spechr[ch] & C_DIGIT)) {
yytext[yyp++] = (usch)ch;
goto ppnum;
} else {
unch(ch);
ch = '.';
}
goto any;
case '\"':
if (tflag && defining)
goto any;
strng:
for (;;) {
if ((ch = inch()) == '\\') {
yytext[yyp++] = (usch)ch;
yytext[yyp++] = (usch)inch();
continue;
} else if (ch == -1) {
break;
} else
yytext[yyp++] = (usch)ch;
if (ch == '\"')
break;
}
yytext[yyp] = 0;
return STRING;
case 'L':
if ((ch = inch()) == '\"' && !tflag) {
yytext[yyp++] = (usch)ch;
goto strng;
} else if (ch == '\'' && !tflag) {
yytext[yyp++] = (usch)ch;
goto chlit;
}
unch(ch);
/* FALLTHROUGH */
/* Yetch, all identifiers */
case 'a': case 'b': case 'c': case 'd': case 'e': case 'f':
case 'g': case 'h': case 'i': case 'j': case 'k': case 'l':
case 'm': case 'n': case 'o': case 'p': case 'q': case 'r':
case 's': case 't': case 'u': case 'v': case 'w': case 'x':
case 'y': case 'z':
case 'A': case 'B': case 'C': case 'D': case 'E': case 'F':
case 'G': case 'H': case 'I': case 'J': case 'K':
case 'M': case 'N': case 'O': case 'P': case 'Q': case 'R':
case 'S': case 'T': case 'U': case 'V': case 'W': case 'X':
case 'Y': case 'Z':
case '_': /* {L}({L}|{D})* */
/* Special hacks */
for (;;) { /* get chars */
if ((ch = inch()) == -1)
break;
if ((spechr[ch] & C_ID)) {
yytext[yyp++] = (usch)ch;
} else {
unch(ch);
break;
}
}
yytext[yyp] = 0; /* need already string */
/* end special hacks */
return IDENT;
default:
any:
yytext[yyp] = 0;
return yytext[0];
} /* endcase */
goto zagain;
}
