/*
* Define a user named label to have the offset of the next opcode.
*
* given:
* name label name
*/
void
definelabel(char *name)
{
register LABEL *lp; /* current label */
long i; /* current label index */
i = findstr(&labelnames, name);
if (i >= 0) {
lp = &labels[i];
if (lp->l_offset >= 0) {
scanerror(T_NULL, "Label \"%s\" is multiply defined",
name);
return;
}
setlabel(lp);
return;
}
if (labelcount >= MAXLABELS) {
scanerror(T_NULL, "Too many labels in use");
return;
}
lp = &labels[labelcount++];
lp->l_chain = -1L;
lp->l_offset = (long)curfunc->f_opcodecount;
lp->l_name = addstr(&labelnames, name);
clearopt();
}
static Boolean getfds(int fd[2], int c, int default0, int default1) {
int n;
fd[0] = default0;
fd[1] = default1;
if (c != '[') {
UNGETC(c);
return TRUE;
}
if ((unsigned int) (n = GETC() - '0') > 9) {
scanerror("expected digit after '['");
return FALSE;
}
while ((unsigned int) (c = GETC() - '0') <= 9)
n = n * 10 + c;
fd[0] = n;
switch (c += '0') {
case '=':
if ((unsigned int) (n = GETC() - '0') > 9) {
if (n != ']' - '0') {
scanerror("expected digit or ']' after '='");
return FALSE;
}
fd[1] = CLOSED;
} else {
while ((unsigned int) (c = GETC() - '0') <= 9)
n = n * 10 + c;
if (c != ']' - '0') {
scanerror("expected ']' after digit");
return FALSE;
}
fd[1] = n;
}
break;
case ']':
break;
default:
scanerror("expected '=' or ']' after digit");
return FALSE;
}
return TRUE;
}
/*
* Check to make sure that all labels are defined.
*/
void
checklabels(void)
{
register LABEL *lp; /* label being checked */
long i; /* counter */
for (i = labelcount, lp = labels; --i >= 0; lp++) {
if (lp->l_offset >= 0)
continue;
scanerror(T_NULL, "Label \"%s\" was never defined",
lp->l_name);
}
}
static void getpair(int c) {
int n;
fd_left = fd_right = UNSET;
if (c != '[') {
ugchar(c);
return;
}
if ((unsigned int) (n = gchar() - '0') > 9) {
scanerror("expected digit after '['");
return;
}
while ((unsigned int) (c = gchar() - '0') <= 9)
n = n * 10 + c;
fd_left = n;
c += '0';
switch (c) {
default:
scanerror("expected '=' or ']' after digit");
return;
case ']':
return;
case '=':
if ((unsigned int) (n = gchar() - '0') > 9) {
if (n != ']' - '0') {
scanerror("expected digit or ']' after '='");
return;
}
fd_right = CLOSED;
} else {
while ((unsigned int) (c = gchar() - '0') <= 9)
n = n * 10 + c;
if (c != ']' - '0') {
scanerror("expected ']' after digit");
return;
}
fd_right = n;
}
}
}
/*
* Add the offset corresponding to the specified user label name to the
* opcode table for a function. If the label is not yet defined, then a
* chain of undefined offsets is built using the offset value, and it
* will be fixed up when the label is defined.
*
* given:
* name user symbol name
*/
void
addlabel(char *name)
{
register LABEL *lp; /* current label */
long i; /* counter */
for (i = labelcount, lp = labels; --i >= 0; lp++) {
if (strcmp(name, lp->l_name))
continue;
uselabel(lp);
return;
}
if (labelcount >= MAXLABELS) {
scanerror(T_NULL, "Too many labels in use");
return;
}
lp = &labels[labelcount++];
lp->l_offset = -1L;
lp->l_chain = -1L;
lp->l_name = addstr(&labelnames, name);
uselabel(lp);
}
int
s_yylex()
{
register c;
register char *cp;
int f;
char delim;
if (yylineno == 0)
incrlineno();
while(1) {
/*
* skip white space
*/
if (c = lookaheadchar ) {
lookaheadchar = 0;
}
else
c = readkey();
cp = yytext;
while (c == ' ' || c == '\t' || c == 0 || c == 12 /* FF */) {
c = readkey();
}
yytext[0] = c; yytext[1] = yytext[2] = 0;
if( isascii(c) && (isalpha( c ) || c == '_') ) {
do {
*cp++ = c;
c = readkey();
} while (isascii(c) && (isalnum(c) || c == '_'));
*cp = 0;
lookaheadchar = c;
c = look_kw(yytext);
clearla();
if (c == 0) {
yylval.strval = allocstring(yytext);
return (YID);
}
return c;
}
else if( isascii(c) && isdigit(c) ) {
f = 0;
do {
*cp++ = c;
c = readkey();
} while (isascii(c) && isdigit(c));
if (c == '.') {
c = readkey();
if (c == '.') {
*cp = 0;
lookaheadchar = YDOTDOT;
yylval.strval = allocstring(yytext);
return (YINT);
}
infpnumb:
f++;
*cp++ = '.';
if (!isascii(c) || !isdigit(c)) {
scanerror("syntax error: digits required after decimal point");
*cp++ = '0';
} else
while (isdigit(c)) {
*cp++ = c;
c = readkey();
}
}
if (c == 'e' || c == 'E') {
f++;
*cp++ = c;
if ((c = lookaheadchar) == 0)
c = readkey();
if (c == '+' || c == '-') {
*cp++ = c;
c = readkey();
}
if (!isascii(c) || !isdigit(c)) {
scanerror("syntax error: digits required in exponent");
*cp++ = '0';
} else
while (isascii(c) && isdigit(c)) {
*cp++ = c;
c = readkey();
}
}
*cp = 0;
lookaheadchar = c;
clearla();
yylval.strval = allocstring(yytext);
if (f)
return (YNUMB);
return (YINT);
}
printt2("Select on char %d - %c\n", c, c );
switch (c) {
case EOF:
return 0;
case ' ':
case '\t':
case 12: /* form feed */
break;
case '"':
case '\'':
*cp++ = delim = c;
do {
do {
c = readkey();
if (c == '\n' || c == EOFCHAR) {
scanerror("syntax error: unmatched quote for string" );
if (cp == yytext)
*cp++ = ' ', cp++;
return YILLCH;
}
*cp++ = c;
} while (c != delim);
c = readkey();
} while (c == delim);
if( c == '^' || c== '#' ) {
par_error( "Can't imbed ^A or #nnn codes in strings. Try concatenating strings together\n" );
}
*--cp = 0;
#ifndef TURBO
if (cp == yytext && delim == '\'') {
scanerror("syntax error: null string not allowed");
*cp++ = ' ';
*cp++ = 0;
}
#endif
lookaheadchar = c;
clearla();
/* len of 2 means 1 char and 1 quote char */
if (delim == '"' || strlen(yytext) != 2) {
yylval.strval = allocstring(yytext);
return (YSTRING);
}
else {
yylval.intval = yytext[1];
return (YCHAR);
}
case '.':
c = readkey();
if (c == '.')
return (YDOTDOT);
if (isdigit(c)) {
scanerror("syntax error: digits required before decimal point");
*cp++ = '0';
goto infpnumb;
}
lookaheadchar = c;
clearla();
return '.';
case '\n':
break;
case '{': /* { ... } comment */
delim = '{';
comment:
c = readkey();
#ifdef TURBO
if (c == '$' && turbo_flag) {
f = scanturbo();
if (f >= 0)
return f;
}
else
#endif
if (c == '+') {
/* Stubs generated by alist, we know they use {} */
f = scanstub();
if (f == YC_BLCOMMENT) {
/* Kludge - throw away to keep grammar LALR.
* Doesn't matter since they will be generated
* in all appropriate places anyway.
*/
continue; /* outer while loop */
}
if (f >= 0)
return f;
}
else {
for (;;) {
if (delim=='{' && c == '}') {
break;
}
if (c == '\n') {
/* Break into one line pieces */
*cp++ = 0;
savecomment(yytext);
cp = yytext;
*cp = 0;
} else {
*cp++ = c;
if (c <= 0) {
/* nonterminated comment */
/* This "can't happen" */
fatal("Bug - nonterm comment");
}
}
c = readkey();
if (delim=='(' && c == ')' && cp[-1] == '*') {
*--cp = 0;
break;
}
}
*cp++ = 0;
}
/*
* Comments generated by the lister for procedure or
* function calls (in parens, ending in =) are ignored.
*/
if (parendepth <= 0 || cp[-2] != '=')
savecomment(yytext);
clearla();
cp = yytext;
*cp = 0;
if (allowcom)
return 0;
break;
case ':':
if ((c=readkey()) == '=') {
*++cp = c;
return YCOLEQUALS;
}
lookaheadchar = c;
clearla();
return ':';
case '(':
if ((c=readkey()) == '*') {
delim = '(';
goto comment;
}
lookaheadchar = c;
clearla();
parendepth++;
return '(';
case ')':
parendepth--;
return ')';
case '$':
while( isxdigit(c = readkey()) )
*++cp = c;
*++cp = 0;
lookaheadchar = c;
if( strlen(yytext) <= 1 )
return YILLCH;
yylval.strval = allocstring(yytext);
return YINT;
case '#':
c = readkey();
if( c == '$' ) {
unsigned hxnum;
while( isxdigit(c = readkey()) )
hxnum = (hxnum << 4) + ((c > '9') ? 9 : 0)
+ (c & 0xf);
lookaheadchar = c;
yylval.intval = hxnum;
return YCHAR;
}
else while (isdigit(c)) {
*++cp = c;
c = readkey();
}
*++cp = 0;
lookaheadchar = c;
yylval.intval = atoi(&yytext[1]);
return YCHAR;
case '^':
c = readkey();
if (strchr("\\_", c)) { /* others include []@^ */
yylval.intval = c & 0x1f;
return YCHAR;
}
else {
lookaheadchar = c;
yylval.intval = '^';
return YPTR;
}
case '@':
yylval.intval = '@';
return YPTR;
case ';':
case ',':
case '=':
case '*':
case '+':
case '/':
case '-':
case '[':
case ']':
case '<':
case '>':
case '_':
case '\\':
case '}': /* for DO..SET */
return c;
case YDOTDOT:
return YDOTDOT;
default:
if (c <= 0)
return (0);
do
lookaheadchar = readkey();
while (lookaheadchar == c);
clearla();
printt1("illegal char in scanner %o\n", c);
return (YILLCH);
}
} /* big while */
}
extern int yylex(void) {
static Boolean dollar = FALSE;
int c;
size_t i; /* The purpose of all these local assignments is to */
const char *meta; /* allow optimizing compilers like gcc to load these */
char *buf = tokenbuf; /* values into registers. On a sparc this is a */
YYSTYPE *y = &yylval; /* win, in code size *and* execution time */
if (goterror) {
goterror = FALSE;
return NL;
}
/* rc variable-names may contain only alnum, '*' and '_', so use dnw if we are scanning one. */
meta = (dollar ? dnw : nw);
dollar = FALSE;
if (newline) {
--input->lineno; /* slight space optimization; print_prompt2() always increments lineno */
print_prompt2();
newline = FALSE;
}
top: while ((c = GETC()) == ' ' || c == '\t')
w = NW;
if (c == EOF)
return ENDFILE;
if (!meta[(unsigned char) c]) { /* it's a word or keyword. */
InsertFreeCaret();
w = RW;
i = 0;
do {
buf[i++] = c;
if (i >= bufsize)
buf = tokenbuf = erealloc(buf, bufsize *= 2);
} while ((c = GETC()) != EOF && !meta[(unsigned char) c]);
UNGETC(c);
buf[i] = '\0';
w = KW;
if (buf[1] == '\0') {
int k = *buf;
if (k == '@' || k == '~')
return k;
} else if (*buf == 'f') {
if (streq(buf + 1, "n")) return FN;
if (streq(buf + 1, "or")) return FOR;
} else if (*buf == 'l') {
if (streq(buf + 1, "ocal")) return LOCAL;
if (streq(buf + 1, "et")) return LET;
} else if (streq(buf, "~~"))
return EXTRACT;
else if (streq(buf, "%closure"))
return CLOSURE;
w = RW;
y->str = gcdup(buf);
return WORD;
}
if (c == '`' || c == '!' || c == '$' || c == '\'') {
InsertFreeCaret();
if (c == '!')
w = KW;
}
switch (c) {
case '!':
return '!';
case '`':
c = GETC();
if (c == '`')
return BACKBACK;
UNGETC(c);
return '`';
case '$':
dollar = TRUE;
switch (c = GETC()) {
case '#': return COUNT;
case '^': return FLAT;
case '&': return PRIM;
default: UNGETC(c); return '$';
}
case '\'':
w = RW;
i = 0;
while ((c = GETC()) != '\'' || (c = GETC()) == '\'') {
buf[i++] = c;
if (c == '\n')
print_prompt2();
if (c == EOF) {
w = NW;
scanerror("eof in quoted string");
return ERROR;
}
if (i >= bufsize)
buf = tokenbuf = erealloc(buf, bufsize *= 2);
}
UNGETC(c);
buf[i] = '\0';
y->str = gcdup(buf);
return QWORD;
case '\\':
if ((c = GETC()) == '\n') {
print_prompt2();
UNGETC(' ');
goto top; /* Pretend it was just another space. */
}
if (c == EOF) {
UNGETC(EOF);
goto badescape;
}
UNGETC(c);
c = '\\';
InsertFreeCaret();
w = RW;
c = GETC();
switch (c) {
case 'a': *buf = '\a'; break;
case 'b': *buf = '\b'; break;
case 'e': *buf = '\033'; break;
case 'f': *buf = '\f'; break;
case 'n': *buf = '\n'; break;
case 'r': *buf = '\r'; break;
case 't': *buf = '\t'; break;
case 'x': case 'X': {
int n = 0;
for (;;) {
c = GETC();
if (!isxdigit(c))
break;
n = (n << 4)
| (c - (isdigit(c) ? '0' : ((islower(c) ? 'a' : 'A') - 0xA)));
}
if (n == 0)
goto badescape;
UNGETC(c);
*buf = n;
break;
}
case '0': case '1': case '2': case '3': case '4': case '5': case '6': case '7': {
int n = 0;
do {
n = (n << 3) | (c - '0');
c = GETC();
} while (isodigit(c));
if (n == 0)
goto badescape;
UNGETC(c);
*buf = n;
break;
}
default:
if (isalnum(c)) {
badescape:
scanerror("bad backslash escape");
return ERROR;
}
*buf = c;
break;
}
buf[1] = 0;
y->str = gcdup(buf);
return QWORD;
case '#':
while ((c = GETC()) != '\n') /* skip comment until newline */
if (c == EOF)
return ENDFILE;
/* FALLTHROUGH */
case '\n':
input->lineno++;
newline = TRUE;
w = NW;
return NL;
case '(':
if (w == RW) /* not keywords, so let & friends work */
c = SUB;
/* FALLTHROUGH */
case ';':
case '^':
case ')':
case '=':
case '{': case '}':
w = NW;
return c;
case '&':
w = NW;
c = GETC();
if (c == '&')
return ANDAND;
UNGETC(c);
return '&';
case '|': {
int p[2];
w = NW;
c = GETC();
if (c == '|')
return OROR;
if (!getfds(p, c, 1, 0))
return ERROR;
if (p[1] == CLOSED) {
scanerror("expected digit after '='"); /* can't close a pipe */
return ERROR;
}
y->tree = mk(nPipe, p[0], p[1]);
return PIPE;
}
{
char *cmd;
int fd[2];
case '<':
fd[0] = 0;
if ((c = GETC()) == '>')
if ((c = GETC()) == '>') {
c = GETC();
cmd = "%open-append";
} else
cmd = "%open-write";
else if (c == '<')
if ((c = GETC()) == '<') {
c = GETC();
cmd = "%here";
} else
cmd = "%heredoc";
else if (c == '=')
return CALL;
else
cmd = "%open";
goto redirection;
case '>':
fd[0] = 1;
if ((c = GETC()) == '>')
if ((c = GETC()) == '<') {
c = GETC();
cmd = "%open-append";
} else
cmd = "%append";
else if (c == '<') {
c = GETC();
cmd = "%open-create";
} else
cmd = "%create";
goto redirection;
redirection:
w = NW;
if (!getfds(fd, c, fd[0], DEFAULT))
return ERROR;
if (fd[1] != DEFAULT) {
y->tree = (fd[1] == CLOSED)
? mkclose(fd[0])
: mkdup(fd[0], fd[1]);
return DUP;
}
y->tree = mkredircmd(cmd, fd[0]);
return REDIR;
}
default:
assert(c != '\0');
w = NW;
return c; /* don't know what it is, let yacc barf on it */
}
}
extern int yylex() {
static bool dollar = FALSE;
bool saw_meta = FALSE;
int c;
SIZE_T i; /* The purpose of all these local assignments is to */
char *meta; /* allow optimizing compilers like gcc to load these */
char *buf = realbuf; /* values into registers. On a sparc this is a */
YYSTYPE *y = &yylval; /* win, in code size *and* execution time */
if (errset) {
errset = FALSE;
return '\n';
}
/* rc variable-names may contain only alnum, '*' and '_', so use dnw if we are scanning one. */
meta = (dollar ? dnw : nw);
dollar = FALSE;
if (newline) {
--lineno; /* slight space optimization; print_prompt2() always increments lineno */
print_prompt2();
newline = FALSE;
}
top: while ((c = gchar()) == ' ' || c == '\t')
w = NW;
if (c == EOF)
return END;
if (!meta[(unsigned char) c]) { /* it's a word or keyword. */
checkfreecaret;
w = RW;
i = 0;
read: do {
buf[i++] = c;
if (c == '?' || c == '[' || c == '*')
saw_meta = TRUE;
if (i >= bufsize)
buf = realbuf = erealloc(buf, bufsize *= 2);
} while ((c = gchar()) != EOF && !meta[(unsigned char) c]);
while (c == '\\') {
if ((c = gchar()) == '\n') {
print_prompt2();
c = ' '; /* Pretend a space was read */
break;
} else {
bs: if (meta != dnw) { /* all words but varnames may have a bslash */
buf[i++] = '\\';
if (i >= bufsize)
buf = realbuf = erealloc(buf, bufsize *= 2);
if (!meta[(unsigned char) c])
goto read;
} else {
ugchar(c);
c = '\\';
break;
}
}
}
ugchar(c);
buf[i] = '\0';
w = KW;
if (i == 2) {
if (*buf == 'i' && buf[1] == 'f') return IF;
if (*buf == 'f' && buf[1] == 'n') return FN;
if (*buf == 'i' && buf[1] == 'n') return IN;
}
if (streq(buf, "for")) return FOR;
if (streq(buf, "else")) return ELSE;
if (streq(buf, "switch")) return SWITCH;
if (streq(buf, "while")) return WHILE;
if (streq(buf, "case")) return CASE;
w = RW;
y->word.w = ncpy(buf);
if (saw_meta) {
char *r, *s;
y->word.m = nalloc(strlen(buf) + 1);
for (r = buf, s = y->word.m; *r != '\0'; r++, s++)
*s = (*r == '?' || *r == '[' || *r == '*');
} else {
y->word.m = NULL;
}
return WORD;
}
if (c == '`' || c == '!' || c == '@' || c == '~' || c == '$' || c == '\'') {
checkfreecaret;
if (c == '!' || c == '@' || c == '~')
w = KW;
}
switch (c) {
case '!':
return BANG;
case '@':
return SUBSHELL;
case '~':
return TWIDDLE;
case '`':
c = gchar();
if (c == '`')
return BACKBACK;
ugchar(c);
return '`';
case '$':
dollar = TRUE;
c = gchar();
if (c == '#')
return COUNT;
if (c == '^')
return FLAT;
ugchar(c);
return '$';
case '\'':
w = RW;
i = 0;
do {
buf[i++] = c;
if (c == '\n')
print_prompt2();
if (c == EOF) {
w = NW;
scanerror("eof in quoted string");
return HUH;
}
if (i >= bufsize)
buf = realbuf = erealloc(buf, bufsize *= 2);
} while ((c = gchar()) != '\'' || (c = gchar()) == '\''); /* quote "'" thus: 'how''s it going?' */
ugchar(c);
buf[i] = '\0';
y->word.w = ncpy(buf);
y->word.m = NULL;
return WORD;
case '\\':
if ((c = gchar()) == '\n') {
print_prompt2();
goto top; /* Pretend it was just another space. */
}
ugchar(c);
c = '\\';
checkfreecaret;
c = gchar();
i = 0;
goto bs;
case '(':
if (w == RW) /* SUB's happen only after real words, not keyowrds, so if () and while () work */
c = SUB;
w = NW;
return c;
case '#':
while ((c = gchar()) != '\n') /* skip comment until newline */
if (c == EOF)
return END;
/* FALLTHROUGH */
case '\n':
lineno++;
newline = TRUE;
/* FALLTHROUGH */
case ';':
case '^':
case ')':
case '=':
case '{': case '}':
w = NW;
return c;
case '&':
w = NW;
c = gchar();
if (c == '&')
return ANDAND;
ugchar(c);
return '&';
case '|':
w = NW;
c = gchar();
if (c == '|')
return OROR;
getpair(c);
if (errset)
return HUH;
if ((y->pipe.left = fd_left) == UNSET)
y->pipe.left = 1; /* default to fd 1 */
if ((y->pipe.right = fd_right) == UNSET)
y->pipe.right = 0; /* default to fd 0 */
if (y->pipe.right == CLOSED) {
scanerror("expected digit after '='"); /* can't close a pipe */
return HUH;
}
return PIPE;
case '>':
c = gchar();
if (c == '>') {
c = gchar();
y->redir.type = rAppend;
} else
y->redir.type = rCreate;
y->redir.fd = 1;
goto common;
case '<':
c = gchar();
if (c == '<') {
c = gchar();
if (c == '<') {
c = gchar();
y->redir.type = rHerestring;
} else {
y->redir.type = rHeredoc;
}
} else
y->redir.type = rFrom;
y->redir.fd = 0;
common:
w = NW;
getpair(c);
if (errset)
return HUH;
if (fd_right == UNSET) { /* redirection, not dup */
if (fd_left != UNSET) {
y->redir.fd = fd_left;
return SREDIR;
}
return (y->redir.type == rFrom || y->redir.type == rCreate) ? REDIR : SREDIR;
} else { /* dup; recast yylval */
y->dup.type = y->redir.type;
y->dup.left = fd_left;
y->dup.right = fd_right;
return DUP;
}
default:
w = NW;
return c; /* don't know what it is, let yacc barf on it */
}
}
