cpp.c

<B>Digital的Unix操作系统VAX 4.2源码</B>

#ifndef lint
static char *sccsid = "@(#)cpp.c	4.1	ULTRIX	7/3/90";
#endif lint

/************************************************************************
 *
 *			Modification History
 *
 *	Jon Reeves, 25-Apr-89
 * 013- Add new features from MIPS V2.0 beta: Pascal comment handling
 *	(if LANGUAGE_PASCAL defined); pragma passing; -mips2 flag passed
 *	Some source recorrelating: 006's INCLDIR is now MAXIDIRS
 *
 *	A note for posterity, since it almost changed here: macro names
 *	cannot be allowed to start with $, since it makes cpp unusable with
 *	assembler immediate constants (like libc syscall jackets).
 *
 *	Jon Reeves, 31-Oct-88
 * 012- Fix eob handling; if eob fired at the end of a // comment, we
 *	infinite looped.
 *
 *	Jon Reeves, 26-Sep-88
 * 011-	Merge in features from MIPS cpp (-I<null>; ignore -v).
 *
 *	Jon Reeves, 14-Jan-88
 * 010- Up 'symsiz' (the size of the symbol table for define constants)
 *	from 2500 to 5000.  (Contrary to comments on change 8, that
 *	change only removed restrictions on the side buffer.  Change 5
 *	is still obsolete.)
 *
 *	Jon Reeves, 03-Dec-87
 * 009-	Merge mod.sources v07i023 patches: add #elif (code from Doug
 *	Gwyn) and, if the -B option is present, recognize //.*\n
 *	comments for C++ (code by Arnold Robbins, Emory U.).
 *	Corrections for line number handling by jlr.
 *
 *	Jon Reeves, 28-May-87
 * 008- Merge BSD 4.3 changes, including: removal of arbitrary limits
 *	(makes 005 obsolete); detect unclosed comments; recognize #ident
 *	for V.3 compatibility; handle # line "file" better; -M same as -Em
 *	(subsumes first part of 004).
 *
 *	Jon Reeves, 27-May-87
 * 007- Copy newlines appearing between macro name and opening paren.
 *	Keeps dbx line counts happy.
 *
 *	David Metsky, 06-May-87
 * 006- Upped the nuber of include directories (-I) from 10 to 32.
 *	Added the defined constant INCLDIR, set to 32.
 *
 *	David Metsky, 08-Jan-87
 * 005- Upped size of SBSIZE from 20*BUFSIZ to 30*BUFSIZ.
 *
 *	Lu  Anne Van de Pas, 01-Apr-86
 * 004- Add support for -Em switch to output include file dependencies 
 *	and increase symsiz from 2000 to 2500 
 *
 *	David L Ballenger, 27-Nov-1984
 * 003-	Increase value of NPREDEF so that more -D and -U options can be
 *	used.
 *
 *	Stephen Reilly, 29-Dec-83
 * 002- Add copyright date
 *
 *	APS, 27-Dec-83
 * 001- Added new defaults symbols bsd4_2 and ultrix
 *
 ***********************************************************************/


/************************************************************************
 *									*
 *			Copyright (c) 1983 - 1989 by			*
 *		Digital Equipment Corporation, Maynard, MA		*
 *			All rights reserved.				*
 *									*
 *   This software is furnished under a license and may be used and	*
 *   copied  only  in accordance with the terms of such license and	*
 *   with the  inclusion  of  the  above  copyright  notice.   This	*
 *   software  or  any  other copies thereof may not be provided or	*
 *   otherwise made available to any other person.  No title to and	*
 *   ownership of the software is hereby transferred.			*
 *									*
 *   This software is  derived  from  software  received  from  the	*
 *   University of California, Berkeley, MIPS Computer Systems, Inc.,	*
 *   and from Bell Laboratories.  Use, duplication, or disclosure is	*
 *   subject to restrictions under license agreements with University	*
 *   of California, MIPS Computer Systems, Inc., and with AT&T.		*
 *									*
 *   The information in this software is subject to change  without	*
 *   notice  and should not be construed as a commitment by Digital	*
 *   Equipment Corporation.						*
 *									*
 *   Digital assumes no responsibility for the use  or  reliability	*
 *   of its software on equipment which is not supplied by Digital.	*
 *									*
 ************************************************************************/

#ifdef FLEXNAMES
#define	NCPS	128
#else
#define	NCPS	8
#endif

# include "stdio.h"
# include "ctype.h"
/* C command
/* written by John F. Reiser
/* July/August 1978
*/

#define STATIC
  
#define FIRSTOPEN -2
#define STDIN 0
#define READ 0
#define WRITE 1
#define SALT '#'
#if !defined BUFSIZ || BUFSIZ < 8192
#undef BUFSIZ
#define BUFSIZ 8192
#endif
#define WORK_AREA 12288

char *pbeg,*pbuf,*pend;
char *outp,*inp;
char *newp;
char cinit;

/* some code depends on whether characters are sign or zero extended */
/*	#if '\377' < 0		not used here, old cpp doesn't understand */
#if pdp11 | vax | mc68000
#define COFF 128
#else
#define COFF 0
#endif

# if gcos
#define ALFSIZ 512	/* alphabet size */
# else
#define ALFSIZ 256	/* alphabet size */
# endif
char macbit[ALFSIZ+11];
char toktyp[ALFSIZ];
#define BLANK 1
#define IDENT 2
#define NUMBR 3

/* a superimposed code is used to reduce the number of calls to the
/* symbol table lookup routine.  (if the kth character of an identifier
/* is 'a' and there are no macro names whose kth character is 'a'
/* then the identifier cannot be a macro name, hence there is no need
/* to look in the symbol table.)  'scw1' enables the test based on
/* single characters and their position in the identifier.  'scw2'
/* enables the test based on adjacent pairs of characters and their
/* position in the identifier.  scw1 typically costs 1 indexed fetch,
/* an AND, and a jump per character of identifier, until the identifier
/* is known as a non-macro name or until the end of the identifier.
/* scw1 is inexpensive.  scw2 typically costs 4 indexed fetches,
/* an add, an AND, and a jump per character of identifier, but it is also
/* slightly more effective at reducing symbol table searches.
/* scw2 usually costs too much because the symbol table search is
/* usually short; but if symbol table search should become expensive,
/* the code is here.
/* using both scw1 and scw2 is of dubious value.
*/
#define scw1 1
#define scw2 0

#if scw2
char t21[ALFSIZ],t22[ALFSIZ],t23[ALFSIZ+NCPS];
#endif

#if scw1
#define b0 1
#define b1 2
#define b2 4
#define b3 8
#define b4 16
#define b5 32
#define b6 64
#define b7 128
#endif

#define IB 1
#define SB 2
#define NB 4
#define CB 8
#define QB 16
#define WB 32
char fastab[ALFSIZ];
char slotab[ALFSIZ];
char *ptrtab;
#define isslo (ptrtab==(slotab+COFF))
#define isid(a)  ((fastab+COFF)[a]&IB)
#define isspc(a) (ptrtab[a]&SB)
#define isnum(a) ((fastab+COFF)[a]&NB)
#define iscom(a) ((fastab+COFF)[a]&CB)
#define isquo(a) ((fastab+COFF)[a]&QB)
#define iswarn(a) ((fastab+COFF)[a]&WB)

#define eob(a) ((a)>=pend)
#define bob(a) (pbeg>=(a))

# define cputc(a,b)	if(!flslvl) putc(a,b)

char buffer[NCPS+BUFSIZ+BUFSIZ+NCPS];

char	*lastcopy;

char *malloc(), *realloc();

# define DROP 0xFE	/* special character not legal ASCII or EBCDIC */
# define WARN DROP
# define SAME 0
# define MAXINC 30      /* was 10. bumped for rel. 1.30 --uk 11/30/87 */
# define MAXFRE 14	/* max buffers of macro pushback */
# define MAXFRM 64	/* max number of formals/actuals to a macro */
# define MAXIDIRS 50	/* max number of -I's, added const --mad 7/23/86 */
                        /* bumped for rel. 1.30--uk 12/1/87  */

static char warnc = WARN;

int mactop,fretop;
char *instack[MAXFRE],*bufstack[MAXFRE],*endbuf[MAXFRE];

int plvl;	/* parenthesis level during scan for macro actuals */
int maclin;	/* line number of macro call requiring actuals */
char *macfil;	/* file name of macro call requiring actuals */
char *macnam;	/* name of macro requiring actuals */
int maclvl;	/* # calls since last decrease in nesting level */
char *macforw;	/* pointer which must be exceeded to decrease nesting level */
int macdam;	/* offset to macforw due to buffer shifting */
/* LFL 3-9-88 */
int pascal_file = 0;		/* True if LANGUAGE_PASCAL is defined */

#if tgp
int tgpscan;	/* flag for dump(); */
#endif

STATIC	int	inctop[MAXINC];
STATIC	char	*fnames[MAXINC];
STATIC	char	*dirnams[MAXINC];	/* actual directory of #include files */
STATIC	int	fins[MAXINC];
STATIC	int	lineno[MAXINC];

STATIC	char	*dirs[MAXIDIRS];	/* -I and <> directories */
char *strdex(), *copy(), *subst(), *trmdir();
struct symtab *stsym();
STATIC	int	fin	= FIRSTOPEN;
STATIC	FILE	*fout	= stdout;
STATIC	int	nd	= 1;
STATIC	int	pflag;	/* don't put out lines "# 12 foo.c" */
int	passcom;	/* don't delete comments */
int	eolcom;		/* allow // ... \n comments */
int	incomment;	/* True if parsing a comment */
STATIC	int rflag;	/* allow macro recursion */
STATIC	int mflag;	/* generate makefile dependencies */
STATIC	char *infile;	/* name of .o file to build dependencies from */
STATIC 	FILE *mout;	/* file to place dependencies on */
#define START 1
#define CONT  2
#define BACK  3
STATIC	int	ifno;
# define NPREDEF 100
STATIC	char *prespc[NPREDEF];
STATIC	char **predef = prespc;
STATIC	char *punspc[NPREDEF];
STATIC	char **prund = punspc;
STATIC	int	exfail;
struct symtab {
	char	*name;
	char	*value;
} *lastsym, *lookup(), *slookup();

# if gcos
#include <setjmp.h>
static jmp_buf env;
# define main	mainpp
# undef exit
# define exit(S)	longjmp(env, 1)
# define open(S,D)	fileno(fopen(S, "r"))
# define close(F)	fclose(_f[F])
extern FILE *_f[];
# define symsiz 8000
# else
# define symsiz 8000
# endif
STATIC	struct symtab stab[symsiz];

STATIC	struct symtab *defloc;
STATIC	struct symtab *udfloc;
STATIC	struct symtab *incloc;
STATIC	struct symtab *ifloc;
STATIC	struct symtab *eliloc;		/* DAG -- added */
STATIC	struct symtab *elsloc;
STATIC	struct symtab *eifloc;
STATIC	struct symtab *ifdloc;
STATIC	struct symtab *ifnloc;
STATIC  struct symtab *pragmaloc;
STATIC	struct symtab *ysysloc;
STATIC	struct symtab *varloc;
STATIC	struct symtab *lneloc;
STATIC	struct symtab *ulnloc;
STATIC	struct symtab *uflloc;
STATIC	struct symtab *identloc;	/* Sys 5r3 compatibility */
STATIC	int	trulvl;
STATIC	int	flslvl;
#define MAX_IF_NESTING	64		/* DAG -- added (must be at least6) */
STATIC	int	ifdone[MAX_IF_NESTING];	/* DAG -- added */

sayline(where)
	int where;
{
	if (mflag && where==START) fprintf(mout, "%s: %s\n", infile, fnames[ifno]);
	if (pflag==0) fprintf(fout,"# %d \"%s\"\n", lineno[ifno], fnames[ifno]);
}

/* data structure guide
/*
/* most of the scanning takes place in the buffer:
/*
/*  (low address)                                             (high address)
/*  pbeg                           pbuf                                 pend
/*  |      <-- BUFSIZ chars -->      |         <-- BUFSIZ chars -->        |
/*  _______________________________________________________________________
/* |_______________________________________________________________________|
/*          |               |               |
/*          |<-- waiting -->|               |<-- waiting -->
/*          |    to be      |<-- current -->|    to be
/*          |    written    |    token      |    scanned
/*          |               |               |
/*          outp            inp             p
/*
/*  *outp   first char not yet written to output file
/*  *inp    first char of current token
/*  *p      first char not yet scanned
/*
/* macro expansion: write from *outp to *inp (chars waiting to be written),
/* ignore from *inp to *p (chars of the macro call), place generated
/* characters in front of *p (in reverse order), update pointers,
/* resume scanning.
/*
/* symbol table pointers point to just beyond the end of macro definitions;
/* the first preceding character is the number of formal parameters.
/* the appearance of a formal in the body of a definition is marked by
/* 2 chars: the char WARN, and a char containing the parameter number.
/* the first char of a definition is preceded by a zero character.
/*
/* when macro expansion attempts to back up over the beginning of the
/* buffer, some characters preceding *pend are saved in a side buffer,
/* the address of the side buffer is put on 'instack', and the rest
/* of the main buffer is moved to the right.  the end of the saved buffer
/* is kept in 'endbuf' since there may be nulls in the saved buffer.
/*
/* similar action is taken when an 'include' statement is processed,
/* except that the main buffer must be completely emptied.  the array
/* element 'inctop[ifno]' records the last side buffer saved when
/* file 'ifno' was included.  these buffers remain dormant while
/* the file is being read, and are reactivated at end-of-file.
/*
/* instack[0 : mactop] holds the addresses of all pending side buffers.
/* instack[inctop[ifno]+1 : mactop-1] holds the addresses of the side
/* buffers which are "live"; the side buffers instack[0 : inctop[ifno]]
/* are dormant, waiting for end-of-file on the current file.
/*
/* space for side buffers is obtained from 'malloc' and is never returned.
/* bufstack[0:fretop-1] holds addresses of side buffers which
/* are available for use.
*/

dump() {
/* write part of buffer which lies between  outp  and  inp .
/* this should be a direct call to 'write', but the system slows to a crawl
/* if it has to do an unaligned copy.  thus we buffer.  this silly loop
/* is 15% of the total time, thus even the 'putc' macro is too slow.
*/
	register char *p1,*p2; register FILE *f;
	if ((p1=outp)==inp || flslvl!=0) return;
#if tgp
#define MAXOUT 80
	if (!tgpscan) {/* scan again to insure <= MAXOUT chars between linefeeds */
		register char c,*pblank; char savc,stopc,brk;
		tgpscan=1; brk=stopc=pblank=0; p2=inp; savc= *p2; *p2='\0';
		while (c= *p1++) {
			if (c=='\\') c= *p1++;
			if (stopc==c) stopc=0;
			else if (c=='"' || c=='\'') stopc=c;
			if (p1-outp>MAXOUT && pblank!=0) {
				*pblank++='\n'; inp=pblank; dump(); brk=1; pblank=0;
			}
			if (c==' ' && stopc==0) pblank=p1-1;
		}
		if (brk) sayline(CONT);
		*p2=savc; inp=p2; p1=outp; tgpscan=0;
	}
#endif
	f=fout;
# if gcos
/* filter out "$ program c" card if first line of input */
/* gmatch is a simple pattern matcher in the GCOS Standard Library */
{	static int gmfirst = 0;
	if (!gmfirst) {
		++gmfirst;
		if (gmatch(p1, "^$*program[ \t]*c*"))
			p1 = strdex(p1, '\n');
	}
}
# endif
	/* Performance improvement - skip the writes if doing makefile - rr */
	if (!mflag) while (p1<inp) putc(*p1++,f);
	outp = inp;
}