/* Various processing of names.
- Copyright (C) 1988, 92, 94, 96, 97, 98, 1999 Free Software Foundation, Inc.
+
+ Copyright (C) 1988, 1992, 1994, 1996, 1997, 1998, 1999, 2000, 2001,
+ 2003, 2004, 2005, 2006, 2007, 2009 Free Software Foundation, Inc.
This program 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 2, or (at your option) any later
+ Free Software Foundation; either version 3, or (at your option) any later
version.
This program is distributed in the hope that it will be useful, but
You should have received a copy of the GNU General Public License along
with this program; if not, write to the Free Software Foundation, Inc.,
- 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
-
-/* Enable GNU extensions in fnmatch.h. */
-#ifndef _GNU_SOURCE
-# define _GNU_SOURCE 1
-#endif
+ 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA. */
-#include "system.h"
+#include <system.h>
-#include <pwd.h>
-#include <grp.h>
#include <fnmatch.h>
+#include <hash.h>
#include <quotearg.h>
#include "common.h"
\f
/* User and group names. */
-extern struct group *getgrnam ();
-extern struct passwd *getpwnam ();
-#if !HAVE_GETPWUID
-extern struct passwd *getpwuid ();
+struct group *getgrnam ();
+struct passwd *getpwnam ();
+#if ! HAVE_DECL_GETPWUID
+struct passwd *getpwuid ();
#endif
-#if !HAVE_GETGRGID
-extern struct group *getgrgid ();
+#if ! HAVE_DECL_GETGRGID
+struct group *getgrgid ();
#endif
/* Make sure you link with the proper libraries if you are running the
This code should also be modified for non-UNIX systems to do something
reasonable. */
-static char cached_uname[UNAME_FIELD_SIZE];
-static char cached_gname[GNAME_FIELD_SIZE];
+static char *cached_uname;
+static char *cached_gname;
static uid_t cached_uid; /* valid only if cached_uname is not empty */
static gid_t cached_gid; /* valid only if cached_gname is not empty */
/* These variables are valid only if nonempty. */
-static char cached_no_such_uname[UNAME_FIELD_SIZE];
-static char cached_no_such_gname[GNAME_FIELD_SIZE];
+static char *cached_no_such_uname;
+static char *cached_no_such_gname;
/* These variables are valid only if nonzero. It's not worth optimizing
the case for weird systems where 0 is not a valid uid or gid. */
static uid_t cached_no_such_uid;
static gid_t cached_no_such_gid;
-/*------------------------------------------.
-| Given UID, find the corresponding UNAME. |
-`------------------------------------------*/
+static void register_individual_file (char const *name);
+/* Given UID, find the corresponding UNAME. */
void
-uid_to_uname (uid_t uid, char uname[UNAME_FIELD_SIZE])
+uid_to_uname (uid_t uid, char **uname)
{
struct passwd *passwd;
if (uid != 0 && uid == cached_no_such_uid)
{
- *uname = '\0';
+ *uname = xstrdup ("");
return;
}
- if (!cached_uname[0] || uid != cached_uid)
+ if (!cached_uname || uid != cached_uid)
{
passwd = getpwuid (uid);
if (passwd)
{
cached_uid = uid;
- strncpy (cached_uname, passwd->pw_name, UNAME_FIELD_SIZE);
+ assign_string (&cached_uname, passwd->pw_name);
}
else
{
cached_no_such_uid = uid;
- *uname = '\0';
+ *uname = xstrdup ("");
return;
}
}
- strncpy (uname, cached_uname, UNAME_FIELD_SIZE);
+ *uname = xstrdup (cached_uname);
}
-/*------------------------------------------.
-| Given GID, find the corresponding GNAME. |
-`------------------------------------------*/
-
+/* Given GID, find the corresponding GNAME. */
void
-gid_to_gname (gid_t gid, char gname[GNAME_FIELD_SIZE])
+gid_to_gname (gid_t gid, char **gname)
{
struct group *group;
if (gid != 0 && gid == cached_no_such_gid)
{
- *gname = '\0';
+ *gname = xstrdup ("");
return;
}
- if (!cached_gname[0] || gid != cached_gid)
+ if (!cached_gname || gid != cached_gid)
{
- setgrent (); /* FIXME: why?! */
group = getgrgid (gid);
if (group)
{
cached_gid = gid;
- strncpy (cached_gname, group->gr_name, GNAME_FIELD_SIZE);
+ assign_string (&cached_gname, group->gr_name);
}
else
{
cached_no_such_gid = gid;
- *gname = '\0';
+ *gname = xstrdup ("");
return;
}
}
- strncpy (gname, cached_gname, GNAME_FIELD_SIZE);
+ *gname = xstrdup (cached_gname);
}
-/*-------------------------------------------------------------------------.
-| Given UNAME, set the corresponding UID and return 1, or else, return 0. |
-`-------------------------------------------------------------------------*/
-
+/* Given UNAME, set the corresponding UID and return 1, or else, return 0. */
int
-uname_to_uid (char uname[UNAME_FIELD_SIZE], uid_t *uidp)
+uname_to_uid (char const *uname, uid_t *uidp)
{
struct passwd *passwd;
- if (cached_no_such_uname[0]
- && strncmp (uname, cached_no_such_uname, UNAME_FIELD_SIZE) == 0)
+ if (cached_no_such_uname
+ && strcmp (uname, cached_no_such_uname) == 0)
return 0;
- if (!cached_uname[0]
+ if (!cached_uname
|| uname[0] != cached_uname[0]
- || strncmp (uname, cached_uname, UNAME_FIELD_SIZE) != 0)
+ || strcmp (uname, cached_uname) != 0)
{
passwd = getpwnam (uname);
if (passwd)
{
cached_uid = passwd->pw_uid;
- strncpy (cached_uname, uname, UNAME_FIELD_SIZE);
+ assign_string (&cached_uname, passwd->pw_name);
}
else
{
- strncpy (cached_no_such_uname, uname, UNAME_FIELD_SIZE);
+ assign_string (&cached_no_such_uname, uname);
return 0;
}
}
return 1;
}
-/*-------------------------------------------------------------------------.
-| Given GNAME, set the corresponding GID and return 1, or else, return 0. |
-`-------------------------------------------------------------------------*/
-
+/* Given GNAME, set the corresponding GID and return 1, or else, return 0. */
int
-gname_to_gid (char gname[GNAME_FIELD_SIZE], gid_t *gidp)
+gname_to_gid (char const *gname, gid_t *gidp)
{
struct group *group;
- if (cached_no_such_gname[0]
- && strncmp (gname, cached_no_such_gname, GNAME_FIELD_SIZE) == 0)
+ if (cached_no_such_gname
+ && strcmp (gname, cached_no_such_gname) == 0)
return 0;
- if (!cached_gname[0]
+ if (!cached_gname
|| gname[0] != cached_gname[0]
- || strncmp (gname, cached_gname, GNAME_FIELD_SIZE) != 0)
+ || strcmp (gname, cached_gname) != 0)
{
group = getgrnam (gname);
if (group)
{
cached_gid = group->gr_gid;
- strncpy (cached_gname, gname, GNAME_FIELD_SIZE);
+ assign_string (&cached_gname, gname);
}
else
{
- strncpy (cached_no_such_gname, gname, GNAME_FIELD_SIZE);
+ assign_string (&cached_no_such_gname, gname);
return 0;
}
}
*gidp = cached_gid;
return 1;
}
-\f
-/* Names from the command call. */
-
-static const char **name_array; /* store an array of names */
-static int allocated_names; /* how big is the array? */
-static int names; /* how many entries does it have? */
-static int name_index; /* how many of the entries have we scanned? */
-/*------------------------.
-| Initialize structures. |
-`------------------------*/
-
-void
-init_names (void)
+\f
+struct name *
+make_name (const char *file_name)
{
- allocated_names = 10;
- name_array = xmalloc (sizeof (const char *) * allocated_names);
- names = 0;
+ struct name *p = xzalloc (sizeof (*p));
+ if (!file_name)
+ file_name = "";
+ p->name = xstrdup (file_name);
+ p->length = strlen (p->name);
+ return p;
}
-/*--------------------------------------------------------------.
-| Add NAME at end of name_array, reallocating it as necessary. |
-`--------------------------------------------------------------*/
-
void
-name_add (const char *name)
+free_name (struct name *p)
{
- if (names == allocated_names)
+ if (p)
{
- allocated_names *= 2;
- name_array =
- xrealloc (name_array, sizeof (const char *) * allocated_names);
+ free (p->name);
+ free (p->caname);
+ free (p);
}
- name_array[names++] = name;
}
+
\f
-/* Names from external name file. */
+/* Names from the command call. */
-static FILE *name_file; /* file to read names from */
-static char *name_buffer; /* buffer to hold the current file name */
-static size_t name_buffer_length; /* allocated length of name_buffer */
+static struct name *namelist; /* first name in list, if any */
+static struct name **nametail = &namelist; /* end of name list */
-/*---.
-| ? |
-`---*/
+/* File name arguments are processed in two stages: first a
+ name_array (see below) is filled, then the names from it
+ are moved into the namelist.
-/* FIXME: I should better check more closely. It seems at first glance that
- is_pattern is only used when reading a file, and ignored for all
- command line arguments. */
+ This awkward process is needed only to implement --same-order option,
+ which is meant to help process large archives on machines with
+ limited memory. With this option on, namelist contains at most one
+ entry, which diminishes the memory consumption.
+
+ However, I very much doubt if we still need this -- Sergey */
-static inline int
-is_pattern (const char *string)
-{
- return strchr (string, '*') || strchr (string, '[') || strchr (string, '?');
-}
+/* A name_array element contains entries of three types: */
-/*-----------------------------------------------------------------------.
-| Set up to gather file names for tar. They can either come from a file |
-| or were saved from decoding arguments. |
-`-----------------------------------------------------------------------*/
+#define NELT_NAME 0 /* File name */
+#define NELT_CHDIR 1 /* Change directory request */
+#define NELT_FMASK 2 /* Change fnmatch options request */
-void
-name_init (int argc, char *const *argv)
+struct name_elt /* A name_array element. */
{
- name_buffer = xmalloc (NAME_FIELD_SIZE + 2);
- name_buffer_length = NAME_FIELD_SIZE;
+ char type; /* Element type, see NELT_* constants above */
+ union
+ {
+ const char *name; /* File or directory name */
+ int matching_flags;/* fnmatch options if type == NELT_FMASK */
+ } v;
+};
- if (files_from_option)
+static struct name_elt *name_array; /* store an array of names */
+static size_t allocated_names; /* how big is the array? */
+static size_t names; /* how many entries does it have? */
+static size_t name_index; /* how many of the entries have we scanned? */
+
+/* Check the size of name_array, reallocating it as necessary. */
+static void
+check_name_alloc ()
+{
+ if (names == allocated_names)
{
- if (!strcmp (files_from_option, "-"))
- {
- request_stdin ("-T");
- name_file = stdin;
- }
- else if (name_file = fopen (files_from_option, "r"), !name_file)
- open_fatal (files_from_option);
+ if (allocated_names == 0)
+ allocated_names = 10; /* Set initial allocation */
+ name_array = x2nrealloc (name_array, &allocated_names,
+ sizeof (name_array[0]));
}
}
-/*---.
-| ? |
-`---*/
-
+/* Add to name_array the file NAME with fnmatch options MATCHING_FLAGS */
void
-name_term (void)
+name_add_name (const char *name, int matching_flags)
{
- free (name_buffer);
- free (name_array);
-}
+ static int prev_flags = 0; /* FIXME: Or EXCLUDE_ANCHORED? */
+ struct name_elt *ep;
-/*---------------------------------------------------------------------.
-| Read the next filename from name_file and null-terminate it. Put it |
-| into name_buffer, reallocating and adjusting name_buffer_length if |
-| necessary. Return 0 at end of file, 1 otherwise. |
-`---------------------------------------------------------------------*/
+ check_name_alloc ();
+ ep = &name_array[names++];
+ if (prev_flags != matching_flags)
+ {
+ ep->type = NELT_FMASK;
+ ep->v.matching_flags = matching_flags;
+ prev_flags = matching_flags;
+ check_name_alloc ();
+ ep = &name_array[names++];
+ }
+ ep->type = NELT_NAME;
+ ep->v.name = name;
+}
-static int
-read_name_from_file (void)
+/* Add to name_array a chdir request for the directory NAME */
+void
+name_add_dir (const char *name)
{
- int character;
- size_t counter = 0;
-
- /* FIXME: getc may be called even if character was EOF the last time here. */
-
- /* FIXME: This + 2 allocation might serve no purpose. */
+ struct name_elt *ep;
+ check_name_alloc ();
+ ep = &name_array[names++];
+ ep->type = NELT_CHDIR;
+ ep->v.name = name;
+}
- while (character = getc (name_file),
- character != EOF && character != filename_terminator)
- {
- if (counter == name_buffer_length)
- {
- name_buffer_length += NAME_FIELD_SIZE;
- name_buffer = xrealloc (name_buffer, name_buffer_length + 2);
- }
- name_buffer[counter++] = character;
- }
+\f
+/* Names from external name file. */
- if (counter == 0 && character == EOF)
- return 0;
+static char *name_buffer; /* buffer to hold the current file name */
+static size_t name_buffer_length; /* allocated length of name_buffer */
- if (counter == name_buffer_length)
- {
- name_buffer_length += NAME_FIELD_SIZE;
- name_buffer = xrealloc (name_buffer, name_buffer_length + 2);
- }
- name_buffer[counter] = '\0';
+/* Set up to gather file names for tar. They can either come from a
+ file or were saved from decoding arguments. */
+void
+name_init (void)
+{
+ name_buffer = xmalloc (NAME_FIELD_SIZE + 2);
+ name_buffer_length = NAME_FIELD_SIZE;
+}
- return 1;
+void
+name_term (void)
+{
+ free (name_buffer);
+ free (name_array);
}
-/*------------------------------------------------------------------------.
-| Get the next name from ARGV or the file of names. Result is in static |
-| storage and can't be relied upon across two calls. |
-| |
-| If CHANGE_DIRS is true, treat a filename of the form "-C" as meaning |
-| that the next filename is the name of a directory to change to. If |
-| `filename_terminator' is NUL, CHANGE_DIRS is effectively always false. |
-`------------------------------------------------------------------------*/
+static int matching_flags; /* exclude_fnmatch options */
-char *
-name_next (int change_dirs)
+/* Get the next NELT_NAME element from name_array. Result is in
+ static storage and can't be relied upon across two calls.
+
+ If CHANGE_DIRS is true, treat any entries of type NELT_CHDIR as
+ the request to change to the given directory.
+
+ Entries of type NELT_FMASK cause updates of the matching_flags
+ value. */
+struct name_elt *
+name_next_elt (int change_dirs)
{
+ static struct name_elt entry;
const char *source;
char *cursor;
- int chdir_flag = 0;
-
- if (filename_terminator == '\0')
- change_dirs = 0;
- while (1)
+ while (name_index != names)
{
- /* Get a name, either from file or from saved arguments. */
-
- if (name_index == names)
+ struct name_elt *ep;
+ size_t source_len;
+
+ ep = &name_array[name_index++];
+ if (ep->type == NELT_FMASK)
{
- if (! name_file)
- break;
- if (! read_name_from_file ())
- break;
+ matching_flags = ep->v.matching_flags;
+ continue;
}
- else
+
+ source = ep->v.name;
+ source_len = strlen (source);
+ if (name_buffer_length < source_len)
{
- source = name_array[name_index++];
- if (strlen (source) > name_buffer_length)
+ do
{
- free (name_buffer);
- name_buffer_length = strlen (source);
- name_buffer = xmalloc (name_buffer_length + 2);
+ name_buffer_length *= 2;
+ if (! name_buffer_length)
+ xalloc_die ();
}
- strcpy (name_buffer, source);
+ while (name_buffer_length < source_len);
+
+ free (name_buffer);
+ name_buffer = xmalloc (name_buffer_length + 2);
}
+ strcpy (name_buffer, source);
/* Zap trailing slashes. */
cursor = name_buffer + strlen (name_buffer) - 1;
- while (cursor > name_buffer && *cursor == '/')
+ while (cursor > name_buffer && ISSLASH (*cursor))
*cursor-- = '\0';
- if (chdir_flag)
+ if (change_dirs && ep->type == NELT_CHDIR)
{
if (chdir (name_buffer) < 0)
chdir_fatal (name_buffer);
- chdir_flag = 0;
}
- else if (change_dirs && strcmp (name_buffer, "-C") == 0)
- chdir_flag = 1;
else
{
- unquote_string (name_buffer);
- return name_buffer;
+ if (unquote_option)
+ unquote_string (name_buffer);
+ if (incremental_option)
+ register_individual_file (name_buffer);
+ entry.type = ep->type;
+ entry.v.name = name_buffer;
+ return &entry;
}
}
- /* No more names in file. */
-
- if (name_file && chdir_flag)
- FATAL_ERROR ((0, 0, _("Missing file name after -C")));
-
- return 0;
+ return NULL;
}
-/*------------------------------.
-| Close the name file, if any. |
-`------------------------------*/
-
-void
-name_close (void)
+const char *
+name_next (int change_dirs)
{
- if (name_file && name_file != stdin)
- if (fclose (name_file) != 0)
- close_error (name_buffer);
+ struct name_elt *nelt = name_next_elt (change_dirs);
+ return nelt ? nelt->v.name : NULL;
}
-/*-------------------------------------------------------------------------.
-| Gather names in a list for scanning. Could hash them later if we really |
-| care. |
-| |
-| If the names are already sorted to match the archive, we just read them |
-| one by one. name_gather reads the first one, and it is called by |
-| name_match as appropriate to read the next ones. At EOF, the last name |
-| read is just left in the buffer. This option lets users of small |
-| machines extract an arbitrary number of files by doing "tar t" and |
-| editing down the list of files. |
-`-------------------------------------------------------------------------*/
+/* Gather names in a list for scanning. Could hash them later if we
+ really care.
+
+ If the names are already sorted to match the archive, we just read
+ them one by one. name_gather reads the first one, and it is called
+ by name_match as appropriate to read the next ones. At EOF, the
+ last name read is just left in the buffer. This option lets users
+ of small machines extract an arbitrary number of files by doing
+ "tar t" and editing down the list of files. */
void
name_gather (void)
{
/* Buffer able to hold a single name. */
- static struct name *buffer;
- static size_t allocated_length;
+ static struct name *buffer = NULL;
- char const *name;
+ struct name_elt *ep;
if (same_order_option)
{
static int change_dir;
- if (allocated_length == 0)
- {
- allocated_length = sizeof (struct name) + NAME_FIELD_SIZE;
- buffer = xmalloc (allocated_length);
- /* FIXME: This memset is overkill, and ugly... */
- memset (buffer, 0, allocated_length);
- }
+ while ((ep = name_next_elt (0)) && ep->type == NELT_CHDIR)
+ change_dir = chdir_arg (xstrdup (ep->v.name));
- while ((name = name_next (0)) && strcmp (name, "-C") == 0)
+ if (ep)
{
- char const *dir = name_next (0);
- if (! dir)
- FATAL_ERROR ((0, 0, _("Missing file name after -C")));
- change_dir = chdir_arg (xstrdup (dir));
- }
-
- if (name)
- {
- buffer->length = strlen (name);
- if (sizeof (struct name) + buffer->length >= allocated_length)
- {
- allocated_length = sizeof (struct name) + buffer->length;
- buffer = xrealloc (buffer, allocated_length);
- }
+ free_name (buffer);
+ buffer = make_name (ep->v.name);
buffer->change_dir = change_dir;
- strncpy (buffer->name, name, buffer->length);
- buffer->name[buffer->length] = 0;
buffer->next = 0;
- buffer->found = 0;
-
- /* FIXME: Poorly named globals, indeed... */
+ buffer->found_count = 0;
+ buffer->matching_flags = matching_flags;
+ buffer->dir_contents = NULL;
+ buffer->parent = NULL;
+
namelist = buffer;
- namelast = namelist;
+ nametail = &namelist->next;
}
+ else if (change_dir)
+ addname (0, change_dir, NULL);
}
else
{
for (;;)
{
int change_dir0 = change_dir;
- while ((name = name_next (0)) && strcmp (name, "-C") == 0)
- {
- char const *dir = name_next (0);
- if (! dir)
- FATAL_ERROR ((0, 0, _("Missing file name after -C")));
- change_dir = chdir_arg (xstrdup (dir));
- }
- if (name)
- addname (name, change_dir);
+ while ((ep = name_next_elt (0)) && ep->type == NELT_CHDIR)
+ change_dir = chdir_arg (xstrdup (ep->v.name));
+
+ if (ep)
+ addname (ep->v.name, change_dir, NULL);
else
{
if (change_dir != change_dir0)
- addname (0, change_dir);
+ addname (0, change_dir, NULL);
break;
}
}
}
}
-/*-----------------------------.
-| Add a name to the namelist. |
-`-----------------------------*/
-
-void
-addname (char const *string, int change_dir)
+/* Add a name to the namelist. */
+struct name *
+addname (char const *string, int change_dir, struct name *parent)
{
- struct name *name;
- size_t length;
-
- length = string ? strlen (string) : 0;
- name = xmalloc (sizeof (struct name) + length);
- memset (name, 0, sizeof (struct name) + length);
- name->next = 0;
-
- if (string)
- {
- name->fake = 0;
- name->length = length;
- memcpy (name->name, string, length + 1);
- }
- else
- name->fake = 1;
+ struct name *name = make_name (string);
- name->found = 0;
- name->regexp = 0; /* assume not a regular expression */
- name->firstch = 1; /* assume first char is literal */
+ name->prev = *nametail;
+ name->next = NULL;
+ name->found_count = 0;
+ name->matching_flags = matching_flags;
name->change_dir = change_dir;
- name->dir_contents = 0;
+ name->dir_contents = NULL;
+ name->parent = parent;
+
+ *nametail = name;
+ nametail = &name->next;
+ return name;
+}
- if (string && is_pattern (string))
+/* Find a match for FILE_NAME (whose string length is LENGTH) in the name
+ list. */
+static struct name *
+namelist_match (char const *file_name, size_t length)
+{
+ struct name *p;
+
+ for (p = namelist; p; p = p->next)
{
- name->regexp = 1;
- if (string[0] == '*' || string[0] == '[' || string[0] == '?')
- name->firstch = 0;
+ if (p->name[0]
+ && exclude_fnmatch (p->name, file_name, p->matching_flags))
+ return p;
}
- if (namelast)
- namelast->next = name;
- namelast = name;
- if (!namelist)
- namelist = name;
+ return NULL;
}
-/*------------------------------------------------------------------------.
-| Return true if and only if name PATH (from an archive) matches any name |
-| from the namelist. |
-`------------------------------------------------------------------------*/
+void
+remname (struct name *name)
+{
+ struct name *p;
+
+ if ((p = name->prev) != NULL)
+ p->next = name->next;
+ else
+ namelist = name->next;
-int
-name_match (const char *path)
+ if ((p = name->next) != NULL)
+ p->prev = name->prev;
+ else
+ nametail = &name->prev;
+}
+
+/* Return true if and only if name FILE_NAME (from an archive) matches any
+ name from the namelist. */
+bool
+name_match (const char *file_name)
{
- size_t length = strlen (path);
+ size_t length = strlen (file_name);
while (1)
{
struct name *cursor = namelist;
if (!cursor)
- return ! files_from_option;
-
- if (cursor->fake)
+ return true;
+
+ if (cursor->name[0] == 0)
{
chdir_do (cursor->change_dir);
namelist = 0;
- return ! files_from_option;
+ nametail = &namelist;
+ return true;
}
- for (; cursor; cursor = cursor->next)
+ cursor = namelist_match (file_name, length);
+ if (cursor)
{
- /* If first chars don't match, quick skip. */
-
- if (cursor->firstch && cursor->name[0] != path[0])
- continue;
-
- if (cursor->regexp
- ? fnmatch (cursor->name, path, FNM_LEADING_DIR) == 0
- : (cursor->length <= length
- && (path[cursor->length] == '\0'
- || path[cursor->length] == '/')
- && memcmp (path, cursor->name, cursor->length) == 0))
+ if (!(ISSLASH (file_name[cursor->length]) && recursion_option)
+ || cursor->found_count == 0)
+ cursor->found_count++; /* remember it matched */
+ if (starting_file_option)
{
- cursor->found = 1; /* remember it matched */
- if (starting_file_option)
- {
- free (namelist);
- namelist = 0;
- }
- chdir_do (cursor->change_dir);
-
- /* We got a match. */
- return 1;
+ free (namelist);
+ namelist = 0;
+ nametail = &namelist;
}
+ chdir_do (cursor->change_dir);
+
+ /* We got a match. */
+ return ISFOUND (cursor);
}
/* Filename from archive not found in namelist. If we have the whole
namelist here, just return 0. Otherwise, read the next name in and
- compare it. If this was the last name, namelist->found will remain
- on. If not, we loop to compare the newly read name. */
+ compare it. If this was the last name, namelist->found_count will
+ remain on. If not, we loop to compare the newly read name. */
- if (same_order_option && namelist->found)
+ if (same_order_option && namelist->found_count)
{
name_gather (); /* read one more */
- if (namelist->found)
- return 0;
+ if (namelist->found_count)
+ return false;
}
else
- return 0;
+ return false;
}
}
-/*------------------------------------------------------------------.
-| Print the names of things in the namelist that were not matched. |
-`------------------------------------------------------------------*/
+/* Returns true if all names from the namelist were processed.
+ P is the stat_info of the most recently processed entry.
+ The decision is postponed until the next entry is read if:
-void
-names_notfound (void)
+ 1) P ended with a slash (i.e. it was a directory)
+ 2) P matches any entry from the namelist *and* represents a subdirectory
+ or a file lying under this entry (in the terms of directory structure).
+
+ This is necessary to handle contents of directories. */
+bool
+all_names_found (struct tar_stat_info *p)
{
- struct name *cursor;
- struct name *next;
+ struct name const *cursor;
+ size_t len;
+
+ if (test_label_option)
+ return true;
+ if (!p->file_name || occurrence_option == 0 || p->had_trailing_slash)
+ return false;
+ len = strlen (p->file_name);
+ for (cursor = namelist; cursor; cursor = cursor->next)
+ {
+ if ((cursor->name[0] && !WASFOUND (cursor))
+ || (len >= cursor->length && ISSLASH (p->file_name[cursor->length])))
+ return false;
+ }
+ return true;
+}
+
+static inline int
+is_pattern (const char *string)
+{
+ return strchr (string, '*') || strchr (string, '[') || strchr (string, '?');
+}
- for (cursor = namelist; cursor; cursor = next)
+static void
+regex_usage_warning (const char *name)
+{
+ static int warned_once = 0;
+
+ if (warn_regex_usage && is_pattern (name))
{
- next = cursor->next;
- if (!cursor->found && !cursor->fake)
- ERROR ((0, 0, _("%s: Not found in archive"),
- quotearg_colon (cursor->name)));
-
- /* We could free the list, but the process is about to die anyway, so
- save some CPU time. Amigas and other similarly broken software
- will need to waste the time, though. */
-
-#ifdef amiga
- if (!same_order_option)
- free (cursor);
-#endif
+ warned_once = 1;
+ WARN ((0, 0,
+ /* TRANSLATORS: The following three msgids form a single sentence.
+ */
+ _("Pattern matching characters used in file names. Please,")));
+ WARN ((0, 0,
+ _("use --wildcards to enable pattern matching, or --no-wildcards to")));
+ WARN ((0, 0,
+ _("suppress this warning.")));
}
+}
+
+/* Print the names of things in the namelist that were not matched. */
+void
+names_notfound (void)
+{
+ struct name const *cursor;
+
+ for (cursor = namelist; cursor; cursor = cursor->next)
+ if (!WASFOUND (cursor) && cursor->name[0])
+ {
+ regex_usage_warning (cursor->name);
+ if (cursor->found_count == 0)
+ ERROR ((0, 0, _("%s: Not found in archive"),
+ quotearg_colon (cursor->name)));
+ else
+ ERROR ((0, 0, _("%s: Required occurrence not found in archive"),
+ quotearg_colon (cursor->name)));
+ }
+
+ /* Don't bother freeing the name list; we're about to exit. */
namelist = 0;
- namelast = 0;
+ nametail = &namelist;
if (same_order_option)
{
- char *name;
+ const char *name;
- while (name = name_next (1), name)
- ERROR ((0, 0, _("%s: Not found in archive"),
- quotearg_colon (name)));
+ while ((name = name_next (1)) != NULL)
+ {
+ regex_usage_warning (name);
+ ERROR ((0, 0, _("%s: Not found in archive"),
+ quotearg_colon (name)));
+ }
}
}
\f
/* Sorting name lists. */
-/* Sort linked LIST of names, of given LENGTH, using COMPARE to order
- names. Return the sorted list. Apart from the type `struct name'
- and the definition of SUCCESSOR, this is a generic list-sorting
- function, but it's too painful to make it both generic and portable
+/* Sort *singly* linked LIST of names, of given LENGTH, using COMPARE
+ to order names. Return the sorted list. Note that after calling
+ this function, the `prev' links in list elements are messed up.
+
+ Apart from the type `struct name' and the definition of SUCCESSOR,
+ this is a generic list-sorting function, but it's too painful to
+ make it both generic and portable
in C. */
static struct name *
-merge_sort (struct name *list, int length,
- int (*compare) (struct name const*, struct name const*))
+merge_sort_sll (struct name *list, int length,
+ int (*compare) (struct name const*, struct name const*))
{
struct name *first_list;
struct name *second_list;
second_list = SUCCESSOR (cursor);
SUCCESSOR (cursor) = 0;
- first_list = merge_sort (first_list, first_length, compare);
- second_list = merge_sort (second_list, second_length, compare);
+ first_list = merge_sort_sll (first_list, first_length, compare);
+ second_list = merge_sort_sll (second_list, second_length, compare);
merge_point = &result;
while (first_list && second_list)
#undef SUCCESSOR
}
+/* Sort doubly linked LIST of names, of given LENGTH, using COMPARE
+ to order names. Return the sorted list. */
+static struct name *
+merge_sort (struct name *list, int length,
+ int (*compare) (struct name const*, struct name const*))
+{
+ struct name *head, *p, *prev;
+ head = merge_sort_sll (list, length, compare);
+ /* Fixup prev pointers */
+ for (prev = NULL, p = head; p; prev = p, p = p->next)
+ p->prev = prev;
+ return head;
+}
+
/* A comparison function for sorting names. Put found names last;
break ties by string comparison. */
static int
-compare_names (struct name const *n1, struct name const *n2)
+compare_names_found (struct name const *n1, struct name const *n2)
{
- int found_diff = n2->found - n1->found;
+ int found_diff = WASFOUND (n2) - WASFOUND (n1);
return found_diff ? found_diff : strcmp (n1->name, n2->name);
}
+
+/* Simple comparison by names. */
+static int
+compare_names (struct name const *n1, struct name const *n2)
+{
+ return strcmp (n1->name, n2->name);
+}
+
\f
-/* Add all the dirs in PATH, which is a directory, to the namelist.
+/* Add all the dirs under NAME, which names a directory, to the namelist.
If any of the files is a directory, recurse on the subdirectory.
- CHANGE_DIR is the number of the directory that PATH is relative to.
- DEVICE is the device not to leave, if the -l option is specified. */
+ DEVICE is the device not to leave, if the -l option is specified.
+ CMDLINE is true, if the NAME appeared on the command line. */
static void
-add_hierarchy_to_namelist (char *path, int change_dir, dev_t device)
+add_hierarchy_to_namelist (struct name *name, dev_t device, bool cmdline)
{
- char *buffer = get_directory_contents (path, device);
-
- {
- struct name *name;
-
- for (name = namelist; name; name = name->next)
- if (strcmp (name->name, path) == 0)
- break;
- if (name)
- name->dir_contents = buffer ? buffer : "\0\0\0\0";
- }
-
- if (buffer)
+ char *file_name = name->name;
+ const char *buffer = scan_directory (file_name, device, cmdline);
+
+ if (! buffer)
+ name->dir_contents = "\0\0\0\0";
+ else
{
- size_t name_length = strlen (path);
+ struct name *child_head = NULL, *child_tail = NULL;
+ size_t name_length = name->length;
size_t allocated_length = (name_length >= NAME_FIELD_SIZE
? name_length + NAME_FIELD_SIZE
: NAME_FIELD_SIZE);
- char *name_buffer = xmalloc (allocated_length + 1);
+ char *namebuf = xmalloc (allocated_length + 1);
/* FIXME: + 2 above? */
- char *string;
+ const char *string;
size_t string_length;
+ int change_dir = name->change_dir;
- strcpy (name_buffer, path);
- if (name_buffer[name_length - 1] != '/')
+ name->dir_contents = buffer;
+ strcpy (namebuf, file_name);
+ if (! ISSLASH (namebuf[name_length - 1]))
{
- name_buffer[name_length++] = '/';
- name_buffer[name_length] = '\0';
+ namebuf[name_length++] = '/';
+ namebuf[name_length] = '\0';
}
for (string = buffer; *string; string += string_length + 1)
string_length = strlen (string);
if (*string == 'D')
{
- if (name_length + string_length >= allocated_length)
+ struct name *np;
+
+ if (allocated_length <= name_length + string_length)
{
- while (name_length + string_length >= allocated_length)
- allocated_length += NAME_FIELD_SIZE;
- name_buffer = xrealloc (name_buffer, allocated_length + 1);
+ do
+ {
+ allocated_length *= 2;
+ if (! allocated_length)
+ xalloc_die ();
+ }
+ while (allocated_length <= name_length + string_length);
+
+ namebuf = xrealloc (namebuf, allocated_length + 1);
}
- strcpy (name_buffer + name_length, string + 1);
- addname (name_buffer, change_dir);
- if (*string == 'D')
- add_hierarchy_to_namelist (name_buffer, change_dir, device);
+ strcpy (namebuf + name_length, string + 1);
+ np = addname (namebuf, change_dir, name);
+ if (!child_head)
+ child_head = np;
+ else
+ child_tail->sibling = np;
+ child_tail = np;
+ add_hierarchy_to_namelist (np, device, false);
}
}
- free (name_buffer);
+ free (namebuf);
+ name->child = child_head;
}
}
\f
+/* Auxiliary functions for hashed table of struct name's. */
+
+static size_t
+name_hash (void const *entry, size_t n_buckets)
+{
+ struct name const *name = entry;
+ return hash_string (name->caname, n_buckets);
+}
+
+/* Compare two directories for equality of their names. */
+static bool
+name_compare (void const *entry1, void const *entry2)
+{
+ struct name const *name1 = entry1;
+ struct name const *name2 = entry2;
+ return strcmp (name1->caname, name2->caname) == 0;
+}
+
+\f
+/* Rebase `name' member of CHILD and all its siblings to
+ the new PARENT. */
+static void
+rebase_child_list (struct name *child, struct name *parent)
+{
+ size_t old_prefix_len = child->parent->length;
+ size_t new_prefix_len = parent->length;
+ char *new_prefix = parent->name;
+
+ for (; child; child = child->sibling)
+ {
+ size_t size = child->length - old_prefix_len + new_prefix_len;
+ char *newp = xmalloc (size + 1);
+ strcpy (newp, new_prefix);
+ strcat (newp, child->name + old_prefix_len);
+ free (child->name);
+ child->name = newp;
+ child->length = size;
+
+ rebase_directory (child->name, old_prefix_len, child->parent->name,
+ new_prefix);
+ }
+}
+
/* Collect all the names from argv[] (or whatever), expand them into a
directory tree, and sort them. This gets only subdirectories, not
all files. */
collect_and_sort_names (void)
{
struct name *name;
- struct name *next_name;
+ struct name *next_name, *prev_name;
int num_names;
struct stat statbuf;
-
+ Hash_table *nametab;
+
name_gather ();
- if (listed_incremental_option)
- read_directory_file ();
-
if (!namelist)
- addname (".", 0);
+ addname (".", 0, NULL);
+ if (listed_incremental_option)
+ {
+ switch (chdir_count ())
+ {
+ case 0:
+ break;
+
+ case 1:
+ if (namelist->change_dir == 0)
+ USAGE_ERROR ((0, 0,
+ _("Using -C option inside file list is not "
+ "allowed with --listed-incremental")));
+ break;
+
+ default:
+ USAGE_ERROR ((0, 0,
+ _("Only one -C option is allowed with "
+ "--listed-incremental")));
+ }
+ chdir_do (namelist->change_dir);
+ read_directory_file ();
+ }
+
for (name = namelist; name; name = next_name)
{
next_name = name->next;
- if (name->found || name->dir_contents)
+ if (name->found_count || name->dir_contents)
continue;
- if (name->regexp) /* FIXME: just skip regexps for now */
+ if (name->matching_flags & EXCLUDE_WILDCARDS)
+ /* NOTE: EXCLUDE_ANCHORED is not relevant here */
+ /* FIXME: just skip regexps for now */
continue;
chdir_do (name->change_dir);
- if (name->fake)
+
+ if (name->name[0] == 0)
continue;
if (deref_stat (dereference_option, name->name, &statbuf) != 0)
{
- stat_error (name->name);
+ stat_diag (name->name);
continue;
}
if (S_ISDIR (statbuf.st_mode))
{
- name->found = 1;
- add_hierarchy_to_namelist (name->name, name->change_dir,
- statbuf.st_dev);
+ name->found_count++;
+ if (name->found_count == 1)
+ add_hierarchy_to_namelist (name, statbuf.st_dev, true);
}
}
num_names = 0;
for (name = namelist; name; name = name->next)
num_names++;
+
namelist = merge_sort (namelist, num_names, compare_names);
- for (name = namelist; name; name = name->next)
- name->found = 0;
-}
+ num_names = 0;
+ nametab = hash_initialize (0, 0,
+ name_hash,
+ name_compare, NULL);
+ for (name = namelist; name; name = next_name)
+ {
+ next_name = name->next;
+ name->caname = normalize_filename (name->name);
+ if (prev_name)
+ {
+ struct name *p = hash_lookup (nametab, name);
+ if (p)
+ {
+ /* Keep the one listed in the command line */
+ if (!name->parent)
+ {
+ if (p->child)
+ rebase_child_list (p->child, name);
+ /* FIXME: remove_directory (p->caname); ? */
+ remname (p);
+ free_name (p);
+ num_names--;
+ }
+ else
+ {
+ if (name->child)
+ rebase_child_list (name->child, p);
+ /* FIXME: remove_directory (name->caname); ? */
+ remname (name);
+ free_name (name);
+ continue;
+ }
+ }
+ }
+ name->found_count = 0;
+ hash_insert (nametab, name);
+ prev_name = name;
+ num_names++;
+ }
+ nametail = &prev_name;
+ hash_free (nametab);
-/*-------------------------------------------------------------------------.
-| This is like name_match, except that it returns a pointer to the name it |
-| matched, and doesn't set FOUND in structure. The caller will have to do |
-| that if it wants to. Oh, and if the namelist is empty, it returns null, |
-| unlike name_match, which returns TRUE. |
-`-------------------------------------------------------------------------*/
+ namelist = merge_sort (namelist, num_names, compare_names_found);
+ if (listed_incremental_option)
+ {
+ for (name = namelist; name && name->name[0] == 0; name++)
+ ;
+ if (name)
+ name->dir_contents = append_incremental_renames (name->dir_contents);
+ }
+}
+
+/* This is like name_match, except that
+ 1. It returns a pointer to the name it matched, and doesn't set FOUND
+ in structure. The caller will have to do that if it wants to.
+ 2. If the namelist is empty, it returns null, unlike name_match, which
+ returns TRUE. */
struct name *
-name_scan (const char *path)
+name_scan (const char *file_name)
{
- size_t length = strlen (path);
+ size_t length = strlen (file_name);
while (1)
{
- struct name *cursor = namelist;
-
- if (!cursor)
- return 0;
-
- for (; cursor; cursor = cursor->next)
- {
- /* If first chars don't match, quick skip. */
-
- if (cursor->firstch && cursor->name[0] != path[0])
- continue;
-
- if (cursor->regexp
- ? fnmatch (cursor->name, path, FNM_LEADING_DIR) == 0
- : (cursor->length <= length
- && (path[cursor->length] == '\0'
- || path[cursor->length] == '/')
- && memcmp (path, cursor->name, cursor->length) == 0))
- return cursor; /* we got a match */
- }
+ struct name *cursor = namelist_match (file_name, length);
+ if (cursor)
+ return cursor;
/* Filename from archive not found in namelist. If we have the whole
namelist here, just return 0. Otherwise, read the next name in and
- compare it. If this was the last name, namelist->found will remain
- on. If not, we loop to compare the newly read name. */
+ compare it. If this was the last name, namelist->found_count will
+ remain on. If not, we loop to compare the newly read name. */
- if (same_order_option && namelist->found)
+ if (same_order_option && namelist && namelist->found_count)
{
name_gather (); /* read one more */
- if (namelist->found)
+ if (namelist->found_count)
return 0;
}
else
}
}
-/*-----------------------------------------------------------------------.
-| This returns a name from the namelist which doesn't have ->found set. |
-| It sets ->found before returning, so successive calls will find and |
-| return all the non-found names in the namelist |
-`-----------------------------------------------------------------------*/
-
+/* This returns a name from the namelist which doesn't have ->found
+ set. It sets ->found before returning, so successive calls will
+ find and return all the non-found names in the namelist. */
struct name *gnu_list_name;
char *
{
if (!gnu_list_name)
gnu_list_name = namelist;
- while (gnu_list_name && (gnu_list_name->found | gnu_list_name->fake))
+ while (gnu_list_name
+ && (gnu_list_name->found_count || gnu_list_name->name[0] == 0))
gnu_list_name = gnu_list_name->next;
if (gnu_list_name)
{
- gnu_list_name->found = 1;
+ gnu_list_name->found_count++;
chdir_do (gnu_list_name->change_dir);
return gnu_list_name->name;
}
return 0;
}
-/*---.
-| ? |
-`---*/
-
void
blank_name_list (void)
{
gnu_list_name = 0;
for (name = namelist; name; name = name->next)
- name->found = 0;
+ name->found_count = 0;
}
-/*---.
-| ? |
-`---*/
-
+/* Yield a newly allocated file name consisting of FILE_NAME concatenated to
+ NAME, with an intervening slash if FILE_NAME does not already end in one. */
char *
-new_name (const char *path, const char *name)
+new_name (const char *file_name, const char *name)
{
- char *buffer = xmalloc (strlen (path) + strlen (name) + 2);
-
- sprintf (buffer, "%s/%s", path, name);
+ size_t file_name_len = strlen (file_name);
+ size_t namesize = strlen (name) + 1;
+ int slash = file_name_len && ! ISSLASH (file_name[file_name_len - 1]);
+ char *buffer = xmalloc (file_name_len + slash + namesize);
+ memcpy (buffer, file_name, file_name_len);
+ buffer[file_name_len] = '/';
+ memcpy (buffer + file_name_len + slash, name, namesize);
return buffer;
}
-/* Return nonzero if file NAME is excluded. Exclude a name if its
- prefix matches a pattern that contains slashes, or if one of its
- components matches a pattern that contains no slashes. */
-int
+/* Return nonzero if file NAME is excluded. */
+bool
excluded_name (char const *name)
{
- char const *p;
- name += FILESYSTEM_PREFIX_LEN (name);
-
- if (excluded_filename (excluded_with_slash, name,
- FNM_FILE_NAME | FNM_LEADING_DIR))
- return 1;
-
- for (p = name; *p; p++)
- if (((p == name || ISSLASH (p[-1])) && !ISSLASH (p[0]))
- && excluded_filename (excluded_without_slash, p,
- FNM_FILE_NAME | FNM_LEADING_DIR))
- return 1;
-
- return 0;
+ return excluded_file_name (excluded, name + FILE_SYSTEM_PREFIX_LEN (name));
}
\f
/* Names to avoid dumping. */
-
-struct avoided_name
-{
- struct avoided_name const *next;
- char name[1];
-};
-
-static struct avoided_name *avoided_names;
+static Hash_table *avoided_name_table;
/* Remember to not archive NAME. */
void
add_avoided_name (char const *name)
{
- struct avoided_name *p = xmalloc (sizeof *p + strlen (name));
- p->next = avoided_names;
- avoided_names = p;
- strcpy (p->name, name);
+ hash_string_insert (&avoided_name_table, name);
}
/* Should NAME be avoided when archiving? */
-int
+bool
is_avoided_name (char const *name)
{
- struct avoided_name const *p;
- for (p = avoided_names; p; p = p->next)
- if (strcmp (p->name, name) == 0)
- return 1;
- return 0;
+ return hash_string_lookup (avoided_name_table, name);
+}
+
+\f
+static Hash_table *individual_file_table;
+
+static void
+register_individual_file (char const *name)
+{
+ struct stat st;
+
+ if (deref_stat (dereference_option, name, &st) != 0)
+ return; /* Will be complained about later */
+ if (S_ISDIR (st.st_mode))
+ return;
+
+ hash_string_insert (&individual_file_table, name);
+}
+
+bool
+is_individual_file (char const *name)
+{
+ return hash_string_lookup (individual_file_table, name);
+}
+
+\f
+
+/* Return the size of the prefix of FILE_NAME that is removed after
+ stripping NUM leading file name components. NUM must be
+ positive. */
+
+size_t
+stripped_prefix_len (char const *file_name, size_t num)
+{
+ char const *p = file_name + FILE_SYSTEM_PREFIX_LEN (file_name);
+ while (ISSLASH (*p))
+ p++;
+ while (*p)
+ {
+ bool slash = ISSLASH (*p);
+ p++;
+ if (slash)
+ {
+ if (--num == 0)
+ return p - file_name;
+ while (ISSLASH (*p))
+ p++;
+ }
+ }
+ return -1;
+}
+\f
+/* Return nonzero if NAME contains ".." as a file name component. */
+bool
+contains_dot_dot (char const *name)
+{
+ char const *p = name + FILE_SYSTEM_PREFIX_LEN (name);
+
+ for (;; p++)
+ {
+ if (p[0] == '.' && p[1] == '.' && (ISSLASH (p[2]) || !p[2]))
+ return 1;
+
+ while (! ISSLASH (*p))
+ {
+ if (! *p++)
+ return 0;
+ }
+ }
}
projects / chaz / tar / blobdiff