parser.c

早期freebsd实现

	*rpp = NULL;
	if (redir) {
		if (n1->type != NSUBSHELL) {
			n2 = (union node *)stalloc(sizeof (struct nredir));
			n2->type = NREDIR;
			n2->nredir.n = n1;
			n1 = n2;
		}
		n1->nredir.redirect = redir;
	}
	return n1;
}


STATIC union node *
simplecmd(rpp, redir) 
	union node **rpp, *redir;
	{
	union node *args, **app;
	union node **orig_rpp = rpp;
	union node *n;

	/* If we don't have any redirections already, then we must reset */
	/* rpp to be the address of the local redir variable.  */
	if (redir == 0)
		rpp = &redir;

	args = NULL;
	app = &args;
	/* 
	 * We save the incoming value, because we need this for shell
	 * functions.  There can not be a redirect or an argument between
	 * the function name and the open parenthesis.  
	 */
	orig_rpp = rpp;

	for (;;) {
		if (readtoken() == TWORD) {
			n = (union node *)stalloc(sizeof (struct narg));
			n->type = NARG;
			n->narg.text = wordtext;
			n->narg.backquote = backquotelist;
			*app = n;
			app = &n->narg.next;
		} else if (lasttoken == TREDIR) {
			*rpp = n = redirnode;
			rpp = &n->nfile.next;
			parsefname();	/* read name of redirection file */
		} else if (lasttoken == TLP && app == &args->narg.next
					    && rpp == orig_rpp) {
			/* We have a function */
			if (readtoken() != TRP)
				synexpect(TRP);
#ifdef notdef
			if (! goodname(n->narg.text))
				synerror("Bad function name");
#endif
			n->type = NDEFUN;
			n->narg.next = command();
			return n;
		} else {
			tokpushback++;
			break;
		}
	}
	*app = NULL;
	*rpp = NULL;
	n = (union node *)stalloc(sizeof (struct ncmd));
	n->type = NCMD;
	n->ncmd.backgnd = 0;
	n->ncmd.args = args;
	n->ncmd.redirect = redir;
	return n;
}


STATIC void
parsefname() {
	union node *n = redirnode;

	if (readtoken() != TWORD)
		synexpect(-1);
	if (n->type == NHERE) {
		struct heredoc *here = heredoc;
		struct heredoc *p;
		int i;

		if (quoteflag == 0)
			n->type = NXHERE;
		TRACE(("Here document %d\n", n->type));
		if (here->striptabs) {
			while (*wordtext == '\t')
				wordtext++;
		}
		if (! noexpand(wordtext) || (i = strlen(wordtext)) == 0 || i > EOFMARKLEN)
			synerror("Illegal eof marker for << redirection");
		rmescapes(wordtext);
		here->eofmark = wordtext;
		here->next = NULL;
		if (heredoclist == NULL)
			heredoclist = here;
		else {
			for (p = heredoclist ; p->next ; p = p->next);
			p->next = here;
		}
	} else if (n->type == NTOFD || n->type == NFROMFD) {
		if (is_digit(wordtext[0]))
			n->ndup.dupfd = digit_val(wordtext[0]);
		else if (wordtext[0] == '-')
			n->ndup.dupfd = -1;
		else
			goto bad;
		if (wordtext[1] != '\0') {
bad:
			synerror("Bad fd number");
		}
	} else {
		n->nfile.fname = (union node *)stalloc(sizeof (struct narg));
		n = n->nfile.fname;
		n->type = NARG;
		n->narg.next = NULL;
		n->narg.text = wordtext;
		n->narg.backquote = backquotelist;
	}
}


/*
 * Input any here documents.
 */

STATIC void
parseheredoc() {
	struct heredoc *here;
	union node *n;

	while (heredoclist) {
		here = heredoclist;
		heredoclist = here->next;
		if (needprompt) {
			setprompt(2);
			needprompt = 0;
		}
		readtoken1(pgetc(), here->here->type == NHERE? SQSYNTAX : DQSYNTAX,
				here->eofmark, here->striptabs);
		n = (union node *)stalloc(sizeof (struct narg));
		n->narg.type = NARG;
		n->narg.next = NULL;
		n->narg.text = wordtext;
		n->narg.backquote = backquotelist;
		here->here->nhere.doc = n;
	}
}

STATIC int
peektoken() {
	int t;

	t = readtoken();
	tokpushback++;
	return (t);
}

STATIC int xxreadtoken();

STATIC int
readtoken() {
	int t;
	int savecheckkwd = checkkwd;
	struct alias *ap;
#ifdef DEBUG
	int alreadyseen = tokpushback;
#endif
	
	top:
	t = xxreadtoken();

	if (checkkwd) {
		/*
		 * eat newlines
		 */
		if (checkkwd == 2) {
			checkkwd = 0;
			while (t == TNL) {
				parseheredoc();
				t = xxreadtoken();
			}
		} else
			checkkwd = 0;
		/*
		 * check for keywords and aliases
		 */
		if (t == TWORD && !quoteflag) {
			register char * const *pp, *s;

			for (pp = (char **)parsekwd; *pp; pp++) {
				if (**pp == *wordtext && equal(*pp, wordtext)) {
					lasttoken = t = pp - parsekwd + KWDOFFSET;
					TRACE(("keyword %s recognized\n", tokname[t]));
					goto out;
				}
			}
			if (ap = lookupalias(wordtext, 1)) {
				pushstring(ap->val, strlen(ap->val), ap);
				checkkwd = savecheckkwd;
				goto top;
			}
		}
out:
		checkkwd = 0;
	}
#ifdef DEBUG
	if (!alreadyseen)
	    TRACE(("token %s %s\n", tokname[t], t == TWORD ? wordtext : ""));
	else
	    TRACE(("reread token %s %s\n", tokname[t], t == TWORD ? wordtext : ""));
#endif
	return (t);
}


/*
 * Read the next input token.
 * If the token is a word, we set backquotelist to the list of cmds in
 *	backquotes.  We set quoteflag to true if any part of the word was
 *	quoted.
 * If the token is TREDIR, then we set redirnode to a structure containing
 *	the redirection.
 * In all cases, the variable startlinno is set to the number of the line
 *	on which the token starts.
 *
 * [Change comment:  here documents and internal procedures]
 * [Readtoken shouldn't have any arguments.  Perhaps we should make the
 *  word parsing code into a separate routine.  In this case, readtoken
 *  doesn't need to have any internal procedures, but parseword does.
 *  We could also make parseoperator in essence the main routine, and
 *  have parseword (readtoken1?) handle both words and redirection.]
 */

#define RETURN(token)	return lasttoken = token

STATIC int
xxreadtoken() {
	register c;

	if (tokpushback) {
		tokpushback = 0;
		return lasttoken;
	}
	if (needprompt) {
		setprompt(2);
		needprompt = 0;
	}
	startlinno = plinno;
	for (;;) {	/* until token or start of word found */
		c = pgetc_macro();
		if (c == ' ' || c == '\t')
			continue;		/* quick check for white space first */
		switch (c) {
		case ' ': case '\t':
			continue;
		case '#':
			while ((c = pgetc()) != '\n' && c != PEOF);
			pungetc();
			continue;
		case '\\':
			if (pgetc() == '\n') {
				startlinno = ++plinno;
				if (doprompt)
					setprompt(2);
				else
					setprompt(0);
				continue;
			}
			pungetc();
			goto breakloop;
		case '\n':
			plinno++;
			needprompt = doprompt;
			RETURN(TNL);
		case PEOF:
			RETURN(TEOF);
		case '&':
			if (pgetc() == '&')
				RETURN(TAND);
			pungetc();
			RETURN(TBACKGND);
		case '|':
			if (pgetc() == '|')
				RETURN(TOR);
			pungetc();
			RETURN(TPIPE);
		case ';':
			if (pgetc() == ';')
				RETURN(TENDCASE);
			pungetc();
			RETURN(TSEMI);
		case '(':
			RETURN(TLP);
		case ')':
			RETURN(TRP);
		default:
			goto breakloop;
		}
	}
breakloop:
	return readtoken1(c, BASESYNTAX, (char *)NULL, 0);
#undef RETURN
}



/*
 * If eofmark is NULL, read a word or a redirection symbol.  If eofmark
 * is not NULL, read a here document.  In the latter case, eofmark is the
 * word which marks the end of the document and striptabs is true if
 * leading tabs should be stripped from the document.  The argument firstc
 * is the first character of the input token or document.
 *
 * Because C does not have internal subroutines, I have simulated them
 * using goto's to implement the subroutine linkage.  The following macros
 * will run code that appears at the end of readtoken1.
 */

#define CHECKEND()	{goto checkend; checkend_return:;}
#define PARSEREDIR()	{goto parseredir; parseredir_return:;}
#define PARSESUB()	{goto parsesub; parsesub_return:;}
#define PARSEBACKQOLD()	{oldstyle = 1; goto parsebackq; parsebackq_oldreturn:;}
#define PARSEBACKQNEW()	{oldstyle = 0; goto parsebackq; parsebackq_newreturn:;}
#define	PARSEARITH()	{goto parsearith; parsearith_return:;}

STATIC int
readtoken1(firstc, syntax, eofmark, striptabs)
	int firstc;
	char const *syntax;
	char *eofmark;
	int striptabs;
	{
	register c = firstc;
	register char *out;
	int len;
	char line[EOFMARKLEN + 1];
	struct nodelist *bqlist;
	int quotef;
	int dblquote;
	int varnest;	/* levels of variables expansion */
	int arinest;	/* levels of arithmetic expansion */
	int parenlevel;	/* levels of parens in arithmetic */
	int oldstyle;
	char const *prevsyntax;	/* syntax before arithmetic */

	startlinno = plinno;
	dblquote = 0;
	if (syntax == DQSYNTAX)
		dblquote = 1;
	quotef = 0;
	bqlist = NULL;
	varnest = 0;
	arinest = 0;
	parenlevel = 0;

	STARTSTACKSTR(out);
	loop: {	/* for each line, until end of word */
#if ATTY
		if (c == '\034' && doprompt
		 && attyset() && ! equal(termval(), "emacs")) {
			attyline();
			if (syntax == BASESYNTAX)
				return readtoken();
			c = pgetc();
			goto loop;
		}
#endif
		CHECKEND();	/* set c to PEOF if at end of here document */
		for (;;) {	/* until end of line or end of word */
			CHECKSTRSPACE(3, out);	/* permit 3 calls to USTPUTC */
			if (parsebackquote && c == '\\') {
				c = pgetc();	/* XXX - compat with old /bin/sh */
				if (c != '\\' && c != '`' && c != '$') {
					pungetc();
					c = '\\';
				}
			}
			switch(syntax[c]) {
			case CNL:	/* '\n' */
				if (syntax == BASESYNTAX)
					goto endword;	/* exit outer loop */
				USTPUTC(c, out);
				plinno++;
				if (doprompt)
					setprompt(2);
				else
					setprompt(0);
				c = pgetc();
				goto loop;		/* continue outer loop */
			case CWORD:
				USTPUTC(c, out);
				break;
			case CCTL:
				if (eofmark == NULL || dblquote)
					USTPUTC(CTLESC, out);
				USTPUTC(c, out);
				break;
			case CBACK:	/* backslash */
				c = pgetc();
				if (c == PEOF) {
					USTPUTC('\\', out);
					pungetc();
				} else if (c == '\n') {
					if (doprompt)
						setprompt(2);
					else
						setprompt(0);
				} else {
					if (dblquote && c != '\\' && c != '`' && c != '$'
							 && (c != '"' || eofmark != NULL))
						USTPUTC('\\', out);
					if (SQSYNTAX[c] == CCTL)
						USTPUTC(CTLESC, out);
					USTPUTC(c, out);
					quotef++;
				}
				break;
			case CSQUOTE:
				syntax = SQSYNTAX;
				break;
			case CDQUOTE:
				syntax = DQSYNTAX;
				dblquote = 1;
				break;
			case CENDQUOTE:
				if (eofmark) {
					USTPUTC(c, out);
				} else {
					if (arinest)
						syntax = ARISYNTAX;
					else
						syntax = BASESYNTAX;
					quotef++;
					dblquote = 0;
				}
				break;
			case CVAR:	/* '$' */
				PARSESUB();		/* parse substitution */
				break;
			case CENDVAR:	/* '}' */
				if (varnest > 0) {
					varnest--;
					USTPUTC(CTLENDVAR, out);
				} else {
					USTPUTC(c, out);
				}
				break;
			case CLP:	/* '(' in arithmetic */
				parenlevel++;