|
DataMuseum.dkPresents historical artifacts from the history of: Commodore CBM-900 |
This is an automatic "excavation" of a thematic subset of
See our Wiki for more about Commodore CBM-900 Excavated with: AutoArchaeologist - Free & Open Source Software. |
top - download
Length: 8469 (0x2115)
Types: TextFile
Notes: UNIX file
Names: »awk1.c«
└─⟦f27320a65⟧ Bits:30001972 Commodore 900 hard disk image with partial source code
└─⟦f4b8d8c84⟧ UNIX V7 Filesystem
└─ ⟦this⟧ »cmd/awk/awk1.c«
/*
* AWK
* Main routine.
* Lexical analysis of input.
* Program input routines.
* Error message routines.
*/
#include "awk.h"
#include "y.tab.h"
YYSTYPE yylval;
char *readclass();
NODE *instring();
NODE *innumber();
/*
* Main routine of AWK.
*/
main(argc, argv)
int argc;
char *argv[];
{
register int i;
register char *ap;
register FILE *fp;
register char *prog = NULL;
register char *progfile;
NODE *yyparse();
awkinit();
while (argc>1 && *argv[1]=='-') {
if (argv[1][1] == '\0')
break; /* - is stdin */
switch (argv[1][1]) {
case 'F':
sassign(FSp, &argv[1][2]);
break;
case 'f':
if (argc < 2)
usage();
if ((pfp = fopen(argv[2], "r")) == NULL)
awkerr("Cannot open %s", argv[2]);
setbuf(pfp, inbuf);
progfile = argv[2];
argv++;
argc--;
break;
case 'y':
yflag++;
break;
default:
usage();
}
argv++;
argc--;
}
if (argc>1 && pfp==NULL) {
prog = argv[1];
argv++;
argc--;
}
if (pfp==NULL && prog==NULL)
usage();
while (argc>1 && parameter(argv[1])) {
argv++;
argc--;
}
pgetinit(prog);
if (pfp != NULL) {
lineno = 1;
sassign(FILENAMEp, progfile);
}
yyparse();
if (pfp != NULL)
fclose(pfp);
fsmapinit(FS);
runflag = 1;
beginflag = 1;
if (setjmp(nextenv) == 0)
awk(codep, NULL, SNULL);
beginflag = 0;
if (argc > 1) {
for (i=1; i<argc; i++) {
ap = argv[i];
if (ap[0]=='-' && ap[1]=='\0')
fp = stdin;
else if ((fp = fopen(ap, "r")) == NULL) {
fprintf(stderr, "awk: cannot open %s\n", ap);
exit(1);
}
awk(codep, fp, ap);
}
} else
awk(codep, stdin, SNULL);
awkexit(0);
}
/*
* Read a parameter from the command line.
* Return non-zero if the string is a parameter.
* We have to fudge a bit to make sure `yylex'
* is properly reset.
*/
parameter(s)
char *s;
{
register char *cp;
register int c;
NODE *left, *right;
cp = s;
if (!isalpha(*cp))
return (0);
for (; (c = *cp)!='='; cp++) {
if (c == '\0')
return (0);
if (isspace(c)) {
*cp = '\0';
continue;
}
if (!(isalpha(c) || isdigit(c)))
return (0);
}
*cp++ = '\0';
left = lookup(s);
while (isspace(c = *cp++))
;
pgetinit(cp);
if (c == '"')
right = instring(c);
else if (isdigit(c) || c=='.')
right = innumber(c);
else
awkerr("non-constant assignment to parameter `%s'", s);
xassign(left, right);
return (1);
}
/*
* Leave awk and run the
* cleanup (`END') phase.
*/
awkexit(s)
int s;
{
register OFILE *ofp;
beginflag = 0;
endflag = 1;
exitflag = 1;
if (setjmp(nextenv) == 0)
awk(codep, NULL, SNULL);
for (ofp = files; ofp < endof(files); ofp++)
if (ofp->of_fp != NULL) {
if (ofp->of_flag & OFPIPE)
pclose(ofp->of_fp); else
fclose(ofp->of_fp);
}
exit(s);
}
/*
* Lexical analyser.
* This must have flags to
* determine when certain
* types of characters are
* special (e.g. reading regular
* expressions)
* Delete multiple newlines.
*/
yylex()
{
static int nlf = 1;
static int prev, next;
register char *cp;
register int c;
register int t;
register NODE *np;
if (lexre)
return (relex());
again:
if (next) {
t = next;
next = 0;
} else if ((t = c = pgetc()) == EOF)
;
else if (!isascii(c))
awkerr("%o is an illegal character", c);
else if (isdigit(c) || c=='.') {
yylval.u_node = innumber(c);
t = NUMBER_;
} else if (isalpha(c)) {
cp = wordbuf;
do {
*cp++ = c;
c = pgetc();
} while (isalpha(c) || isdigit(c));
pungetc(c);
*cp = '\0';
np = lookup(wordbuf);
if (np->t_op == AKEYW)
t = np->t_INT;
else if (np->t_op == AFUNC) {
t = FUNCTION_;
yylval.u_node = np;
} else {
t = ID_;
yylval.u_node = np;
}
} else switch (c) {
case '"':
yylval.u_node = instring(c);
t = STRING_;
break;
case '#':
while ((c = pgetc())!=EOF && c!='\n')
;
goto again;
case '{':
brlevel++;
break;
case '}':
if (--brlevel < 0) {
awkerr("Unbalanced braces");
brlevel = 0;
}
next = '}';
t = ';';
break;
case '!':
if (checkop('~'))
t = NMATCH_;
else if (checkop('='))
t = NE_;
break;
case '|':
if (checkop('|'))
t = OROR_;
break;
case '&':
if (checkop('&'))
t = ANDAND_;
break;
case '+':
if (checkop('='))
t = ASADD_;
else if (checkop('+'))
t = INC_;
break;
case '-':
if (checkop('='))
t = ASSUB_;
else if (checkop('-'))
t = DEC_;
break;
case '*':
if (checkop('='))
t = ASMUL_;
break;
case '/':
if (checkop('='))
t = ASDIV_;
break;
case '%':
if (checkop('='))
t = ASDIV_;
break;
case '>':
if (checkop('='))
t = GE_;
else if (checkop('>'))
t = FAPPEND_;
else if (outflag)
t = FOUT_;
break;
case '<':
if (checkop('='))
t = LE_;
break;
case '=':
if (checkop('='))
t = EQ_;
break;
case '\n':
if (nlf || nlskip)
goto again;
if (brlevel)
t = ';';
break;
case ' ':
case '\t':
goto again;
}
nlf = nlskip = 0;
if (t==ELSE_ && prev!=';' && prev!='\n' && prev!='}') {
next = t;
t = ';';
} else if (t == '\n')
nlf++;
if (t==';') {
outflag = 0;
if (prev == ';')
goto again;
}
return (prev = t);
}
/*
* Part of lexical analyser
* for reading regular expressions.
*/
relex()
{
register int c;
register int n, max;
switch (c = pgetc()) {
case '?':
return (REZOCL_);
case '*':
return (RECLOS_);
case '+':
return (RENECL_);
case '^':
return (REBOL_);
case '$':
return (REEOL_);
case '.':
return (REANY_);
case '|':
return (REOR_);
case '(':
case ')':
case '/':
return (c);
case '[':
yylval.u_charp = readclass();
return (RECLASS_);
case EOF:
case '\n':
awkerr("Non-terminated regular expression");
case '\\':
c = pgetc();
if (c>='0' && c<='7') {
n = 0;
max = 3;
do {
n = (n<<3) + c - '0';
if ((c=pgetc())<'0' || c>'7')
break;
} while (--max);
pungetc(c);
c = n;
}
default:
yylval.u_char = c;
return (RECHAR_);
}
}
/*
* Check for dual-character operators.
* such as `+=' and return 1 if so.
* This looks for a match on the next
* character `nc'.
*/
checkop(nc)
int nc;
{
register int c;
if ((c = pgetc()) == nc)
return (1);
pungetc(c);
return (0);
}
/*
* Read in a character class from
* a regular expression (called from relex).
*/
char *
readclass()
{
register char *cc;
register c, i, pc;
int comp;
cc = xalloc(NCLASS);
for (i=0; i<NCLASS; i++)
cc[i] = 0;
if ((c = pgetc()) != '^') {
comp = 0;
pungetc(c);
} else
comp = 1;
pc = EOF;
while ((c = pgetc()) != ']') {
if (c==EOF || c=='\n')
awkerr("Non-terminated character class");
if (c=='-' && pc!=EOF) {
if ((c = pgetc()) == ']')
break;
for (i=pc; i<=c; i++)
cc[i/NBPC] |= 1<<(i%NBPC);
pc = EOF;
} else {
cc[c/NBPC] |= 1<<(c%NBPC);
pc = c;
}
}
if (comp)
for (i=0; i<NCLASS; i++)
cc[i] ^= -1;
return (cc);
}
/*
* Return a string read from
* the program file.
* The string is to be terminated
* with `ec'.
*/
NODE *
instring(ec)
register int ec;
{
register char *cp;
register int c;
register int octal, nc;
cp = wordbuf;
while ((c = pgetc()) != ec) {
if (c=='\n' || c==EOF)
awkerr("%s in string", c==EOF ? "EOF" : "Newline");
if (c == '\\') {
switch (c = pgetc()) {
case 'r':
c = '\r';
break;
case 'n':
c = '\n';
break;
case 'b':
c = '\b';
break;
case 't':
c = '\t';
break;
default:
if (c>='0' && c<='7') {
octal = 0;
nc = 3;
for (;;) {
octal = octal*8 + c - '0';
if (--nc <= 0)
break;
c = pgetc();
if (c==ec || c<'0' || c>'7') {
pungetc(c);
break;
}
}
c = octal;
}
}
}
if (cp < &wordbuf[NWORD-1])
*cp++ = c;
else {
awkerr("Character string too long");
break;
}
}
*cp = '\0';
cp = xalloc(strlen(wordbuf)+sizeof(char));
strcpy(cp, wordbuf);
return (snode(cp, T_ALLOC));
}
/*
* Read a number from the
* program file.
*/
NODE *
innumber(c)
register int c;
{
register char *np;
register int floatflag = 0;
np = wordbuf;
for (;;) {
if (c == '.')
floatflag++;
*np++ = c;
c = pgetc();
if (c=='e' || c=='E') {
floatflag++;
*np++ = c;
if ((c = pgetc())=='-' || c=='+') {
*np++ = c;
c = pgetc();
}
do {
*np++ = c;
c = pgetc();
} while (c>='0' && c<='9');
break;
}
if (c!='.' && !(c>='0' && c<='9'))
break;
}
pungetc(c);
*np = '\0';
return (floatflag ? fnode(stof(wordbuf)) : inode(stoi(wordbuf)));
}