glob.c

android-w.song.android.widget

/* glob.c -- file-name wildcard pattern matching for Bash.

   Copyright (C) 1985-2009 Free Software Foundation, Inc.

   This file is part of GNU Bash, the Bourne-Again SHell.
   
   Bash is free software: you can redistribute it and/or modify
   it under the terms of the GNU General Public License as published by
   the Free Software Foundation, either version 3 of the License, or
   (at your option) any later version.

   Bash is distributed in the hope that it will be useful,
   but WITHOUT ANY WARRANTY; without even the implied warranty of
   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
   GNU General Public License for more details.

   You should have received a copy of the GNU General Public License
   along with Bash.  If not, see <http://www.gnu.org/licenses/>.
*/

/* To whomever it may concern: I have never seen the code which most
   Unix programs use to perform this function.  I wrote this from scratch
   based on specifications for the pattern matching.  --RMS.  */

#include <config.h>

#if !defined (__GNUC__) && !defined (HAVE_ALLOCA_H) && defined (_AIX)
  #pragma alloca
#endif /* _AIX && RISC6000 && !__GNUC__ */

#include "bashtypes.h"

#if defined (HAVE_UNISTD_H)
#  include <unistd.h>
#endif

#include "bashansi.h"
#include "posixdir.h"
#include "posixstat.h"
#include "shmbutil.h"
#include "xmalloc.h"

#include "filecntl.h"
#if !defined (F_OK)
#  define F_OK 0
#endif

#include "stdc.h"
#include "memalloc.h"

#include "shell.h"

#include "glob.h"
#include "strmatch.h"

#if !defined (HAVE_BCOPY) && !defined (bcopy)
#  define bcopy(s, d, n) ((void) memcpy ((d), (s), (n)))
#endif /* !HAVE_BCOPY && !bcopy */

#if !defined (NULL)
#  if defined (__STDC__)
#    define NULL ((void *) 0)
#  else
#    define NULL 0x0
#  endif /* __STDC__ */
#endif /* !NULL */

#if !defined (FREE)
#  define FREE(x)	if (x) free (x)
#endif

/* Don't try to alloca() more than this much memory for `struct globval'
   in glob_vector() */
#ifndef ALLOCA_MAX
#  define ALLOCA_MAX	100000
#endif

struct globval
  {
    struct globval *next;
    char *name;
  };

extern void throw_to_top_level __P((void));
extern int sh_eaccess __P((char *, int));
extern char *sh_makepath __P((const char *, const char *, int));

extern int extended_glob;

/* Global variable which controls whether or not * matches .*.
   Non-zero means don't match .*.  */
int noglob_dot_filenames = 1;

/* Global variable which controls whether or not filename globbing
   is done without regard to case. */
int glob_ignore_case = 0;

/* Global variable to return to signify an error in globbing. */
char *glob_error_return;

static struct globval finddirs_error_return;

/* Some forward declarations. */
static int skipname __P((char *, char *, int));
#if HANDLE_MULTIBYTE
static int mbskipname __P((char *, char *, int));
#endif
#if HANDLE_MULTIBYTE
static void udequote_pathname __P((char *));
static void wdequote_pathname __P((char *));
#else
#  define dequote_pathname udequote_pathname
#endif
static void dequote_pathname __P((char *));
static int glob_testdir __P((char *));
static char **glob_dir_to_array __P((char *, char **, int));

/* Compile `glob_loop.c' for single-byte characters. */
#define CHAR	unsigned char
#define INT	int
#define L(CS)	CS
#define INTERNAL_GLOB_PATTERN_P internal_glob_pattern_p
#include "glob_loop.c"

/* Compile `glob_loop.c' again for multibyte characters. */
#if HANDLE_MULTIBYTE

#define CHAR	wchar_t
#define INT	wint_t
#define L(CS)	L##CS
#define INTERNAL_GLOB_PATTERN_P internal_glob_wpattern_p
#include "glob_loop.c"

#endif /* HANDLE_MULTIBYTE */

/* And now a function that calls either the single-byte or multibyte version
   of internal_glob_pattern_p. */
int
glob_pattern_p (pattern)
     const char *pattern;
{
#if HANDLE_MULTIBYTE
  size_t n;
  wchar_t *wpattern;
  int r;

  if (MB_CUR_MAX == 1)
    return (internal_glob_pattern_p ((unsigned char *)pattern));

  /* Convert strings to wide chars, and call the multibyte version. */
  n = xdupmbstowcs (&wpattern, NULL, pattern);
  if (n == (size_t)-1)
    /* Oops.  Invalid multibyte sequence.  Try it as single-byte sequence. */
    return (internal_glob_pattern_p ((unsigned char *)pattern));

  r = internal_glob_wpattern_p (wpattern);
  free (wpattern);

  return r;
#else
  return (internal_glob_pattern_p (pattern));
#endif
}

/* Return 1 if DNAME should be skipped according to PAT.  Mostly concerned
   with matching leading `.'. */

static int
skipname (pat, dname, flags)
     char *pat;
     char *dname;
     int flags;
{
  /* If a leading dot need not be explicitly matched, and the pattern
     doesn't start with a `.', don't match `.' or `..' */
  if (noglob_dot_filenames == 0 && pat[0] != '.' &&
	(pat[0] != '\\' || pat[1] != '.') &&
	(dname[0] == '.' &&
	  (dname[1] == '\0' || (dname[1] == '.' && dname[2] == '\0'))))
    return 1;

  /* If a dot must be explicity matched, check to see if they do. */
  else if (noglob_dot_filenames && dname[0] == '.' && pat[0] != '.' &&
	(pat[0] != '\\' || pat[1] != '.'))
    return 1;

  return 0;
}

#if HANDLE_MULTIBYTE
/* Return 1 if DNAME should be skipped according to PAT.  Handles multibyte
   characters in PAT and DNAME.  Mostly concerned with matching leading `.'. */

static int
mbskipname (pat, dname, flags)
     char *pat, *dname;
     int flags;
{
  int ret;
  wchar_t *pat_wc, *dn_wc;
  size_t pat_n, dn_n;

  pat_n = xdupmbstowcs (&pat_wc, NULL, pat);
  dn_n = xdupmbstowcs (&dn_wc, NULL, dname);

  ret = 0;
  if (pat_n != (size_t)-1 && dn_n !=(size_t)-1)
    {
      /* If a leading dot need not be explicitly matched, and the
	 pattern doesn't start with a `.', don't match `.' or `..' */
      if (noglob_dot_filenames == 0 && pat_wc[0] != L'.' &&
	    (pat_wc[0] != L'\\' || pat_wc[1] != L'.') &&
	    (dn_wc[0] == L'.' &&
	      (dn_wc[1] == L'\0' || (dn_wc[1] == L'.' && dn_wc[2] == L'\0'))))
	ret = 1;

      /* If a leading dot must be explicity matched, check to see if the
	 pattern and dirname both have one. */
     else if (noglob_dot_filenames && dn_wc[0] == L'.' &&
	   pat_wc[0] != L'.' &&
	   (pat_wc[0] != L'\\' || pat_wc[1] != L'.'))
	ret = 1;
    }

  FREE (pat_wc);
  FREE (dn_wc);

  return ret;
}
#endif /* HANDLE_MULTIBYTE */

/* Remove backslashes quoting characters in PATHNAME by modifying PATHNAME. */
static void
udequote_pathname (pathname)
     char *pathname;
{
  register int i, j;

  for (i = j = 0; pathname && pathname[i]; )
    {
      if (pathname[i] == '\\')
	i++;

      pathname[j++] = pathname[i++];

      if (pathname[i - 1] == 0)
	break;
    }
  if (pathname)
    pathname[j] = '\0';
}

#if HANDLE_MULTIBYTE
/* Remove backslashes quoting characters in PATHNAME by modifying PATHNAME. */
static void
wdequote_pathname (pathname)
     char *pathname;
{
  mbstate_t ps;
  size_t len, n;
  wchar_t *wpathname;
  int i, j;
  wchar_t *orig_wpathname;

  len = strlen (pathname);
  /* Convert the strings into wide characters.  */
  n = xdupmbstowcs (&wpathname, NULL, pathname);
  if (n == (size_t) -1)
    /* Something wrong. */
    return;
  orig_wpathname = wpathname;

  for (i = j = 0; wpathname && wpathname[i]; )
    {
      if (wpathname[i] == L'\\')
	i++;

      wpathname[j++] = wpathname[i++];

      if (wpathname[i - 1] == L'\0')
	break;
    }
  if (wpathname)
    wpathname[j] = L'\0';

  /* Convert the wide character string into unibyte character set. */
  memset (&ps, '\0', sizeof(mbstate_t));
  n = wcsrtombs(pathname, (const wchar_t **)&wpathname, len, &ps);
  pathname[len] = '\0';

  /* Can't just free wpathname here; wcsrtombs changes it in many cases. */
  free (orig_wpathname);
}

static void
dequote_pathname (pathname)
     char *pathname;
{
  if (MB_CUR_MAX > 1)
    wdequote_pathname (pathname);
  else
    udequote_pathname (pathname);
}
#endif /* HANDLE_MULTIBYTE */

/* Test whether NAME exists. */

#if defined (HAVE_LSTAT)
#  define GLOB_TESTNAME(name)  (lstat (name, &finfo))
#else /* !HAVE_LSTAT */
#  if !defined (AFS)
#    define GLOB_TESTNAME(name)  (sh_eaccess (name, F_OK))
#  else /* AFS */
#    define GLOB_TESTNAME(name)  (access (name, F_OK))
#  endif /* AFS */
#endif /* !HAVE_LSTAT */

/* Return 0 if DIR is a directory, -1 otherwise. */
static int
glob_testdir (dir)
     char *dir;
{
  struct stat finfo;

/*itrace("glob_testdir: testing %s", dir);*/
  if (stat (dir, &finfo) < 0)
    return (-1);

  if (S_ISDIR (finfo.st_mode) == 0)
    return (-1);

  return (0);
}

/* Recursively scan SDIR for directories matching PAT (PAT is always `**').
   FLAGS is simply passed down to the recursive call to glob_vector.  Returns
   a list of matching directory names.  EP, if non-null, is set to the last
   element of the returned list.  NP, if non-null, is set to the number of
   directories in the returned list.  These two variables exist for the
   convenience of the caller (always glob_vector). */
static struct globval *
finddirs (pat, sdir, flags, ep, np)
     char *pat;
     char *sdir;
     int flags;
     struct globval **ep;
     int *np;
{
  char **r, *n;
  int ndirs;
  struct globval *ret, *e, *g;

/*itrace("finddirs: pat = `%s' sdir = `%s' flags = 0x%x", pat, sdir, flags);*/
  e = ret = 0;
  r = glob_vector (pat, sdir, flags);
  if (r == 0 || r[0] == 0)
    {
      if (np)
	*np = 0;
      if (ep)
        *ep = 0;
      if (r && r != &glob_error_return)
	free (r);
      return (struct globval *)0;
    }
  for (ndirs = 0; r[ndirs] != 0; ndirs++)
    {
      g = (struct globval *) malloc (sizeof (struct globval));
      if (g == 0)
	{
	  while (ret)		/* free list built so far */
	    {
	      g = ret->next;
	      free (ret);
	      ret = g;
	    }

	  free (r);
	  if (np)
	    *np = 0;
	  if (ep)
	    *ep = 0;
	  return (&finddirs_error_return);
	}
      if (e == 0)
	e = g;

      g->next = ret;
      ret = g;

      g->name = r[ndirs];
    }

  free (r);
  if (ep)
    *ep = e;
  if (np)
    *np = ndirs;

  return ret;
}

     	
/* Return a vector of names of files in directory DIR
   whose names match glob pattern PAT.
   The names are not in any particular order.
   Wildcards at the beginning of PAT do not match an initial period.

   The vector is terminated by an element that is a null pointer.

   To free the space allocated, first free the vector's elements,
   then free the vector.

   Return 0 if cannot get enough memory to hold the pointer
   and the names.

   Return -1 if cannot access directory DIR.
   Look in errno for more information.  */

char **
glob_vector (pat, dir, flags)
     char *pat;
     char *dir;
     int flags;
{
  DIR *d;
  register struct dirent *dp;
  struct globval *lastlink, *e, *dirlist;
  register struct globval *nextlink;
  register char *nextname, *npat, *subdir;
  unsigned int count;
  int lose, skip, ndirs, isdir, sdlen, add_current, patlen;
  register char **name_vector;
  register unsigned int i;
  int mflags;		/* Flags passed to strmatch (). */
  int pflags;		/* flags passed to sh_makepath () */
  int nalloca;
  struct globval *firstmalloc, *tmplink;
  char *convfn;

  lastlink = 0;
  count = lose = skip = add_current = 0;

  firstmalloc = 0;
  nalloca = 0;

/*itrace("glob_vector: pat = `%s' dir = `%s' flags = 0x%x", pat, dir, flags);*/
  /* If PAT is empty, skip the loop, but return one (empty) filename. */
  if (pat == 0 || *pat == '\0')
    {
      if (glob_testdir (dir) < 0)
	return ((char **) &glob_error_return);

      nextlink = (struct globval *)alloca (sizeof (struct globval));
      if (nextlink == NULL)
	return ((char **) NULL);

      nextlink->next = (struct globval *)0;
      nextname = (char *) malloc (1);
      if (nextname == 0)
	lose = 1;
      else
	{
	  lastlink = nextlink;
	  nextlink->name = nextname;
	  nextname[0] = '\0';
	  count = 1;
	}

      skip = 1;
    }

  patlen = strlen (pat);

  /* If the filename pattern (PAT) does not contain any globbing characters,
     we can dispense with reading the directory, and just see if there is
     a filename `DIR/PAT'.  If there is, and we can access it, just make the
     vector to return and bail immediately. */
  if (skip == 0 && glob_pattern_p (pat) == 0)
    {
      int dirlen;
      struct stat finfo;

      if (glob_testdir (dir) < 0)
	return ((char **) &glob_error_return);

      dirlen = strlen (dir);
      nextname = (char *)malloc (dirlen + patlen + 2);
      npat = (char *)malloc (patlen + 1);
      if (nextname == 0 || npat == 0)
	lose = 1;
      else
	{
	  strcpy (npat, pat);
	  dequote_pathname (npat);

	  strcpy (nextname, dir);
	  nextname[dirlen++] = '/';
	  strcpy (nextname + dirlen, npat);

	  if (GLOB_TESTNAME (nextname) >= 0)
	    {
	      free (nextname);
	      nextlink = (struct globval *)alloca (sizeof (struct globval));
	      if (nextlink)
		{
		  nextlink->next = (struct globval *)0;
		  lastlink = nextlink;
		  nextlink->name = npat;
		  count = 1;
		}
	      else
		lose = 1;
	    }
	  else
	    {
	      free (nextname);
	      free (npat);
	    }
	}

      skip = 1;
    }

  if (skip == 0)
    {
      /* Open the directory, punting immediately if we cannot.  If opendir
	 is not robust (i.e., it opens non-directories successfully), test
	 that DIR is a directory and punt if it's not. */
#if defined (OPENDIR_NOT_ROBUST)
      if (glob_testdir (dir) < 0)
	return ((char **) &glob_error_return);
#endif

      d = opendir (dir);
      if (d == NULL)
	return ((char **) &glob_error_return);

      /* Compute the flags that will be passed to strmatch().  We don't
	 need to do this every time through the loop. */
      mflags = (noglob_dot_filenames ? FNM_PERIOD : 0) | FNM_PATHNAME;

#ifdef FNM_CASEFOLD
      if (glob_ignore_case)
	mflags |= FNM_CASEFOLD;
#endif

      if (extended_glob)
	mflags |= FNM_EXTMATCH;

      add_current = ((flags & (GX_ALLDIRS|GX_ADDCURDIR)) == (GX_ALLDIRS|GX_ADDCURDIR));