misc.c

这是一个同样来自贝尔实验室的和UNIX有着渊源的操作系统, 其简洁的设计和实现易于我们学习和理解

/*
 * Miscellaneous functions
 */

#include "sh.h"
#include <ctype.h>	/* for FILECHCONV */
#ifdef HAVE_LIMITS_H
# include <limits.h>
#endif

#ifndef UCHAR_MAX
# define UCHAR_MAX	0xFF
#endif

short ctypes [UCHAR_MAX+1];	/* type bits for unsigned char */

static int	do_gmatch ARGS((const unsigned char *s, const unsigned char *p,
			const unsigned char *se, const unsigned char *pe,
			int isfile));
static const unsigned char *cclass ARGS((const unsigned char *p, int sub));

/*
 * Fast character classes
 */
void
setctypes(s, t)
	register const char *s;
	register int t;
{
	register int i;

	if (t & C_IFS) {
		for (i = 0; i < UCHAR_MAX+1; i++)
			ctypes[i] &= ~C_IFS;
		ctypes[0] |= C_IFS; /* include \0 in C_IFS */
	}
	while (*s != 0)
		ctypes[(unsigned char) *s++] |= t;
}

void
initctypes()
{
	register int c;

	for (c = 'a'; c <= 'z'; c++)
		ctypes[c] |= C_ALPHA;
	for (c = 'A'; c <= 'Z'; c++)
		ctypes[c] |= C_ALPHA;
	ctypes['_'] |= C_ALPHA;
	setctypes("0123456789", C_DIGIT);
	setctypes(" \t\n|&;<>()", C_LEX1); /* \0 added automatically */
	setctypes("*@#!$-?", C_VAR1);
	setctypes(" \t\n", C_IFSWS);
	setctypes("=-+?", C_SUBOP1);
	setctypes("#%", C_SUBOP2);
	setctypes(" \n\t\"#$&'()*;<>?[\\`|", C_QUOTE);
}

/* convert unsigned long to base N string */

char *
ulton(n, base)
	register unsigned long n;
	int base;
{
	register char *p;
	static char buf [20];

	p = &buf[sizeof(buf)];
	*--p = '\0';
	do {
		*--p = "0123456789ABCDEF"[n%base];
		n /= base;
	} while (n != 0);
	return p;
}

char *
str_save(s, ap)
	register const char *s;
	Area *ap;
{
	return s ? strcpy((char*) alloc((size_t)strlen(s)+1, ap), s) : NULL;
}

/* Allocate a string of size n+1 and copy upto n characters from the possibly
 * null terminated string s into it.  Always returns a null terminated string
 * (unless n < 0).
 */
char *
str_nsave(s, n, ap)
	register const char *s;
	int n;
	Area *ap;
{
	char *ns;

	if (n < 0)
		return 0;
	ns = alloc(n + 1, ap);
	ns[0] = '\0';
	return strncat(ns, s, n);
}

/* called from expand.h:XcheckN() to grow buffer */
char *
Xcheck_grow_(xsp, xp, more)
	XString *xsp;
	char *xp;
	int more;
{
	char *old_beg = xsp->beg;

	xsp->len += more > xsp->len ? more : xsp->len;
	xsp->beg = aresize(xsp->beg, xsp->len + 8, xsp->areap);
	xsp->end = xsp->beg + xsp->len;
	return xsp->beg + (xp - old_beg);
}

const struct option options[] = {
	/* Special cases (see parse_args()): -A, -o, -s.
	 * Options are sorted by their longnames - the order of these
	 * entries MUST match the order of sh_flag F* enumerations in sh.h.
	 */
	{ "allexport",	'a',		OF_ANY },
#ifdef BRACE_EXPAND
	{ "braceexpand",  0,		OF_ANY }, /* non-standard */
#endif
	{ "bgnice",	  0,		OF_ANY },
	{ (char *) 0, 	'c',	    OF_CMDLINE },
	{ "errexit",	'e',		OF_ANY },
	{ "ignoreeof",	  0,		OF_ANY },
	{ "interactive",'i',	    OF_CMDLINE },
	{ "keyword",	'k',		OF_ANY },
	{ "login",	'l',	    OF_CMDLINE },
	{ "markdirs",	'X',		OF_ANY },
#ifdef JOBS
	{ "monitor",	'm',		OF_ANY },
#else /* JOBS */
	{ (char *) 0,	'm',		     0 }, /* so FMONITOR not ifdef'd */
#endif /* JOBS */
	{ "noclobber",	'C',		OF_ANY },
	{ "noexec",	'n',		OF_ANY },
	{ "noglob",	'f',		OF_ANY },
	{ "nohup",	  0,		OF_ANY },
	{ "nointeractive", 'I',		OF_CMDLINE },
	{ "nolog",	  0,		OF_ANY }, /* no effect */
#ifdef	JOBS
	{ "notify",	'b',		OF_ANY },
#endif	/* JOBS */
	{ "nounset",	'u',		OF_ANY },
	{ "physical",	  0,		OF_ANY }, /* non-standard */
	{ "posix",	  0,		OF_ANY }, /* non-standard */
	{ "privileged",	'p',		OF_ANY },
	{ "restricted",	'r',	    OF_CMDLINE },
	{ "stdin",	's',	    OF_CMDLINE }, /* pseudo non-standard */
	{ "trackall",	'h',		OF_ANY },
	{ "verbose",	'v',		OF_ANY },
	{ "xtrace",	'x',		OF_ANY },
	/* Anonymous flags: used internally by shell only
	 * (not visable to user)
	 */
	{ (char *) 0,	0,		OF_INTERNAL }, /* FTALKING_I */
};

/*
 * translate -o option into F* constant (also used for test -o option)
 */
int
option(n)
	const char *n;
{
	int i;

	for (i = 0; i < NELEM(options); i++)
		if (options[i].name && strcmp(options[i].name, n) == 0)
			return i;

	return -1;
}

struct options_info {
	int opt_width;
	struct {
		const char *name;
		int	flag;
	} opts[NELEM(options)];
};

static char *options_fmt_entry ARGS((void *arg, int i, char *buf, int buflen));
static void printoptions ARGS((int verbose));

/* format a single select menu item */
static char *
options_fmt_entry(arg, i, buf, buflen)
	void *arg;
	int i;
	char *buf;
	int buflen;
{
	struct options_info *oi = (struct options_info *) arg;

	shf_snprintf(buf, buflen, "%-*s %s",
		oi->opt_width, oi->opts[i].name,
		Flag(oi->opts[i].flag) ? "on" : "off");
	return buf;
}

static void
printoptions(verbose)
	int verbose;
{
	int i;

	if (verbose) {
		struct options_info oi;
		int n, len;

		/* verbose version */
		shprintf("Current option settings\n");

		for (i = n = oi.opt_width = 0; i < NELEM(options); i++)
			if (options[i].name) {
				len = strlen(options[i].name);
				oi.opts[n].name = options[i].name;
				oi.opts[n++].flag = i;
				if (len > oi.opt_width)
					oi.opt_width = len;
			}
		print_columns(shl_stdout, n, options_fmt_entry, &oi,
			      oi.opt_width + 5);
	} else {
		/* short version ala ksh93 */
		shprintf("set");
		for (i = 0; i < NELEM(options); i++)
			if (Flag(i) && options[i].name)
				shprintf(" -o %s", options[i].name);
		shprintf(newline);
	}
}

char *
getoptions()
{
	int i;
	char m[(int) FNFLAGS + 1];
	register char *cp = m;

	for (i = 0; i < NELEM(options); i++)
		if (options[i].c && Flag(i))
			*cp++ = options[i].c;
	*cp = 0;
	return str_save(m, ATEMP);
}

/* change a Flag(*) value; takes care of special actions */
void
change_flag(f, what, newval)
	enum sh_flag f;	/* flag to change */
	int what;	/* what is changing the flag (command line vs set) */
	int newval;
{
	int oldval;

	oldval = Flag(f);
	Flag(f) = newval;
	/* Turning off -p? */
	if (f == FPRIVILEGED && oldval && !newval) {
		setuid(ksheuid = getuid());
		setgid(getgid());
	} else if (f == FPOSIX && newval) {
#ifdef BRACE_EXPAND
		Flag(FBRACEEXPAND) = 0
#endif /* BRACE_EXPAND */
		;
	}
	/* Changing interactive flag? */
	if (f == FTALKING) {
		if ((what == OF_CMDLINE || what == OF_SET) && procpid == kshpid)
			Flag(FTALKING_I) = newval;
	} else if(f == FNOTTALKING) {
		if ((what == OF_CMDLINE || what == OF_SET) && procpid == kshpid)
			Flag(FTALKING_I) = !newval;
	}
}

/* parse command line & set command arguments.  returns the index of
 * non-option arguments, -1 if there is an error.
 */
int
parse_args(argv, what, setargsp)
	char **argv;
	int	what;		/* OF_CMDLINE or OF_SET */
	int	*setargsp;
{
	static char cmd_opts[NELEM(options) + 3]; /* o:\0 */
	static char set_opts[NELEM(options) + 5]; /* Ao;s\0 */
	char *opts;
	char *array = (char *) 0;
	Getopt go;
	int i, optc, set, sortargs = 0, arrayset = 0;

	/* First call?  Build option strings... */
	if (cmd_opts[0] == '\0') {
		char *p, *q;

		strcpy(cmd_opts, "o:"); /* see cmd_opts[] declaration */
		p = cmd_opts + strlen(cmd_opts);
		strcpy(set_opts, "A:o;s"); /* see set_opts[] declaration */
		q = set_opts + strlen(set_opts);
		for (i = 0; i < NELEM(options); i++) {
			if (options[i].c) {
				if (options[i].flags & OF_CMDLINE)
					*p++ = options[i].c;
				if (options[i].flags & OF_SET)
					*q++ = options[i].c;
			}
		}
		*p = '\0';
		*q = '\0';
	}

	if (what == OF_CMDLINE) {
		char *p;
		/* Set FLOGIN before parsing options so user can clear
		 * flag using +l.
		 */
		Flag(FLOGIN) = (argv[0][0] == '-'
				|| ((p = ksh_strrchr_dirsep(argv[0]))
				     && *++p == '-'));
		opts = cmd_opts;
	} else
		opts = set_opts;
	ksh_getopt_reset(&go, GF_ERROR|GF_PLUSOPT);
	while ((optc = ksh_getopt(argv, &go, opts)) != EOF) {
		set = (go.info & GI_PLUS) ? 0 : 1;
		switch (optc) {
		  case 'A':
			arrayset = set ? 1 : -1;
			array = go.optarg;
			break;

		  case 'o':
			if (go.optarg == (char *) 0) {
				/* lone -o: print options
				 *
				 * Note that on the command line, -o requires
				 * an option (ie, can't get here if what is
				 * OF_CMDLINE).
				 */
				printoptions(set);
				break;
			}
			i = option(go.optarg);
			if (i >= 0 && set == Flag(i))
				/* Don't check the context if the flag
				 * isn't changing - makes "set -o interactive"
				 * work if you're already interactive.  Needed
				 * if the output of "set +o" is to be used.
				 */
				;
			else if (i >= 0 && (options[i].flags & what))
				change_flag((enum sh_flag) i, what, set);
			else {
				bi_errorf("%s: bad option", go.optarg);
				return -1;
			}
			break;

		  case '?':
			return -1;

		  default:
			/* -s: sort positional params (at&t ksh stupidity) */
			if (what == OF_SET && optc == 's') {
				sortargs = 1;
				break;
			}
			for (i = 0; i < NELEM(options); i++)
				if (optc == options[i].c
				    && (what & options[i].flags))
				{
					change_flag((enum sh_flag) i, what,
						    set);
					break;
				}
			if (i == NELEM(options)) {
				internal_errorf(1, "parse_args: `%c'", optc);
				return -1; /* not reached */
			}
		}
	}
	if (!(go.info & GI_MINUSMINUS) && argv[go.optind]
	    && (argv[go.optind][0] == '-' || argv[go.optind][0] == '+')
	    && argv[go.optind][1] == '\0')
	{
		/* lone - clears -v and -x flags */
		if (argv[go.optind][0] == '-' && !Flag(FPOSIX))
			Flag(FVERBOSE) = Flag(FXTRACE) = 0;
		/* set skips lone - or + option */
		go.optind++;
	}
	if (setargsp)
		/* -- means set $#/$* even if there are no arguments */
		*setargsp = !arrayset && ((go.info & GI_MINUSMINUS)
					  || argv[go.optind]);

	if (arrayset && (!*array || *skip_varname(array, FALSE))) {
		bi_errorf("%s: is not an identifier", array);
		return -1;
	}
	if (sortargs) {
		for (i = go.optind; argv[i]; i++)
			;
		qsortp((void **) &argv[go.optind], (size_t) (i - go.optind),
			xstrcmp);
	}
	if (arrayset) {
		set_array(array, arrayset, argv + go.optind);
		for (; argv[go.optind]; go.optind++)
			;
	}

	return go.optind;
}

/* parse a decimal number: returns 0 if string isn't a number, 1 otherwise */
int
getn(as, ai)
	const char *as;
	int *ai;
{
	const char *s;
	register int n;
	int sawdigit = 0;

	s = as;
	if (*s == '-' || *s == '+')
		s++;
	for (n = 0; digit(*s); s++, sawdigit = 1)
		n = n * 10 + (*s - '0');
	*ai = (*as == '-') ? -n : n;
	if (*s || !sawdigit)
		return 0;
	return 1;
}

/* getn() that prints error */
int
bi_getn(as, ai)
	const char *as;
	int *ai;
{
	int rv = getn(as, ai);

	if (!rv)
		bi_errorf("%s: bad number", as);
	return rv;
}

/* -------- gmatch.c -------- */

/*
 * int gmatch(string, pattern)
 * char *string, *pattern;
 *
 * Match a pattern as in sh(1).
 * pattern character are prefixed with MAGIC by expand.
 */

int
gmatch(s, p, isfile)
	const char *s, *p;
	int isfile;
{
	const char *se, *pe;

	if (s == NULL || p == NULL)
		return 0;
	se = s + strlen(s);
	pe = p + strlen(p);
	/* isfile is false iff no syntax check has been done on
	 * the pattern.  If check fails, just to a strcmp().
	 */
	if (!isfile && !has_globbing(p, pe)) {
		int len = pe - p + 1;
		char tbuf[64];
		char *t = len <= sizeof(tbuf) ? tbuf
				: (char *) alloc(len, ATEMP);
		debunk(t, p);
		return !strcmp(t, s);
	}
	return do_gmatch((const unsigned char *) s, (const unsigned char *) se,
			 (const unsigned char *) p, (const unsigned char *) pe,
			 isfile);
}

/* Returns if p is a syntacticly correct globbing pattern, false
 * if it contains no pattern characters or if there is a syntax error.
 * Syntax errors are:
 *	- [ with no closing ]
 *	- imballenced $(...) expression
 *	- [...] and *(...) not nested (eg, [a$(b|]c), *(a[b|c]d))
 */
/*XXX
- if no magic,
	if dest given, copy to dst
	return ?
- if magic && (no globbing || syntax error)
	debunk to dst
	return ?
- return ?
*/
int
has_globbing(xp, xpe)
	const char *xp, *xpe;
{
	const unsigned char *p = (const unsigned char *) xp;
	const unsigned char *pe = (const unsigned char *) xpe;
	int c;
	int nest = 0, bnest = 0;
	int saw_glob = 0;
	int in_bracket = 0; /* inside [...] */

	for (; p < pe; p++) {
		if (!ISMAGIC(*p))
			continue;
		if ((c = *++p) == '*' || c == '?')
			saw_glob = 1;
		else if (c == '[') {
			if (!in_bracket) {
				saw_glob = 1;
				in_bracket = 1;
				if (ISMAGIC(p[1]) && p[2] == NOT)
					p += 2;
				if (ISMAGIC(p[1]) && p[2] == ']')
					p += 2;
			}
			/* XXX Do we need to check ranges here? POSIX Q */
		} else if (c == ']') {
			if (in_bracket) {
				if (bnest)		/* [a*(b]) */
					return 0;
				in_bracket = 0;
			}
		} else if ((c & 0x80) && strchr("*+?@! ", c & 0x7f)) {
			saw_glob = 1;
			if (in_bracket)
				bnest++;
			else
				nest++;
		} else if (c == '|') {
			if (in_bracket && !bnest)	/* *(a[foo|bar]) */
				return 0;
		} else if (c == /*(*/ ')') {
			if (in_bracket) {
				if (!bnest--)		/* *(a[b)c] */
					return 0;
			} else if (nest)
				nest--;
		}
		/* else must be a MAGIC-MAGIC, or MAGIC-!, MAGIC--, MAGIC-]
			 MAGIC-{, MAGIC-,, MAGIC-} */
	}
	return saw_glob && !in_bracket && !nest;
}

/* Function must return either 0 or 1 (assumed by code for 0x80|'!') */
static int
do_gmatch(s, se, p, pe, isfile)
	const unsigned char *s, *p;
	const unsigned char *se, *pe;
	int isfile;
{
	register int sc, pc;
	const unsigned char *prest, *psub, *pnext;
	const unsigned char *srest;

	if (s == NULL || p == NULL)
		return 0;
	while (p < pe) {
		pc = *p++;
		sc = s < se ? *s : '\0';
		s++;
		if (isfile) {
			sc = FILECHCONV(sc);
			pc = FILECHCONV(pc);
		}
		if (!ISMAGIC(pc)) {
			if (sc != pc)
				return 0;
			continue;
		}
		switch (*p++) {
		  case '[':
			if (sc == 0 || (p = cclass(p, sc)) == NULL)
				return 0;
			break;

		  case '?':
			if (sc == 0)
				return 0;
			break;

		  case '*':
			if (p == pe)
				return 1;
			s--;
			do {
				if (do_gmatch(s, se, p, pe, isfile))
					return 1;
			} while (s++ < se);
			return 0;

		  /*
		   * [*+?@!](pattern|pattern|..)
		   *
		   * Not ifdef'd KSH as this is needed for ${..%..}, etc.
		   */
		  case 0x80|'+': /* matches one or more times */
		  case 0x80|'*': /* matches zero or more times */
			if (!(prest = pat_scan(p, pe, 0)))
				return 0;
			s--;
			/* take care of zero matches */
			if (p[-1] == (0x80 | '*')
			    && do_gmatch(s, se, prest, pe, isfile))
				return 1;
			for (psub = p; ; psub = pnext) {
				pnext = pat_scan(psub, pe, 1);
				for (srest = s; srest <= se; srest++) {
					if (do_gmatch(s, srest,
						psub, pnext - 2, isfile)
					    && (do_gmatch(srest, se,
							  prest, pe, isfile)
						|| (s != srest
						    && do_gmatch(srest, se,
							p - 2, pe, isfile))))
						return 1;
				}
				if (pnext == prest)
					break;
			}
			return 0;

		  case 0x80|'?': /* matches zero or once */
		  case 0x80|'@': /* matches one of the patterns */
		  case 0x80|' ': /* simile for @ */
			if (!(prest = pat_scan(p, pe, 0)))
				return 0;
			s--;
			/* Take care of zero matches */