X-Git-Url: https://git.dogcows.com/gitweb?a=blobdiff_plain;f=src%2Flist.c;h=e88d53b711f6390ef86d8e7b8422bbd5a451fef0;hb=9ded0a0567b58bfbdf10a8dc98e0158e486466ac;hp=cb3862bc1def4fba9600ccbf66c5b26811d8ebdf;hpb=c18999bad51b56995a293cb4d6936315b5a6bdff;p=chaz%2Ftar diff --git a/src/list.c b/src/list.c index cb3862b..e88d53b 100644 --- a/src/list.c +++ b/src/list.c @@ -1,5 +1,8 @@ /* List a tar archive, with support routines for reading a tar archive. - Copyright 1988,92,93,94,96,97,98,1999 Free Software Foundation, Inc. + + Copyright 1988, 1992, 1993, 1994, 1996, 1997, 1998, 1999, 2000, + 2001 Free Software Foundation, Inc. + Written by John Gilmore, on 1985-08-26. This program is free software; you can redistribute it and/or modify it @@ -22,11 +25,6 @@ #include "system.h" #include -#include -#ifndef time -time_t time (); -#endif - #include "common.h" #define max(a, b) ((a) < (b) ? (b) : (a)) @@ -34,6 +32,10 @@ time_t time (); union block *current_header; /* points to current archive header */ struct stat current_stat; /* stat struct corresponding */ 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 PARAMS ((const char *, size_t, const char *, uintmax_t, uintmax_t)); @@ -61,10 +63,7 @@ base64_init (void) base64_map[(int) base_64_digits[i]] = i; } -/*-----------------------------------. -| Main loop for reading an archive. | -`-----------------------------------*/ - +/* Main loop for reading an archive. */ void read_and (void (*do_something) ()) { @@ -78,7 +77,7 @@ read_and (void (*do_something) ()) while (1) { prev_status = status; - status = read_header (); + status = read_header (0); switch (status) { case HEADER_STILL_UNREAD: @@ -98,8 +97,7 @@ read_and (void (*do_something) ()) < newer_mtime_option)) || excluded_name (current_file_name)) { - char save_typeflag = current_header->header.typeflag; - switch (save_typeflag) + switch (current_header->header.typeflag) { case GNUTYPE_VOLHDR: case GNUTYPE_MULTIVOL: @@ -112,22 +110,11 @@ read_and (void (*do_something) ()) quotearg_colon (current_file_name))); /* Fall through. */ default: - - /* Skip past it in the archive. */ - - set_next_block_after (current_header); - if (current_header->oldgnu_header.isextended) - skip_extended_headers (); - - /* Skip to the next header on the archive. */ - - if (save_typeflag != DIRTYPE) - skip_file (current_stat.st_size); + skip_member (); continue; } } - apply_nonancestor_delayed_set_stat (current_file_name); (*do_something) (); continue; @@ -179,15 +166,11 @@ read_and (void (*do_something) ()) break; } - apply_delayed_set_stat (); close_archive (); names_notfound (); /* print names not found */ } -/*---------------------------------------------. -| Print a header block, based on tar options. | -`---------------------------------------------*/ - +/* Print a header block, based on tar options. */ void list_archive (void) { @@ -244,34 +227,27 @@ list_archive (void) } - /* Skip past the header in the archive, and past any extended headers. */ - - set_next_block_after (current_header); - if (current_header->oldgnu_header.isextended) - skip_extended_headers (); - if (multi_volume_option) assign_string (&save_name, current_file_name); - /* Skip to the next header on the archive. */ - - skip_file (current_stat.st_size); + skip_member (); if (multi_volume_option) assign_string (&save_name, 0); } -/*-----------------------------------------------------------------------. -| 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.st_size. | -| | -| Return 1 for success, 0 if the checksum is bad, EOF on eof, 2 for a | -| block full of zeros (EOF marker). | -| | -| You must always set_next_block_after(current_header) to skip past the | -| header which this routine reads. | -`-----------------------------------------------------------------------*/ +/* 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.st_size. + + Return 1 for success, 0 if the checksum is bad, EOF on eof, 2 for a + block full of zeros (EOF marker). + + If RAW_EXTENDED_HEADERS is nonzero, do not automagically fold the + GNU long name and link headers into later headers. + + You must always set_next_block_after(current_header) to skip past + the header which this routine reads. */ /* The standard BSD tar sources create the checksum by adding up the bytes in the header as type char. I think the type char was unsigned @@ -283,7 +259,7 @@ list_archive (void) computes two checksums -- signed and unsigned. */ enum read_header -read_header (void) +read_header (bool raw_extended_headers) { size_t i; int unsigned_sum; /* the POSIX one :-) */ @@ -292,11 +268,14 @@ read_header (void) uintmax_t parsed_sum; char *p; union block *header; - char **longp; + union block *header_copy; char *bp; union block *data_block; size_t size, written; - static char *next_long_name, *next_long_link; + union block *next_long_name = 0; + union block *next_long_link = 0; + size_t next_long_name_blocks; + size_t next_long_link_blocks; while (1) { @@ -311,7 +290,7 @@ read_header (void) for (i = sizeof *header; i-- != 0;) { unsigned_sum += (unsigned char) *p; - signed_sum += signed_char (*p++); + signed_sum += (signed char) (*p++); } if (unsigned_sum == 0) @@ -322,7 +301,7 @@ read_header (void) for (i = sizeof header->header.chksum; i-- != 0;) { unsigned_sum -= (unsigned char) header->header.chksum[i]; - signed_sum -= signed_char (header->header.chksum[i]); + signed_sum -= (signed char) (header->header.chksum[i]); } unsigned_sum += ' ' * sizeof header->header.chksum; signed_sum += ' ' * sizeof header->header.chksum; @@ -349,19 +328,38 @@ read_header (void) if (header->header.typeflag == GNUTYPE_LONGNAME || header->header.typeflag == GNUTYPE_LONGLINK) { - longp = ((header->header.typeflag == GNUTYPE_LONGNAME) - ? &next_long_name - : &next_long_link); + if (raw_extended_headers) + return HEADER_SUCCESS_EXTENDED; + else + { + size_t name_size = current_stat.st_size; + size = name_size - name_size % BLOCKSIZE + 2 * BLOCKSIZE; + if (name_size != current_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); + next_long_name = header_copy; + next_long_name_blocks = size / BLOCKSIZE; + } + else + { + if (next_long_link) + free (next_long_link); + next_long_link = header_copy; + next_long_link_blocks = size / BLOCKSIZE; + } set_next_block_after (header); - if (*longp) - free (*longp); - size = current_stat.st_size; - if (size != current_stat.st_size) - xalloc_die (); - bp = *longp = xmalloc (size); - - for (; size > 0; size -= written) + *header_copy = *header; + bp = header_copy->buffer + BLOCKSIZE; + + for (size -= BLOCKSIZE; size > 0; size -= written) { data_block = find_next_block (); if (! data_block) @@ -379,76 +377,89 @@ read_header (void) (data_block->buffer + written - 1)); } + *bp = '\0'; + /* Loop! */ } else { - char *name; - struct posix_header *h = ¤t_header->header; + char const *name; + struct posix_header const *h = ¤t_header->header; char namebuf[sizeof h->prefix + 1 + NAME_FIELD_SIZE + 1]; - name = next_long_name; - if (! name) + if (recent_long_name) + free (recent_long_name); + + if (next_long_name) + { + name = next_long_name->buffer + BLOCKSIZE; + recent_long_name = next_long_name; + recent_long_name_blocks = next_long_name_blocks; + } + else { /* Accept file names as specified by POSIX.1-1996 section 10.1.1. */ - int posix_header = strcmp (h->magic, TMAGIC) == 0; char *np = namebuf; - if (posix_header && h->prefix[0]) + if (h->prefix[0] && strcmp (h->magic, TMAGIC) == 0) { memcpy (np, h->prefix, sizeof h->prefix); 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'; name = namebuf; + recent_long_name = 0; + recent_long_name_blocks = 0; } assign_string (¤t_file_name, name); - if (next_long_name) + + if (recent_long_link) + free (recent_long_link); + + if (next_long_link) { - free (next_long_name); - next_long_name = 0; + name = next_long_link->buffer + BLOCKSIZE; + recent_long_link = next_long_link; + recent_long_link_blocks = next_long_link_blocks; } - - name = next_long_link; - if (! name) + else { memcpy (namebuf, h->linkname, sizeof h->linkname); namebuf[sizeof h->linkname] = '\0'; name = namebuf; + recent_long_link = 0; + recent_long_link_blocks = 0; } assign_string (¤t_link_name, name); - if (next_long_link) - { - free (next_long_link); - next_long_link = 0; - } return HEADER_SUCCESS; } } } -/*-------------------------------------------------------------------------. -| Decode things from a file HEADER block into STAT_INFO, also setting | -| *FORMAT_POINTER depending on the header block format. If DO_USER_GROUP, | -| decode the user/group information (this is useful for extraction, but | -| waste time when merely listing). | -| | -| read_header() has already decoded the checksum and length, so we don't. | -| | -| This routine should *not* be called twice for the same block, since the | -| two calls might use different DO_USER_GROUP values and thus might end up | -| with different uid/gid for the two calls. If anybody wants the uid/gid | -| they should decode it first, and other callers should decode it without | -| uid/gid before calling a routine, e.g. print_header, that assumes | -| decoded data. | -`-------------------------------------------------------------------------*/ +/* Decode things from a file HEADER block into STAT_INFO, also setting + *FORMAT_POINTER depending on the header block format. If + DO_USER_GROUP, decode the user/group information (this is useful + for extraction, but waste time when merely listing). + read_header() has already decoded the checksum and length, so we don't. + + This routine should *not* be called twice for the same block, since + the two calls might use different DO_USER_GROUP values and thus + might end up with different uid/gid for the two calls. If anybody + wants the uid/gid they should decode it first, and other callers + should decode it without uid/gid before calling a routine, + e.g. print_header, that assumes decoded data. */ void decode_header (union block *header, struct stat *stat_info, enum archive_format *format_pointer, int do_user_group) @@ -514,12 +525,11 @@ decode_header (union block *header, struct stat *stat_info, } } -/*------------------------------------------------------------------------. -| 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. | -`------------------------------------------------------------------------*/ - +/* 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. Return -1 on error, diagnosing the error if TYPE is + nonzero. */ static uintmax_t from_header (char const *where0, size_t digs, char const *type, uintmax_t minus_minval, uintmax_t maxval) @@ -569,7 +579,7 @@ from_header (char const *where0, size_t digs, char const *type, nonzero digit is 1, we can't recover the original value reliably; so do this only if the digit is 2 or more. This catches the common case of 32-bit negative time stamps. */ - if ((overflow || maxval < value) && '2' <= *where1) + if ((overflow || maxval < value) && '2' <= *where1 && type) { /* Compute the negative of the input value, assuming two's complement. */ @@ -601,76 +611,76 @@ from_header (char const *where0, size_t digs, char const *type, if (overflow) { - ERROR ((0, 0, - _("Archive octal value %.*s is out of %s range"), - (int) (where - where1), where1, type)); + if (type) + ERROR ((0, 0, + _("Archive octal value %.*s is out of %s range"), + (int) (where - where1), where1, type)); return -1; } } - else if (type) + else if (*where == '-' || *where == '+') { - /* The following forms cannot appear as checksums, so we don't - check for them if TYPE is null. */ - - 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) + { + warned_once = 1; + WARN ((0, 0, + _("Archive contains obsolescent base-64 headers"))); + } + negative = *where++ == '-'; + while (where != lim + && (dig = base64_map[(unsigned char) *where]) < 64) { - /* 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 (value << LG_64 >> LG_64 != value) { - warned_once = 1; - WARN ((0, 0, - _("Archive contains obsolescent base-64 headers"))); - } - negative = *where++ == '-'; - while (where != lim - && (dig = base64_map[(unsigned char) *where]) < 64) - { - if (value << LG_64 >> LG_64 != value) - { - ERROR ((0, 0, - _("Archive signed base-64 string `%.*s' is out of %s range"), - (int) digs, where0, type)); - return -1; - } - value = (value << LG_64) | dig; - where++; + char *string = alloca (digs + 1); + memcpy (string, where0, digs); + string[digs] = '\0'; + if (type) + ERROR ((0, 0, + _("Archive signed base-64 string %s is out of %s range"), + quote (string), type)); + return -1; } + value = (value << LG_64) | dig; + where++; } - else if (*where == '\200' /* positive base-256 */ - || *where == '\377' /* negative base-256 */) + } + else if (*where == '\200' /* positive base-256 */ + || *where == '\377' /* negative base-256 */) + { + /* Parse base-256 output. A nonnegative number N is + represented as (256**DIGS)/2 + N; a negative number -N is + represented as (256**DIGS) - N, i.e. as two's complement. + The representation guarantees that the leading bit is + always on, so that we don't confuse this format with the + others (assuming ASCII bytes of 8 bits or more). */ + int signbit = *where & (1 << (LG_256 - 2)); + uintmax_t topbits = (((uintmax_t) - signbit) + << (CHAR_BIT * sizeof (uintmax_t) + - LG_256 - (LG_256 - 2))); + value = (*where++ & ((1 << (LG_256 - 2)) - 1)) - signbit; + for (;;) { - /* Parse base-256 output. A nonnegative number N is - represented as (256**DIGS)/2 + N; a negative number -N is - represented as (256**DIGS) - N, i.e. as two's complement. - The representation guarantees that the leading bit is - always on, so that we don't confuse this format with the - others (assuming ASCII bytes of 8 bits or more). */ - int signbit = *where & (1 << (LG_256 - 2)); - uintmax_t topbits = (((uintmax_t) - signbit) - << (CHAR_BIT * sizeof (uintmax_t) - - LG_256 - (LG_256 - 2))); - value = (*where++ & ((1 << (LG_256 - 2)) - 1)) - signbit; - for (;;) + value = (value << LG_256) + (unsigned char) *where++; + if (where == lim) + break; + if (((value << LG_256 >> LG_256) | topbits) != value) { - value = (value << LG_256) + (unsigned char) *where++; - if (where == lim) - break; - if (((value << LG_256 >> LG_256) | topbits) != value) - { - ERROR ((0, 0, - _("Archive base-256 value is out of %s range"), - type)); - return -1; - } + if (type) + ERROR ((0, 0, + _("Archive base-256 value is out of %s range"), + type)); + return -1; } - negative = signbit; - if (negative) - value = -value; } + negative = signbit; + if (negative) + value = -value; } if (where != lim && *where && !ISSPACE ((unsigned char) *where)) @@ -783,12 +793,9 @@ size_from_header (const char *p, size_t s) time_t time_from_header (const char *p, size_t s) { - time_t t = from_header (p, s, "time_t", - - (uintmax_t) TYPE_MINIMUM (time_t), - (uintmax_t) TYPE_MAXIMUM (time_t)); - if (start_time < t && time (0) < t) - WARN ((0, 0, _("Archive contains future timestamp %s"), tartime (t))); - return t; + return from_header (p, s, "time_t", + - (uintmax_t) TYPE_MINIMUM (time_t), + (uintmax_t) TYPE_MAXIMUM (time_t)); } uid_t @@ -807,10 +814,8 @@ uintmax_from_header (const char *p, size_t s) } -/*----------------------------------------------------------------------. -| Format O as a null-terminated decimal string into BUF _backwards_; | -| return pointer to start of result. | -`----------------------------------------------------------------------*/ +/* 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) { @@ -836,8 +841,8 @@ tartime (time_t t) if (p) { char const *time_stamp = p + 4; - for (p += 16; p[4] != '\n'; p++) - p[0] = p[4]; + for (p += 16; p[3] != '\n'; p++) + p[0] = p[3]; p[0] = '\0'; return time_stamp; } @@ -867,16 +872,15 @@ tartime (time_t t) return p; } -/*-------------------------------------------------------------------------. -| Actually print it. | -| | -| Plain and fancy file header block logging. Non-verbose just prints the | -| name, e.g. for "tar t" or "tar x". This should just contain file names, | -| so it can be fed back into tar with xargs or the "-T" option. The | -| verbose option can give a bunch of info, one line per file. I doubt | -| anybody tries to parse its format, or if they do, they shouldn't. Unix | -| tar is pretty random here anyway. | -`-------------------------------------------------------------------------*/ +/* Actually print it. + + Plain and fancy file header block logging. Non-verbose just prints + the name, e.g. for "tar t" or "tar x". This should just contain + file names, so it can be fed back into tar with xargs or the "-T" + option. The verbose option can give a bunch of info, one line per + file. I doubt anybody tries to parse its format, or if they do, + 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... */ @@ -985,17 +989,45 @@ print_header (void) && !numeric_owner_option) user = current_header->header.uname; else - user = STRINGIFY_BIGINT (UINTMAX_FROM_HEADER - (current_header->header.uid), - uform); + { + /* 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)); + if (u != -1) + user = STRINGIFY_BIGINT (u, uform); + else + { + sprintf (uform, "%ld", + (long) UID_FROM_HEADER (current_header->header.uid)); + user = uform; + } + } if (*current_header->header.gname && current_format != V7_FORMAT && !numeric_owner_option) group = current_header->header.gname; else - group = STRINGIFY_BIGINT (UINTMAX_FROM_HEADER - (current_header->header.gid), - gform); + { + /* 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)); + if (g != -1) + group = STRINGIFY_BIGINT (g, gform); + else + { + sprintf (gform, "%ld", + (long) GID_FROM_HEADER (current_header->header.gid)); + group = gform; + } + } /* Format the file size or major/minor device numbers. */ @@ -1043,8 +1075,13 @@ print_header (void) break; default: - fprintf (stdlis, _(" unknown file type `%c'\n"), - current_header->header.typeflag); + { + char type_string[2]; + type_string[0] = current_header->header.typeflag; + type_string[1] = '\0'; + fprintf (stdlis, _(" unknown file type %s\n"), + quote (type_string)); + } break; case AREGTYPE: @@ -1079,10 +1116,7 @@ print_header (void) fflush (stdlis); } -/*--------------------------------------------------------------. -| Print a similar line when we make a directory automatically. | -`--------------------------------------------------------------*/ - +/* Print a similar line when we make a directory automatically. */ void print_for_mkdir (char *pathname, int length, mode_t mode) { @@ -1107,10 +1141,7 @@ print_for_mkdir (char *pathname, int length, mode_t mode) } } -/*--------------------------------------------------------. -| Skip over SIZE bytes of data in blocks in the archive. | -`--------------------------------------------------------*/ - +/* Skip over SIZE bytes of data in blocks in the archive. */ void skip_file (off_t size) { @@ -1135,21 +1166,26 @@ skip_file (off_t size) } } -/*---. -| ? | -`---*/ - +/* Skip the current member in the archive. */ void -skip_extended_headers (void) +skip_member (void) { - union block *exhdr; + char save_typeflag = current_header->header.typeflag; + set_next_block_after (current_header); - do + if (current_header->oldgnu_header.isextended) { - exhdr = find_next_block (); - if (!exhdr) - FATAL_ERROR ((0, 0, _("Unexpected EOF in archive"))); - set_next_block_after (exhdr); + union block *exhdr; + do + { + exhdr = find_next_block (); + if (!exhdr) + FATAL_ERROR ((0, 0, _("Unexpected EOF in archive"))); + set_next_block_after (exhdr); + } + while (exhdr->sparse_header.isextended); } - while (exhdr->sparse_header.isextended); + + if (save_typeflag != DIRTYPE) + skip_file (current_stat.st_size); }