Message-Id: <199611070507.WAA06016@pooh.pageplus.com>
To: new-httpd@hyperreal.com
Subject: Re: call for votes: move code into subdirs before release? 
In-reply-to: Your message of "04 Nov 1996 18:24:31 GMT."
             <57n2wxpvkw.fsf@tees.elsevier.co.uk>
MIME-Version: 1.0
Content-Type: multipart/mixed; boundary="----- =_aaaaaaaaaa0"
Content-ID: <6006.847343237.0@pooh.pageplus.com>
Date: Wed, 06 Nov 1996 22:07:56 -0700
From: Howard Fear <hsf@pooh.pageplus.com>
Sender: owner-new-httpd@apache.org
Precedence: bulk
Reply-To: new-httpd@hyperreal.com

------- =_aaaaaaaaaa0
Content-Type: text/plain; charset="us-ascii"
Content-ID: <6006.847343237.1@pooh.pageplus.com>

Paul Richards writes:
> I started to look into this but I need whoever did this to provide me
> with the original version of the lib that they used and a set of
> diffs. I can get the diffs myself if you tell me where the original
> version is. 

See attached files.
(I can't remember if I sent the original Makefile or not, so I've
included it as well)

> If we don't do this properly merging in updates to the original
> library will get messy.

Henry's next version will be a complete rewrite.

--
Howard Fear      email1: howard_fear@pageplus.com
                 email2: howard_fear@redcape.com
                 http://www.pageplus.com/~hsf/


------- =_aaaaaaaaaa0
Content-Type: text/plain; charset="us-ascii"
Content-ID: <6006.847343237.2@pooh.pageplus.com>
Content-Description: regex2.h

/*
 * First, the stuff that ends up in the outside-world include file
 = typedef off_t regoff_t;
 = typedef struct {
 = 	int re_magic;
 = 	size_t re_nsub;		// number of parenthesized subexpressions
 = 	const char *re_endp;	// end pointer for REG_PEND
 = 	struct re_guts *re_g;	// none of your business :-)
 = } regex_t;
 = typedef struct {
 = 	regoff_t rm_so;		// start of match
 = 	regoff_t rm_eo;		// end of match
 = } regmatch_t;
 */
/*
 * internals of regex_t
 */
#define	MAGIC1	((('r'^0200)<<8) | 'e')

/*
 * The internal representation is a *strip*, a sequence of
 * operators ending with an endmarker.  (Some terminology etc. is a
 * historical relic of earlier versions which used multiple strips.)
 * Certain oddities in the representation are there to permit running
 * the machinery backwards; in particular, any deviation from sequential
 * flow must be marked at both its source and its destination.  Some
 * fine points:
 *
 * - OPLUS_ and O_PLUS are *inside* the loop they create.
 * - OQUEST_ and O_QUEST are *outside* the bypass they create.
 * - OCH_ and O_CH are *outside* the multi-way branch they create, while
 *   OOR1 and OOR2 are respectively the end and the beginning of one of
 *   the branches.  Note that there is an implicit OOR2 following OCH_
 *   and an implicit OOR1 preceding O_CH.
 *
 * In state representations, an operator's bit is on to signify a state
 * immediately *preceding* "execution" of that operator.
 */
typedef unsigned long sop;	/* strip operator */
typedef long sopno;
#define	OPRMASK	0xf8000000
#define	OPDMASK	0x07ffffff
#define	OPSHIFT	((unsigned)27)
#define	OP(n)	((n)&OPRMASK)
#define	OPND(n)	((n)&OPDMASK)
#define	SOP(op, opnd)	((op)|(opnd))
/* operators			   meaning	operand			*/
/*						(back, fwd are offsets)	*/
#define	OEND	(1<<OPSHIFT)	/* endmarker	-			*/
#define	OCHAR	(2<<OPSHIFT)	/* character	unsigned char		*/
#define	OBOL	(3<<OPSHIFT)	/* left anchor	-			*/
#define	OEOL	(4<<OPSHIFT)	/* right anchor	-			*/
#define	OANY	(5<<OPSHIFT)	/* .		-			*/
#define	OANYOF	(6<<OPSHIFT)	/* [...]	set number		*/
#define	OBACK_	(7<<OPSHIFT)	/* begin \d	paren number		*/
#define	O_BACK	(8<<OPSHIFT)	/* end \d	paren number		*/
#define	OPLUS_	(9<<OPSHIFT)	/* + prefix	fwd to suffix		*/
#define	O_PLUS	(10<<OPSHIFT)	/* + suffix	back to prefix		*/
#define	OQUEST_	(11<<OPSHIFT)	/* ? prefix	fwd to suffix		*/
#define	O_QUEST	(12<<OPSHIFT)	/* ? suffix	back to prefix		*/
#define	OLPAREN	(13<<OPSHIFT)	/* (		fwd to )		*/
#define	ORPAREN	(14<<OPSHIFT)	/* )		back to (		*/
#define	OCH_	(15<<OPSHIFT)	/* begin choice	fwd to OOR2		*/
#define	OOR1	(16<<OPSHIFT)	/* | pt. 1	back to OOR1 or OCH_	*/
#define	OOR2	(17<<OPSHIFT)	/* | pt. 2	fwd to OOR2 or O_CH	*/
#define	O_CH	(18<<OPSHIFT)	/* end choice	back to OOR1		*/
#define	OBOW	(19<<OPSHIFT)	/* begin word	-			*/
#define	OEOW	(20<<OPSHIFT)	/* end word	-			*/

/*
 * Structure for [] character-set representation.  Character sets are
 * done as bit vectors, grouped 8 to a byte vector for compactness.
 * The individual set therefore has both a pointer to the byte vector
 * and a mask to pick out the relevant bit of each byte.  A hash code
 * simplifies testing whether two sets could be identical.
 *
 * This will get trickier for multicharacter collating elements.  As
 * preliminary hooks for dealing with such things, we also carry along
 * a string of multi-character elements, and decide the size of the
 * vectors at run time.
 */
typedef struct {
	uch *ptr;		/* -> uch [csetsize] */
	uch mask;		/* bit within array */
	uch hash;		/* hash code */
	size_t smultis;
	char *multis;		/* -> char[smulti]  ab\0cd\0ef\0\0 */
} cset;
/* note that CHadd and CHsub are unsafe, and CHIN doesn't yield 0/1 */
#define	CHadd(cs, c)	((cs)->ptr[(uch)(c)] |= (cs)->mask, (cs)->hash += (c))
#define	CHsub(cs, c)	((cs)->ptr[(uch)(c)] &= ~(cs)->mask, (cs)->hash -= (c))
#define	CHIN(cs, c)	((cs)->ptr[(uch)(c)] & (cs)->mask)
#define	MCadd(p, cs, cp)	mcadd(p, cs, cp)	/* regcomp() internal fns */
#define	MCsub(p, cs, cp)	mcsub(p, cs, cp)
#define	MCin(p, cs, cp)	mcin(p, cs, cp)

/* stuff for character categories */
typedef unsigned char cat_t;

/*
 * main compiled-expression structure
 */
struct re_guts {
	int magic;
#		define	MAGIC2	((('R'^0200)<<8)|'E')
	sop *strip;		/* malloced area for strip */
	int csetsize;		/* number of bits in a cset vector */
	int ncsets;		/* number of csets in use */
	cset *sets;		/* -> cset [ncsets] */
	uch *setbits;		/* -> uch[csetsize][ncsets/CHAR_BIT] */
	int cflags;		/* copy of regcomp() cflags argument */
	sopno nstates;		/* = number of sops */
	sopno firststate;	/* the initial OEND (normally 0) */
	sopno laststate;	/* the final OEND */
	int iflags;		/* internal flags */
#		define	USEBOL	01	/* used ^ */
#		define	USEEOL	02	/* used $ */
#		define	BAD	04	/* something wrong */
	int nbol;		/* number of ^ used */
	int neol;		/* number of $ used */
	int ncategories;	/* how many character categories */
	cat_t *categories;	/* ->catspace[-CHAR_MIN] */
	char *must;		/* match must contain this string */
	int mlen;		/* length of must */
	size_t nsub;		/* copy of re_nsub */
	int backrefs;		/* does it use back references? */
	sopno nplus;		/* how deep does it nest +s? */
	/* catspace must be last */
	cat_t catspace[1];	/* actually [NC] */
};

/* misc utilities */
#define	OUT	(CHAR_MAX+1)	/* a non-character value */
#define	ISWORD(c)	(isalnum(c) || (c) == '_')

------- =_aaaaaaaaaa0
Content-Type: text/plain; charset="us-ascii"
Content-ID: <6006.847343237.3@pooh.pageplus.com>
Content-Description: utils.h

/* utility definitions */
#define	DUPMAX		_POSIX2_RE_DUP_MAX	/* xxx is this right? */
#define	INFINITY	(DUPMAX + 1)
#define	NC		(CHAR_MAX - CHAR_MIN + 1)
typedef unsigned char uch;

/* switch off assertions (if not already off) if no REDEBUG */
#ifndef REDEBUG
#ifndef NDEBUG
#define	NDEBUG	/* no assertions please */
#endif
#endif
#include <assert.h>

/* for old systems with bcopy() but no memmove() */
#ifdef USEBCOPY
#define	memmove(d, s, c)	bcopy(s, d, c)
#endif

------- =_aaaaaaaaaa0
Content-Type: text/plain; charset="us-ascii"
Content-ID: <6006.847343237.4@pooh.pageplus.com>
Content-Description: Makefile

# You probably want to take -DREDEBUG out of CFLAGS, and put something like
# -O in, *after* testing (-DREDEBUG strengthens testing by enabling a lot of
# internal assertion checking and some debugging facilities).
# Put -Dconst= in for a pre-ANSI compiler.
# Do not take -DPOSIX_MISTAKE out.
# REGCFLAGS isn't important to you (it's for my use in some special contexts).
CFLAGS=-I. -DPOSIX_MISTAKE -DREDEBUG $(REGCFLAGS)

# If you have a pre-ANSI compiler, put -o into MKHFLAGS.  If you want
# the Berkeley __P macro, put -b in.
MKHFLAGS=

# Flags for linking but not compiling, if any.
LDFLAGS=

# Extra libraries for linking, if any.
LIBS=

# Internal stuff, should not need changing.
OBJPRODN=regcomp.o regexec.o regerror.o regfree.o
OBJS=$(OBJPRODN) split.o debug.o main.o
H=cclass.h cname.h regex2.h utils.h
REGSRC=regcomp.c regerror.c regexec.c regfree.c
ALLSRC=$(REGSRC) engine.c debug.c main.c split.c

# Stuff that matters only if you're trying to lint the package.
LINTFLAGS=-I. -Dstatic= -Dconst= -DREDEBUG
LINTC=regcomp.c regexec.c regerror.c regfree.c debug.c main.c
JUNKLINT=possible pointer alignment|null effect

# arrangements to build forward-reference header files
.SUFFIXES:	.ih .h
.c.ih:
	sh ./mkh $(MKHFLAGS) -p $< >$@

default:	r

lib:	purge $(OBJPRODN)
	rm -f libregex.a
	ar crv libregex.a $(OBJPRODN)

purge:
	rm -f *.o

# stuff to build regex.h
REGEXH=regex.h
REGEXHSRC=regex2.h $(REGSRC)
$(REGEXH):	$(REGEXHSRC) mkh
	sh ./mkh $(MKHFLAGS) -i _REGEX_H_ $(REGEXHSRC) >regex.tmp
	cmp -s regex.tmp regex.h 2>/dev/null || cp regex.tmp regex.h
	rm -f regex.tmp

# dependencies
$(OBJPRODN) debug.o:	utils.h regex.h regex2.h
regcomp.o:	cclass.h cname.h regcomp.ih
regexec.o:	engine.c engine.ih
regerror.o:	regerror.ih
debug.o:	debug.ih
main.o:	main.ih

# tester
re:	$(OBJS)
	$(CC) $(CFLAGS) $(LDFLAGS) $(OBJS) $(LIBS) -o $@

# regression test
r:	re tests
	./re <tests
	./re -el <tests
	./re -er <tests

# 57 variants, and other stuff, for development use -- not useful to you
ra:	./re tests
	-./re <tests
	-./re -el <tests
	-./re -er <tests

rx:	./re tests
	./re -x <tests
	./re -x -el <tests
	./re -x -er <tests

t:	./re tests
	-time ./re <tests
	-time ./re -cs <tests
	-time ./re -el <tests
	-time ./re -cs -el <tests

l:	$(LINTC)
	lint $(LINTFLAGS) -h $(LINTC) 2>&1 | egrep -v '$(JUNKLINT)' | tee lint

fullprint:
	ti README WHATSNEW notes todo | list
	ti *.h | list
	list *.c
	list regex.3 regex.7

print:
	ti README WHATSNEW notes todo | list
	ti *.h | list
	list reg*.c engine.c


mf.tmp:	Makefile
	sed '/^REGEXH=/s/=.*/=regex.h/' Makefile | sed '/#DEL$$/d' >$@

DTRH=cclass.h cname.h regex2.h utils.h
PRE=COPYRIGHT README WHATSNEW
POST=mkh regex.3 regex.7 tests $(DTRH) $(ALLSRC) fake/*.[ch]
FILES=$(PRE) Makefile $(POST)
DTR=$(PRE) Makefile=mf.tmp $(POST)
dtr:	$(FILES) mf.tmp
	makedtr $(DTR) >$@
	rm mf.tmp

cio:	$(FILES)
	cio $(FILES)

rdf:	$(FILES)
	rcsdiff -c $(FILES) 2>&1 | p

# various forms of cleanup
tidy:
	rm -f junk* core core.* *.core dtr *.tmp lint

clean:	tidy
	rm -f *.o *.s *.ih re libregex.a

# don't do this one unless you know what you're doing
spotless:	clean
	rm -f mkh regex.h

------- =_aaaaaaaaaa0--