X-Git-Url: https://git.dogcows.com/gitweb?p=chaz%2Ftar;a=blobdiff_plain;f=src%2Flist.c;h=d46be651cb3752b0f1fe2d0fc29d79a8a256feb3;hp=a62c706261a58ca142f19e9fb532c802b4328a6b;hb=45ccda119355a1087450039a250359c1d0de0d08;hpb=368172bf255c539fdbf2fd74328514c7e0acde8f diff --git a/src/list.c b/src/list.c index a62c706..d46be65 100644 --- a/src/list.c +++ b/src/list.c @@ -1,43 +1,55 @@ /* List a tar archive, with support routines for reading a tar archive. - Copyright (C) 1988, 1992, 1993, 1994, 1996, 1997, 1998, 1999, 2000, - 2001, 2003, 2004 Free Software Foundation, Inc. + Copyright 1988, 1992-1994, 1996-2001, 2003-2007, 2010, 2012-2014 Free + Software Foundation, Inc. - Written by John Gilmore, on 1985-08-26. + This file is part of GNU tar. - 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 - version. + GNU tar 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. - This program 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. + GNU tar 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 this program; if not, write to the Free Software Foundation, Inc., - 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */ + You should have received a copy of the GNU General Public License + along with this program. If not, see . -/* Define to non-zero for forcing old ctime format instead of ISO format. */ -#undef USE_OLD_CTIME + Written by John Gilmore, on 1985-08-26. */ #include +#include #include #include "common.h" -#define max(a, b) ((a) < (b) ? (b) : (a)) - union block *current_header; /* points to current archive header */ enum archive_format current_format; /* recognized format */ union block *recent_long_name; /* recent long name header and contents */ union block *recent_long_link; /* likewise, for long link */ size_t recent_long_name_blocks; /* number of blocks in recent_long_name */ size_t recent_long_link_blocks; /* likewise, for long link */ - -static uintmax_t from_header (const char *, size_t, const char *, - uintmax_t, uintmax_t, bool); +static union block *recent_global_header; /* Recent global header block */ + +#define GID_FROM_HEADER(where) gid_from_header (where, sizeof (where)) +#define MAJOR_FROM_HEADER(where) major_from_header (where, sizeof (where)) +#define MINOR_FROM_HEADER(where) minor_from_header (where, sizeof (where)) +#define MODE_FROM_HEADER(where, hbits) \ + mode_from_header (where, sizeof (where), hbits) +#define TIME_FROM_HEADER(where) time_from_header (where, sizeof (where)) +#define UID_FROM_HEADER(where) uid_from_header (where, sizeof (where)) + +static gid_t gid_from_header (const char *buf, size_t size); +static major_t major_from_header (const char *buf, size_t size); +static minor_t minor_from_header (const char *buf, size_t size); +static mode_t mode_from_header (const char *buf, size_t size, bool *hbits); +static time_t time_from_header (const char *buf, size_t size); +static uid_t uid_from_header (const char *buf, size_t size); +static intmax_t from_header (const char *, size_t, const char *, + intmax_t, uintmax_t, bool, bool); /* Base 64 digits; see Internet RFC 2045 Table 1. */ static char const base_64_digits[64] = @@ -62,12 +74,100 @@ base64_init (void) base64_map[(int) base_64_digits[i]] = i; } +static char * +decode_xform (char *file_name, void *data) +{ + int type = *(int*)data; + + switch (type) + { + case XFORM_SYMLINK: + /* FIXME: It is not quite clear how and to which extent are the symbolic + links subject to filename transformation. In the absence of another + solution, symbolic links are exempt from component stripping and + name suffix normalization, but subject to filename transformation + proper. */ + return file_name; + + case XFORM_LINK: + file_name = safer_name_suffix (file_name, true, absolute_names_option); + break; + + case XFORM_REGFILE: + file_name = safer_name_suffix (file_name, false, absolute_names_option); + break; + } + + if (strip_name_components) + { + size_t prefix_len = stripped_prefix_len (file_name, + strip_name_components); + if (prefix_len == (size_t) -1) + prefix_len = strlen (file_name); + file_name += prefix_len; + } + return file_name; +} + +static bool +transform_member_name (char **pinput, int type) +{ + return transform_name_fp (pinput, type, decode_xform, &type); +} + +static void +enforce_one_top_level (char **pfile_name) +{ + char *file_name = *pfile_name; + char *p; + + for (p = file_name; *p && (ISSLASH (*p) || *p == '.'); p++) + ; + + if (!*p) + return; + + if (strncmp (p, one_top_level_dir, strlen (one_top_level_dir)) == 0) + { + int pos = strlen (one_top_level_dir); + if (ISSLASH (p[pos]) || p[pos] == 0) + return; + } + + *pfile_name = new_name (one_top_level_dir, file_name); + normalize_filename_x (*pfile_name); + free (file_name); +} + +void +transform_stat_info (int typeflag, struct tar_stat_info *stat_info) +{ + if (typeflag == GNUTYPE_VOLHDR) + /* Name transformations don't apply to volume headers. */ + return; + + transform_member_name (&stat_info->file_name, XFORM_REGFILE); + switch (typeflag) + { + case SYMTYPE: + transform_member_name (&stat_info->link_name, XFORM_SYMLINK); + break; + + case LNKTYPE: + transform_member_name (&stat_info->link_name, XFORM_LINK); + } + + if (one_top_level_option) + enforce_one_top_level (¤t_stat_info.file_name); +} + /* Main loop for reading an archive. */ void read_and (void (*do_something) (void)) { enum read_header status = HEADER_STILL_UNREAD; enum read_header prev_status; + struct timespec mtime; base64_init (); name_gather (); @@ -77,9 +177,9 @@ read_and (void (*do_something) (void)) { prev_status = status; tar_stat_destroy (¤t_stat_info); - xheader_destroy (&extended_header); - status = read_header (false); + status = read_header (¤t_header, ¤t_stat_info, + read_header_auto); switch (status) { case HEADER_STILL_UNREAD: @@ -90,26 +190,25 @@ read_and (void (*do_something) (void)) /* Valid header. We should decode next field (mode) first. Ensure incoming names are null terminated. */ - + decode_header (current_header, ¤t_stat_info, + ¤t_format, 1); if (! name_match (current_stat_info.file_name) || (NEWER_OPTION_INITIALIZED (newer_mtime_option) /* FIXME: We get mtime now, and again later; this causes duplicate diagnostics if header.mtime is bogus. */ - && ((current_stat_info.stat.st_mtime + && ((mtime.tv_sec = TIME_FROM_HEADER (current_header->header.mtime)), -#ifdef ST_MTIM_NSEC /* FIXME: Grab fractional time stamps from extended header. */ - current_stat_info.stat.st_mtim.ST_MTIM_NSEC = 0, -#endif - OLDER_STAT_TIME (current_stat_info.stat, m))) + mtime.tv_nsec = 0, + current_stat_info.mtime = mtime, + OLDER_TAR_STAT_TIME (current_stat_info, m))) || excluded_name (current_stat_info.file_name)) { switch (current_header->header.typeflag) { case GNUTYPE_VOLHDR: case GNUTYPE_MULTIVOL: - case GNUTYPE_NAMES: break; case DIRTYPE: @@ -118,13 +217,13 @@ read_and (void (*do_something) (void)) quotearg_colon (current_stat_info.file_name))); /* Fall through. */ default: - decode_header (current_header, - ¤t_stat_info, ¤t_format, 0); skip_member (); continue; } } + transform_stat_info (current_header->header.typeflag, + ¤t_stat_info); (*do_something) (); continue; @@ -142,11 +241,13 @@ read_and (void (*do_something) (void)) { char buf[UINTMAX_STRSIZE_BOUND]; - status = read_header (false); + status = read_header (¤t_header, ¤t_stat_info, + read_header_auto); if (status == HEADER_ZERO_BLOCK) break; - WARN ((0, 0, _("A lone zero block at %s"), - STRINGIFY_BIGINT (current_block_ordinal (), buf))); + WARNOPT (WARN_ALONE_ZERO_BLOCK, + (0, 0, _("A lone zero block at %s"), + STRINGIFY_BIGINT (current_block_ordinal (), buf))); break; } status = prev_status; @@ -207,63 +308,23 @@ read_and (void (*do_something) (void)) void list_archive (void) { - /* Print the header block. */ + off_t block_ordinal = current_block_ordinal (); - decode_header (current_header, ¤t_stat_info, ¤t_format, 0); + /* Print the header block. */ if (verbose_option) - print_header (¤t_stat_info, -1); + print_header (¤t_stat_info, current_header, block_ordinal); - if (incremental_option && current_header->header.typeflag == GNUTYPE_DUMPDIR) + if (incremental_option) { - off_t size; - size_t written, check; - union block *data_block; - - set_next_block_after (current_header); - if (multi_volume_option) + if (verbose_option > 2) { - assign_string (&save_name, current_stat_info.file_name); - save_totsize = current_stat_info.stat.st_size; + if (is_dumpdir (¤t_stat_info)) + list_dumpdir (current_stat_info.dumpdir, + dumpdir_size (current_stat_info.dumpdir)); } - for (size = current_stat_info.stat.st_size; size > 0; size -= written) - { - if (multi_volume_option) - save_sizeleft = size; - data_block = find_next_block (); - if (!data_block) - { - ERROR ((0, 0, _("Unexpected EOF in archive"))); - break; /* FIXME: What happens, then? */ - } - written = available_space_after (data_block); - if (written > size) - written = size; - errno = 0; - check = fwrite (data_block->buffer, sizeof (char), written, stdlis); - set_next_block_after ((union block *) - (data_block->buffer + written - 1)); - if (check != written) - { - write_error_details (current_stat_info.file_name, check, written); - skip_file (size - written); - break; - } - } - if (multi_volume_option) - assign_string (&save_name, 0); - fputc ('\n', stdlis); - fflush (stdlis); - return; - } - if (multi_volume_option) - assign_string (&save_name, current_stat_info.file_name); - skip_member (); - - if (multi_volume_option) - assign_string (&save_name, 0); } /* Check header checksum */ @@ -283,9 +344,9 @@ tar_checksum (union block *header, bool silent) int unsigned_sum = 0; /* the POSIX one :-) */ int signed_sum = 0; /* the Sun one :-( */ int recorded_sum; - uintmax_t parsed_sum; + int parsed_sum; char *p; - + p = header->buffer; for (i = sizeof *header; i-- != 0;) { @@ -308,13 +369,12 @@ tar_checksum (union block *header, bool silent) parsed_sum = from_header (header->header.chksum, sizeof header->header.chksum, 0, - (uintmax_t) 0, - (uintmax_t) TYPE_MAXIMUM (int), silent); - if (parsed_sum == (uintmax_t) -1) + 0, INT_MAX, true, silent); + if (parsed_sum < 0) return HEADER_FAILURE; recorded_sum = parsed_sum; - + if (unsigned_sum != recorded_sum && signed_sum != recorded_sum) return HEADER_FAILURE; @@ -322,22 +382,30 @@ tar_checksum (union block *header, bool silent) } /* Read a block that's supposed to be a header block. Return its - address in "current_header", and if it is good, the file's size in - current_stat_info.stat.st_size. + address in *RETURN_BLOCK, and if it is good, the file's size + and names (file name, link name) in *INFO. + + Return one of enum read_header describing the status of the + operation. - Return 1 for success, 0 if the checksum is bad, EOF on eof, 2 for a - block full of zeros (EOF marker). + The MODE parameter instructs read_header what to do with special + header blocks, i.e.: extended POSIX, GNU long name or long link, + etc.: - If RAW_EXTENDED_HEADERS is nonzero, do not automagically fold the - GNU long name and link headers into later headers. + read_header_auto process them automatically, + read_header_x_raw when a special header is read, return + HEADER_SUCCESS_EXTENDED without actually + processing the header, + read_header_x_global when a POSIX global header is read, + decode it and return HEADER_SUCCESS_EXTENDED. - You must always set_next_block_after(current_header) to skip past + You must always set_next_block_after(*return_block) to skip past the header which this routine reads. */ enum read_header -read_header (bool raw_extended_headers) +read_header (union block **return_block, struct tar_stat_info *info, + enum read_header_mode mode) { - char *p; union block *header; union block *header_copy; char *bp; @@ -345,15 +413,15 @@ read_header (bool raw_extended_headers) size_t size, written; union block *next_long_name = 0; union block *next_long_link = 0; - size_t next_long_name_blocks; - size_t next_long_link_blocks; + size_t next_long_name_blocks = 0; + size_t next_long_link_blocks = 0; while (1) { enum read_header status; - + header = find_next_block (); - current_header = header; + *return_block = header; if (!header) return HEADER_END_OF_FILE; @@ -363,39 +431,45 @@ read_header (bool raw_extended_headers) /* Good block. Decode file size and return. */ if (header->header.typeflag == LNKTYPE) - current_stat_info.stat.st_size = 0; /* links 0 size on tape */ + info->stat.st_size = 0; /* links 0 size on tape */ else - current_stat_info.stat.st_size = OFF_FROM_HEADER (header->header.size); + { + info->stat.st_size = OFF_FROM_HEADER (header->header.size); + if (info->stat.st_size < 0) + return HEADER_FAILURE; + } if (header->header.typeflag == GNUTYPE_LONGNAME || header->header.typeflag == GNUTYPE_LONGLINK || header->header.typeflag == XHDTYPE - || header->header.typeflag == XGLTYPE) + || header->header.typeflag == XGLTYPE + || header->header.typeflag == SOLARIS_XHDTYPE) { - if (raw_extended_headers) + if (mode == read_header_x_raw) return HEADER_SUCCESS_EXTENDED; else if (header->header.typeflag == GNUTYPE_LONGNAME || header->header.typeflag == GNUTYPE_LONGLINK) { - size_t name_size = current_stat_info.stat.st_size; - size = name_size - name_size % BLOCKSIZE + 2 * BLOCKSIZE; - if (name_size != current_stat_info.stat.st_size - || size < name_size) + size_t name_size = info->stat.st_size; + size_t n = name_size % BLOCKSIZE; + size = name_size + BLOCKSIZE; + if (n) + size += BLOCKSIZE - n; + + if (name_size != info->stat.st_size || size < name_size) xalloc_die (); header_copy = xmalloc (size + 1); if (header->header.typeflag == GNUTYPE_LONGNAME) { - if (next_long_name) - free (next_long_name); + free (next_long_name); next_long_name = header_copy; next_long_name_blocks = size / BLOCKSIZE; } else { - if (next_long_link) - free (next_long_link); + free (next_long_link); next_long_link = header_copy; next_long_link_blocks = size / BLOCKSIZE; } @@ -424,12 +498,25 @@ read_header (bool raw_extended_headers) *bp = '\0'; } - else if (header->header.typeflag == XHDTYPE) - xheader_read (header, OFF_FROM_HEADER (header->header.size)); + else if (header->header.typeflag == XHDTYPE + || header->header.typeflag == SOLARIS_XHDTYPE) + xheader_read (&info->xhdr, header, + OFF_FROM_HEADER (header->header.size)); else if (header->header.typeflag == XGLTYPE) { - xheader_read (header, OFF_FROM_HEADER (header->header.size)); - xheader_decode_global (); + struct xheader xhdr; + + if (!recent_global_header) + recent_global_header = xmalloc (sizeof *recent_global_header); + memcpy (recent_global_header, header, + sizeof *recent_global_header); + memset (&xhdr, 0, sizeof xhdr); + xheader_read (&xhdr, header, + OFF_FROM_HEADER (header->header.size)); + xheader_decode_global (&xhdr); + xheader_destroy (&xhdr); + if (mode == read_header_x_global) + return HEADER_SUCCESS_EXTENDED; } /* Loop! */ @@ -438,11 +525,10 @@ read_header (bool raw_extended_headers) else { char const *name; - struct posix_header const *h = ¤t_header->header; + struct posix_header const *h = &header->header; char namebuf[sizeof h->prefix + 1 + NAME_FIELD_SIZE + 1]; - if (recent_long_name) - free (recent_long_name); + free (recent_long_name); if (next_long_name) { @@ -462,11 +548,6 @@ read_header (bool raw_extended_headers) np[sizeof h->prefix] = '\0'; np += strlen (np); *np++ = '/'; - - /* Prevent later references to current_header from - mistakenly treating this as an old GNU header. - This assignment invalidates h->prefix. */ - current_header->oldgnu_header.isextended = 0; } memcpy (np, h->name, sizeof h->name); np[sizeof h->name] = '\0'; @@ -474,12 +555,11 @@ read_header (bool raw_extended_headers) recent_long_name = 0; recent_long_name_blocks = 0; } - assign_string (¤t_stat_info.orig_file_name, name); - assign_string (¤t_stat_info.file_name, name); - current_stat_info.had_trailing_slash = strip_trailing_slashes (current_stat_info.file_name); + assign_string (&info->orig_file_name, name); + assign_string (&info->file_name, name); + info->had_trailing_slash = strip_trailing_slashes (info->file_name); - if (recent_long_link) - free (recent_long_link); + free (recent_long_link); if (next_long_link) { @@ -495,7 +575,7 @@ read_header (bool raw_extended_headers) recent_long_link = 0; recent_long_link_blocks = 0; } - assign_string (¤t_stat_info.link_name, name); + assign_string (&info->link_name, name); return HEADER_SUCCESS; } @@ -522,6 +602,8 @@ decode_header (union block *header, struct tar_stat_info *stat_info, enum archive_format *format_pointer, int do_user_group) { enum archive_format format; + bool hbits; + mode_t mode = MODE_FROM_HEADER (header->header.mode, &hbits); if (strcmp (header->header.magic, TMAGIC) == 0) { @@ -531,34 +613,43 @@ decode_header (union block *header, struct tar_stat_info *stat_info, && ISOCTAL (header->star_header.ctime[0]) && header->star_header.ctime[11] == ' ') format = STAR_FORMAT; - else if (extended_header.size) + else if (stat_info->xhdr.size) format = POSIX_FORMAT; else format = USTAR_FORMAT; } - else if (strcmp (header->header.magic, OLDGNU_MAGIC) == 0) - format = OLDGNU_FORMAT; + else if (strcmp (header->buffer + offsetof (struct posix_header, magic), + OLDGNU_MAGIC) + == 0) + format = hbits ? OLDGNU_FORMAT : GNU_FORMAT; else format = V7_FORMAT; *format_pointer = format; - stat_info->stat.st_mode = MODE_FROM_HEADER (header->header.mode); - stat_info->stat.st_mtime = TIME_FROM_HEADER (header->header.mtime); + stat_info->stat.st_mode = mode; + stat_info->mtime.tv_sec = TIME_FROM_HEADER (header->header.mtime); + stat_info->mtime.tv_nsec = 0; assign_string (&stat_info->uname, header->header.uname[0] ? header->header.uname : NULL); assign_string (&stat_info->gname, header->header.gname[0] ? header->header.gname : NULL); - stat_info->devmajor = MAJOR_FROM_HEADER (header->header.devmajor); - stat_info->devminor = MINOR_FROM_HEADER (header->header.devminor); - stat_info->stat.st_atime = start_time; - stat_info->stat.st_ctime = start_time; + xheader_xattr_init (stat_info); if (format == OLDGNU_FORMAT && incremental_option) { - stat_info->stat.st_atime = TIME_FROM_HEADER (header->oldgnu_header.atime); - stat_info->stat.st_ctime = TIME_FROM_HEADER (header->oldgnu_header.ctime); + stat_info->atime.tv_sec = TIME_FROM_HEADER (header->oldgnu_header.atime); + stat_info->ctime.tv_sec = TIME_FROM_HEADER (header->oldgnu_header.ctime); + stat_info->atime.tv_nsec = stat_info->ctime.tv_nsec = 0; + } + else if (format == STAR_FORMAT) + { + stat_info->atime.tv_sec = TIME_FROM_HEADER (header->star_header.atime); + stat_info->ctime.tv_sec = TIME_FROM_HEADER (header->star_header.ctime); + stat_info->atime.tv_nsec = stat_info->ctime.tv_nsec = 0; } + else + stat_info->atime = stat_info->ctime = start_time; if (format == V7_FORMAT) { @@ -568,13 +659,6 @@ decode_header (union block *header, struct tar_stat_info *stat_info, } else { - - if (format == STAR_FORMAT) - { - stat_info->stat.st_atime = TIME_FROM_HEADER (header->star_header.atime); - stat_info->stat.st_ctime = TIME_FROM_HEADER (header->star_header.ctime); - } - if (do_user_group) { /* FIXME: Decide if this should somewhat depend on -p. */ @@ -594,8 +678,9 @@ decode_header (union block *header, struct tar_stat_info *stat_info, { case BLKTYPE: case CHRTYPE: - stat_info->stat.st_rdev = makedev (stat_info->devmajor, - stat_info->devminor); + stat_info->stat.st_rdev = + makedev (MAJOR_FROM_HEADER (header->header.devmajor), + MINOR_FROM_HEADER (header->header.devminor)); break; default: @@ -612,23 +697,42 @@ decode_header (union block *header, struct tar_stat_info *stat_info, stat_info->is_sparse = true; } else - stat_info->is_sparse = false; + { + stat_info->is_sparse = false; + if (((current_format == GNU_FORMAT + || current_format == OLDGNU_FORMAT) + && current_header->header.typeflag == GNUTYPE_DUMPDIR) + || stat_info->dumpdir) + stat_info->is_dumpdir = true; + } } + /* Convert buffer at WHERE0 of size DIGS from external format to - uintmax_t. The data is of type TYPE. The buffer must represent a - value in the range -MINUS_MINVAL through MAXVAL. DIGS must be - positive. SILENT=true inhibits printing diagnostic messages. - Return -1 on error, diagnosing the error if TYPE is - nonzero. */ -static uintmax_t + intmax_t. DIGS must be positive. If TYPE is nonnull, the data are + of type TYPE. The buffer must represent a value in the range + MINVAL through MAXVAL; if the mathematically correct result V would + be greater than INTMAX_MAX, return a negative integer V such that + (uintmax_t) V yields the correct result. If OCTAL_ONLY, allow only octal + numbers instead of the other GNU extensions. Return -1 on error, + diagnosing the error if TYPE is nonnull and if !SILENT. */ +#if ! (INTMAX_MAX <= UINTMAX_MAX && - (INTMAX_MIN + 1) <= UINTMAX_MAX) +# error "from_header internally represents intmax_t as uintmax_t + sign" +#endif +#if ! (UINTMAX_MAX / 2 <= INTMAX_MAX) +# error "from_header returns intmax_t to represent uintmax_t" +#endif +static intmax_t from_header (char const *where0, size_t digs, char const *type, - uintmax_t minus_minval, uintmax_t maxval, bool silent) + intmax_t minval, uintmax_t maxval, + bool octal_only, bool silent) { uintmax_t value; + uintmax_t uminval = minval; + uintmax_t minus_minval = - uminval; char const *where = where0; char const *lim = where + digs; - int negative = 0; + bool negative = false; /* Accommodate buggy tar of unknown vintage, which outputs leading NUL if the previous field overflows. */ @@ -641,11 +745,13 @@ from_header (char const *where0, size_t digs, char const *type, { if (type && !silent) ERROR ((0, 0, + /* TRANSLATORS: %s is type of the value (gid_t, uid_t, + etc.) */ _("Blanks in header where numeric %s value expected"), type)); return -1; } - if (!ISSPACE ((unsigned char) *where)) + if (!isspace ((unsigned char) *where)) break; where++; } @@ -654,14 +760,14 @@ from_header (char const *where0, size_t digs, char const *type, if (ISODIGIT (*where)) { char const *where1 = where; - uintmax_t overflow = 0; + bool overflow = false; for (;;) { value += *where++ - '0'; if (where == lim || ! ISODIGIT (*where)) break; - overflow |= value ^ (value << LG_8 >> LG_8); + overflow |= value != (value << LG_8 >> LG_8); value <<= LG_8; } @@ -685,7 +791,7 @@ from_header (char const *where0, size_t digs, char const *type, if (where == lim || ! ISODIGIT (*where)) break; digit = *where - '0'; - overflow |= value ^ (value << LG_8 >> LG_8); + overflow |= value != (value << LG_8 >> LG_8); value <<= LG_8; } value++; @@ -695,9 +801,10 @@ from_header (char const *where0, size_t digs, char const *type, { if (!silent) WARN ((0, 0, + /* TRANSLATORS: Second %s is a type name (gid_t,uid_t,etc.) */ _("Archive octal value %.*s is out of %s range; assuming two's complement"), (int) (where - where1), where1, type)); - negative = 1; + negative = true; } } @@ -705,24 +812,30 @@ from_header (char const *where0, size_t digs, char const *type, { if (type && !silent) ERROR ((0, 0, + /* TRANSLATORS: Second %s is a type name (gid_t,uid_t,etc.) */ _("Archive octal value %.*s is out of %s range"), (int) (where - where1), where1, type)); return -1; } } + else if (octal_only) + { + /* Suppress the following extensions. */ + } else if (*where == '-' || *where == '+') { /* Parse base-64 output produced only by tar test versions 1.13.6 (1999-08-11) through 1.13.11 (1999-08-23). Support for this will be withdrawn in future releases. */ int dig; - static int warned_once; - if (! warned_once) + if (!silent) { - warned_once = 1; - if (!silent) - WARN ((0, 0, - _("Archive contains obsolescent base-64 headers"))); + static bool warned_once; + if (! warned_once) + { + warned_once = true; + WARN ((0, 0, _("Archive contains obsolescent base-64 headers"))); + } } negative = *where++ == '-'; while (where != lim @@ -771,12 +884,12 @@ from_header (char const *where0, size_t digs, char const *type, return -1; } } - negative = signbit; + negative = signbit != 0; if (negative) value = -value; } - if (where != lim && *where && !ISSPACE ((unsigned char) *where)) + if (where != lim && *where && !isspace ((unsigned char) *where)) { if (type) { @@ -791,9 +904,10 @@ from_header (char const *where0, size_t digs, char const *type, while (where0 != lim && ! lim[-1]) lim--; - quotearg_buffer (buf, sizeof buf, where0, lim - where, o); + quotearg_buffer (buf, sizeof buf, where0, lim - where0, o); if (!silent) ERROR ((0, 0, + /* TRANSLATORS: Second %s is a type name (gid_t,uid_t,etc.) */ _("Archive contains %.*s where numeric %s value expected"), (int) sizeof buf, buf, type)); } @@ -802,7 +916,7 @@ from_header (char const *where0, size_t digs, char const *type, } if (value <= (negative ? minus_minval : maxval)) - return negative ? -value : value; + return represent_uintmax (negative ? -value : value); if (type && !silent) { @@ -815,7 +929,8 @@ from_header (char const *where0, size_t digs, char const *type, *--value_string = '-'; if (minus_minval) *--minval_string = '-'; - ERROR ((0, 0, _("Archive value %s is out of %s range %s.%s"), + /* TRANSLATORS: Second %s is type name (gid_t,uid_t,etc.) */ + ERROR ((0, 0, _("Archive value %s is out of %s range %s..%s"), value_string, type, minval_string, STRINGIFY_BIGINT (maxval, maxval_buf))); } @@ -823,50 +938,52 @@ from_header (char const *where0, size_t digs, char const *type, return -1; } -gid_t +static gid_t gid_from_header (const char *p, size_t s) { return from_header (p, s, "gid_t", - - (uintmax_t) TYPE_MINIMUM (gid_t), - (uintmax_t) TYPE_MAXIMUM (gid_t), - false); + TYPE_MINIMUM (gid_t), TYPE_MAXIMUM (gid_t), + false, false); } -major_t +static major_t major_from_header (const char *p, size_t s) { return from_header (p, s, "major_t", - - (uintmax_t) TYPE_MINIMUM (major_t), - (uintmax_t) TYPE_MAXIMUM (major_t), false); + TYPE_MINIMUM (major_t), TYPE_MAXIMUM (major_t), + false, false); } -minor_t +static minor_t minor_from_header (const char *p, size_t s) { return from_header (p, s, "minor_t", - - (uintmax_t) TYPE_MINIMUM (minor_t), - (uintmax_t) TYPE_MAXIMUM (minor_t), false); + TYPE_MINIMUM (minor_t), TYPE_MAXIMUM (minor_t), + false, false); } -mode_t -mode_from_header (const char *p, size_t s) +/* Convert P to the file mode, as understood by tar. + Set *HBITS if there are any unrecognized bits. */ +static mode_t +mode_from_header (const char *p, size_t s, bool *hbits) { - /* Do not complain about unrecognized mode bits. */ - unsigned u = from_header (p, s, "mode_t", - - (uintmax_t) TYPE_MINIMUM (mode_t), - TYPE_MAXIMUM (uintmax_t), false); - return ((u & TSUID ? S_ISUID : 0) - | (u & TSGID ? S_ISGID : 0) - | (u & TSVTX ? S_ISVTX : 0) - | (u & TUREAD ? S_IRUSR : 0) - | (u & TUWRITE ? S_IWUSR : 0) - | (u & TUEXEC ? S_IXUSR : 0) - | (u & TGREAD ? S_IRGRP : 0) - | (u & TGWRITE ? S_IWGRP : 0) - | (u & TGEXEC ? S_IXGRP : 0) - | (u & TOREAD ? S_IROTH : 0) - | (u & TOWRITE ? S_IWOTH : 0) - | (u & TOEXEC ? S_IXOTH : 0)); + intmax_t u = from_header (p, s, "mode_t", + INTMAX_MIN, UINTMAX_MAX, + false, false); + mode_t mode = ((u & TSUID ? S_ISUID : 0) + | (u & TSGID ? S_ISGID : 0) + | (u & TSVTX ? S_ISVTX : 0) + | (u & TUREAD ? S_IRUSR : 0) + | (u & TUWRITE ? S_IWUSR : 0) + | (u & TUEXEC ? S_IXUSR : 0) + | (u & TGREAD ? S_IRGRP : 0) + | (u & TGWRITE ? S_IWGRP : 0) + | (u & TGEXEC ? S_IXGRP : 0) + | (u & TOREAD ? S_IROTH : 0) + | (u & TOWRITE ? S_IWOTH : 0) + | (u & TOEXEC ? S_IXOTH : 0)); + *hbits = (u & ~07777) != 0; + return mode; } off_t @@ -874,96 +991,89 @@ off_from_header (const char *p, size_t s) { /* Negative offsets are not allowed in tar files, so invoke from_header with minimum value 0, not TYPE_MINIMUM (off_t). */ - return from_header (p, s, "off_t", (uintmax_t) 0, - (uintmax_t) TYPE_MAXIMUM (off_t), false); + return from_header (p, s, "off_t", + 0, TYPE_MAXIMUM (off_t), + false, false); } -size_t -size_from_header (const char *p, size_t s) -{ - return from_header (p, s, "size_t", (uintmax_t) 0, - (uintmax_t) TYPE_MAXIMUM (size_t), false); -} - -time_t +static time_t time_from_header (const char *p, size_t s) { return from_header (p, s, "time_t", - - (uintmax_t) TYPE_MINIMUM (time_t), - (uintmax_t) TYPE_MAXIMUM (time_t), false); + TYPE_MINIMUM (time_t), TYPE_MAXIMUM (time_t), + false, false); } -uid_t +static uid_t uid_from_header (const char *p, size_t s) { return from_header (p, s, "uid_t", - - (uintmax_t) TYPE_MINIMUM (uid_t), - (uintmax_t) TYPE_MAXIMUM (uid_t), false); + TYPE_MINIMUM (uid_t), TYPE_MAXIMUM (uid_t), + false, false); } uintmax_t uintmax_from_header (const char *p, size_t s) { - return from_header (p, s, "uintmax_t", (uintmax_t) 0, - TYPE_MAXIMUM (uintmax_t), false); + return from_header (p, s, "uintmax_t", 0, UINTMAX_MAX, false, false); } -/* Format O as a null-terminated decimal string into BUF _backwards_; - return pointer to start of result. */ -char * -stringify_uintmax_t_backwards (uintmax_t o, char *buf) -{ - *--buf = '\0'; - do - *--buf = '0' + (int) (o % 10); - while ((o /= 10) != 0); - return buf; -} - /* Return a printable representation of T. The result points to static storage that can be reused in the next call to this - function, to ctime, or to asctime. */ + function, to ctime, or to asctime. If FULL_TIME, then output the + time stamp to its full resolution; otherwise, just output it to + 1-minute resolution. */ char const * -tartime (time_t t) +tartime (struct timespec t, bool full_time) { + enum { fraclen = sizeof ".FFFFFFFFF" - 1 }; static char buffer[max (UINTMAX_STRSIZE_BOUND + 1, - INT_STRLEN_BOUND (int) + 16)]; + INT_STRLEN_BOUND (int) + 16) + + fraclen]; + struct tm *tm; + time_t s = t.tv_sec; + int ns = t.tv_nsec; + bool negative = s < 0; char *p; -#if USE_OLD_CTIME - p = ctime (&t); - if (p) + if (negative && ns != 0) { - char const *time_stamp = p + 4; - for (p += 16; p[3] != '\n'; p++) - p[0] = p[3]; - p[0] = '\0'; - return time_stamp; + s++; + ns = 1000000000 - ns; } -#else - /* Use ISO 8610 format. See: - http://www.cl.cam.ac.uk/~mgk25/iso-time.html */ - struct tm *tm = utc_option ? gmtime (&t) : localtime (&t); + + tm = utc_option ? gmtime (&s) : localtime (&s); if (tm) { - sprintf (buffer, "%04ld-%02d-%02d %02d:%02d:%02d", - tm->tm_year + 1900L, tm->tm_mon + 1, tm->tm_mday, - tm->tm_hour, tm->tm_min, tm->tm_sec); + if (full_time) + { + sprintf (buffer, "%04ld-%02d-%02d %02d:%02d:%02d", + tm->tm_year + 1900L, tm->tm_mon + 1, tm->tm_mday, + tm->tm_hour, tm->tm_min, tm->tm_sec); + code_ns_fraction (ns, buffer + strlen (buffer)); + } + else + sprintf (buffer, "%04ld-%02d-%02d %02d:%02d", + tm->tm_year + 1900L, tm->tm_mon + 1, tm->tm_mday, + tm->tm_hour, tm->tm_min); return buffer; } -#endif /* The time stamp cannot be broken down, most likely because it is out of range. Convert it as an integer, right-adjusted in a field with the same width as the usual - 19-byte 4-year ISO time format. */ - p = stringify_uintmax_t_backwards (t < 0 ? - (uintmax_t) t : (uintmax_t) t, - buffer + sizeof buffer); - if (t < 0) + 4-year ISO time format. */ + p = umaxtostr (negative ? - (uintmax_t) s : s, + buffer + sizeof buffer - UINTMAX_STRSIZE_BOUND - fraclen); + if (negative) *--p = '-'; - while (buffer + sizeof buffer - 19 - 1 < p) + while ((buffer + sizeof buffer - sizeof "YYYY-MM-DD HH:MM" + + (full_time ? sizeof ":SS.FFFFFFFFF" - 1 : 0)) + < p) *--p = ' '; + if (full_time) + code_ns_fraction (ns, buffer + sizeof buffer - 1 - fraclen); return p; } @@ -977,35 +1087,43 @@ tartime (time_t t) they shouldn't. Unix tar is pretty random here anyway. */ -/* FIXME: Note that print_header uses the globals HEAD, HSTAT, and - HEAD_STANDARD, which must be set up in advance. Not very clean.. */ +/* Width of "user/group size", with initial value chosen + heuristically. This grows as needed, though this may cause some + stairstepping in the output. Make it too small and the output will + almost always look ragged. Make it too large and the output will + be spaced out too far. */ +static int ugswidth = 19; -/* UGSWIDTH starts with 18, so with user and group names <= 8 chars, the - columns never shift during the listing. */ -#define UGSWIDTH 18 -static int ugswidth = UGSWIDTH; /* maximum width encountered so far */ +/* Width of printed time stamps. It grows if longer time stamps are + found (typically, those with nanosecond resolution). Like + USGWIDTH, some stairstepping may occur. */ +static int datewidth = sizeof "YYYY-MM-DD HH:MM" - 1; -/* DATEWIDTH is the number of columns taken by the date and time fields. */ -#if USE_OLD_CDATE -# define DATEWIDTH 19 -#else -# define DATEWIDTH 18 -#endif +static bool volume_label_printed = false; -void -print_header (struct tar_stat_info *st, off_t block_ordinal) +static void +simple_print_header (struct tar_stat_info *st, union block *blk, + off_t block_ordinal) { - char modes[11]; + char modes[12]; char const *time_stamp; - char *temp_name = st->orig_file_name ? st->orig_file_name : st->file_name; + int time_stamp_len; + char *temp_name; /* These hold formatted ints. */ - char uform[UINTMAX_STRSIZE_BOUND], gform[UINTMAX_STRSIZE_BOUND]; + char uform[max (INT_BUFSIZE_BOUND (intmax_t), UINTMAX_STRSIZE_BOUND)]; + char gform[sizeof uform]; char *user, *group; char size[2 * UINTMAX_STRSIZE_BOUND]; /* holds formatted size or major,minor */ char uintbuf[UINTMAX_STRSIZE_BOUND]; int pad; + int sizelen; + + if (show_transformed_names_option) + temp_name = st->file_name ? st->file_name : st->orig_file_name; + else + temp_name = st->orig_file_name ? st->orig_file_name : st->file_name; if (block_number_option) { @@ -1021,16 +1139,20 @@ print_header (struct tar_stat_info *st, off_t block_ordinal) if (verbose_option <= 1) { /* Just the fax, mam. */ - fprintf (stdlis, "%s\n", quotearg (temp_name)); + fputs (quotearg (temp_name), stdlis); + if (show_transformed_names_option && st->had_trailing_slash) + fputc ('/', stdlis); + fputc ('\n', stdlis); } else { /* File type and modes. */ modes[0] = '?'; - switch (current_header->header.typeflag) + switch (blk->header.typeflag) { case GNUTYPE_VOLHDR: + volume_label_printed = true; modes[0] = 'V'; break; @@ -1038,22 +1160,16 @@ print_header (struct tar_stat_info *st, off_t block_ordinal) modes[0] = 'M'; break; - case GNUTYPE_NAMES: - modes[0] = 'N'; - break; - case GNUTYPE_LONGNAME: case GNUTYPE_LONGLINK: modes[0] = 'L'; - ERROR ((0, 0, _("Visible longname error"))); + ERROR ((0, 0, _("Unexpected long name header"))); break; case GNUTYPE_SPARSE: case REGTYPE: case AREGTYPE: - modes[0] = '-'; - if (temp_name[strlen (temp_name) - 1] == '/') - modes[0] = 'd'; + modes[0] = st->had_trailing_slash ? 'd' : '-'; break; case LNKTYPE: modes[0] = 'h'; @@ -1081,15 +1197,23 @@ print_header (struct tar_stat_info *st, off_t block_ordinal) break; } - decode_mode (st->stat.st_mode, modes + 1); + pax_decode_mode (st->stat.st_mode, modes + 1); + + /* extended attributes: GNU `ls -l'-like preview */ + xattrs_print_char (st, modes + 10); /* Time stamp. */ - time_stamp = tartime (st->stat.st_mtime); + time_stamp = tartime (st->mtime, full_time_option); + time_stamp_len = strlen (time_stamp); + if (datewidth < time_stamp_len) + datewidth = time_stamp_len; /* User and group names. */ - if (st->uname && current_format != V7_FORMAT + if (st->uname + && st->uname[0] + && current_format != V7_FORMAT && !numeric_owner_option) user = st->uname; else @@ -1097,22 +1221,18 @@ print_header (struct tar_stat_info *st, off_t block_ordinal) /* Try parsing it as an unsigned integer first, and as a uid_t if that fails. This method can list positive user ids that are too large to fit in a uid_t. */ - uintmax_t u = from_header (current_header->header.uid, - sizeof current_header->header.uid, 0, - (uintmax_t) 0, - (uintmax_t) TYPE_MAXIMUM (uintmax_t), - false); - if (u != -1) - user = STRINGIFY_BIGINT (u, uform); - else - { - sprintf (uform, "%ld", - (long) UID_FROM_HEADER (current_header->header.uid)); - user = uform; - } + uintmax_t u = from_header (blk->header.uid, + sizeof blk->header.uid, 0, + 0, UINTMAX_MAX, + false, false); + user = (u != -1 + ? STRINGIFY_BIGINT (u, uform) + : imaxtostr (UID_FROM_HEADER (blk->header.uid), uform)); } - if (st->gname && current_format != V7_FORMAT + if (st->gname + && st->gname[0] + && current_format != V7_FORMAT && !numeric_owner_option) group = st->gname; else @@ -1120,24 +1240,18 @@ print_header (struct tar_stat_info *st, off_t block_ordinal) /* Try parsing it as an unsigned integer first, and as a gid_t if that fails. This method can list positive group ids that are too large to fit in a gid_t. */ - uintmax_t g = from_header (current_header->header.gid, - sizeof current_header->header.gid, 0, - (uintmax_t) 0, - (uintmax_t) TYPE_MAXIMUM (uintmax_t), - false); - if (g != -1) - group = STRINGIFY_BIGINT (g, gform); - else - { - sprintf (gform, "%ld", - (long) GID_FROM_HEADER (current_header->header.gid)); - group = gform; - } + uintmax_t g = from_header (blk->header.gid, + sizeof blk->header.gid, 0, + 0, UINTMAX_MAX, + false, false); + group = (g != -1 + ? STRINGIFY_BIGINT (g, gform) + : imaxtostr (GID_FROM_HEADER (blk->header.gid), gform)); } /* Format the file size or major/minor device numbers. */ - switch (current_header->header.typeflag) + switch (blk->header.typeflag) { case CHRTYPE: case BLKTYPE: @@ -1156,16 +1270,20 @@ print_header (struct tar_stat_info *st, off_t block_ordinal) /* Figure out padding and print the whole line. */ - pad = strlen (user) + strlen (group) + strlen (size) + 1; + sizelen = strlen (size); + pad = strlen (user) + 1 + strlen (group) + 1 + sizelen; if (pad > ugswidth) ugswidth = pad; - fprintf (stdlis, "%s %s/%s %*s%s %s", - modes, user, group, ugswidth - pad, "", size, time_stamp); + fprintf (stdlis, "%s %s/%s %*s %-*s", + modes, user, group, ugswidth - pad + sizelen, size, + datewidth, time_stamp); fprintf (stdlis, " %s", quotearg (temp_name)); + if (show_transformed_names_option && st->had_trailing_slash) + fputc ('/', stdlis); - switch (current_header->header.typeflag) + switch (blk->header.typeflag) { case SYMTYPE: fprintf (stdlis, " -> %s\n", quotearg (st->link_name)); @@ -1178,7 +1296,7 @@ print_header (struct tar_stat_info *st, off_t block_ordinal) default: { char type_string[2]; - type_string[0] = current_header->header.typeflag; + type_string[0] = blk->header.typeflag; type_string[1] = '\0'; fprintf (stdlis, _(" unknown file type %s\n"), quote (type_string)); @@ -1212,17 +1330,48 @@ print_header (struct tar_stat_info *st, off_t block_ordinal) case GNUTYPE_MULTIVOL: strcpy (size, STRINGIFY_BIGINT - (UINTMAX_FROM_HEADER (current_header->oldgnu_header.offset), + (UINTMAX_FROM_HEADER (blk->oldgnu_header.offset), uintbuf)); fprintf (stdlis, _("--Continued at byte %s--\n"), size); break; - - case GNUTYPE_NAMES: - fprintf (stdlis, _("--Mangled file names--\n")); - break; } } fflush (stdlis); + xattrs_print (st); +} + + +static void +print_volume_label (void) +{ + struct tar_stat_info vstat; + union block vblk; + enum archive_format dummy; + + memset (&vblk, 0, sizeof (vblk)); + vblk.header.typeflag = GNUTYPE_VOLHDR; + if (recent_global_header) + memcpy (vblk.header.mtime, recent_global_header->header.mtime, + sizeof vblk.header.mtime); + tar_stat_init (&vstat); + assign_string (&vstat.file_name, "."); + decode_header (&vblk, &vstat, &dummy, 0); + assign_string (&vstat.file_name, volume_label); + simple_print_header (&vstat, &vblk, 0); + tar_stat_destroy (&vstat); +} + +void +print_header (struct tar_stat_info *st, union block *blk, + off_t block_ordinal) +{ + if (current_format == POSIX_FORMAT && !volume_label_printed && volume_label) + { + print_volume_label (); + volume_label_printed = true; + } + + simple_print_header (st, blk, block_ordinal); } /* Print a similar line when we make a directory automatically. */ @@ -1236,7 +1385,7 @@ print_for_mkdir (char *dirname, int length, mode_t mode) /* File type and modes. */ modes[0] = 'd'; - decode_mode (mode, modes + 1); + pax_decode_mode (mode, modes + 1); if (block_number_option) { @@ -1245,8 +1394,8 @@ print_for_mkdir (char *dirname, int length, mode_t mode) STRINGIFY_BIGINT (current_block_ordinal (), buf)); } - fprintf (stdlis, "%s %*s %.*s\n", modes, ugswidth + DATEWIDTH, - _("Creating directory:"), length, quotearg (dirname)); + fprintf (stdlis, "%s %*s %s\n", modes, ugswidth + 1 + datewidth, + _("Creating directory:"), quotearg (dirname)); } } @@ -1256,25 +1405,19 @@ skip_file (off_t size) { union block *x; - if (multi_volume_option) - { - save_totsize = size; - save_sizeleft = size; - } + /* FIXME: Make sure mv_begin_read is always called before it */ if (seekable_archive) { off_t nblk = seek_archive (size); if (nblk >= 0) - { - size -= nblk * BLOCKSIZE; - if (multi_volume_option) /* Argh.. */ - save_sizeleft -= nblk * BLOCKSIZE; - } + size -= nblk * BLOCKSIZE; else seekable_archive = false; } - + + mv_size_left (size); + while (size > 0) { x = find_next_block (); @@ -1283,8 +1426,7 @@ skip_file (off_t size) set_next_block_after (x); size -= BLOCKSIZE; - if (multi_volume_option) - save_sizeleft -= BLOCKSIZE; + mv_size_left (size); } } @@ -1293,13 +1435,49 @@ skip_file (off_t size) void skip_member (void) { - char save_typeflag = current_header->header.typeflag; - set_next_block_after (current_header); - - assign_string (&save_name, current_stat_info.file_name); + if (!current_stat_info.skipped) + { + char save_typeflag = current_header->header.typeflag; + set_next_block_after (current_header); - if (current_stat_info.is_sparse) - sparse_skip_file (¤t_stat_info); - else if (save_typeflag != DIRTYPE) - skip_file (current_stat_info.stat.st_size); + mv_begin_read (¤t_stat_info); + + if (current_stat_info.is_sparse) + sparse_skip_file (¤t_stat_info); + else if (save_typeflag != DIRTYPE) + skip_file (current_stat_info.stat.st_size); + + mv_end (); + } +} + +void +test_archive_label (void) +{ + base64_init (); + name_gather (); + + open_archive (ACCESS_READ); + if (read_header (¤t_header, ¤t_stat_info, read_header_auto) + == HEADER_SUCCESS) + { + decode_header (current_header, + ¤t_stat_info, ¤t_format, 0); + if (current_header->header.typeflag == GNUTYPE_VOLHDR) + assign_string (&volume_label, current_header->header.name); + + if (volume_label) + { + if (verbose_option) + print_volume_label (); + if (!name_match (volume_label) && multi_volume_option) + { + char *s = drop_volume_label_suffix (volume_label); + name_match (s); + free (s); + } + } + } + close_archive (); + label_notfound (); }