diff options
| -rw-r--r-- | ChangeLog | 101 | ||||
| -rw-r--r-- | INSTALL | 28 | ||||
| -rw-r--r-- | Makefile.in | 39 | ||||
| -rw-r--r-- | NEWS | 23 | ||||
| -rw-r--r-- | README | 37 | ||||
| -rw-r--r-- | archive_reader.cc | 224 | ||||
| -rw-r--r-- | archive_reader.h | 122 | ||||
| -rw-r--r-- | arg_parser.cc | 30 | ||||
| -rw-r--r-- | arg_parser.h | 69 | ||||
| -rw-r--r-- | common.cc | 31 | ||||
| -rw-r--r-- | common_decode.cc | 36 | ||||
| -rwxr-xr-x | configure | 27 | ||||
| -rw-r--r-- | create.cc | 142 | ||||
| -rw-r--r-- | create_lz.cc | 83 | ||||
| -rw-r--r-- | decode_lz.cc | 536 | ||||
| -rw-r--r-- | delete.cc | 77 | ||||
| -rw-r--r-- | delete_lz.cc | 148 | ||||
| -rw-r--r-- | doc/tarlz.1 | 24 | ||||
| -rw-r--r-- | doc/tarlz.info | 1038 | ||||
| -rw-r--r-- | doc/tarlz.texi | 346 | ||||
| -rw-r--r-- | exclude.cc | 24 | ||||
| -rw-r--r-- | extended.cc | 29 | ||||
| -rw-r--r-- | extract.cc | 575 | ||||
| -rw-r--r-- | list_lz.cc | 586 | ||||
| -rw-r--r-- | lzip_index.cc | 109 | ||||
| -rw-r--r-- | lzip_index.h | 30 | ||||
| -rw-r--r-- | main.cc | 389 | ||||
| -rw-r--r-- | tarlz.h | 231 | ||||
| -rwxr-xr-x | testsuite/check.sh | 101 |
29 files changed, 2949 insertions, 2286 deletions
@@ -1,27 +1,40 @@ +2020-07-30 Antonio Diaz Diaz <antonio@gnu.org> + + * Version 0.17 released. + * New option '--mtime'. + * New option '-p, --preserve-permissions'. + * Implement multi-threaded '-d, --diff'. + * list_lz.cc: Rename to decode_lz.cc. + * main.cc (main): Report an error if a file name is empty or if the + archive is specified more than once. + * lzip_index.cc: Improve messages for corruption in last header. + * Don't #include <sys/sysmacros.h> when compiling on BSD. + * tarlz.texi: New chapter 'Internal structure of tarlz'. + 2019-10-08 Antonio Diaz Diaz <antonio@gnu.org> * Version 0.16 released. - * extract.cc (extract_member): Fixed call order of chown, chmod. + * extract.cc (extract_member): Fix call order of chown, chmod. * delete_lz.cc (delete_members_lz): Return 2 if collective member. * main.cc: Set a valid invocation_name even if argc == 0. * #include <sys/sysmacros.h> unconditionally. - * tarlz.texi: Added new chapter 'Portable character set'. + * tarlz.texi: New chapter 'Portable character set'. 2019-04-11 Antonio Diaz Diaz <antonio@gnu.org> * Version 0.15 released. - * Added new option '--delete' (uncompressed and --no-solid archives). - * list_lz.cc: Fixed MT listing of archives with format violations. + * New option '--delete' (from uncompressed and --no-solid archives). + * list_lz.cc: Fix MT listing of archives with format violations. 2019-03-12 Antonio Diaz Diaz <antonio@gnu.org> * Version 0.14 released. - * Added new option '--exclude'. - * Added new option '-h, --dereference'. + * New option '--exclude'. + * New option '-h, --dereference'. * Short option name '-h' no longer means '--help'. - * create.cc: Implemented '-A, --concatenate', '-r, --append' to + * create.cc: Implement '-A, --concatenate' and '-r, --append' to uncompressed archives and to standard output. - * main.cc: Ported option '--out-slots' from plzip. + * main.cc: Port option '--out-slots' from plzip. 2019-02-27 Antonio Diaz Diaz <antonio@gnu.org> @@ -33,59 +46,59 @@ 2019-02-22 Antonio Diaz Diaz <antonio@gnu.org> * Version 0.12 released. - * create.cc (fill_headers): Fixed use of st_rdev instead of st_dev. + * create.cc (fill_headers): Fix use of st_rdev instead of st_dev. * Save just numerical uid/gid if user or group not in database. * extract.cc (format_member_name): Print devmajor and devminor. - * Added new option '-d, --diff'. - * Added new option '--ignore-ids'. + * New option '-d, --diff'. + * New option '--ignore-ids'. * extract.cc: Fast '-t, --list' on seekable uncompressed archives. 2019-02-13 Antonio Diaz Diaz <antonio@gnu.org> * Version 0.11 released. - * extract.cc (archive_read): Fixed endless loop with empty lz file. - * Implemented multi-threaded '-c, --create' and '-r, --append'. + * extract.cc (archive_read): Fix endless loop with empty lz file. + * Implement multi-threaded '-c, --create' and '-r, --append'. * '--bsolid' is now the default compression granularity. * create.cc (remove_leading_dotslash): Remember more than one prefix. - * tarlz.texi: Added new chapter 'Minimum archive sizes'. + * tarlz.texi: New chapter 'Minimum archive sizes'. 2019-01-31 Antonio Diaz Diaz <antonio@gnu.org> * Version 0.10 released. - * Added new option '--bsolid'. - * Added new option '-B, --data-size'. + * New option '--bsolid'. + * New option '-B, --data-size'. * create.cc: Set ustar name to zero if extended header is used. 2019-01-22 Antonio Diaz Diaz <antonio@gnu.org> * Version 0.9 released. - * Implemented multi-threaded '-t, --list'. - * Added new option '-n, --threads'. + * Implement multi-threaded '-t, --list'. + * New option '-n, --threads'. * Recognize global pax headers. Ignore them for now. * strtoul has been replaced with length-safe parsers. - * tarlz.texi: Added new chapter 'Limitations of parallel tar decoding'. + * tarlz.texi: New chapter 'Limitations of parallel tar decoding'. 2018-12-16 Antonio Diaz Diaz <antonio@gnu.org> * Version 0.8 released. - * Added new option '--anonymous' (--owner=root --group=root). + * New option '--anonymous' (--owner=root --group=root). * extract.cc (decode): 'tarlz -xf foo ./bar' now extracts 'bar'. * create.cc: Set to zero most fields in extended headers. - * tarlz.texi: Added new chapter 'Amendments to pax format'. + * tarlz.texi: New chapter 'Amendments to pax format'. 2018-11-23 Antonio Diaz Diaz <antonio@gnu.org> * Version 0.7 released. - * Added new option '--keep-damaged'. - * Added new option '--no-solid'. + * New option '--keep-damaged'. + * New option '--no-solid'. * create.cc (archive_write): Minimize dictionary size. - * create.cc: Detect and skip archive in '-A', '-c' and '-r'. + * create.cc: Detect and skip archive in '-A', '-c', and '-r'. * main.cc (show_version): Show the version of lzlib being used. 2018-10-19 Antonio Diaz Diaz <antonio@gnu.org> * Version 0.6 released. - * Added new option '-A, --concatenate'. + * New option '-A, --concatenate'. * Option '--ignore-crc' replaced with '--missing-crc'. * create.cc (add_member): Test that uid, gid, mtime, devmajor and devminor are in ustar range. @@ -95,32 +108,32 @@ 2018-09-29 Antonio Diaz Diaz <antonio@gnu.org> * Version 0.5 released. - * Implemented simplified posix pax format. - * Implemented CRC32-C (Castagnoli) of the extended header data. - * Added new option '--ignore-crc'. - * Added missing #includes for major, minor and makedev. - * tarlz.texi: Documented the new archive format. + * Implement simplified posix pax format. + * Implement CRC32-C (Castagnoli) of the extended header data. + * New option '--ignore-crc'. + * Add missing #includes for major, minor and makedev. + * tarlz.texi: Document the new archive format. 2018-04-23 Antonio Diaz Diaz <antonio@gnu.org> * Version 0.4 released. - * Added some missing #includes. + * Add some missing #includes. * main.cc: Open files in binary mode on OS2. 2018-03-19 Antonio Diaz Diaz <antonio@gnu.org> * Version 0.3 released. - * Project renamed to 'tarlz' from 'pmtar' (Poor Man's Tar). - * Added new option '-C, --directory'. - * Implemented lzip compression of members at archive creation. - * Added new option '-r, --append'. - * Added new options '--owner', '--group'. - * Added new options '--asolid', '--dsolid', '--solid'. - * Implemented file appending to compressed archive. - * Implemented transparent decompression of the archive. - * Implemented skipping over damaged (un)compressed members. - * Implemented recursive extraction/listing of directories. - * Implemented verbose extract/list output. + * Rename project to 'tarlz' from 'pmtar' (Poor Man's Tar). + * New option '-C, --directory'. + * Implement lzip compression of members at archive creation. + * New option '-r, --append'. + * New options '--owner' and '--group'. + * New options '--asolid', '--dsolid', and '--solid'. + * Implement file appending to compressed archive. + * Implement transparent decompression of the archive. + * Implement skipping over damaged (un)compressed members. + * Implement recursive extraction/listing of directories. + * Implement verbose extract/list output. * tarlz.texi: New file. 2014-01-22 Antonio Diaz Diaz <antonio@gnu.org> @@ -133,8 +146,8 @@ * Version 0.1 released. -Copyright (C) 2013-2019 Antonio Diaz Diaz. +Copyright (C) 2013-2020 Antonio Diaz Diaz. This file is a collection of facts, and thus it is not copyrightable, -but just in case, you have unlimited permission to copy, distribute and +but just in case, you have unlimited permission to copy, distribute, and modify it. @@ -1,9 +1,10 @@ Requirements ------------ -You will need a C++ compiler and the lzlib compression library installed. +You will need a C++11 compiler and the compression library lzlib installed. +(gcc 3.3.6 or newer is recommended). I use gcc 6.1.0 and 4.1.2, but the code should compile with any standards compliant compiler. -Lzlib must be version 1.0 or newer, but --keep-damaged requires lzlib 1.11 +Lzlib must be version 1.8 or newer, but --keep-damaged requires lzlib 1.11 or newer to recover as much data as possible from each damaged member. Gcc is available at http://gcc.gnu.org. Lzlib is available at http://www.nongnu.org/lzip/lzlib.html. @@ -28,7 +29,10 @@ the main archive. To link against a lzlib not installed in a standard place, use: - ./configure CPPFLAGS='-I<dir_of_lzlib.h>' LDFLAGS='-L<dir_of_liblz.a>' + ./configure CPPFLAGS='-I <includedir>' LDFLAGS='-L <libdir>' + + (Replace <includedir> with the directory containing the file lzlib.h, + and <libdir> with the directory containing the file liblz.a). 3. Run make. @@ -40,21 +44,21 @@ the main archive. documentation. Or type 'make install-compress', which additionally compresses the - info manual and the man page after installation. (Installing - compressed docs may become the default in the future). + info manual and the man page after installation. + (Installing compressed docs may become the default in the future). - You can install only the program, the info manual or the man page by - typing 'make install-bin', 'make install-info' or 'make install-man' + You can install only the program, the info manual, or the man page by + typing 'make install-bin', 'make install-info', or 'make install-man' respectively. Another way ----------- You can also compile tarlz into a separate directory. -To do this, you must use a version of 'make' that supports the 'VPATH' -variable, such as GNU 'make'. 'cd' to the directory where you want the +To do this, you must use a version of 'make' that supports the variable +'VPATH', such as GNU 'make'. 'cd' to the directory where you want the object files and executables to go and run the 'configure' script. -'configure' automatically checks for the source code in '.', in '..' and +'configure' automatically checks for the source code in '.', in '..', and in the directory that 'configure' is in. 'configure' recognizes the option '--srcdir=DIR' to control where to @@ -65,7 +69,7 @@ After running 'configure', you can run 'make' and 'make install' as explained above. -Copyright (C) 2013-2019 Antonio Diaz Diaz. +Copyright (C) 2013-2020 Antonio Diaz Diaz. This file is free documentation: you have unlimited permission to copy, -distribute and modify it. +distribute, and modify it. diff --git a/Makefile.in b/Makefile.in index 9697efe..44f0d86 100644 --- a/Makefile.in +++ b/Makefile.in @@ -8,9 +8,9 @@ LIBS = -llz -lpthread SHELL = /bin/sh CAN_RUN_INSTALLINFO = $(SHELL) -c "install-info --version" > /dev/null 2>&1 -objs = arg_parser.o lzip_index.o common.o common_decode.o create.o \ - create_lz.o delete.o delete_lz.o exclude.o extended.o extract.o \ - list_lz.o main.o +objs = arg_parser.o lzip_index.o archive_reader.o common.o common_decode.o \ + create.o create_lz.o delete.o delete_lz.o exclude.o extended.o \ + extract.o decode_lz.o main.o .PHONY : all install install-bin install-info install-man \ @@ -30,20 +30,21 @@ main.o : main.cc %.o : %.cc $(CXX) $(CPPFLAGS) $(CXXFLAGS) -c -o $@ $< -$(objs) : Makefile -arg_parser.o : arg_parser.h -common.o : tarlz.h -common_decode.o : arg_parser.h tarlz.h -create.o : arg_parser.h tarlz.h -create_lz.o : arg_parser.h tarlz.h -delete.o : arg_parser.h lzip_index.h tarlz.h -delete_lz.o : arg_parser.h lzip_index.h tarlz.h -exclude.o : tarlz.h -extended.o : tarlz.h -extract.o : arg_parser.h lzip_index.h tarlz.h -list_lz.o : arg_parser.h lzip_index.h tarlz.h -lzip_index.o : lzip_index.h tarlz.h -main.o : arg_parser.h tarlz.h +$(objs) : Makefile +arg_parser.o : arg_parser.h +archive_reader.o : tarlz.h lzip_index.h archive_reader.h +common.o : arg_parser.h tarlz.h +common_decode.o : arg_parser.h tarlz.h +create.o : arg_parser.h tarlz.h +create_lz.o : arg_parser.h tarlz.h +delete.o : arg_parser.h tarlz.h lzip_index.h +delete_lz.o : arg_parser.h tarlz.h lzip_index.h +exclude.o : tarlz.h +extended.o : tarlz.h +extract.o : arg_parser.h tarlz.h lzip_index.h archive_reader.h +decode_lz.o : arg_parser.h tarlz.h lzip_index.h archive_reader.h +lzip_index.o : tarlz.h lzip_index.h +main.o : arg_parser.h tarlz.h doc : info man @@ -82,7 +83,7 @@ install-info : -rm -f "$(DESTDIR)$(infodir)/$(pkgname).info"* $(INSTALL_DATA) $(VPATH)/doc/$(pkgname).info "$(DESTDIR)$(infodir)/$(pkgname).info" -if $(CAN_RUN_INSTALLINFO) ; then \ - install-info --info-dir="$(DESTDIR)$(infodir)" "$(DESTDIR)$(infodir)/$(pkgname).info" ; \ + install-info --info-dir="$(DESTDIR)$(infodir)" "$(DESTDIR)$(infodir)/$(pkgname).info" ; \ fi install-info-compress : install-info @@ -103,7 +104,7 @@ uninstall-bin : uninstall-info : -if $(CAN_RUN_INSTALLINFO) ; then \ - install-info --info-dir="$(DESTDIR)$(infodir)" --remove "$(DESTDIR)$(infodir)/$(pkgname).info" ; \ + install-info --info-dir="$(DESTDIR)$(infodir)" --remove "$(DESTDIR)$(infodir)/$(pkgname).info" ; \ fi -rm -f "$(DESTDIR)$(infodir)/$(pkgname).info"* @@ -1,11 +1,20 @@ -Changes in version 0.16: +Changes in version 0.17: -'chown' and 'chmod' are now called in the right order on extracion to -preserve the S_ISUID and S_ISGID bits of executable files. +The new option '--mtime' has been added. -The return value of '--delete' when failing to delete a tar member not -individually compressed has been fixed. It returned 0, but should be 2. +The new option '-p, --preserve-permissions' has been added. -The header <sys/sysmacros.h> is now #included unconditionally. +Multi-threaded '-d, --diff' has been implemented. See chapters 'Internal +structure of tarlz' and 'Limitations of parallel tar decoding' in the manual +for details. -The new chapter 'Portable character set' has been added to the manual. +Tarlz now reports an error if a file name is empty (tarlz -tf foo "") or if +the archive is specified more than once. + +Tarlz now reports corruption or truncation of the last header in a +multimenber file specifically instead of showing the generic message "Last +member in input file is truncated or corrupt." + +The header <sys/sysmacros.h> is now not #included when compiling on BSD. + +The new chapter 'Internal structure of tarlz' has been added to the manual. @@ -1,7 +1,7 @@ Description Tarlz is a massively parallel (multi-threaded) combined implementation of -the tar archiver and the lzip compressor. Tarlz creates, lists and extracts +the tar archiver and the lzip compressor. Tarlz creates, lists, and extracts archives in a simplified and safer variant of the POSIX pax format compressed with lzip, keeping the alignment between tar members and lzip members. The resulting multimember tar.lz archive is fully backward @@ -18,7 +18,8 @@ because it does not keep the members aligned. Tarlz can create tar archives with five levels of compression granularity; per file (--no-solid), per block (--bsolid, default), per directory -(--dsolid), appendable solid (--asolid), and solid (--solid). +(--dsolid), appendable solid (--asolid), and solid (--solid). It can also +create uncompressed tar archives. Of course, compressing each file (or each directory) individually can't achieve a compression ratio as high as compressing solidly the whole tar @@ -31,16 +32,16 @@ archive, but it has the following advantages: member), and unwanted members can be deleted from the archive. Just like an uncompressed tar archive. - * It is a safe POSIX-style backup format. In case of corruption, - tarlz can extract all the undamaged members from the tar.lz - archive, skipping over the damaged members, just like the standard - (uncompressed) tar. Moreover, the option '--keep-damaged' can be - used to recover as much data as possible from each damaged member, - and lziprecover can be used to recover some of the damaged members. + * It is a safe POSIX-style backup format. In case of corruption, tarlz + can extract all the undamaged members from the tar.lz archive, + skipping over the damaged members, just like the standard + (uncompressed) tar. Moreover, the option '--keep-damaged' can be used + to recover as much data as possible from each damaged member, and + lziprecover can be used to recover some of the damaged members. - * A multimember tar.lz archive is usually smaller than the - corresponding solidly compressed tar.gz archive, except when - individually compressing files smaller than about 32 KiB. + * A multimember tar.lz archive is usually smaller than the corresponding + solidly compressed tar.gz archive, except when compressing files + smaller than about 32 KiB individually. Note that the POSIX pax format has a serious flaw. The metadata stored in pax extended records are not protected by any kind of check sequence. @@ -54,19 +55,19 @@ Metadata like file name and file size must be always protected in an archive format because of the adverse effects of undetected corruption in them, potentially much worse that undetected corruption in the data. Even more so in the case of pax because the amount of metadata it stores is potentially -large, making undetected corruption more probable. +large, making undetected corruption and archiver misbehavior more probable. Headers and metadata must be protected separately from data because the integrity checking of lzip may not be able to detect the corruption before the metadata has been used, for example, to create a new file in the wrong place. -Because of the above, tarlz protects the extended records with a CRC in a -way compatible with standard tar tools. +Because of the above, tarlz protects the extended records with a Cyclic +Redundancy Check (CRC) in a way compatible with standard tar tools. Tarlz does not understand other tar formats like gnu, oldgnu, star or v7. -'tarlz -tf archive.tar.lz > /dev/null' can be used to verify that the format -of the archive is compatible with tarlz. +The command 'tarlz -tf archive.tar.lz > /dev/null' can be used to verify +that the format of the archive is compatible with tarlz. The diagram below shows the correspondence between each tar member (formed by one or two headers plus optional data) in the tar archive and each lzip @@ -84,10 +85,10 @@ tar.lz +===============+=================================================+========+ -Copyright (C) 2013-2019 Antonio Diaz Diaz. +Copyright (C) 2013-2020 Antonio Diaz Diaz. This file is free documentation: you have unlimited permission to copy, -distribute and modify it. +distribute, and modify it. The file Makefile.in is a data file used by configure to produce the Makefile. It has the same copyright owner and permissions that configure diff --git a/archive_reader.cc b/archive_reader.cc new file mode 100644 index 0000000..496c33b --- /dev/null +++ b/archive_reader.cc @@ -0,0 +1,224 @@ +/* Tarlz - Archiver with multimember lzip compression + Copyright (C) 2013-2020 Antonio Diaz Diaz. + + 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 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. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <http://www.gnu.org/licenses/>. +*/ + +#define _FILE_OFFSET_BITS 64 + +#include <algorithm> +#include <cerrno> +#include <climits> +#include <cstring> +#include <string> +#include <vector> +#include <stdint.h> +#include <unistd.h> +#include <lzlib.h> + +#include "tarlz.h" +#include "lzip_index.h" +#include "archive_reader.h" + + +int Archive_reader_base::parse_records( Extended & extended, + const Tar_header header, + Resizable_buffer & rbuf, + const bool permissive ) + { + const long long edsize = parse_octal( header + size_o, size_l ); + const long long bufsize = round_up( edsize ); + if( edsize <= 0 || edsize >= 1LL << 33 || bufsize >= INT_MAX ) + return 1; // overflow or no extended data + if( !rbuf.resize( bufsize ) ) return 1; // extended records buffer + int retval = read( (uint8_t *)rbuf(), bufsize ); + if( retval == 0 && !extended.parse( rbuf(), edsize, permissive ) ) + retval = 2; + return retval; + } + + +/* Read 'size' uncompressed bytes, decompressing the input if needed. + Return value: 0 = OK, 1 = damaged member, 2 = fatal error. */ +int Archive_reader::read( uint8_t * const buf, const int size ) + { + if( fatal_ ) return err( 2 ); + if( first_read ) // check format + { + first_read = false; + uncompressed_seekable = ad.seekable && !ad.indexed && + ad.lzip_index.file_size() > 3 * header_size; + if( size != header_size ) + internal_error( "size != header_size on first call." ); + const int rd = readblock( ad.infd, buf, size ); + if( rd != size && errno ) + return err( 2, "Error reading archive", errno, rd ); + const Lzip_header & header = (*(const Lzip_header *)buf); + const bool islz = ( rd >= min_member_size && header.verify_magic() && + header.verify_version() && + isvalid_ds( header.dictionary_size() ) ); + const bool istar = ( rd == size && verify_ustar_chksum( buf ) ); + const bool iseof = + ( !islz && !istar && rd == size && block_is_zero( buf, size ) ); + bool maybe_lz = islz; // maybe corrupt tar.lz + if( !islz && !istar && !iseof ) // corrupt or invalid format + { + const bool lz_ext = has_lz_ext( ad.name ); + show_file_error( ad.namep, lz_ext ? posix_lz_msg : posix_msg ); + if( lz_ext && rd >= min_member_size ) maybe_lz = true; + else return err( 1 ); + } + if( !maybe_lz ) // uncompressed + { if( rd == size ) return 0; + return err( 2, "EOF reading archive", 0, rd ); } + uncompressed_seekable = false; // compressed + decoder = LZ_decompress_open(); + if( !decoder || LZ_decompress_errno( decoder ) != LZ_ok ) + return err( 2, mem_msg ); + if( LZ_decompress_write( decoder, buf, rd ) != rd ) + internal_error( "library error (LZ_decompress_write)." ); + const int ret = read( buf, size ); if( ret != 0 ) return ret; + if( verify_ustar_chksum( buf ) || block_is_zero( buf, size ) ) return 0; + return err( 2, islz ? posix_lz_msg : "" ); + } + + if( !decoder ) // uncompressed + { + const int rd = readblock( ad.infd, buf, size ); + if( rd == size ) return 0; else return err( 2, end_msg, 0, rd ); + } + const int ibuf_size = 16384; + uint8_t ibuf[ibuf_size]; + int sz = 0; + while( sz < size ) + { + const int rd = LZ_decompress_read( decoder, buf + sz, size - sz ); + if( rd < 0 ) + { + const unsigned long long old_pos = LZ_decompress_total_in_size( decoder ); + if( LZ_decompress_sync_to_member( decoder ) < 0 ) + internal_error( "library error (LZ_decompress_sync_to_member)." ); + e_skip_ = true; set_error_status( 2 ); + const unsigned long long new_pos = LZ_decompress_total_in_size( decoder ); + // lzlib < 1.8 does not update total_in_size when syncing to member + if( new_pos >= old_pos && new_pos < LLONG_MAX ) + return err( 1, "", 0, sz, true ); + return err( 2, "Skipping to next header failed. " + "Lzlib 1.8 or newer required.", 0, sz ); + } + if( rd == 0 && LZ_decompress_finished( decoder ) == 1 ) + { return err( 2, end_msg, 0, sz ); } + sz += rd; + if( sz < size && !at_eof && LZ_decompress_write_size( decoder ) > 0 ) + { + const int rsize = std::min( ibuf_size, LZ_decompress_write_size( decoder ) ); + const int rd = readblock( ad.infd, ibuf, rsize ); + if( LZ_decompress_write( decoder, ibuf, rd ) != rd ) + internal_error( "library error (LZ_decompress_write)." ); + if( rd < rsize ) + { + at_eof = true; LZ_decompress_finish( decoder ); + if( errno ) return err( 2, "Error reading archive.", errno, sz ); + } + } + } + return 0; + } + + +int Archive_reader::skip_member( const Extended & extended ) + { + long long rest = round_up( extended.file_size() ); // size + padding + if( uncompressed_seekable && lseek( ad.infd, rest, SEEK_CUR ) > 0 ) + return 0; + const int bufsize = 32 * header_size; + uint8_t buf[bufsize]; + while( rest > 0 ) // skip tar member + { + const int rsize = ( rest >= bufsize ) ? bufsize : rest; + const int ret = read( buf, rsize ); + if( ret != 0 ) return ret; + rest -= rsize; + } + return 0; + } + + +void Archive_reader_i::set_member( const long i ) + { + LZ_decompress_reset( decoder ); // prepare for new member + data_pos_ = ad.lzip_index.dblock( i ).pos(); + mdata_end = ad.lzip_index.dblock( i ).end(); + archive_pos = ad.lzip_index.mblock( i ).pos(); + member_id = i; + } + + +/* Read 'size' decompressed bytes from the archive. + Return value: 0 = OK, 1 = damaged member, 2 = fatal error. */ +int Archive_reader_i::read( uint8_t * const buf, const int size ) + { + int sz = 0; + + while( sz < size ) + { + const int rd = LZ_decompress_read( decoder, buf + sz, size - sz ); + if( rd < 0 ) + return err( 1, LZ_strerror( LZ_decompress_errno( decoder ) ) ); + if( rd == 0 && LZ_decompress_finished( decoder ) == 1 ) + return err( 2, end_msg ); + sz += rd; data_pos_ += rd; + if( sz < size && LZ_decompress_write_size( decoder ) > 0 ) + { + const long long ibuf_size = 16384; + uint8_t ibuf[ibuf_size]; + const long long member_end = ad.lzip_index.mblock( member_id ).end(); + const long long rest = ( ( archive_pos < member_end ) ? + member_end : ad.lzip_index.cdata_size() ) - archive_pos; + const int rsize = std::min( LZ_decompress_write_size( decoder ), + (int)std::min( ibuf_size, rest ) ); + if( rsize <= 0 ) LZ_decompress_finish( decoder ); + else + { + const int rd = preadblock( ad.infd, ibuf, rsize, archive_pos ); + if( LZ_decompress_write( decoder, ibuf, rd ) != rd ) + internal_error( "library error (LZ_decompress_write)." ); + archive_pos += rd; + if( rd < rsize ) + { + LZ_decompress_finish( decoder ); + if( errno ) return err( 2, "Error reading archive" ); + } + } + } + } + return 0; + } + + +int Archive_reader_i::skip_member( const Extended & extended ) + { + long long rest = round_up( extended.file_size() ); // size + padding + if( data_pos_ + rest == mdata_end ) { data_pos_ = mdata_end; return 0; } + const int bufsize = 32 * header_size; + uint8_t buf[bufsize]; + while( rest > 0 ) // skip tar member + { + const int rsize = ( rest >= bufsize ) ? bufsize : rest; + const int ret = read( buf, rsize ); + if( ret != 0 ) return ret; + rest -= rsize; + } + return 0; + } diff --git a/archive_reader.h b/archive_reader.h new file mode 100644 index 0000000..a867c0f --- /dev/null +++ b/archive_reader.h @@ -0,0 +1,122 @@ +/* Tarlz - Archiver with multimember lzip compression + Copyright (C) 2013-2020 Antonio Diaz Diaz. + + 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 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. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <http://www.gnu.org/licenses/>. +*/ + +struct Archive_descriptor + { + const std::string name; + const char * const namep; // printable archive name + const int infd; + const Lzip_index lzip_index; + const bool seekable; + const bool indexed; + + Archive_descriptor( const std::string & archive_name ) + : name( archive_name ), namep( name.empty() ? "(stdin)" : name.c_str() ), + infd( name.empty() ? STDIN_FILENO : open_instream( name ) ), + lzip_index( infd, true, false ), + seekable( lseek( infd, 0, SEEK_SET ) == 0 ), + indexed( seekable && lzip_index.retval() == 0 ) {} + }; + + +class Archive_reader_base // base of serial and indexed readers + { +public: + const Archive_descriptor & ad; +protected: + LZ_Decoder * decoder; // destructor closes it if needed + const char * e_msg_; // message for show_file_error + int e_code_; // copy of errno + int e_size_; // partial size read in case of read error + bool e_skip_; // corrupt header skipped + bool fatal_; + + int err( const int retval, const char * const msg = "", const int code = 0, + const int size = 0, const bool skip = false ) + { e_msg_ = msg; e_code_ = code; e_size_ = size; e_skip_ = skip; + if( retval == 2 ) { fatal_ = true; } return retval; } + + Archive_reader_base( const Archive_descriptor & d ) + : ad( d ), decoder( 0 ), e_msg_( "" ), e_code_( 0 ), e_size_( 0 ), + e_skip_( false ), fatal_( false ) {} + +public: + virtual ~Archive_reader_base() + { if( decoder != 0 ) LZ_decompress_close( decoder ); } + + const char * e_msg() const { return e_msg_; } + int e_code() const { return e_code_; } + int e_size() const { return e_size_; } + bool e_skip() const { return e_skip_; } + bool fatal() const { return fatal_; } + + /* Read 'size' uncompressed bytes, decompressing the input if needed. + Return value: 0 = OK, 1 = damaged member, 2 = fatal error. + If !OK, fills all the e_* variables. */ + virtual int read( uint8_t * const buf, const int size ) = 0; + + int parse_records( Extended & extended, const Tar_header header, + Resizable_buffer & rbuf, const bool permissive ); + }; + + +class Archive_reader : public Archive_reader_base // serial reader + { + bool first_read; + bool uncompressed_seekable; // value set by first read call + bool at_eof; + +public: + Archive_reader( const Archive_descriptor & d ) + : Archive_reader_base( d ), first_read( true ), + uncompressed_seekable( false ), at_eof( false ) {} + + int read( uint8_t * const buf, const int size ); + int skip_member( const Extended & extended ); + }; + + +/* If the archive is compressed seekable (indexed), several indexed readers + can be constructed sharing the same Archive_descriptor, for example to + decode the archive in parallel. +*/ +class Archive_reader_i : public Archive_reader_base // indexed reader + { + long long data_pos_; + long long mdata_end; + long long archive_pos; // current position in archive for pread + long member_id; // current member unless reading beyond + +public: + Archive_reader_i( const Archive_descriptor & d ) + : Archive_reader_base( d ), + data_pos_( 0 ), mdata_end( 0 ), archive_pos( 0 ), member_id( 0 ) + { + decoder = LZ_decompress_open(); + if( !decoder || LZ_decompress_errno( decoder ) != LZ_ok ) + { LZ_decompress_close( decoder ); decoder = 0; fatal_ = true; } + } + + long long data_pos() const { return data_pos_; } + bool at_member_end() const { return data_pos_ == mdata_end; } + + // Resets decoder and sets position to the start of the member. + void set_member( const long i ); + + int read( uint8_t * const buf, const int size ); + int skip_member( const Extended & extended ); + }; diff --git a/arg_parser.cc b/arg_parser.cc index ea32fde..b843c09 100644 --- a/arg_parser.cc +++ b/arg_parser.cc @@ -1,20 +1,20 @@ -/* Arg_parser - POSIX/GNU command line argument parser. (C++ version) - Copyright (C) 2006-2019 Antonio Diaz Diaz. +/* Arg_parser - POSIX/GNU command line argument parser. (C++ version) + Copyright (C) 2006-2020 Antonio Diaz Diaz. - This library is free software. Redistribution and use in source and - binary forms, with or without modification, are permitted provided - that the following conditions are met: + This library is free software. Redistribution and use in source and + binary forms, with or without modification, are permitted provided + that the following conditions are met: - 1. Redistributions of source code must retain the above copyright - notice, this list of conditions and the following disclaimer. + 1. Redistributions of source code must retain the above copyright + notice, this list of conditions, and the following disclaimer. - 2. Redistributions in binary form must reproduce the above copyright - notice, this list of conditions and the following disclaimer in the - documentation and/or other materials provided with the distribution. + 2. Redistributions in binary form must reproduce the above copyright + notice, this list of conditions, and the following disclaimer in the + documentation and/or other materials provided with the distribution. - This library 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. + This library 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. */ #include <cstring> @@ -167,7 +167,7 @@ Arg_parser::Arg_parser( const int argc, const char * const argv[], else non_options.push_back( argv[argind++] ); } } - if( error_.size() ) data.clear(); + if( !error_.empty() ) data.clear(); else { for( unsigned i = 0; i < non_options.size(); ++i ) @@ -190,7 +190,7 @@ Arg_parser::Arg_parser( const char * const opt, const char * const arg, { if( opt[2] ) parse_long_option( opt, arg, options, argind ); } else parse_short_option( opt, arg, options, argind ); - if( error_.size() ) data.clear(); + if( !error_.empty() ) data.clear(); } else data.push_back( Record( opt ) ); } diff --git a/arg_parser.h b/arg_parser.h index ceb9933..d9a4af0 100644 --- a/arg_parser.h +++ b/arg_parser.h @@ -1,43 +1,43 @@ -/* Arg_parser - POSIX/GNU command line argument parser. (C++ version) - Copyright (C) 2006-2019 Antonio Diaz Diaz. +/* Arg_parser - POSIX/GNU command line argument parser. (C++ version) + Copyright (C) 2006-2020 Antonio Diaz Diaz. - This library is free software. Redistribution and use in source and - binary forms, with or without modification, are permitted provided - that the following conditions are met: + This library is free software. Redistribution and use in source and + binary forms, with or without modification, are permitted provided + that the following conditions are met: - 1. Redistributions of source code must retain the above copyright - notice, this list of conditions and the following disclaimer. + 1. Redistributions of source code must retain the above copyright + notice, this list of conditions, and the following disclaimer. - 2. Redistributions in binary form must reproduce the above copyright - notice, this list of conditions and the following disclaimer in the - documentation and/or other materials provided with the distribution. + 2. Redistributions in binary form must reproduce the above copyright + notice, this list of conditions, and the following disclaimer in the + documentation and/or other materials provided with the distribution. - This library 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. + This library 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. */ -/* Arg_parser reads the arguments in 'argv' and creates a number of - option codes, option arguments and non-option arguments. +/* Arg_parser reads the arguments in 'argv' and creates a number of + option codes, option arguments, and non-option arguments. - In case of error, 'error' returns a non-empty error message. + In case of error, 'error' returns a non-empty error message. - 'options' is an array of 'struct Option' terminated by an element - containing a code which is zero. A null name means a short-only - option. A code value outside the unsigned char range means a - long-only option. + 'options' is an array of 'struct Option' terminated by an element + containing a code which is zero. A null name means a short-only + option. A code value outside the unsigned char range means a + long-only option. - Arg_parser normally makes it appear as if all the option arguments - were specified before all the non-option arguments for the purposes - of parsing, even if the user of your program intermixed option and - non-option arguments. If you want the arguments in the exact order - the user typed them, call 'Arg_parser' with 'in_order' = true. + Arg_parser normally makes it appear as if all the option arguments + were specified before all the non-option arguments for the purposes + of parsing, even if the user of your program intermixed option and + non-option arguments. If you want the arguments in the exact order + the user typed them, call 'Arg_parser' with 'in_order' = true. - The argument '--' terminates all options; any following arguments are - treated as non-option arguments, even if they begin with a hyphen. + The argument '--' terminates all options; any following arguments are + treated as non-option arguments, even if they begin with a hyphen. - The syntax for optional option arguments is '-<short_option><argument>' - (without whitespace), or '--<long_option>=<argument>'. + The syntax for optional option arguments is '-<short_option><argument>' + (without whitespace), or '--<long_option>=<argument>'. */ class Arg_parser @@ -61,6 +61,7 @@ private: explicit Record( const char * const arg ) : code( 0 ), argument( arg ) {} }; + const std::string empty_arg; std::string error_; std::vector< Record > data; @@ -73,17 +74,17 @@ public: Arg_parser( const int argc, const char * const argv[], const Option options[], const bool in_order = false ); - // Restricted constructor. Parses a single token and argument (if any) + // Restricted constructor. Parses a single token and argument (if any). Arg_parser( const char * const opt, const char * const arg, const Option options[] ); const std::string & error() const { return error_; } - // The number of arguments parsed (may be different from argc) + // The number of arguments parsed. May be different from argc. int arguments() const { return data.size(); } - // If code( i ) is 0, argument( i ) is a non-option. - // Else argument( i ) is the option's argument (or empty). + /* If code( i ) is 0, argument( i ) is a non-option. + Else argument( i ) is the option's argument (or empty). */ int code( const int i ) const { if( i >= 0 && i < arguments() ) return data[i].code; @@ -93,6 +94,6 @@ public: const std::string & argument( const int i ) const { if( i >= 0 && i < arguments() ) return data[i].argument; - else return error_; + else return empty_arg; } }; @@ -1,18 +1,18 @@ -/* Tarlz - Archiver with multimember lzip compression - Copyright (C) 2013-2019 Antonio Diaz Diaz. +/* Tarlz - Archiver with multimember lzip compression + Copyright (C) 2013-2020 Antonio Diaz Diaz. - 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 of the License, or - (at your option) any later version. + 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 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. + 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. - You should have received a copy of the GNU General Public License - along with this program. If not, see <http://www.gnu.org/licenses/>. + You should have received a copy of the GNU General Public License + along with this program. If not, see <http://www.gnu.org/licenses/>. */ #define _FILE_OFFSET_BITS 64 @@ -28,6 +28,7 @@ #include <stdint.h> #include <unistd.h> +#include "arg_parser.h" #include "tarlz.h" @@ -147,3 +148,9 @@ int writeblock( const int fd, const uint8_t * const buf, const int size ) } return sz; } + + +bool nonempty_arg( const Arg_parser & parser, const int i ) + { + return ( parser.code( i ) == 0 && !parser.argument( i ).empty() ); + } diff --git a/common_decode.cc b/common_decode.cc index 791d6e5..6b108c1 100644 --- a/common_decode.cc +++ b/common_decode.cc @@ -1,18 +1,18 @@ -/* Tarlz - Archiver with multimember lzip compression - Copyright (C) 2013-2019 Antonio Diaz Diaz. +/* Tarlz - Archiver with multimember lzip compression + Copyright (C) 2013-2020 Antonio Diaz Diaz. - 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 of the License, or - (at your option) any later version. + 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 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. + 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. - You should have received a copy of the GNU General Public License - along with this program. If not, see <http://www.gnu.org/licenses/>. + You should have received a copy of the GNU General Public License + along with this program. If not, see <http://www.gnu.org/licenses/>. */ #define _FILE_OFFSET_BITS 64 @@ -180,18 +180,18 @@ bool show_member_name( const Extended & extended, const Tar_header header, } -bool check_skip_filename( const Arg_parser & parser, +bool check_skip_filename( const Cl_options & cl_opts, std::vector< char > & name_pending, - const char * const filename, const int filenames ) + const char * const filename ) { if( Exclude::excluded( filename ) ) return true; // skip excluded files - bool skip = filenames > 0; + bool skip = cl_opts.filenames > 0; if( skip ) - for( int i = 0; i < parser.arguments(); ++i ) - if( !parser.code( i ) && parser.argument( i ).size() ) + for( int i = 0; i < cl_opts.parser.arguments(); ++i ) + if( nonempty_arg( cl_opts.parser, i ) ) { const char * const name = - remove_leading_dotslash( parser.argument( i ).c_str() ); + remove_leading_dotslash( cl_opts.parser.argument( i ).c_str() ); if( compare_prefix_dir( name, filename ) || compare_tslash( name, filename ) ) { skip = false; name_pending[i] = false; break; } @@ -1,12 +1,12 @@ #! /bin/sh # configure script for Tarlz - Archiver with multimember lzip compression -# Copyright (C) 2013-2019 Antonio Diaz Diaz. +# Copyright (C) 2013-2020 Antonio Diaz Diaz. # # This configure script is free software: you have unlimited permission -# to copy, distribute and modify it. +# to copy, distribute, and modify it. pkgname=tarlz -pkgversion=0.16 +pkgversion=0.17 progname=tarlz srctrigger=doc/${pkgname}.texi @@ -26,11 +26,7 @@ CXXFLAGS='-Wall -W -O2' LDFLAGS= # checking whether we are using GNU C++. -/bin/sh -c "${CXX} --version" > /dev/null 2>&1 || - { - CXX=c++ - CXXFLAGS=-O2 - } +/bin/sh -c "${CXX} --version" > /dev/null 2>&1 || { CXX=c++ ; CXXFLAGS=-O2 ; } # Loop over all args args= @@ -42,11 +38,12 @@ while [ $# != 0 ] ; do shift # Add the argument quoted to args - args="${args} \"${option}\"" + if [ -z "${args}" ] ; then args="\"${option}\"" + else args="${args} \"${option}\"" ; fi # Split out the argument for options that take them case ${option} in - *=*) optarg=`echo ${option} | sed -e 's,^[^=]*=,,;s,/$,,'` ;; + *=*) optarg=`echo "${option}" | sed -e 's,^[^=]*=,,;s,/$,,'` ;; esac # Process the options @@ -125,7 +122,7 @@ if [ -z "${srcdir}" ] ; then if [ ! -r "${srcdir}/${srctrigger}" ] ; then srcdir=.. ; fi if [ ! -r "${srcdir}/${srctrigger}" ] ; then ## the sed command below emulates the dirname command - srcdir=`echo $0 | sed -e 's,[^/]*$,,;s,/$,,;s,^$,.,'` + srcdir=`echo "$0" | sed -e 's,[^/]*$,,;s,/$,,;s,^$,.,'` fi fi @@ -148,7 +145,7 @@ if [ -z "${no_create}" ] ; then # Run this file to recreate the current configuration. # # This script is free software: you have unlimited permission -# to copy, distribute and modify it. +# to copy, distribute, and modify it. exec /bin/sh $0 ${args} --no-create EOF @@ -170,11 +167,11 @@ echo "LDFLAGS = ${LDFLAGS}" rm -f Makefile cat > Makefile << EOF # Makefile for Tarlz - Archiver with multimember lzip compression -# Copyright (C) 2013-2019 Antonio Diaz Diaz. +# Copyright (C) 2013-2020 Antonio Diaz Diaz. # This file was generated automatically by configure. Don't edit. # # This Makefile is free software: you have unlimited permission -# to copy, distribute and modify it. +# to copy, distribute, and modify it. pkgname = ${pkgname} pkgversion = ${pkgversion} @@ -194,5 +191,5 @@ EOF cat "${srcdir}/Makefile.in" >> Makefile echo "OK. Now you can run make." -echo "If make fails, verify that the lzlib compression library is correctly" +echo "If make fails, verify that the compression library lzlib is correctly" echo "installed (see INSTALL)." @@ -1,18 +1,18 @@ -/* Tarlz - Archiver with multimember lzip compression - Copyright (C) 2013-2019 Antonio Diaz Diaz. +/* Tarlz - Archiver with multimember lzip compression + Copyright (C) 2013-2020 Antonio Diaz Diaz. - 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 of the License, or - (at your option) any later version. + 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 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. + 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. - You should have received a copy of the GNU General Public License - along with this program. If not, see <http://www.gnu.org/licenses/>. + You should have received a copy of the GNU General Public License + along with this program. If not, see <http://www.gnu.org/licenses/>. */ #define _FILE_OFFSET_BITS 64 @@ -30,7 +30,10 @@ #include <unistd.h> #include <sys/stat.h> #include <sys/types.h> +#if !defined __FreeBSD__ && !defined __OpenBSD__ && !defined __NetBSD__ && \ + !defined __DragonFly__ && !defined __APPLE__ #include <sys/sysmacros.h> // for major, minor +#endif #include <ftw.h> #include <grp.h> #include <pwd.h> @@ -40,16 +43,10 @@ #include "tarlz.h" -const CRC32 crc32c( true ); - -int cl_owner = -1; // global vars needed by add_member -int cl_group = -1; -int cl_data_size = 0; -Solidity solidity = bsolid; - namespace { -LZ_Encoder * encoder = 0; // local vars needed by add_member +const Cl_options * gcl_opts = 0; // local vars needed by add_member +LZ_Encoder * encoder = 0; const char * archive_namep = 0; unsigned long long partial_data_size = 0; // size of current block Resizable_buffer grbuf; // extended header + data @@ -82,8 +79,7 @@ public: bool option_C_after_relative_filename( const Arg_parser & parser ) { for( int i = 0; i < parser.arguments(); ++i ) - if( !parser.code( i ) && parser.argument( i ).size() && - parser.argument( i )[0] != '/' ) // relative_filename + if( nonempty_arg( parser, i ) && parser.argument( i )[0] != '/' ) while( ++i < parser.arguments() ) if( parser.code( i ) == 'C' ) return true; return false; @@ -92,7 +88,8 @@ bool option_C_after_relative_filename( const Arg_parser & parser ) /* Check archive type. Return position of EOF blocks or -1 if failure. If remove_eof, leave fd file pos at beginning of the EOF blocks. - Else, leave fd file pos at 0. */ + Else, leave fd file pos at 0. +*/ long long check_appendable( const int fd, const bool remove_eof ) { struct stat st; // fd must be regular @@ -146,7 +143,8 @@ long long check_appendable( const int fd, const bool remove_eof ) /* Skip all tar headers. Return position of EOF blocks or -1 if failure. If remove_eof, leave fd file pos at beginning of the EOF blocks. - Else, leave fd file pos at 0. */ + Else, leave fd file pos at 0. +*/ long long check_uncompressed_appendable( const int fd, const bool remove_eof ) { struct stat st; // fd must be regular @@ -278,7 +276,7 @@ int add_member( const char * const filename, const struct stat *, const int infd = file_size ? open_instream( filename ) : -1; if( file_size && infd < 0 ) { set_error_status( 1 ); return 0; } - if( encoder && solidity == bsolid && + if( encoder && gcl_opts->solidity == bsolid && block_is_full( extended, file_size, partial_data_size ) && !archive_write( 0, 0 ) ) return 1; @@ -313,7 +311,8 @@ int add_member( const char * const filename, const struct stat *, if( close( infd ) != 0 ) { show_file_error( filename, "Error closing file", errno ); return 1; } } - if( encoder && solidity == no_solid && !archive_write( 0, 0 ) ) return 1; + if( encoder && gcl_opts->solidity == no_solid && !archive_write( 0, 0 ) ) + return 1; if( verbosity >= 1 ) std::fprintf( stderr, "%s\n", filename ); return 0; } @@ -382,7 +381,8 @@ bool write_eof_records( const int outfd, const bool compressed ) /* Removes any amount of leading "./" and '/' strings from filename. - Optionally also removes prefixes containing a ".." component. */ + Optionally also removes prefixes containing a ".." component. +*/ const char * remove_leading_dotslash( const char * const filename, const bool dotdot ) { @@ -418,7 +418,7 @@ bool fill_headers( const char * const filename, Extended & extended, Tar_header header, long long & file_size, const int flag ) { struct stat st; - if( hstat( filename, &st ) != 0 ) + if( hstat( filename, &st, gcl_opts->dereference ) != 0 ) { show_file_error( filename, "Can't stat input file", errno ); set_error_status( 1 ); return false; } if( file_is_the_archive( st ) ) @@ -431,14 +431,15 @@ bool fill_headers( const char * const filename, Extended & extended, print_octal( header + mode_o, mode_l - 1, mode & ( S_ISUID | S_ISGID | S_ISVTX | S_IRWXU | S_IRWXG | S_IRWXO ) ); - const uid_t uid = ( cl_owner >= 0 ) ? (uid_t)cl_owner : st.st_uid; - const gid_t gid = ( cl_group >= 0 ) ? (gid_t)cl_group : st.st_gid; + const uid_t uid = (gcl_opts->owner >= 0) ? (uid_t)gcl_opts->owner : st.st_uid; + const gid_t gid = (gcl_opts->group >= 0) ? (gid_t)gcl_opts->group : st.st_gid; if( uid >= 2 << 20 || gid >= 2 << 20 ) { show_file_error( filename, "uid or gid is larger than 2_097_151." ); set_error_status( 1 ); return false; } print_octal( header + uid_o, uid_l - 1, uid ); print_octal( header + gid_o, gid_l - 1, gid ); - const unsigned long long mtime = (st.st_mtime >= 0) ? st.st_mtime : 0; + const unsigned long long mtime = ( gcl_opts->mtime >= 0 ) ? gcl_opts->mtime : + ( ( st.st_mtime >= 0 ) ? st.st_mtime : 0 ); if( mtime >= 1ULL << 33 ) { show_file_error( filename, "mtime is out of ustar range [0, 8_589_934_591]." ); set_error_status( 1 ); return false; } @@ -522,7 +523,7 @@ bool block_is_full( const Extended & extended, { const unsigned long long member_size = // may overflow 'long long' header_size + extended.full_size() + round_up( file_size ); - const unsigned long long target_size = cl_data_size; + const unsigned long long target_size = gcl_opts->data_size; if( partial_data_size >= target_size || ( partial_data_size >= min_data_size && partial_data_size + member_size / 2 > target_size ) ) @@ -573,15 +574,14 @@ bool has_lz_ext( const std::string & name ) } -int concatenate( const std::string & archive_name, const Arg_parser & parser, - const int filenames ) +int concatenate( const Cl_options & cl_opts ) { - if( !filenames ) + if( cl_opts.filenames <= 0 ) { if( verbosity >= 1 ) show_error( "Nothing to concatenate." ); return 0; } - const bool to_stdout = archive_name.empty(); - archive_namep = to_stdout ? "(stdout)" : archive_name.c_str(); + const bool to_stdout = cl_opts.archive_name.empty(); + archive_namep = to_stdout ? "(stdout)" : cl_opts.archive_name.c_str(); const int outfd = - to_stdout ? STDOUT_FILENO : open_outstream( archive_name, false ); + to_stdout ? STDOUT_FILENO : open_outstream( cl_opts.archive_name, false ); if( outfd < 0 ) return 1; if( !to_stdout && !file_is_the_archive.init( outfd ) ) { show_file_error( archive_namep, "Can't stat", errno ); return 1; } @@ -589,7 +589,7 @@ int concatenate( const std::string & archive_name, const Arg_parser & parser, if( to_stdout ) compressed = -1; // unknown else { - compressed = has_lz_ext( archive_name ); // default value + compressed = has_lz_ext( cl_opts.archive_name ); // default value long long pos = check_appendable( outfd, true ); if( pos > 0 ) compressed = true; else if( pos < 0 ) @@ -606,11 +606,10 @@ int concatenate( const std::string & archive_name, const Arg_parser & parser, int retval = 0; bool eof_pending = false; - for( int i = 0; i < parser.arguments(); ++i ) // copy archives + for( int i = 0; i < cl_opts.parser.arguments(); ++i ) // copy archives { - if( parser.code( i ) ) continue; // skip options - if( parser.argument( i ).empty() ) continue; // skip empty names - const char * const filename = parser.argument( i ).c_str(); + if( !nonempty_arg( cl_opts.parser, i ) ) continue; // skip opts, empty names + const char * const filename = cl_opts.parser.argument( i ).c_str(); if( Exclude::excluded( filename ) ) continue; // skip excluded files const int infd = open_instream( filename ); if( infd < 0 ) { retval = 1; break; } @@ -649,10 +648,7 @@ int concatenate( const std::string & archive_name, const Arg_parser & parser, } -int encode( const std::string & archive_name, const Arg_parser & parser, - const int filenames, const int level, const int num_workers, - const int out_slots, const int debug_level, const bool append, - const bool dereference ) +int encode( Cl_options & cl_opts ) { struct Lzma_options { @@ -671,15 +667,17 @@ int encode( const std::string & archive_name, const Arg_parser & parser, { 1 << 24, 68 }, // -7 { 3 << 23, 132 }, // -8 { 1 << 25, 273 } }; // -9 - const bool compressed = ( level >= 0 && level <= 9 ); - const bool to_stdout = archive_name.empty(); - archive_namep = to_stdout ? "(stdout)" : archive_name.c_str(); + const bool compressed = ( cl_opts.level >= 0 && cl_opts.level <= 9 ); + const bool to_stdout = cl_opts.archive_name.empty(); + archive_namep = to_stdout ? "(stdout)" : cl_opts.archive_name.c_str(); + gcl_opts = &cl_opts; - if( !to_stdout && !compressed && has_lz_ext( archive_name ) ) + if( !to_stdout && !compressed && has_lz_ext( cl_opts.archive_name ) ) { show_file_error( archive_namep, "Uncompressed mode incompatible with .lz extension." ); return 2; } - if( !filenames ) + const bool append = cl_opts.program_mode == m_append; + if( cl_opts.filenames <= 0 ) { if( !append && !to_stdout ) // create archive { show_error( "Cowardly refusing to create an empty archive.", 0, true ); @@ -691,10 +689,11 @@ int encode( const std::string & archive_name, const Arg_parser & parser, if( to_stdout ) // create/append to stdout goutfd = STDOUT_FILENO; else if( !append ) // create archive - { if( ( goutfd = open_outstream( archive_name ) ) < 0 ) return 1; } + { if( ( goutfd = open_outstream( cl_opts.archive_name ) ) < 0 ) return 1; } else // append to archive { - if( ( goutfd = open_outstream( archive_name, false ) ) < 0 ) return 1; + if( ( goutfd = open_outstream( cl_opts.archive_name, false ) ) < 0 ) + return 1; if( compressed && check_appendable( goutfd, true ) < 0 ) { show_file_error( archive_namep, "This does not look like an appendable tar.lz archive." ); return 2; } @@ -708,24 +707,24 @@ int encode( const std::string & archive_name, const Arg_parser & parser, if( compressed ) { - const int dictionary_size = option_mapping[level].dictionary_size; - if( cl_data_size <= 0 ) + const int dictionary_size = option_mapping[cl_opts.level].dictionary_size; + if( cl_opts.data_size <= 0 ) { - if( level == 0 ) cl_data_size = 1 << 20; - else cl_data_size = 2 * dictionary_size; + if( cl_opts.level == 0 ) cl_opts.data_size = 1 << 20; + else cl_opts.data_size = 2 * dictionary_size; } /* CWD is not per-thread; multi-threaded --create can't be used if a -C option appears after a relative filename in the command line. */ - if( solidity != asolid && solidity != solid && num_workers > 0 && - !option_C_after_relative_filename( parser ) ) + if( cl_opts.solidity != asolid && cl_opts.solidity != solid && + cl_opts.num_workers > 0 && + !option_C_after_relative_filename( cl_opts.parser ) ) { // show_file_error( archive_namep, "Multi-threaded --create" ); - return encode_lz( archive_namep, parser, dictionary_size, - option_mapping[level].match_len_limit, num_workers, - goutfd, out_slots, debug_level, dereference ); + return encode_lz( cl_opts, archive_namep, dictionary_size, + option_mapping[cl_opts.level].match_len_limit, goutfd ); } encoder = LZ_compress_open( dictionary_size, - option_mapping[level].match_len_limit, LLONG_MAX ); + option_mapping[cl_opts.level].match_len_limit, LLONG_MAX ); if( !encoder || LZ_compress_errno( encoder ) != LZ_ok ) { if( !encoder || LZ_compress_errno( encoder ) == LZ_mem_error ) @@ -737,16 +736,16 @@ int encode( const std::string & archive_name, const Arg_parser & parser, } int retval = 0; - for( int i = 0; i < parser.arguments(); ++i ) // parse command line + for( int i = 0; i < cl_opts.parser.arguments(); ++i ) // parse command line { - const int code = parser.code( i ); - const std::string & arg = parser.argument( i ); + const int code = cl_opts.parser.code( i ); + const std::string & arg = cl_opts.parser.argument( i ); const char * filename = arg.c_str(); if( code == 'C' && chdir( filename ) != 0 ) { show_file_error( filename, "Error changing working directory", errno ); retval = 1; break; } if( code ) continue; // skip options - if( parser.argument( i ).empty() ) continue; // skip empty names + if( cl_opts.parser.argument( i ).empty() ) continue; // skip empty names std::string deslashed; // arg without trailing slashes unsigned len = arg.size(); while( len > 1 && arg[len-1] == '/' ) --len; @@ -758,9 +757,9 @@ int encode( const std::string & archive_name, const Arg_parser & parser, { show_file_error( filename, "Can't stat input file", errno ); set_error_status( 1 ); } else if( ( retval = nftw( filename, add_member, 16, - dereference ? 0 : FTW_PHYS ) ) != 0 ) + cl_opts.dereference ? 0 : FTW_PHYS ) ) != 0 ) break; // write error - else if( encoder && solidity == dsolid && !archive_write( 0, 0 ) ) + else if( encoder && cl_opts.solidity == dsolid && !archive_write( 0, 0 ) ) { retval = 1; break; } } @@ -770,7 +769,8 @@ int encode( const std::string & archive_name, const Arg_parser & parser, uint8_t buf[bufsize]; std::memset( buf, 0, bufsize ); if( encoder && - ( solidity == asolid || ( solidity == bsolid && partial_data_size ) ) && + ( cl_opts.solidity == asolid || + ( cl_opts.solidity == bsolid && partial_data_size ) ) && !archive_write( 0, 0 ) ) retval = 1; // flush encoder else if( !archive_write( buf, bufsize ) || ( encoder && !archive_write( 0, 0 ) ) ) retval = 1; diff --git a/create_lz.cc b/create_lz.cc index 9cfdedd..a6a7146 100644 --- a/create_lz.cc +++ b/create_lz.cc @@ -1,18 +1,18 @@ -/* Tarlz - Archiver with multimember lzip compression - Copyright (C) 2013-2019 Antonio Diaz Diaz. +/* Tarlz - Archiver with multimember lzip compression + Copyright (C) 2013-2020 Antonio Diaz Diaz. - 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 of the License, or - (at your option) any later version. + 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 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. + 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. - You should have received a copy of the GNU General Public License - along with this program. If not, see <http://www.gnu.org/licenses/>. + You should have received a copy of the GNU General Public License + along with this program. If not, see <http://www.gnu.org/licenses/>. */ #define _FILE_OFFSET_BITS 64 @@ -39,9 +39,10 @@ namespace { +const Cl_options * gcl_opts = 0; // local vars needed by add_member_lz enum { max_packet_size = 1 << 20 }; class Packet_courier; -Packet_courier * courierp = 0; // local vars needed by add_member_lz +Packet_courier * courierp = 0; unsigned long long partial_data_size = 0; // size of current block @@ -156,7 +157,7 @@ public: If filename.empty() (end of lzip member token), move to next queue. */ void receive_packet( const Ipacket * const ipacket ) { - if( ipacket->filename.size() ) + if( !ipacket->filename.empty() ) slot_tally.get_slot(); // wait for a free slot xlock( &imutex ); ipacket_queues[receive_worker_id].push( ipacket ); @@ -184,7 +185,7 @@ public: } xunlock( &imutex ); if( ipacket ) - { if( ipacket->filename.size() ) slot_tally.leave_slot(); } + { if( !ipacket->filename.empty() ) slot_tally.leave_slot(); } else { // notify muxer when last worker exits @@ -270,13 +271,13 @@ int add_member_lz( const char * const filename, const struct stat *, if( !fill_headers( filename, *extended, header, file_size, flag ) ) { delete[] header; delete extended; return 0; } - if( solidity == bsolid && + if( gcl_opts->solidity == bsolid && block_is_full( *extended, file_size, partial_data_size ) ) courierp->receive_packet( new Ipacket ); // end of group courierp->receive_packet( new Ipacket( filename, file_size, extended, header ) ); - if( solidity == no_solid ) // one tar member per group + if( gcl_opts->solidity == no_solid ) // one tar member per group courierp->receive_packet( new Ipacket ); if( verbosity >= 1 ) std::fprintf( stderr, "%s\n", filename ); return 0; @@ -285,31 +286,30 @@ int add_member_lz( const char * const filename, const struct stat *, struct Grouper_arg { + const Cl_options * cl_opts; Packet_courier * courier; - const Arg_parser * parser; - bool dereference; }; /* Package metadata of the files to be archived and pass them to the - courier for distribution to workers. */ + courier for distribution to workers. +*/ extern "C" void * grouper( void * arg ) { const Grouper_arg & tmp = *(const Grouper_arg *)arg; + const Cl_options & cl_opts = *tmp.cl_opts; Packet_courier & courier = *tmp.courier; - const Arg_parser & parser = *tmp.parser; - const bool dereference = tmp.dereference; - for( int i = 0; i < parser.arguments(); ++i ) // parse command line + for( int i = 0; i < cl_opts.parser.arguments(); ++i ) // parse command line { - const int code = parser.code( i ); - const std::string & arg = parser.argument( i ); + const int code = cl_opts.parser.code( i ); + const std::string & arg = cl_opts.parser.argument( i ); const char * filename = arg.c_str(); if( code == 'C' && chdir( filename ) != 0 ) { show_file_error( filename, "Error changing working directory", errno ); cleanup_and_fail(); } if( code ) continue; // skip options - if( parser.argument( i ).empty() ) continue; // skip empty names + if( cl_opts.parser.argument( i ).empty() ) continue; // skip empty names std::string deslashed; // arg without trailing slashes unsigned len = arg.size(); while( len > 1 && arg[len-1] == '/' ) --len; @@ -321,13 +321,13 @@ extern "C" void * grouper( void * arg ) { show_file_error( filename, "Can't stat input file", errno ); set_error_status( 1 ); } else if( nftw( filename, add_member_lz, 16, - dereference ? 0 : FTW_PHYS ) != 0 ) - cleanup_and_fail(); // write error or oom - else if( solidity == dsolid ) // end of group + cl_opts.dereference ? 0 : FTW_PHYS ) != 0 ) + cleanup_and_fail(); // write error or OOM + else if( cl_opts.solidity == dsolid ) // end of group courier.receive_packet( new Ipacket ); } - if( solidity == bsolid && partial_data_size ) // finish last block + if( cl_opts.solidity == bsolid && partial_data_size ) // finish last block { partial_data_size = 0; courierp->receive_packet( new Ipacket ); } courier.finish(); // no more packets to send return 0; @@ -336,7 +336,8 @@ extern "C" void * grouper( void * arg ) /* Writes ibuf to encoder. To minimize dictionary size, it does not read from encoder until encoder's input buffer is full or finish is true. - Sends opacket to courier and allocates new obuf each time obuf is full. */ + Sends opacket to courier and allocates new obuf each time obuf is full. +*/ void loop_encode( const uint8_t * const ibuf, const int isize, uint8_t * & obuf, int & opos, Packet_courier & courier, LZ_Encoder * const encoder, const int worker_id, @@ -395,7 +396,8 @@ struct Worker_arg /* Get ipackets from courier, compress headers and file data, and give the - opackets produced to courier. */ + opackets produced to courier. +*/ extern "C" void * cworker( void * arg ) { const Worker_arg & tmp = *(const Worker_arg *)arg; @@ -502,7 +504,8 @@ extern "C" void * cworker( void * arg ) /* Get from courier the processed and sorted packets, and write - their contents to the output archive. */ + their contents to the output archive. +*/ void muxer( Packet_courier & courier, const int outfd ) { while( true ) @@ -521,24 +524,24 @@ void muxer( Packet_courier & courier, const int outfd ) // init the courier, then start the grouper and the workers and call the muxer -int encode_lz( const char * const archive_namep, const Arg_parser & parser, +int encode_lz( const Cl_options & cl_opts, const char * const archive_namep, const int dictionary_size, const int match_len_limit, - const int num_workers, const int outfd, const int out_slots, - const int debug_level, const bool dereference ) + const int outfd ) { const int in_slots = 65536; // max small files (<=512B) in 64 MiB + const int num_workers = cl_opts.num_workers; const int total_in_slots = ( INT_MAX / num_workers >= in_slots ) ? num_workers * in_slots : INT_MAX; + gcl_opts = &cl_opts; /* If an error happens after any threads have been started, exit must be called before courier goes out of scope. */ - Packet_courier courier( num_workers, total_in_slots, out_slots ); + Packet_courier courier( num_workers, total_in_slots, cl_opts.out_slots ); courierp = &courier; // needed by add_member_lz Grouper_arg grouper_arg; + grouper_arg.cl_opts = &cl_opts; grouper_arg.courier = &courier; - grouper_arg.parser = &parser; - grouper_arg.dereference = dereference; pthread_t grouper_thread; int errcode = pthread_create( &grouper_thread, 0, grouper, &grouper_arg ); @@ -582,7 +585,7 @@ int encode_lz( const char * const archive_namep, const Arg_parser & parser, { show_file_error( archive_namep, "Error closing archive", errno ); retval = 1; } - if( debug_level & 1 ) + if( cl_opts.debug_level & 1 ) std::fprintf( stderr, "any worker tried to consume from grouper %8u times\n" "any worker had to wait %8u times\n" diff --git a/decode_lz.cc b/decode_lz.cc new file mode 100644 index 0000000..71c699b --- /dev/null +++ b/decode_lz.cc @@ -0,0 +1,536 @@ +/* Tarlz - Archiver with multimember lzip compression + Copyright (C) 2013-2020 Antonio Diaz Diaz. + + 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 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. + + You should have received a copy of the GNU General Public License + along with this program. If not, see <http://www.gnu.org/licenses/>. +*/ + +#define _FILE_OFFSET_BITS 64 + +#include <algorithm> +#include <cerrno> +#include <climits> +#include <cstdio> +#include <cstdlib> +#include <cstring> +#include <queue> +#include <string> +#include <vector> +#include <pthread.h> +#include <stdint.h> +#include <unistd.h> +#include <sys/stat.h> +#include <lzlib.h> + +#include "arg_parser.h" +#include "tarlz.h" +#include "lzip_index.h" +#include "archive_reader.h" + +/* When a problem is detected by any worker: + - the worker requests mastership and returns. + - the courier discards new packets received or collected. + - the other workers return. + - the muxer drains the queue and returns. */ + +/* Returns the number of bytes really read. + If (returned value < size) and (errno == 0), means EOF was reached. +*/ +int preadblock( const int fd, uint8_t * const buf, const int size, + const long long pos ) + { + int sz = 0; + errno = 0; + while( sz < size ) + { + const int n = pread( fd, buf + sz, size - sz, pos + sz ); + if( n > 0 ) sz += n; + else if( n == 0 ) break; // EOF + else if( errno != EINTR ) break; + errno = 0; + } + return sz; + } + + +namespace { + +/* Returns the number of bytes really written. + If (returned value < size), it is always an error. +*//* +int pwriteblock( const int fd, const uint8_t * const buf, const int size, + const long long pos ) + { + int sz = 0; + errno = 0; + while( sz < size ) + { + const int n = pwrite( fd, buf + sz, size - sz, pos + sz ); + if( n > 0 ) sz += n; + else if( n < 0 && errno != EINTR ) break; + errno = 0; + } + return sz; + } +*/ + +const char * const other_msg = "Other worker found an error."; + +struct Packet // member name and metadata or error message + { + enum Status { ok, member_done, diag, error }; + long member_id; // lzip member containing the header of this tar member + std::string line; // member name and metadata ready to print, if any + Status status; // diagnostics and errors go to stderr + Packet( const long i, const char * const msg, const Status s = ok ) + : member_id( i ), line( msg ), status( s ) {} + }; + + +class Packet_courier // moves packets around + { +public: + unsigned ocheck_counter; + unsigned owait_counter; +private: + long error_member_id; // first lzip member with error/misalign/eof + int deliver_worker_id; // worker queue currently delivering packets + int master_worker_id; // worker in charge if error/misalignment/eof + std::vector< std::queue< const Packet * > > opacket_queues; + int num_working; // number of workers still running + const int num_workers; // number of workers + const unsigned out_slots; // max output packets per queue + pthread_mutex_t omutex; + pthread_cond_t oav_or_exit; // output packet available or all workers exited + std::vector< pthread_cond_t > slot_av; // output slot available + pthread_cond_t check_master; + bool eof_found_; + + Packet_courier( const Packet_courier & ); // declared as private + void operator=( const Packet_courier & ); // declared as private + +public: + Packet_courier( const int workers, const int slots ) + : ocheck_counter( 0 ), owait_counter( 0 ), + error_member_id( -1 ), deliver_worker_id( 0 ), master_worker_id( -1 ), + opacket_queues( workers ), num_working( workers ), + num_workers( workers ), out_slots( slots ), slot_av( workers ), + eof_found_( false ) + { + xinit_mutex( &omutex ); xinit_cond( &oav_or_exit ); + for( unsigned i = 0; i < slot_av.size(); ++i ) xinit_cond( &slot_av[i] ); + xinit_cond( &check_master ); + } + + ~Packet_courier() + { + xdestroy_cond( &check_master ); + for( unsigned i = 0; i < slot_av.size(); ++i ) xdestroy_cond( &slot_av[i] ); + xdestroy_cond( &oav_or_exit ); xdestroy_mutex( &omutex ); + } + + bool eof_found() const { return eof_found_; } + void report_eof() { eof_found_ = true; } + + bool mastership_granted() const { return master_worker_id >= 0; } + + bool request_mastership( const long member_id, const int worker_id ) + { + xlock( &omutex ); + if( mastership_granted() ) // already granted + { xunlock( &omutex ); return ( master_worker_id == worker_id ); } + if( error_member_id < 0 || error_member_id > member_id ) + error_member_id = member_id; + while( !mastership_granted() && ( worker_id != deliver_worker_id || + !opacket_queues[deliver_worker_id].empty() ) ) + xwait( &check_master, &omutex ); + if( !mastership_granted() && worker_id == deliver_worker_id && + opacket_queues[deliver_worker_id].empty() ) + { + master_worker_id = worker_id; // grant mastership + for( int i = 0; i < num_workers; ++i ) // delete all packets + while( !opacket_queues[i].empty() ) + opacket_queues[i].pop(); + xbroadcast( &check_master ); + xunlock( &omutex ); + return true; + } + xunlock( &omutex ); + return false; // mastership granted to another worker + } + + void worker_finished() + { + // notify muxer when last worker exits + xlock( &omutex ); + if( --num_working == 0 ) xsignal( &oav_or_exit ); + xunlock( &omutex ); + } + + /* Collect a packet from a worker. + If a packet is rejected, the worker must terminate. */ + bool collect_packet( const long member_id, const int worker_id, + const char * const msg, + const Packet::Status status = Packet::ok ) + { + const Packet * const opacket = new Packet( member_id, msg, status ); + xlock( &omutex ); + if( ( mastership_granted() && master_worker_id != worker_id ) || + ( error_member_id >= 0 && error_member_id < opacket->member_id ) ) + { xunlock( &omutex ); delete opacket; return false; } // reject packet + while( opacket_queues[worker_id].size() >= out_slots ) + xwait( &slot_av[worker_id], &omutex ); + opacket_queues[worker_id].push( opacket ); + if( worker_id == deliver_worker_id ) xsignal( &oav_or_exit ); + xunlock( &omutex ); + return true; + } + + /* Deliver a packet to muxer. + If packet.status == Packet::member_done, move to next queue. + If packet.line.empty(), wait again (empty lzip member). */ + const Packet * deliver_packet() + { + const Packet * opacket = 0; + xlock( &omutex ); + ++ocheck_counter; + while( true ) + { + while( opacket_queues[deliver_worker_id].empty() && num_working > 0 ) + { + ++owait_counter; + if( !mastership_granted() && error_member_id >= 0 ) + xbroadcast( &check_master ); // mastership requested not yet granted + xwait( &oav_or_exit, &omutex ); + } + if( opacket_queues[deliver_worker_id].empty() ) break; + opacket = opacket_queues[deliver_worker_id].front(); + opacket_queues[deliver_worker_id].pop(); + if( opacket_queues[deliver_worker_id].size() + 1 == out_slots ) + xsignal( &slot_av[deliver_worker_id] ); + if( opacket->status == Packet::member_done && !mastership_granted() ) + { if( ++deliver_worker_id >= num_workers ) deliver_worker_id = 0; } + if( !opacket->line.empty() ) break; + delete opacket; opacket = 0; + } + xunlock( &omutex ); + return opacket; + } + + bool finished() // all packets delivered to muxer + { + if( num_working != 0 ) return false; + for( int i = 0; i < num_workers; ++i ) + if( !opacket_queues[i].empty() ) return false; + return true; + } + }; + + +const char * skip_member_lz( Archive_reader_i & ar, Packet_courier & courier, + const Extended & extended, const long member_id, + const int worker_id ) + { + if( ar.skip_member( extended ) != 0 ) return ar.e_msg(); + if( !courier.collect_packet( member_id, worker_id, "", + ar.at_member_end() ? Packet::member_done : Packet::ok ) ) + return other_msg; + return 0; + } + + +const char * compare_member_lz( const Cl_options & cl_opts, + Archive_reader_i & ar, Packet_courier & courier, + const Extended & extended, const Tar_header header, + Resizable_buffer & rbuf, const long member_id, + const int worker_id ) + { + if( verbosity < 1 ) rbuf()[0] = 0; + else if( !format_member_name( extended, header, rbuf, verbosity > 1 ) ) + return mem_msg; + std::string estr, ostr; + const bool stat_differs = + !compare_file_type( estr, ostr, cl_opts, extended, header ); + if( ( rbuf()[0] && !courier.collect_packet( member_id, worker_id, rbuf(), + Packet::ok ) ) || + ( estr.size() && !courier.collect_packet( member_id, worker_id, + estr.c_str(), Packet::diag ) ) || + ( ostr.size() && !courier.collect_packet( member_id, worker_id, + ostr.c_str(), Packet::ok ) ) || + ( extended.file_size() <= 0 && ar.at_member_end() && + !courier.collect_packet( member_id, worker_id, "", Packet::member_done ) ) ) + return other_msg; + if( extended.file_size() <= 0 ) return 0; + const Typeflag typeflag = (Typeflag)header[typeflag_o]; + if( ( typeflag != tf_regular && typeflag != tf_hiperf ) || stat_differs ) + return skip_member_lz( ar, courier, extended, member_id, worker_id ); + // else compare file contents + const char * const filename = extended.path().c_str(); + const int infd2 = open_instream( filename ); + if( infd2 < 0 ) { set_error_status( 1 ); + return skip_member_lz( ar, courier, extended, member_id, worker_id ); } + int retval = compare_file_contents( estr, ostr, ar, extended.file_size(), + filename, infd2 ); + if( retval ) return ar.e_msg(); + if( ( estr.size() && !courier.collect_packet( member_id, worker_id, + estr.c_str(), Packet::diag ) ) || + ( ostr.size() && !courier.collect_packet( member_id, worker_id, + ostr.c_str(), Packet::ok ) ) || + ( ar.at_member_end() && + !courier.collect_packet( member_id, worker_id, "", Packet::member_done ) ) ) + return other_msg; + return 0; + } + + +const char * list_member_lz( Archive_reader_i & ar, Packet_courier & courier, + const Extended & extended, const Tar_header header, + Resizable_buffer & rbuf, const long member_id, + const int worker_id ) + { + if( verbosity < 0 ) rbuf()[0] = 0; + else if( !format_member_name( extended, header, rbuf, verbosity > 0 ) ) + return mem_msg; + const int ret = ar.skip_member( extended ); // print name even on error + if( !courier.collect_packet( member_id, worker_id, rbuf(), + ar.at_member_end() ? Packet::member_done : Packet::ok ) ) + return other_msg; + if( ret != 0 ) return ar.e_msg(); + return 0; + } + + +struct Worker_arg + { + const Cl_options * cl_opts; + const Archive_descriptor * ad; + Packet_courier * courier; + std::vector< char > * name_pending; + int worker_id; + int num_workers; + }; + + +/* Read lzip members from archive, decode their tar members, and give the + packets produced to courier. +*/ +extern "C" void * dworker( void * arg ) + { + const Worker_arg & tmp = *(const Worker_arg *)arg; + const Cl_options & cl_opts = *tmp.cl_opts; + const Archive_descriptor & ad = *tmp.ad; + Packet_courier & courier = *tmp.courier; + std::vector< char > & name_pending = *tmp.name_pending; + const int worker_id = tmp.worker_id; + const int num_workers = tmp.num_workers; + + bool master = false; + Resizable_buffer rbuf; + Archive_reader_i ar( ad ); // 1 of N parallel readers + if( !rbuf.size() || ar.fatal() ) + { if( courier.request_mastership( worker_id, worker_id ) ) + courier.collect_packet( worker_id, worker_id, mem_msg, Packet::error ); + goto done; } + + for( long i = worker_id; !master && i < ad.lzip_index.members(); i += num_workers ) + { + if( ad.lzip_index.dblock( i ).size() <= 0 ) // empty lzip member + { + if( courier.collect_packet( i, worker_id, "", Packet::member_done ) ) + continue; else break; + } + + long long data_end = ad.lzip_index.dblock( i ).end(); + Extended extended; // metadata from extended records + bool prev_extended = false; // prev header was extended + ar.set_member( i ); // prepare for new member + while( true ) // process one tar header per iteration + { + if( ar.data_pos() >= data_end ) // dblock.end or udata_size + { + if( ar.data_pos() == data_end && !prev_extended ) break; + // member end exceeded or ends in extended, process rest of file + if( !courier.request_mastership( i, worker_id ) ) goto done; + master = true; + if( data_end >= ad.lzip_index.udata_size() ) + { courier.collect_packet( i, worker_id, end_msg, Packet::error ); + goto done; } + data_end = ad.lzip_index.udata_size(); + if( ar.data_pos() == data_end && !prev_extended ) break; + } + Tar_header header; + const int ret = ar.read( header, header_size ); + if( ret != 0 ) + { if( courier.request_mastership( i, worker_id ) ) + courier.collect_packet( i, worker_id, ar.e_msg(), Packet::error ); + goto done; } + if( !verify_ustar_chksum( header ) ) + { + if( !courier.request_mastership( i, worker_id ) ) goto done; + if( block_is_zero( header, header_size ) ) // EOF + { + if( !prev_extended || cl_opts.permissive ) courier.report_eof(); + else courier.collect_packet( i, worker_id, fv_msg1, Packet::error ); + goto done; + } + courier.collect_packet( i, worker_id, ( ar.data_pos() > header_size ) ? + bad_hdr_msg : posix_lz_msg, Packet::error ); + goto done; + } + + const Typeflag typeflag = (Typeflag)header[typeflag_o]; + if( typeflag == tf_global ) + { + const char * msg = 0; + Extended dummy; // global headers are parsed and ignored + if( prev_extended && !cl_opts.permissive ) msg = fv_msg2; + else if( ar.parse_records( dummy, header, rbuf, true ) != 0 ) + msg = gblrec_msg; + else + { + if( ar.data_pos() == data_end && // end of lzip member or EOF + !courier.collect_packet( i, worker_id, "", Packet::member_done ) ) + goto done; + continue; + } + if( courier.request_mastership( i, worker_id ) ) + courier.collect_packet( i, worker_id, msg, Packet::error ); + goto done; + } + if( typeflag == tf_extended ) + { + const char * msg = 0; + if( prev_extended && !cl_opts.permissive ) msg = fv_msg3; + else if( ar.parse_records( extended, header, rbuf, + cl_opts.permissive ) != 0 ) msg = extrec_msg; + else if( !extended.crc_present() && cl_opts.missing_crc ) + msg = mcrc_msg; + else { prev_extended = true; continue; } + if( courier.request_mastership( i, worker_id ) ) + courier.collect_packet( i, worker_id, msg, Packet::error ); + goto done; + } + prev_extended = false; + + extended.fill_from_ustar( header ); // copy metadata from header + + const char * msg; + if( check_skip_filename( cl_opts, name_pending, extended.path().c_str() ) ) + msg = skip_member_lz( ar, courier, extended, i, worker_id ); + else if( cl_opts.program_mode == m_list ) + msg = list_member_lz( ar, courier, extended, header, rbuf, i, worker_id ); + else msg = compare_member_lz( cl_opts, ar, courier, extended, header, + rbuf, i, worker_id ); + if( msg ) + { if( courier.request_mastership( i, worker_id ) ) + courier.collect_packet( i, worker_id, msg, Packet::error ); + goto done; } + extended.reset(); + } + } +done: + courier.worker_finished(); + return 0; + } + + +/* Get from courier the processed and sorted packets, and print + the member lines on stdout or the diagnostics and errors on stderr. +*/ +void muxer( const char * const archive_namep, Packet_courier & courier ) + { + bool error = false; + while( !error ) + { + const Packet * const opacket = courier.deliver_packet(); + if( !opacket ) break; // queue is empty. all workers exited + + switch( opacket->status ) + { + case Packet::error: + show_file_error( archive_namep, opacket->line.c_str() ); + error = true; break; + case Packet::diag: std::fputs( opacket->line.c_str(), stderr ); break; + default: if( opacket->line.size() ) + { std::fputs( opacket->line.c_str(), stdout ); std::fflush( stdout ); } + } + delete opacket; + } + if( !error && !courier.eof_found() ) // no worker found EOF blocks + { show_file_error( archive_namep, end_msg ); error = true; } + if( error ) cleanup_and_fail( 2 ); + } + +} // end namespace + + +// init the courier, then start the workers and call the muxer. +int decode_lz( const Cl_options & cl_opts, const Archive_descriptor & ad, + std::vector< char > & name_pending ) + { + const int out_slots = 65536; // max small files (<=512B) in 64 MiB + const int num_workers = // limited to number of members + std::min( (long)cl_opts.num_workers, ad.lzip_index.members() ); + + /* If an error happens after any threads have been started, exit must be + called before courier goes out of scope. */ + Packet_courier courier( num_workers, out_slots ); + + Worker_arg * worker_args = new( std::nothrow ) Worker_arg[num_workers]; + pthread_t * worker_threads = new( std::nothrow ) pthread_t[num_workers]; + if( !worker_args || !worker_threads ) { show_error( mem_msg ); return 1; } + for( int i = 0; i < num_workers; ++i ) + { + worker_args[i].cl_opts = &cl_opts; + worker_args[i].ad = &ad; + worker_args[i].courier = &courier; + worker_args[i].name_pending = &name_pending; + worker_args[i].worker_id = i; + worker_args[i].num_workers = num_workers; + const int errcode = + pthread_create( &worker_threads[i], 0, dworker, &worker_args[i] ); + if( errcode ) + { show_error( "Can't create worker threads", errcode ); cleanup_and_fail(); } + } + + muxer( ad.namep, courier ); + + for( int i = num_workers - 1; i >= 0; --i ) + { + const int errcode = pthread_join( worker_threads[i], 0 ); + if( errcode ) + { show_error( "Can't join worker threads", errcode ); cleanup_and_fail(); } + } + delete[] worker_threads; + delete[] worker_args; + + int retval = 0; + if( close( ad.infd ) != 0 ) + { show_file_error( ad.namep, "Error closing archive", errno ); retval = 1; } + + if( retval == 0 ) + for( int i = 0; i < cl_opts.parser.arguments(); ++i ) + if( nonempty_arg( cl_opts.parser, i ) && name_pending[i] ) + { show_file_error( cl_opts.parser.argument( i ).c_str(), + "Not found in archive." ); retval = 1; } + + if( cl_opts.debug_level & 1 ) + std::fprintf( stderr, + "muxer tried to consume from workers %8u times\n" + "muxer had to wait %8u times\n", + courier.ocheck_counter, + courier.owait_counter ); + + if( !courier.finished() ) internal_error( "courier not finished." ); + return final_exit_status( retval, cl_opts.program_mode != m_diff ); + } @@ -1,18 +1,18 @@ -/* Tarlz - Archiver with multimember lzip compression - Copyright (C) 2013-2019 Antonio Diaz Diaz. +/* Tarlz - Archiver with multimember lzip compression + Copyright (C) 2013-2020 Antonio Diaz Diaz. - 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 of the License, or - (at your option) any later version. + 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 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. + 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. - You should have received a copy of the GNU General Public License - along with this program. If not, see <http://www.gnu.org/licenses/>. + You should have received a copy of the GNU General Public License + along with this program. If not, see <http://www.gnu.org/licenses/>. */ #define _FILE_OFFSET_BITS 64 @@ -31,8 +31,8 @@ #include <lzlib.h> #include "arg_parser.h" -#include "lzip_index.h" #include "tarlz.h" +#include "lzip_index.h" namespace { @@ -90,7 +90,7 @@ int tail_copy( const char * const archive_namep, const Arg_parser & parser, retval = 1; } if( retval == 0 ) for( int i = 0; i < parser.arguments(); ++i ) - if( !parser.code( i ) && parser.argument( i ).size() && name_pending[i] ) + if( nonempty_arg( parser, i ) && name_pending[i] ) { show_file_error( parser.argument( i ).c_str(), "Not found in archive." ); retval = 1; @@ -100,32 +100,31 @@ int tail_copy( const char * const archive_namep, const Arg_parser & parser, /* Deleting from a corrupt archive must not worsen the corruption. Stop and - tail-copy as soon as corruption is found. */ -int delete_members( const std::string & archive_name, const Arg_parser & parser, - const int filenames, const bool missing_crc, - const bool permissive ) + tail-copy as soon as corruption is found. +*/ +int delete_members( const Cl_options & cl_opts ) { - if( !filenames ) + if( cl_opts.filenames <= 0 ) { if( verbosity >= 1 ) show_error( "Nothing to delete." ); return 0; } - if( archive_name.empty() ) + if( cl_opts.archive_name.empty() ) { show_error( "Deleting from stdin not implemented yet." ); return 1; } - const char * const archive_namep = archive_name.c_str(); - const int infd = open_instream( archive_name ); + const char * const archive_namep = cl_opts.archive_name.c_str(); + const int infd = open_instream( cl_opts.archive_name ); if( infd < 0 ) return 1; - const int outfd = open_outstream( archive_name, false ); + const int outfd = open_outstream( cl_opts.archive_name, false ); if( outfd < 0 ) { close( infd ); return 1; } // mark member names to be deleted - std::vector< char > name_pending( parser.arguments(), false ); - for( int i = 0; i < parser.arguments(); ++i ) - if( !parser.code( i ) && parser.argument( i ).size() && - !Exclude::excluded( parser.argument( i ).c_str() ) ) + std::vector< char > name_pending( cl_opts.parser.arguments(), false ); + for( int i = 0; i < cl_opts.parser.arguments(); ++i ) + if( nonempty_arg( cl_opts.parser, i ) && + !Exclude::excluded( cl_opts.parser.argument( i ).c_str() ) ) name_pending[i] = true; const Lzip_index lzip_index( infd, true, false ); // only regular files if( lzip_index.retval() == 0 ) // compressed - return delete_members_lz( archive_namep, parser, name_pending, lzip_index, - filenames, infd, outfd, missing_crc, permissive ); + return delete_members_lz( cl_opts, archive_namep, name_pending, lzip_index, + infd, outfd ); if( lseek( infd, 0, SEEK_SET ) != 0 ) { show_file_error( archive_namep, "Archive is not seekable." ); return 1; } if( lzip_index.file_size() < 3 * header_size ) @@ -153,7 +152,7 @@ int delete_members( const std::string & archive_name, const Arg_parser & parser, { if( block_is_zero( header, header_size ) ) // EOF { - if( prev_extended && !permissive ) + if( prev_extended && !cl_opts.permissive ) { show_file_error( archive_namep, fv_msg1 ); retval = 2; } break; } @@ -164,7 +163,7 @@ int delete_members( const std::string & archive_name, const Arg_parser & parser, const Typeflag typeflag = (Typeflag)header[typeflag_o]; if( typeflag == tf_global ) { - if( prev_extended && !permissive ) + if( prev_extended && !cl_opts.permissive ) { show_file_error( archive_namep, fv_msg2 ); retval = 2; break; } Extended dummy; // global headers are parsed and ignored if( !parse_records( infd, dummy, header, rbuf, true ) ) @@ -173,11 +172,11 @@ int delete_members( const std::string & archive_name, const Arg_parser & parser, } if( typeflag == tf_extended ) { - if( prev_extended && !permissive ) + if( prev_extended && !cl_opts.permissive ) { show_file_error( archive_namep, fv_msg3 ); retval = 2; break; } - if( !parse_records( infd, extended, header, rbuf, permissive ) ) + if( !parse_records( infd, extended, header, rbuf, cl_opts.permissive ) ) { show_file_error( archive_namep, extrec_msg ); retval = 2; break; } - else if( !extended.crc_present() && missing_crc ) + else if( !extended.crc_present() && cl_opts.missing_crc ) { show_file_error( archive_namep, mcrc_msg ); retval = 2; break; } prev_extended = true; continue; @@ -187,16 +186,14 @@ int delete_members( const std::string & archive_name, const Arg_parser & parser, extended.fill_from_ustar( header ); // copy metadata from header { // skip member - long long rest = extended.file_size(); - const int rem = rest % header_size; - if( rem ) rest += header_size - rem; // padding + long long rest = round_up( extended.file_size() ); // size + padding if( lseek( infd, rest, SEEK_CUR ) <= 0 ) { show_file_error( archive_namep, "Seek error", errno ); retval = 1; break; } } - if( !check_skip_filename( parser, name_pending, extended.path().c_str(), - filenames ) ) // delete tar member + // delete tar member + if( !check_skip_filename( cl_opts, name_pending, extended.path().c_str() ) ) { if( !show_member_name( extended, header, 1, rbuf ) ) { retval = 1; break; } @@ -218,6 +215,6 @@ int delete_members( const std::string & archive_name, const Arg_parser & parser, extended.reset(); } - return tail_copy( archive_namep, parser, name_pending, lzip_index, + return tail_copy( archive_namep, cl_opts.parser, name_pending, lzip_index, istream_pos, infd, outfd, retval ); } diff --git a/delete_lz.cc b/delete_lz.cc index 6888520..961645f 100644 --- a/delete_lz.cc +++ b/delete_lz.cc @@ -1,18 +1,18 @@ -/* Tarlz - Archiver with multimember lzip compression - Copyright (C) 2013-2019 Antonio Diaz Diaz. +/* Tarlz - Archiver with multimember lzip compression + Copyright (C) 2013-2020 Antonio Diaz Diaz. - 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 of the License, or - (at your option) any later version. + 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 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. + 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. - You should have received a copy of the GNU General Public License - along with this program. If not, see <http://www.gnu.org/licenses/>. + You should have received a copy of the GNU General Public License + along with this program. If not, see <http://www.gnu.org/licenses/>. */ #define _FILE_OFFSET_BITS 64 @@ -31,18 +31,103 @@ #include <lzlib.h> #include "arg_parser.h" -#include "lzip_index.h" #include "tarlz.h" +#include "lzip_index.h" + + +/* Read 'size' decompressed bytes from the archive. + Return value: 0 = OK, 1 = damaged member, 2 = fatal error. */ +int archive_read_lz( LZ_Decoder * const decoder, const int infd, + long long & file_pos, const long long member_end, + const long long cdata_size, uint8_t * const buf, + const int size, const char ** msg ) + { + int sz = 0; + + while( sz < size ) + { + const int rd = LZ_decompress_read( decoder, buf + sz, size - sz ); + if( rd < 0 ) + { *msg = LZ_strerror( LZ_decompress_errno( decoder ) ); return 1; } + if( rd == 0 && LZ_decompress_finished( decoder ) == 1 ) + { *msg = end_msg; return 2; } + sz += rd; + if( sz < size && LZ_decompress_write_size( decoder ) > 0 ) + { + const long long ibuf_size = 16384; + uint8_t ibuf[ibuf_size]; + const long long rest = ( file_pos < member_end ) ? + member_end - file_pos : cdata_size - file_pos; + const int rsize = std::min( LZ_decompress_write_size( decoder ), + (int)std::min( ibuf_size, rest ) ); + if( rsize <= 0 ) LZ_decompress_finish( decoder ); + else + { + const int rd = preadblock( infd, ibuf, rsize, file_pos ); + if( LZ_decompress_write( decoder, ibuf, rd ) != rd ) + internal_error( "library error (LZ_decompress_write)." ); + file_pos += rd; + if( rd < rsize ) + { + LZ_decompress_finish( decoder ); + if( errno ) { *msg = "Error reading archive"; return 2; } + } + } + } + } + return 0; + } + + +int parse_records_lz( LZ_Decoder * const decoder, const int infd, + long long & file_pos, const long long member_end, + const long long cdata_size, long long & data_pos, + Extended & extended, const Tar_header header, + Resizable_buffer & rbuf, const char ** msg, + const bool permissive ) + { + const long long edsize = parse_octal( header + size_o, size_l ); + const long long bufsize = round_up( edsize ); + if( edsize <= 0 || edsize >= 1LL << 33 || bufsize >= INT_MAX ) + return 1; // overflow or no extended data + if( !rbuf.resize( bufsize ) ) return 1; // extended records buffer + int retval = archive_read_lz( decoder, infd, file_pos, member_end, + cdata_size, (uint8_t *)rbuf(), bufsize, msg ); + if( retval == 0 ) + { if( extended.parse( rbuf(), edsize, permissive ) ) data_pos += bufsize; + else retval = 2; } + return retval; + } + + +int skip_member_lz( LZ_Decoder * const decoder, const int infd, + long long & file_pos, const long long member_end, + const long long cdata_size, long long & data_pos, + long long rest, const char ** msg ) + { + const int bufsize = 32 * header_size; + uint8_t buf[bufsize]; + while( rest > 0 ) // skip tar member + { + const int rsize = ( rest >= bufsize ) ? bufsize : rest; + const int ret = archive_read_lz( decoder, infd, file_pos, member_end, + cdata_size, buf, rsize, msg ); + if( ret != 0 ) return ret; + data_pos += rsize; + rest -= rsize; + } + return 0; + } /* Deleting from a corrupt archive must not worsen the corruption. Stop and - tail-copy as soon as corruption is found. */ -int delete_members_lz( const char * const archive_namep, - const Arg_parser & parser, + tail-copy as soon as corruption is found. +*/ +int delete_members_lz( const Cl_options & cl_opts, + const char * const archive_namep, std::vector< char > & name_pending, const Lzip_index & lzip_index, - const int filenames, const int infd, const int outfd, - const bool missing_crc, const bool permissive ) + const int infd, const int outfd ) { Resizable_buffer rbuf; LZ_Decoder * const decoder = LZ_decompress_open(); @@ -87,7 +172,7 @@ int delete_members_lz( const char * const archive_namep, { if( block_is_zero( header, header_size ) ) // EOF { - if( prev_extended && !permissive ) + if( prev_extended && !cl_opts.permissive ) { show_file_error( archive_namep, fv_msg1 ); retval = 2; } goto done; } @@ -100,48 +185,47 @@ int delete_members_lz( const char * const archive_namep, const Typeflag typeflag = (Typeflag)header[typeflag_o]; if( typeflag == tf_global ) { - if( prev_extended && !permissive ) + if( prev_extended && !cl_opts.permissive ) { show_file_error( archive_namep, fv_msg2 ); retval = 2; goto done; } Extended dummy; // global headers are parsed and ignored retval = parse_records_lz( decoder, infd, file_pos, member_end, cdata_size, data_pos, dummy, header, rbuf, &msg, true ); if( retval == 0 ) continue; - show_file_error( archive_namep, msg ? msg : gblrec_msg ); + show_file_error( archive_namep, gblrec_msg ); goto done; } if( typeflag == tf_extended ) { - if( prev_extended && !permissive ) { msg = fv_msg3; retval = 2; } + if( prev_extended && !cl_opts.permissive ) + { msg = fv_msg3; retval = 2; } else retval = parse_records_lz( decoder, infd, file_pos, member_end, cdata_size, data_pos, extended, header, - rbuf, &msg, permissive ); - if( retval == 0 && !extended.crc_present() && missing_crc ) + rbuf, &msg, cl_opts.permissive ); + if( retval == 0 && !extended.crc_present() && cl_opts.missing_crc ) { msg = mcrc_msg; retval = 2; } if( retval == 0 ) { prev_extended = true; continue; } - show_file_error( archive_namep, msg ? msg : extrec_msg ); + show_file_error( archive_namep, extrec_msg ); goto done; } prev_extended = false; extended.fill_from_ustar( header ); // copy metadata from header - long long rest = extended.file_size(); - const int rem = rest % header_size; - if( rem ) rest += header_size - rem; // padding + long long rest = round_up( extended.file_size() ); // size + padding if( data_pos + rest >= mdata_end ) data_pos += rest; else // skip tar member if( ( retval = skip_member_lz( decoder, infd, file_pos, member_end, cdata_size, data_pos, rest, &msg ) ) != 0 ) goto done; - if( !check_skip_filename( parser, name_pending, extended.path().c_str(), - filenames ) ) // delete tar member + // delete tar member + if( !check_skip_filename( cl_opts, name_pending, extended.path().c_str() ) ) { // verify that members match if( member_begin != mdata_pos || data_pos != mdata_end ) { show_file_error( extended.path().c_str(), - "Can't delete: not individually compressed." ); + "Can't delete: not compressed individually." ); retval2 = 2; extended.reset(); continue; } if( !show_member_name( extended, header, 1, rbuf ) ) { retval = 1; goto done; } @@ -163,6 +247,6 @@ done: if( LZ_decompress_close( decoder ) < 0 && !retval ) { show_error( "LZ_decompress_close failed." ); retval = 1; } // tail copy keeps trailing data - return tail_copy( archive_namep, parser, name_pending, lzip_index, + return tail_copy( archive_namep, cl_opts.parser, name_pending, lzip_index, istream_pos, infd, outfd, retval ); } diff --git a/doc/tarlz.1 b/doc/tarlz.1 index 83325dd..cf0f659 100644 --- a/doc/tarlz.1 +++ b/doc/tarlz.1 @@ -1,5 +1,5 @@ .\" DO NOT MODIFY THIS FILE! It was generated by help2man 1.46.1. -.TH TARLZ "1" "October 2019" "tarlz 0.16" "User Commands" +.TH TARLZ "1" "July 2020" "tarlz 0.17" "User Commands" .SH NAME tarlz \- creates tar archives with multimember lzip compression .SH SYNOPSIS @@ -64,9 +64,15 @@ use archive file <archive> \fB\-h\fR, \fB\-\-dereference\fR follow symlinks; archive the files they point to .TP +\fB\-\-mtime=\fR<date> +use <date> as mtime for files added to archive +.TP \fB\-n\fR, \fB\-\-threads=\fR<n> set number of (de)compression threads [2] .TP +\fB\-p\fR, \fB\-\-preserve\-permissions\fR +don't subtract the umask on extraction +.TP \fB\-q\fR, \fB\-\-quiet\fR suppress all messages .TP @@ -107,10 +113,10 @@ create solidly compressed archive equivalent to '\-\-owner=root \fB\-\-group\fR=\fI\,root\/\fR' .TP \fB\-\-owner=\fR<owner> -use <owner> name/ID for files added +use <owner> name/ID for files added to archive .TP \fB\-\-group=\fR<group> -use <group> name/ID for files added +use <group> name/ID for files added to archive .TP \fB\-\-keep\-damaged\fR don't delete partially extracted files @@ -121,17 +127,17 @@ exit with error status if missing extended CRC \fB\-\-out\-slots=\fR<n> number of 1 MiB output packets buffered [64] .PP -Exit status: 0 for a normal exit, 1 for environmental problems (file -not found, invalid flags, I/O errors, etc), 2 to indicate a corrupt or -invalid input file, 3 for an internal consistency error (eg, bug) which -caused tarlz to panic. +Exit status: 0 for a normal exit, 1 for environmental problems (file not +found, files differ, invalid flags, I/O errors, etc), 2 to indicate a +corrupt or invalid input file, 3 for an internal consistency error (eg, bug) +which caused tarlz to panic. .SH "REPORTING BUGS" Report bugs to lzip\-bug@nongnu.org .br Tarlz home page: http://www.nongnu.org/lzip/tarlz.html .SH COPYRIGHT -Copyright \(co 2019 Antonio Diaz Diaz. -Using lzlib 1.11 +Copyright \(co 2020 Antonio Diaz Diaz. +Using lzlib 1.12\-rc1a License GPLv2+: GNU GPL version 2 or later <http://gnu.org/licenses/gpl.html> .br This is free software: you are free to change and redistribute it. diff --git a/doc/tarlz.info b/doc/tarlz.info index edcee44..e2c61db 100644 --- a/doc/tarlz.info +++ b/doc/tarlz.info @@ -11,7 +11,7 @@ File: tarlz.info, Node: Top, Next: Introduction, Up: (dir) Tarlz Manual ************ -This manual is for Tarlz (version 0.16, 8 October 2019). +This manual is for Tarlz (version 0.17, 30 July 2020). * Menu: @@ -20,17 +20,18 @@ This manual is for Tarlz (version 0.16, 8 October 2019). * Portable character set:: POSIX portable filename character set * File format:: Detailed format of the compressed archive * Amendments to pax format:: The reasons for the differences with pax -* Multi-threaded tar:: Limitations of parallel tar decoding +* Program design:: Internal structure of tarlz +* Multi-threaded decoding:: Limitations of parallel tar decoding * Minimum archive sizes:: Sizes required for full multi-threaded speed * Examples:: A small tutorial with examples * Problems:: Reporting bugs * Concept index:: Index of concepts - Copyright (C) 2013-2019 Antonio Diaz Diaz. + Copyright (C) 2013-2020 Antonio Diaz Diaz. This manual is free documentation: you have unlimited permission to -copy, distribute and modify it. +copy, distribute, and modify it. File: tarlz.info, Node: Introduction, Next: Invoking tarlz, Prev: Top, Up: Top @@ -38,25 +39,26 @@ File: tarlz.info, Node: Introduction, Next: Invoking tarlz, Prev: Top, Up: T 1 Introduction ************** -Tarlz is a massively parallel (multi-threaded) combined implementation -of the tar archiver and the lzip compressor. Tarlz creates, lists and -extracts archives in a simplified and safer variant of the POSIX pax -format compressed with lzip, keeping the alignment between tar members -and lzip members. The resulting multimember tar.lz archive is fully -backward compatible with standard tar tools like GNU tar, which treat -it like any other tar.lz archive. Tarlz can append files to the end of -such compressed archives. +Tarlz is a massively parallel (multi-threaded) combined implementation of +the tar archiver and the lzip compressor. Tarlz creates, lists and extracts +archives in a simplified and safer variant of the POSIX pax format +compressed with lzip, keeping the alignment between tar members and lzip +members. The resulting multimember tar.lz archive is fully backward +compatible with standard tar tools like GNU tar, which treat it like any +other tar.lz archive. Tarlz can append files to the end of such compressed +archives. Keeping the alignment between tar members and lzip members has two -advantages. It adds an indexed lzip layer on top of the tar archive, -making it possible to decode the archive safely in parallel. It also -minimizes the amount of data lost in case of corruption. Compressing a -tar archive with plzip may even double the amount of files lost for -each lzip member damaged because it does not keep the members aligned. +advantages. It adds an indexed lzip layer on top of the tar archive, making +it possible to decode the archive safely in parallel. It also minimizes the +amount of data lost in case of corruption. Compressing a tar archive with +plzip may even double the amount of files lost for each lzip member damaged +because it does not keep the members aligned. Tarlz can create tar archives with five levels of compression granularity; per file (--no-solid), per block (--bsolid, default), per -directory (--dsolid), appendable solid (--asolid), and solid (--solid). +directory (--dsolid), appendable solid (--asolid), and solid (--solid). It +can also create uncompressed tar archives. Of course, compressing each file (or each directory) individually can't achieve a compression ratio as high as compressing solidly the whole tar @@ -69,22 +71,22 @@ archive, but it has the following advantages: member), and unwanted members can be deleted from the archive. Just like an uncompressed tar archive. - * It is a safe POSIX-style backup format. In case of corruption, - tarlz can extract all the undamaged members from the tar.lz - archive, skipping over the damaged members, just like the standard - (uncompressed) tar. Moreover, the option '--keep-damaged' can be - used to recover as much data as possible from each damaged member, - and lziprecover can be used to recover some of the damaged members. + * It is a safe POSIX-style backup format. In case of corruption, tarlz + can extract all the undamaged members from the tar.lz archive, + skipping over the damaged members, just like the standard + (uncompressed) tar. Moreover, the option '--keep-damaged' can be used + to recover as much data as possible from each damaged member, and + lziprecover can be used to recover some of the damaged members. - * A multimember tar.lz archive is usually smaller than the - corresponding solidly compressed tar.gz archive, except when - individually compressing files smaller than about 32 KiB. + * A multimember tar.lz archive is usually smaller than the corresponding + solidly compressed tar.gz archive, except when compressing files + smaller than about 32 KiB individually. - Tarlz protects the extended records with a CRC in a way compatible -with standard tar tools. *Note crc32::. + Tarlz protects the extended records with a Cyclic Redundancy Check (CRC) +in a way compatible with standard tar tools. *Note crc32::. - Tarlz does not understand other tar formats like 'gnu', 'oldgnu', -'star' or 'v7'. 'tarlz -tf archive.tar.lz > /dev/null' can be used to + Tarlz does not understand other tar formats like 'gnu', 'oldgnu', 'star' +or 'v7'. The command 'tarlz -tf archive.tar.lz > /dev/null' can be used to verify that the format of the archive is compatible with tarlz. @@ -97,27 +99,28 @@ The format for running tarlz is: tarlz [OPTIONS] [FILES] -All operations except '--concatenate' operate on whole trees if any -FILE is a directory. +All operations except '--concatenate' operate on whole trees if any FILE is +a directory. - On archive creation or appending tarlz archives the files specified, -but removes from member names any leading and trailing slashes and any -file name prefixes containing a '..' component. On extraction, leading -and trailing slashes are also removed from member names, and archive -members containing a '..' component in the file name are skipped. Tarlz -detects when the archive being created or enlarged is among the files -to be dumped, appended or concatenated, and skips it. + On archive creation or appending tarlz archives the files specified, but +removes from member names any leading and trailing slashes and any file name +prefixes containing a '..' component. On extraction, leading and trailing +slashes are also removed from member names, and archive members containing +a '..' component in the file name are skipped. Tarlz detects when the +archive being created or enlarged is among the files to be dumped, appended +or concatenated, and skips it. On extraction and listing, tarlz removes leading './' strings from member names in the archive or given in the command line, so that -'tarlz -xf foo ./bar baz' extracts members 'bar' and './baz' from -archive 'foo'. +'tarlz -xf foo ./bar baz' extracts members 'bar' and './baz' from archive +'foo'. - If several compression levels or '--*solid' options are given, the -last setting is used. For example '-9 --solid --uncompressed -1' is -equivalent to '-1 --solid' + If several compression levels or '--*solid' options are given, the last +setting is used. For example '-9 --solid --uncompressed -1' is equivalent +to '-1 --solid' - tarlz supports the following options: + tarlz supports the following options: *Note Argument syntax: +(arg_parser)Argument syntax. '--help' Print an informative help message describing the options and exit. @@ -129,24 +132,23 @@ equivalent to '-1 --solid' '-A' '--concatenate' - Append one or more archives to the end of an archive. All the - archives involved must be regular (seekable) files, and must be - either all compressed or all uncompressed. Compressed and - uncompressed archives can't be mixed. Compressed archives must be - multimember lzip files with the two end-of-file blocks plus any - zero padding contained in the last lzip member of each archive. - The intermediate end-of-file blocks are removed as each new archive - is concatenated. If the archive is uncompressed, tarlz parses and - skips tar headers until it finds the end-of-file blocks. Exit with - status 0 without modifying the archive if no FILES have been - specified. + Append one or more archives to the end of an archive. All the archives + involved must be regular (seekable) files, and must be either all + compressed or all uncompressed. Compressed and uncompressed archives + can't be mixed. Compressed archives must be multimember lzip files + with the two end-of-file blocks plus any zero padding contained in the + last lzip member of each archive. The intermediate end-of-file blocks + are removed as each new archive is concatenated. If the archive is + uncompressed, tarlz parses and skips tar headers until it finds the + end-of-file blocks. Exit with status 0 without modifying the archive + if no FILES have been specified. '-B BYTES' '--data-size=BYTES' - Set target size of input data blocks for the '--bsolid' option. - *Note --bsolid::. Valid values range from 8 KiB to 1 GiB. Default - value is two times the dictionary size, except for option '-0' - where it defaults to 1 MiB. *Note Minimum archive sizes::. + Set target size of input data blocks for the option '--bsolid'. *Note + --bsolid::. Valid values range from 8 KiB to 1 GiB. Default value is + two times the dictionary size, except for option '-0' where it + defaults to 1 MiB. *Note Minimum archive sizes::. '-c' '--create' @@ -154,90 +156,104 @@ equivalent to '-1 --solid' '-C DIR' '--directory=DIR' - Change to directory DIR. When creating or appending, the position - of each '-C' option in the command line is significant; it will - change the current working directory for the following FILES until - a new '-C' option appears in the command line. When extracting or - comparing, all the '-C' options are executed in sequence before - reading the archive. Listing ignores any '-C' options specified. - DIR is relative to the then current working directory, perhaps - changed by a previous '-C' option. - - Note that a process can only have one current working directory - (CWD). Therefore multi-threading can't be used to create an - archive if a '-C' option appears after a relative file name in the - command line. + Change to directory DIR. When creating or appending, the position of + each '-C' option in the command line is significant; it will change the + current working directory for the following FILES until a new '-C' + option appears in the command line. When extracting or comparing, all + the '-C' options are executed in sequence before reading the archive. + Listing ignores any '-C' options specified. DIR is relative to the + then current working directory, perhaps changed by a previous '-C' + option. + + Note that a process can only have one current working directory (CWD). + Therefore multi-threading can't be used to create an archive if a '-C' + option appears after a relative file name in the command line. '-d' '--diff' - Compare and report differences between archive and file system. - For each tar member in the archive, verify that the corresponding - file in the file system exists and is of the same type (regular - file, directory, etc). Report on standard output the differences - found in type, mode (permissions), owner and group IDs, - modification time, file size, file contents (of regular files), - target (of symlinks) and device number (of block/character special - files). - - As tarlz removes leading slashes from member names, the '-C' - option may be used in combination with '--diff' when absolute file - names were used on archive creation: 'tarlz -C / -d'. - Alternatively, tarlz may be run from the root directory to perform - the comparison. + Compare and report differences between archive and file system. For + each tar member in the archive, verify that the corresponding file in + the file system exists and is of the same type (regular file, + directory, etc). Report on standard output the differences found in + type, mode (permissions), owner and group IDs, modification time, file + size, file contents (of regular files), target (of symlinks) and + device number (of block/character special files). + + As tarlz removes leading slashes from member names, the option '-C' may + be used in combination with '--diff' when absolute file names were used + on archive creation: 'tarlz -C / -d'. Alternatively, tarlz may be run + from the root directory to perform the comparison. '--ignore-ids' - Make '--diff' ignore differences in owner and group IDs. This - option is useful when comparing an '--anonymous' archive. + Make '--diff' ignore differences in owner and group IDs. This option is + useful when comparing an '--anonymous' archive. '--delete' - Delete the specified files and directories from an archive in - place. It currently can delete only from uncompressed archives and - from archives with individually compressed files ('--no-solid' - archives). Note that files of about '--data-size' or larger are - compressed individually even if '--bsolid' is used, and can - therefore be deleted. Tarlz takes care to not delete a tar member - unless it is possible to do so. For example it won't try to delete - a tar member that is not individually compressed. To delete a - directory without deleting the files under it, use - 'tarlz --delete -f foo --exclude='dir/*' dir'. Deleting in place - may be dangerous. A corrupt archive, a power cut, or an I/O error - may cause data loss. + Delete files and directories from an archive in place. It currently can + delete only from uncompressed archives and from archives with files + compressed individually ('--no-solid' archives). Note that files of + about '--data-size' or larger are compressed individually even if + '--bsolid' is used, and can therefore be deleted. Tarlz takes care to + not delete a tar member unless it is possible to do so. For example it + won't try to delete a tar member that is not compressed individually. + Even in the case of finding a corrupt member after having deleted some + member(s), tarlz stops and copies the rest of the file as soon as + corruption is found, leaving it just as corrupt as it was, but not + worse. + + To delete a directory without deleting the files under it, use + 'tarlz --delete -f foo --exclude='dir/*' dir'. Deleting in place may + be dangerous. A corrupt archive, a power cut, or an I/O error may cause + data loss. '--exclude=PATTERN' - Exclude files matching a shell pattern like '*.o'. A file is - considered to match if any component of the file name matches. For - example, '*.o' matches 'foo.o', 'foo.o/bar' and 'foo/bar.o'. If - PATTERN contains a '/', it matches a corresponding '/' in the file - name. For example, 'foo/*.o' matches 'foo/bar.o'. + Exclude files matching a shell pattern like '*.o'. A file is considered + to match if any component of the file name matches. For example, '*.o' + matches 'foo.o', 'foo.o/bar' and 'foo/bar.o'. If PATTERN contains a + '/', it matches a corresponding '/' in the file name. For example, + 'foo/*.o' matches 'foo/bar.o'. '-f ARCHIVE' '--file=ARCHIVE' - Use archive file ARCHIVE. '-' used as an ARCHIVE argument reads - from standard input or writes to standard output. + Use archive file ARCHIVE. A hyphen '-' used as an ARCHIVE argument + reads from standard input or writes to standard output. '-h' '--dereference' - Follow symbolic links during archive creation, appending or - comparison. Archive or compare the files they point to instead of - the links themselves. + Follow symbolic links during archive creation, appending or comparison. + Archive or compare the files they point to instead of the links + themselves. + +'--mtime=DATE' + When creating or appending, use DATE as the modification time for + files added to the archive instead of their actual modification times. + The value of DATE may be either '@' followed by the number of seconds + since the epoch, or a date in format 'YYYY-MM-DD HH:MM:SS', or the + name of an existing file starting with '.' or '/'. In the latter case, + the modification time of that file is used. '-n N' '--threads=N' Set the number of (de)compression threads, overriding the system's - default. Valid values range from 0 to "as many as your system can - support". A value of 0 disables threads entirely. If this option - is not used, tarlz tries to detect the number of processors in the - system and use it as default value. 'tarlz --help' shows the - system's default value. See the note about multi-threaded archive - creation in the '-C' option above. Multi-threaded extraction of - files from an archive is not yet implemented. *Note - Multi-threaded tar::. - - Note that the number of usable threads is limited during - compression to ceil( uncompressed_size / data_size ) (*note - Minimum archive sizes::), and during decompression to the number - of lzip members in the tar.lz archive, which you can find by - running 'lzip -lv archive.tar.lz'. + default. Valid values range from 0 to "as many as your system can + support". A value of 0 disables threads entirely. If this option is + not used, tarlz tries to detect the number of processors in the system + and use it as default value. 'tarlz --help' shows the system's default + value. See the note about multi-threaded archive creation in the + option '-C' above. Multi-threaded extraction of files from an archive + is not yet implemented. *Note Multi-threaded decoding::. + + Note that the number of usable threads is limited during compression to + ceil( uncompressed_size / data_size ) (*note Minimum archive sizes::), + and during decompression to the number of lzip members in the tar.lz + archive, which you can find by running 'lzip -lv archive.tar.lz'. + +'-p' +'--preserve-permissions' + On extraction, set file permissions as they appear in the archive. + This is the default behavior when tarlz is run by the superuser. The + default for other users is to subtract the umask of the user running + tarlz from the permissions specified in the archive. '-q' '--quiet' @@ -245,20 +261,18 @@ equivalent to '-1 --solid' '-r' '--append' - Append files to the end of an archive. The archive must be a - regular (seekable) file either compressed or uncompressed. - Compressed members can't be appended to an uncompressed archive, - nor vice versa. If the archive is compressed, it must be a - multimember lzip file with the two end-of-file blocks plus any - zero padding contained in the last lzip member of the archive. It - is possible to append files to an archive with a different - compression granularity. Appending works as follows; first the - end-of-file blocks are removed, then the new members are appended, - and finally two new end-of-file blocks are appended to the - archive. If the archive is uncompressed, tarlz parses and skips - tar headers until it finds the end-of-file blocks. Exit with - status 0 without modifying the archive if no FILES have been - specified. + Append files to the end of an archive. The archive must be a regular + (seekable) file either compressed or uncompressed. Compressed members + can't be appended to an uncompressed archive, nor vice versa. If the + archive is compressed, it must be a multimember lzip file with the two + end-of-file blocks plus any zero padding contained in the last lzip + member of the archive. It is possible to append files to an archive + with a different compression granularity. Appending works as follows; + first the end-of-file blocks are removed, then the new members are + appended, and finally two new end-of-file blocks are appended to the + archive. If the archive is uncompressed, tarlz parses and skips tar + headers until it finds the end-of-file blocks. Exit with status 0 + without modifying the archive if no FILES have been specified. '-t' '--list' @@ -272,15 +286,18 @@ equivalent to '-1 --solid' '-x' '--extract' Extract files from an archive. If FILES are given, extract only the - FILES given. Else extract all the files in the archive. To extract - a directory without extracting the files under it, use - 'tarlz -xf foo --exclude='dir/*' dir'. + FILES given. Else extract all the files in the archive. To extract a + directory without extracting the files under it, use + 'tarlz -xf foo --exclude='dir/*' dir'. Tarlz will not make any special + effort to extract a file over an incompatible type of file. For + example, extracting a link over a directory will usually fail. + (Principle of least surprise). '-0 .. -9' - Set the compression level for '--create' and '--append'. The - default compression level is '-6'. Like lzip, tarlz also minimizes - the dictionary size of the lzip members it creates, reducing the - amount of memory required for decompression. + Set the compression level for '--create' and '--append'. The default + compression level is '-6'. Like lzip, tarlz also minimizes the + dictionary size of the lzip members it creates, reducing the amount of + memory required for decompression. Level Dictionary size Match length limit -0 64 KiB 16 bytes @@ -296,95 +313,94 @@ equivalent to '-1 --solid' '--uncompressed' With '--create', don't compress the tar archive created. Create an - uncompressed tar archive instead. With '--append', don't compress - the new members appended to the tar archive. Compressed members - can't be appended to an uncompressed archive, nor vice versa. + uncompressed tar archive instead. With '--append', don't compress the + new members appended to the tar archive. Compressed members can't be + appended to an uncompressed archive, nor vice versa. '--asolid' When creating or appending to a compressed archive, use appendable solid compression. All the files being added to the archive are - compressed into a single lzip member, but the end-of-file blocks - are compressed into a separate lzip member. This creates a solidly - compressed appendable archive. Solid archives can't be created - nor decoded in parallel. + compressed into a single lzip member, but the end-of-file blocks are + compressed into a separate lzip member. This creates a solidly + compressed appendable archive. Solid archives can't be created nor + decoded in parallel. '--bsolid' When creating or appending to a compressed archive, use block - compression. Tar members are compressed together in a lzip member - until they approximate a target uncompressed size. The size can't - be exact because each solidly compressed data block must contain - an integer number of tar members. Block compression is the default + compression. Tar members are compressed together in a lzip member + until they approximate a target uncompressed size. The size can't be + exact because each solidly compressed data block must contain an + integer number of tar members. Block compression is the default because it improves compression ratio for archives with many files smaller than the block size. This option allows tarlz revert to - default behavior if, for example, it is invoked through an alias - like 'tar='tarlz --solid''. *Note --data-size::, to set the target - block size. + default behavior if, for example, it is invoked through an alias like + 'tar='tarlz --solid''. *Note --data-size::, to set the target block + size. '--dsolid' - When creating or appending to a compressed archive, compress each - file specified in the command line separately in its own lzip - member, and use solid compression for each directory specified in - the command line. The end-of-file blocks are compressed into a - separate lzip member. This creates a compressed appendable archive - with a separate lzip member for each file or top-level directory - specified. + When creating or appending to a compressed archive, compress each file + specified in the command line separately in its own lzip member, and + use solid compression for each directory specified in the command + line. The end-of-file blocks are compressed into a separate lzip + member. This creates a compressed appendable archive with a separate + lzip member for each file or top-level directory specified. '--no-solid' - When creating or appending to a compressed archive, compress each - file separately in its own lzip member. The end-of-file blocks are + When creating or appending to a compressed archive, compress each file + separately in its own lzip member. The end-of-file blocks are compressed into a separate lzip member. This creates a compressed appendable archive with a lzip member for each file. '--solid' When creating or appending to a compressed archive, use solid - compression. The files being added to the archive, along with the + compression. The files being added to the archive, along with the end-of-file blocks, are compressed into a single lzip member. The resulting archive is not appendable. No more files can be later - appended to the archive. Solid archives can't be created nor - decoded in parallel. + appended to the archive. Solid archives can't be created nor decoded + in parallel. '--anonymous' Equivalent to '--owner=root --group=root'. '--owner=OWNER' - When creating or appending, use OWNER for files added to the - archive. If OWNER is not a valid user name, it is decoded as a - decimal numeric user ID. + When creating or appending, use OWNER for files added to the archive. + If OWNER is not a valid user name, it is decoded as a decimal numeric + user ID. '--group=GROUP' - When creating or appending, use GROUP for files added to the - archive. If GROUP is not a valid group name, it is decoded as a - decimal numeric group ID. + When creating or appending, use GROUP for files added to the archive. + If GROUP is not a valid group name, it is decoded as a decimal numeric + group ID. '--keep-damaged' Don't delete partially extracted files. If a decompression error - happens while extracting a file, keep the partial data extracted. - Use this option to recover as much data as possible from each - damaged member. + happens while extracting a file, keep the partial data extracted. Use + this option to recover as much data as possible from each damaged + member. '--missing-crc' - Exit with error status 2 if the CRC of the extended records is - missing. When this option is used, tarlz detects any corruption - in the extended records (only limited by CRC collisions). But note - that a corrupt 'GNU.crc32' keyword, for example 'GNU.crc33', is - reported as a missing CRC instead of as a corrupt record. This - misleading 'Missing CRC' message is the consequence of a flaw in - the POSIX pax format; i.e., the lack of a mandatory check sequence - in the extended records. *Note crc32::. + Exit with error status 2 if the CRC of the extended records is missing. + When this option is used, tarlz detects any corruption in the extended + records (only limited by CRC collisions). But note that a corrupt + 'GNU.crc32' keyword, for example 'GNU.crc33', is reported as a missing + CRC instead of as a corrupt record. This misleading 'Missing CRC' + message is the consequence of a flaw in the POSIX pax format; i.e., + the lack of a mandatory check sequence in the extended records. *Note + crc32::. '--out-slots=N' Number of 1 MiB output packets buffered per worker thread during multi-threaded creation or appending to compressed archives. - Increasing the number of packets may increase compression speed if - the files being archived are larger than 64 MiB compressed, but - requires more memory. Valid values range from 1 to 1024. The - default value is 64. + Increasing the number of packets may increase compression speed if the + files being archived are larger than 64 MiB compressed, but requires + more memory. Valid values range from 1 to 1024. The default value is + 64. - Exit status: 0 for a normal exit, 1 for environmental problems (file -not found, invalid flags, I/O errors, etc), 2 to indicate a corrupt or -invalid input file, 3 for an internal consistency error (eg, bug) which -caused tarlz to panic. + Exit status: 0 for a normal exit, 1 for environmental problems (file not +found, files differ, invalid flags, I/O errors, etc), 2 to indicate a +corrupt or invalid input file, 3 for an internal consistency error (eg, bug) +which caused tarlz to panic. File: tarlz.info, Node: Portable character set, Next: File format, Prev: Invoking tarlz, Up: Top @@ -398,8 +414,11 @@ The set of characters from which portable file names are constructed. a b c d e f g h i j k l m n o p q r s t u v w x y z 0 1 2 3 4 5 6 7 8 9 . _ - - The last three characters are the period, underscore, and -hyphen-minus characters, respectively. + The last three characters are the period, underscore, and hyphen-minus +characters, respectively. + + File names are identifiers. Therefore, archiving works better when file +names use only the portable character set without spaces added. File: tarlz.info, Node: File format, Next: Amendments to pax format, Prev: Portable character set, Up: Top @@ -408,11 +427,13 @@ File: tarlz.info, Node: File format, Next: Amendments to pax format, Prev: Po ************* In the diagram below, a box like this: + +---+ | | <-- the vertical bars might be missing +---+ represents one byte; a box like this: + +==============+ | | +==============+ @@ -422,43 +443,43 @@ bytes (for example 512). A tar.lz file consists of a series of lzip members (compressed data -sets). The members simply appear one after another in the file, with no +sets). The members simply appear one after another in the file, with no additional information before, between, or after them. - Each lzip member contains one or more tar members in a simplified -POSIX pax interchange format. The only pax typeflag value supported by -tarlz (in addition to the typeflag values defined by the ustar format) -is 'x'. The pax format is an extension on top of the ustar format that -removes the size limitations of the ustar format. + Each lzip member contains one or more tar members in a simplified POSIX +pax interchange format. The only pax typeflag value supported by tarlz (in +addition to the typeflag values defined by the ustar format) is 'x'. The +pax format is an extension on top of the ustar format that removes the size +limitations of the ustar format. Each tar member contains one file archived, and is represented by the following sequence: - * An optional extended header block with extended header records. - This header block is of the form described in pax header block, - with a typeflag value of 'x'. The extended header records are - included as the data for this header block. + * An optional extended header block with extended header records. This + header block is of the form described in pax header block, with a + typeflag value of 'x'. The extended header records are included as the + data for this header block. * A header block in ustar format that describes the file. Any fields - defined in the preceding optional extended header records override - the associated fields in this header block for this file. + defined in the preceding optional extended header records override the + associated fields in this header block for this file. * Zero or more blocks that contain the contents of the file. Each tar member must be contiguously stored in a lzip member for the -parallel decoding operations like '--list' to work. If any tar member -is split over two or more lzip members, the archive must be decoded -sequentially. *Note Multi-threaded tar::. +parallel decoding operations like '--list' to work. If any tar member is +split over two or more lzip members, the archive must be decoded +sequentially. *Note Multi-threaded decoding::. - At the end of the archive file there are two 512-byte blocks filled -with binary zeros, interpreted as an end-of-archive indicator. These EOF -blocks are either compressed in a separate lzip member or compressed -along with the tar members contained in the last lzip member. + At the end of the archive file there are two 512-byte blocks filled with +binary zeros, interpreted as an end-of-archive indicator. These EOF blocks +are either compressed in a separate lzip member or compressed along with +the tar members contained in the last lzip member. The diagram below shows the correspondence between each tar member -(formed by one or two headers plus optional data) in the tar archive -and each lzip member in the resulting multimember tar.lz archive, when -per file compression is used: *Note File format: (lzip)File format. +(formed by one or two headers plus optional data) in the tar archive and +each lzip member in the resulting multimember tar.lz archive, when per file +compression is used: *Note File format: (lzip)File format. tar +========+======+=================+===============+========+======+========+ @@ -474,62 +495,62 @@ tar.lz 4.1 Pax header block ==================== -The pax header block is identical to the ustar header block described -below except that the typeflag has the value 'x' (extended). The size -field is the size of the extended header data in bytes. Most other -fields in the pax header block are zeroed on archive creation to -prevent trouble if the archive is read by an ustar tool, and are -ignored by tarlz on archive extraction. *Note flawed-compat::. +The pax header block is identical to the ustar header block described below +except that the typeflag has the value 'x' (extended). The field 'size' is +the size of the extended header data in bytes. Most other fields in the pax +header block are zeroed on archive creation to prevent trouble if the +archive is read by an ustar tool, and are ignored by tarlz on archive +extraction. *Note flawed-compat::. The pax extended header data consists of one or more records, each of them constructed as follows: '"%d %s=%s\n", <length>, <keyword>, <value>' - The <length>, <blank>, <keyword>, <equals-sign>, and <newline> in the -record must be limited to the portable character set. The <length> field -contains the decimal length of the record in bytes, including the -trailing <newline>. The <value> field is stored as-is, without -conversion to UTF-8 nor any other transformation. + The fields <length> and <keyword> in the record must be limited to the +portable character set (*note Portable character set::). The field <length> +contains the decimal length of the record in bytes, including the trailing +newline. The field <value> is stored as-is, without conversion to UTF-8 nor +any other transformation. The fields are separated by the ASCII characters +space, equal-sign, and newline. - These are the <keyword> fields currently supported by tarlz: + These are the <keyword> values currently supported by tarlz: 'linkpath' The pathname of a link being created to another file, of any type, - previously archived. This record overrides the linkname field in - the following ustar header block. The following ustar header block + previously archived. This record overrides the field 'linkname' in the + following ustar header block. The following ustar header block determines the type of link created. If typeflag of the following - header block is 1, it will be a hard link. If typeflag is 2, it - will be a symbolic link and the linkpath value will be used as the - contents of the symbolic link. + header block is 1, it will be a hard link. If typeflag is 2, it will + be a symbolic link and the linkpath value will be used as the contents + of the symbolic link. 'path' - The pathname of the following file. This record overrides the name - and prefix fields in the following ustar header block. + The pathname of the following file. This record overrides the fields + 'name' and 'prefix' in the following ustar header block. 'size' The size of the file in bytes, expressed as a decimal number using digits from the ISO/IEC 646:1991 (ASCII) standard. This record - overrides the size field in the following ustar header block. The - size record is used only for files with a size value greater than + overrides the size field in the following ustar header block. The size + record is used only for files with a size value greater than 8_589_934_591 (octal 77777777777). This is 2^33 bytes or larger. 'GNU.crc32' - CRC32-C (Castagnoli) of the extended header data excluding the 8 - bytes representing the CRC <value> itself. The <value> is - represented as 8 hexadecimal digits in big endian order, - '22 GNU.crc32=00000000\n'. The keyword of the CRC record is - protected by the CRC to guarante that corruption is always detected - (except in case of CRC collision). A CRC was chosen because a - checksum is too weak for a potentially large list of variable - sized records. A checksum can't detect simple errors like the - swapping of two bytes. + CRC32-C (Castagnoli) of the extended header data excluding the 8 bytes + representing the CRC <value> itself. The <value> is represented as 8 + hexadecimal digits in big endian order, '22 GNU.crc32=00000000\n'. The + keyword of the CRC record is protected by the CRC to guarante that + corruption is always detected (except in case of CRC collision). A CRC + was chosen because a checksum is too weak for a potentially large list + of variable sized records. A checksum can't detect simple errors like + the swapping of two bytes. 4.2 Ustar header block ====================== -The ustar header block has a length of 512 bytes and is structured as -shown in the following table. All lengths and offsets are in decimal. +The ustar header block has a length of 512 bytes and is structured as shown +in the following table. All lengths and offsets are in decimal. Field Name Offset Length (in bytes) name 0 100 @@ -549,34 +570,33 @@ devmajor 329 8 devminor 337 8 prefix 345 155 - All characters in the header block are coded using the ISO/IEC -646:1991 (ASCII) standard, except in fields storing names for files, -users, and groups. For maximum portability between implementations, -names should only contain characters from the portable character set. -But if an implementation supports the use of characters outside of '/' -and the portable character set in names for files, users, and groups, -tarlz will use the byte values in these names unmodified. + All characters in the header block are coded using the ISO/IEC 646:1991 +(ASCII) standard, except in fields storing names for files, users, and +groups. For maximum portability between implementations, names should only +contain characters from the portable character set (*note Portable +character set::), but if an implementation supports the use of characters +outside of '/' and the portable character set in names for files, users, +and groups, tarlz will use the byte values in these names unmodified. - The fields name, linkname, and prefix are null-terminated character -strings except when all characters in the array contain non-null + The fields 'name', 'linkname', and 'prefix' are null-terminated +character strings except when all characters in the array contain non-null characters including the last character. - The name and the prefix fields produce the pathname of the file. A -new pathname is formed, if prefix is not an empty string (its first -character is not null), by concatenating prefix (up to the first null -character), a <slash> character, and name; otherwise, name is used -alone. In either case, name is terminated at the first null character. -If prefix begins with a null character, it is ignored. In this manner, -pathnames of at most 256 characters can be supported. If a pathname does -not fit in the space provided, an extended record is used to store the -pathname. + The fields 'prefix' and 'name' produce the pathname of the file. A new +pathname is formed, if prefix is not an empty string (its first character +is not null), by concatenating prefix (up to the first null character), a +slash character, and name; otherwise, name is used alone. In either case, +name is terminated at the first null character. If prefix begins with a +null character, it is ignored. In this manner, pathnames of at most 256 +characters can be supported. If a pathname does not fit in the space +provided, an extended record is used to store the pathname. - The linkname field does not use the prefix to produce a pathname. If -the linkname does not fit in the 100 characters provided, an extended -record is used to store the linkname. + The field 'linkname' does not use the prefix to produce a pathname. If +the linkname does not fit in the 100 characters provided, an extended record +is used to store the linkname. - The mode field provides 12 access permission bits. The following -table shows the symbolic name of each bit and its octal value: + The field 'mode' provides 12 access permission bits. The following table +shows the symbolic name of each bit and its octal value: Bit Name Value Bit Name Value Bit Name Value --------------------------------------------------- @@ -585,30 +605,28 @@ S_IRUSR 00400 S_IWUSR 00200 S_IXUSR 00100 S_IRGRP 00040 S_IWGRP 00020 S_IXGRP 00010 S_IROTH 00004 S_IWOTH 00002 S_IXOTH 00001 - The uid and gid fields are the user and group ID of the owner and + The fields 'uid' and 'gid' are the user and group IDs of the owner and group of the file, respectively. - The size field contains the octal representation of the size of the -file in bytes. If the typeflag field specifies a file of type '0' -(regular file) or '7' (high performance regular file), the number of -logical records following the header is (size / 512) rounded to the next -integer. For all other values of typeflag, tarlz either sets the size -field to 0 or ignores it, and does not store or expect any logical -records following the header. If the file size is larger than -8_589_934_591 bytes (octal 77777777777), an extended record is used to -store the file size. - - The mtime field contains the octal representation of the modification -time of the file at the time it was archived, obtained from the stat() -function. - - The chksum field contains the octal representation of the value of -the simple sum of all bytes in the header logical record. Each byte in -the header is treated as an unsigned value. When calculating the -checksum, the chksum field is treated as if it were all <space> -characters. - - The typeflag field contains a single character specifying the type of + The field 'size' contains the octal representation of the size of the +file in bytes. If the field 'typeflag' specifies a file of type '0' +(regular file) or '7' (high performance regular file), the number of logical +records following the header is (size / 512) rounded to the next integer. +For all other values of typeflag, tarlz either sets the size field to 0 or +ignores it, and does not store or expect any logical records following the +header. If the file size is larger than 8_589_934_591 bytes +(octal 77777777777), an extended record is used to store the file size. + + The field 'mtime' contains the octal representation of the modification +time of the file at the time it was archived, obtained from the function +'stat'. + + The field 'chksum' contains the octal representation of the value of the +simple sum of all bytes in the header logical record. Each byte in the +header is treated as an unsigned value. When calculating the checksum, the +chksum field is treated as if it were all space characters. + + The field 'typeflag' contains a single character specifying the type of file archived: ''0'' @@ -621,9 +639,9 @@ file archived: Symbolic link. ''3', '4'' - Character special file and block special file respectively. In - this case the devmajor and devminor fields contain information - defining the device in unspecified format. + Character special file and block special file respectively. In this + case the fields 'devmajor' and 'devminor' contain information defining + the device in unspecified format. ''5'' Directory. @@ -632,157 +650,233 @@ file archived: FIFO special file. ''7'' - Reserved to represent a file to which an implementation has - associated some high-performance attribute. Tarlz treats this type - of file as a regular file (type 0). + Reserved to represent a file to which an implementation has associated + some high-performance attribute. Tarlz treats this type of file as a + regular file (type 0). - The magic field contains the ASCII null-terminated string "ustar". -The version field contains the characters "00" (0x30,0x30). The fields -uname, and gname are null-terminated character strings except when all + The field 'magic' contains the ASCII null-terminated string "ustar". The +field 'version' contains the characters "00" (0x30,0x30). The fields +'uname' and 'gname' are null-terminated character strings except when all characters in the array contain non-null characters including the last character. Each numeric field contains a leading space- or zero-filled, optionally null-terminated octal number using digits from the ISO/IEC -646:1991 (ASCII) standard. Tarlz is able to decode numeric fields 1 -byte longer than standard ustar by not requiring a terminating null -character. +646:1991 (ASCII) standard. Tarlz is able to decode numeric fields 1 byte +longer than standard ustar by not requiring a terminating null character. -File: tarlz.info, Node: Amendments to pax format, Next: Multi-threaded tar, Prev: File format, Up: Top +File: tarlz.info, Node: Amendments to pax format, Next: Program design, Prev: File format, Up: Top 5 The reasons for the differences with pax ****************************************** -Tarlz creates safe archives that allow the reliable detection of -invalid or corrupt metadata during decoding even when the integrity -checking of lzip can't be used because the lzip members are only -decompressed partially, as it happens in parallel '--list' and -'--extract'. In order to achieve this goal, tarlz makes some changes to -the variant of the pax format that it uses. This chapter describes -these changes and the concrete reasons to implement them. +Tarlz creates safe archives that allow the reliable detection of invalid or +corrupt metadata during decoding even when the integrity checking of lzip +can't be used because the lzip members are only decompressed partially, as +it happens in parallel '--diff', '--list', and '--extract'. In order to +achieve this goal, tarlz makes some changes to the variant of the pax +format that it uses. This chapter describes these changes and the concrete +reasons to implement them. 5.1 Add a CRC of the extended records ===================================== -The POSIX pax format has a serious flaw. The metadata stored in pax -extended records are not protected by any kind of check sequence. -Corruption in a long file name may cause the extraction of the file in -the wrong place without warning. Corruption in a large file size may -cause the truncation of the file or the appending of garbage to the -file, both followed by a spurious warning about a corrupt header far -from the place of the undetected corruption. +The POSIX pax format has a serious flaw. The metadata stored in pax extended +records are not protected by any kind of check sequence. Corruption in a +long file name may cause the extraction of the file in the wrong place +without warning. Corruption in a large file size may cause the truncation of +the file or the appending of garbage to the file, both followed by a +spurious warning about a corrupt header far from the place of the undetected +corruption. Metadata like file name and file size must be always protected in an -archive format because of the adverse effects of undetected corruption -in them, potentially much worse that undetected corruption in the data. -Even more so in the case of pax because the amount of metadata it -stores is potentially large, making undetected corruption more probable. +archive format because of the adverse effects of undetected corruption in +them, potentially much worse that undetected corruption in the data. Even +more so in the case of pax because the amount of metadata it stores is +potentially large, making undetected corruption and archiver misbehavior +more probable. - Headers and metadata must be protected separately from data because -the integrity checking of lzip may not be able to detect the corruption -before the metadata has been used, for example, to create a new file in -the wrong place. + Headers and metadata must be protected separately from data because the +integrity checking of lzip may not be able to detect the corruption before +the metadata has been used, for example, to create a new file in the wrong +place. - Because of the above, tarlz protects the extended records with a CRC -in a way compatible with standard tar tools. *Note key_crc32::. + Because of the above, tarlz protects the extended records with a CRC in a +way compatible with standard tar tools. *Note key_crc32::. 5.2 Remove flawed backward compatibility ======================================== -In order to allow the extraction of pax archives by a tar utility -conforming to the POSIX-2:1993 standard, POSIX.1-2008 recommends -selecting extended header field values that allow such tar to create a -regular file containing the extended header records as data. This -approach is broken because if the extended header is needed because of -a long file name, the name and prefix fields will be unable to contain -the full pathname of the file. Therefore the files corresponding to -both the extended header and the overridden ustar header will be -extracted using truncated file names, perhaps overwriting existing -files or directories. It may be a security risk to extract a file with -a truncated file name. +In order to allow the extraction of pax archives by a tar utility conforming +to the POSIX-2:1993 standard, POSIX.1-2008 recommends selecting extended +header field values that allow such tar to create a regular file containing +the extended header records as data. This approach is broken because if the +extended header is needed because of a long file name, the fields 'prefix' +and 'name' will be unable to contain the full pathname of the file. +Therefore the files corresponding to both the extended header and the +overridden ustar header will be extracted using truncated file names, +perhaps overwriting existing files or directories. It may be a security risk +to extract a file with a truncated file name. To avoid this problem, tarlz writes extended headers with all fields -zeroed except size, chksum, typeflag, magic and version. This prevents -old tar programs from extracting the extended records as a file in the -wrong place. Tarlz also sets to zero those fields of the ustar header -overridden by extended records. +zeroed except size, chksum, typeflag, magic and version. This prevents old +tar programs from extracting the extended records as a file in the wrong +place. Tarlz also sets to zero those fields of the ustar header overridden +by extended records. - If an extended header is required for any reason (for example a file -size larger than 8 GiB or a link name longer than 100 bytes), tarlz -moves the file name also to the extended header to prevent an ustar -tool from trying to extract the file or link. This also makes easier -during parallel decoding the detection of a tar member split between -two lzip members at the boundary between the extended header and the -ustar header. + If an extended header is required for any reason (for example a file size +larger than 8 GiB or a link name longer than 100 bytes), tarlz moves the +file name also to the extended header to prevent an ustar tool from trying +to extract the file or link. This also makes easier during parallel decoding +the detection of a tar member split between two lzip members at the boundary +between the extended header and the ustar header. 5.3 As simple as possible (but not simpler) =========================================== -The tarlz format is mainly ustar. Extended pax headers are used only -when needed because the length of a file name or link name, or the size -of a file exceed the limits of the ustar format. Adding extended -headers to each member just to record subsecond timestamps seems -wasteful for a backup format. Moreover, minimizing the overhead may -help recovering the archive with lziprecover in case of corruption. +The tarlz format is mainly ustar. Extended pax headers are used only when +needed because the length of a file name or link name, or the size of a file +exceed the limits of the ustar format. Adding 1 KiB of extended headers to +each member just to record subsecond timestamps seems wasteful for a backup +format. Moreover, minimizing the overhead may help recovering the archive +with lziprecover in case of corruption. - Global pax headers are tolerated, but not supported; they are parsed -and ignored. Some operations may not behave as expected if the archive -contains global headers. + Global pax headers are tolerated, but not supported; they are parsed and +ignored. Some operations may not behave as expected if the archive contains +global headers. 5.4 Avoid misconversions to/from UTF-8 ====================================== -There is no portable way to tell what charset a text string is coded -into. Therefore, tarlz stores all fields representing text strings -unmodified, without conversion to UTF-8 nor any other transformation. -This prevents accidental double UTF-8 conversions. If the need arises -this behavior will be adjusted with a command line option in the future. +There is no portable way to tell what charset a text string is coded into. +Therefore, tarlz stores all fields representing text strings unmodified, +without conversion to UTF-8 nor any other transformation. This prevents +accidental double UTF-8 conversions. If the need arises this behavior will +be adjusted with a command line option in the future. -File: tarlz.info, Node: Multi-threaded tar, Next: Minimum archive sizes, Prev: Amendments to pax format, Up: Top +File: tarlz.info, Node: Program design, Next: Multi-threaded decoding, Prev: Amendments to pax format, Up: Top + +6 Internal structure of tarlz +***************************** + +The parts of tarlz related to sequential processing of the archive are more +or less similar to any other tar and won't be described here. The +interesting parts described here are those related to Multi-threaded +processing. + + The structure of the part of tarlz performing Multi-threaded archive +creation is somewhat similar to that of plzip with the added complication of +the solidity levels. A grouper thread and several worker threads are +created, acting the main thread as muxer (multiplexer) thread. A "packet +courier" takes care of data transfers among threads and limits the maximum +number of data blocks (packets) being processed simultaneously. + + The grouper traverses the directory tree, groups together the metadata of +the files to be archived in each lzip member, and distributes them to the +workers. The workers compress the metadata received from the grouper along +with the file data read from the file system. The muxer collects processed +packets from the workers, and writes them to the archive. + +,--------, +| data|---> to each worker below +| | ,------------, +| file | ,-->| worker 0 |--, +| system | | `------------' | +| | ,---------, | ,------------, | ,-------, ,---------, +|metadata|--->| grouper |-+-->| worker 1 |--+-->| muxer |-->| archive | +`--------' `---------' | `------------' | `-------' `---------' + | ... | + | ,------------, | + `-->| worker N-1 |--' + `------------' + + Decoding an archive is somewhat similar to how plzip decompresses a +regular file to standard output, with the differences that it is not the +data but only messages what is written to stdout/stderr, and that each +worker may access files in the file system either to read them (diff) or +write them (extract). As in plzip, each worker reads members directly from +the archive. + +,--------, +| file |<---> data to/from each worker below +| system | +`--------' + ,------------, + ,-->| worker 0 |--, + | `------------' | +,---------, | ,------------, | ,-------, ,--------, +| archive |-+-->| worker 1 |--+-->| muxer |-->| stdout | +`---------' | `------------' | `-------' | stderr | + | ... | `--------' + | ,------------, | + `-->| worker N-1 |--' + `------------' + + As misaligned tar.lz archives can't be decoded in parallel, and the +misalignment can't be detected until after decoding has started, a +"mastership request" mechanism has been designed that allows the decoding to +continue instead of signalling an error. + + During parallel decoding, if a worker finds a misalignment, it requests +mastership to decode the rest of the archive. When mastership is requested, +an error_member_id is set, and all subsequently received packets with +member_id > error_member_id are rejected. All workers requesting mastership +are blocked at the request_mastership call until mastership is granted. +Mastership is granted to the delivering worker when its queue is empty to +make sure that all preceding packets have been processed. When mastership is +granted, all packets are deleted and all subsequently received packets not +coming from the master are rejected. + + If a worker can't continue decoding for any cause (for example lack of +memory or finding a split tar member at the beginning of a lzip member), it +requests mastership to print an error and terminate the program. Only if +some other worker requests mastership in a previous lzip member can this +error be avoided. -6 Limitations of parallel tar decoding + +File: tarlz.info, Node: Multi-threaded decoding, Next: Minimum archive sizes, Prev: Program design, Up: Top + +7 Limitations of parallel tar decoding ************************************** Safely decoding an arbitrary tar archive in parallel is impossible. For -example, if a tar archive containing another tar archive is decoded -starting from some position other than the beginning, there is no way -to know if the first header found there belongs to the outer tar -archive or to the inner tar archive. Tar is a format inherently serial; -it was designed for tapes. +example, if a tar archive containing another tar archive is decoded starting +from some position other than the beginning, there is no way to know if the +first header found there belongs to the outer tar archive or to the inner +tar archive. Tar is a format inherently serial; it was designed for tapes. In the case of compressed tar archives, the start of each compressed -block determines one point through which the tar archive can be decoded -in parallel. Therefore, in tar.lz archives the decoding operations -can't be parallelized if the tar members are not aligned with the lzip -members. Tar archives compressed with plzip can't be decoded in -parallel because tar and plzip do not have a way to align both sets of -members. Certainly one can decompress one such archive with a -multi-threaded tool like plzip, but the increase in speed is not as -large as it could be because plzip must serialize the decompressed data -and pass them to tar, which decodes them sequentially, one tar member -at a time. +block determines one point through which the tar archive can be decoded in +parallel. Therefore, in tar.lz archives the decoding operations can't be +parallelized if the tar members are not aligned with the lzip members. Tar +archives compressed with plzip can't be decoded in parallel because tar and +plzip do not have a way to align both sets of members. Certainly one can +decompress one such archive with a multi-threaded tool like plzip, but the +increase in speed is not as large as it could be because plzip must +serialize the decompressed data and pass them to tar, which decodes them +sequentially, one tar member at a time. On the other hand, if the tar.lz archive is created with a tool like tarlz, which can guarantee the alignment between tar members and lzip -members because it controls both archiving and compression, then the -lzip format becomes an indexed layer on top of the tar archive which -makes possible decoding it safely in parallel. +members because it controls both archiving and compression, then the lzip +format becomes an indexed layer on top of the tar archive which makes +possible decoding it safely in parallel. - Tarlz is able to automatically decode aligned and unaligned -multimember tar.lz archives, keeping backwards compatibility. If tarlz -finds a member misalignment during multi-threaded decoding, it switches -to single-threaded mode and continues decoding the archive. Currently -only the '--list' option is able to do multi-threaded decoding. + Tarlz is able to automatically decode aligned and unaligned multimember +tar.lz archives, keeping backwards compatibility. If tarlz finds a member +misalignment during multi-threaded decoding, it switches to single-threaded +mode and continues decoding the archive. Currently only the options +'--diff' and '--list' are able to do multi-threaded decoding. If the files in the archive are large, multi-threaded '--list' on a regular (seekable) tar.lz archive can be hundreds of times faster than -sequential '--list' because, in addition to using several processors, -it only needs to decompress part of each lzip member. See the following +sequential '--list' because, in addition to using several processors, it +only needs to decompress part of each lzip member. See the following example listing the Silesia corpus on a dual core machine: tarlz -9 --no-solid -cf silesia.tar.lz silesia @@ -790,29 +884,33 @@ example listing the Silesia corpus on a dual core machine: time plzip -cd silesia.tar.lz | tar -tf - (3.256s) time tarlz -tf silesia.tar.lz (0.020s) + On the other hand, multi-threaded '--list' won't detect corruption in +the tar member data because it only decodes the part of each lzip member +corresponding to the tar member header. + -File: tarlz.info, Node: Minimum archive sizes, Next: Examples, Prev: Multi-threaded tar, Up: Top +File: tarlz.info, Node: Minimum archive sizes, Next: Examples, Prev: Multi-threaded decoding, Up: Top -7 Minimum archive sizes required for multi-threaded block compression +8 Minimum archive sizes required for multi-threaded block compression ********************************************************************* When creating or appending to a compressed archive using multi-threaded -block compression, tarlz puts tar members together in blocks and -compresses as many blocks simultaneously as worker threads are chosen, -creating a multimember compressed archive. - - For this to work as expected (and roughly multiply the compression -speed by the number of available processors), the uncompressed archive -must be at least as large as the number of worker threads times the -block size (*note --data-size::). Else some processors will not get any -data to compress, and compression will be proportionally slower. The -maximum speed increase achievable on a given archive is limited by the -ratio (uncompressed_size / data_size). For example, a tarball the size -of gcc or linux will scale up to 10 or 12 processors at level -9. - - The following table shows the minimum uncompressed archive size -needed for full use of N processors at a given compression level, using -the default data size for each level: +block compression, tarlz puts tar members together in blocks and compresses +as many blocks simultaneously as worker threads are chosen, creating a +multimember compressed archive. + + For this to work as expected (and roughly multiply the compression speed +by the number of available processors), the uncompressed archive must be at +least as large as the number of worker threads times the block size (*note +--data-size::). Else some processors will not get any data to compress, and +compression will be proportionally slower. The maximum speed increase +achievable on a given archive is limited by the ratio +(uncompressed_size / data_size). For example, a tarball the size of gcc or +linux will scale up to 10 or 14 processors at level -9. + + The following table shows the minimum uncompressed archive size needed +for full use of N processors at a given compression level, using the default +data size for each level: Processors 2 4 8 16 64 256 ------------------------------------------------------------------ @@ -831,7 +929,7 @@ Level File: tarlz.info, Node: Examples, Next: Problems, Prev: Minimum archive sizes, Up: Top -8 A small tutorial with examples +9 A small tutorial with examples ******************************** Example 1: Create a multimember compressed archive 'archive.tar.lz' @@ -840,33 +938,33 @@ containing files 'a', 'b' and 'c'. tarlz -cf archive.tar.lz a b c -Example 2: Append files 'd' and 'e' to the multimember compressed -archive 'archive.tar.lz'. +Example 2: Append files 'd' and 'e' to the multimember compressed archive +'archive.tar.lz'. tarlz -rf archive.tar.lz d e -Example 3: Create a solidly compressed appendable archive -'archive.tar.lz' containing files 'a', 'b' and 'c'. Then append files -'d' and 'e' to the archive. +Example 3: Create a solidly compressed appendable archive 'archive.tar.lz' +containing files 'a', 'b' and 'c'. Then append files 'd' and 'e' to the +archive. tarlz --asolid -cf archive.tar.lz a b c tarlz --asolid -rf archive.tar.lz d e Example 4: Create a compressed appendable archive containing directories -'dir1', 'dir2' and 'dir3' with a separate lzip member per directory. -Then append files 'a', 'b', 'c', 'd' and 'e' to the archive, all of -them contained in a single lzip member. The resulting archive -'archive.tar.lz' contains 5 lzip members (including the EOF member). +'dir1', 'dir2' and 'dir3' with a separate lzip member per directory. Then +append files 'a', 'b', 'c', 'd' and 'e' to the archive, all of them +contained in a single lzip member. The resulting archive 'archive.tar.lz' +contains 5 lzip members (including the EOF member). tarlz --dsolid -cf archive.tar.lz dir1 dir2 dir3 tarlz --asolid -rf archive.tar.lz a b c d e -Example 5: Create a solidly compressed archive 'archive.tar.lz' -containing files 'a', 'b' and 'c'. Note that no more files can be later -appended to the archive. +Example 5: Create a solidly compressed archive 'archive.tar.lz' containing +files 'a', 'b' and 'c'. Note that no more files can be later appended to +the archive. tarlz --solid -cf archive.tar.lz a b c @@ -876,8 +974,8 @@ Example 6: Extract all files from archive 'archive.tar.lz'. tarlz -xf archive.tar.lz -Example 7: Extract files 'a' and 'c', and the whole tree under -directory 'dir1' from archive 'archive.tar.lz'. +Example 7: Extract files 'a' and 'c', and the whole tree under directory +'dir1' from archive 'archive.tar.lz'. tarlz -xf archive.tar.lz a c dir1 @@ -890,17 +988,17 @@ Example 8: Copy the contents of directory 'sourcedir' to the directory File: tarlz.info, Node: Problems, Next: Concept index, Prev: Examples, Up: Top -9 Reporting bugs -**************** +10 Reporting bugs +***************** -There are probably bugs in tarlz. There are certainly errors and -omissions in this manual. If you report them, they will get fixed. If -you don't, no one will ever know about them and they will remain unfixed -for all eternity, if not longer. +There are probably bugs in tarlz. There are certainly errors and omissions +in this manual. If you report them, they will get fixed. If you don't, no +one will ever know about them and they will remain unfixed for all +eternity, if not longer. If you find a bug in tarlz, please send electronic mail to -<lzip-bug@nongnu.org>. Include the version number, which you can find -by running 'tarlz --version'. +<lzip-bug@nongnu.org>. Include the version number, which you can find by +running 'tarlz --version'. File: tarlz.info, Node: Concept index, Prev: Problems, Up: Top @@ -911,41 +1009,41 @@ Concept index * Menu: -* Amendments to pax format: Amendments to pax format. - (line 6) -* bugs: Problems. (line 6) -* examples: Examples. (line 6) -* file format: File format. (line 6) -* getting help: Problems. (line 6) -* introduction: Introduction. (line 6) -* invoking: Invoking tarlz. (line 6) -* minimum archive sizes: Minimum archive sizes. (line 6) -* options: Invoking tarlz. (line 6) -* parallel tar decoding: Multi-threaded tar. (line 6) -* portable character set: Portable character set. - (line 6) -* usage: Invoking tarlz. (line 6) -* version: Invoking tarlz. (line 6) +* Amendments to pax format: Amendments to pax format. (line 6) +* bugs: Problems. (line 6) +* examples: Examples. (line 6) +* file format: File format. (line 6) +* getting help: Problems. (line 6) +* introduction: Introduction. (line 6) +* invoking: Invoking tarlz. (line 6) +* minimum archive sizes: Minimum archive sizes. (line 6) +* options: Invoking tarlz. (line 6) +* parallel tar decoding: Multi-threaded decoding. (line 6) +* portable character set: Portable character set. (line 6) +* program design: Program design. (line 6) +* usage: Invoking tarlz. (line 6) +* version: Invoking tarlz. (line 6) Tag Table: Node: Top223 -Node: Introduction1155 -Node: Invoking tarlz3841 -Ref: --data-size6006 -Ref: --bsolid13287 -Node: Portable character set16917 -Node: File format17420 -Ref: key_crc3222248 -Node: Amendments to pax format27647 -Ref: crc3228304 -Ref: flawed-compat29564 -Node: Multi-threaded tar32198 -Node: Minimum archive sizes34737 -Node: Examples36870 -Node: Problems38587 -Node: Concept index39113 +Node: Introduction1212 +Node: Invoking tarlz3982 +Ref: --data-size6193 +Ref: --bsolid14608 +Node: Portable character set18244 +Node: File format18887 +Ref: key_crc3223812 +Node: Amendments to pax format29271 +Ref: crc3229935 +Ref: flawed-compat31220 +Node: Program design33865 +Node: Multi-threaded decoding37756 +Node: Minimum archive sizes40492 +Node: Examples42630 +Node: Problems44345 +Node: Concept index44873 End Tag Table diff --git a/doc/tarlz.texi b/doc/tarlz.texi index 25acaf8..00116ee 100644 --- a/doc/tarlz.texi +++ b/doc/tarlz.texi @@ -6,8 +6,8 @@ @finalout @c %**end of header -@set UPDATED 8 October 2019 -@set VERSION 0.16 +@set UPDATED 30 July 2020 +@set VERSION 0.17 @dircategory Data Compression @direntry @@ -40,7 +40,8 @@ This manual is for Tarlz (version @value{VERSION}, @value{UPDATED}). * Portable character set:: POSIX portable filename character set * File format:: Detailed format of the compressed archive * Amendments to pax format:: The reasons for the differences with pax -* Multi-threaded tar:: Limitations of parallel tar decoding +* Program design:: Internal structure of tarlz +* Multi-threaded decoding:: Limitations of parallel tar decoding * Minimum archive sizes:: Sizes required for full multi-threaded speed * Examples:: A small tutorial with examples * Problems:: Reporting bugs @@ -48,10 +49,10 @@ This manual is for Tarlz (version @value{VERSION}, @value{UPDATED}). @end menu @sp 1 -Copyright @copyright{} 2013-2019 Antonio Diaz Diaz. +Copyright @copyright{} 2013-2020 Antonio Diaz Diaz. This manual is free documentation: you have unlimited permission -to copy, distribute and modify it. +to copy, distribute, and modify it. @node Introduction @@ -77,7 +78,8 @@ because it does not keep the members aligned. Tarlz can create tar archives with five levels of compression granularity; per file (---no-solid), per block (---bsolid, default), per directory -(---dsolid), appendable solid (---asolid), and solid (---solid). +(---dsolid), appendable solid (---asolid), and solid (---solid). It can also +create uncompressed tar archives. @noindent Of course, compressing each file (or each directory) individually can't @@ -105,16 +107,16 @@ and lziprecover can be used to recover some of the damaged members. @item A multimember tar.lz archive is usually smaller than the corresponding solidly compressed tar.gz archive, except when -individually compressing files smaller than about 32 KiB. +compressing files smaller than about 32 KiB individually. @end itemize -Tarlz protects the extended records with a CRC in a way compatible with -standard tar tools. @xref{crc32}. +Tarlz protects the extended records with a Cyclic Redundancy Check (CRC) in +a way compatible with standard tar tools. @xref{crc32}. Tarlz does not understand other tar formats like @samp{gnu}, @samp{oldgnu}, -@samp{star} or @samp{v7}. @w{@samp{tarlz -tf archive.tar.lz > /dev/null}} -can be used to verify that the format of the archive is compatible with -tarlz. +@samp{star} or @samp{v7}. The command +@w{@samp{tarlz -tf archive.tar.lz > /dev/null}} can be used to verify that +the format of the archive is compatible with tarlz. @node Invoking tarlz @@ -151,7 +153,11 @@ If several compression levels or @samp{--*solid} options are given, the last setting is used. For example @w{@samp{-9 --solid --uncompressed -1}} is equivalent to @samp{-1 --solid} -tarlz supports the following options: +tarlz supports the following +@uref{http://www.nongnu.org/arg-parser/manual/arg_parser_manual.html#Argument-syntax,,options}: +@ifnothtml +@xref{Argument syntax,,,arg_parser}. +@end ifnothtml @table @code @item --help @@ -177,7 +183,7 @@ modifying the archive if no @var{files} have been specified. @anchor{--data-size} @item -B @var{bytes} @itemx --data-size=@var{bytes} -Set target size of input data blocks for the @samp{--bsolid} option. +Set target size of input data blocks for the option @samp{--bsolid}. @xref{--bsolid}. Valid values range from @w{8 KiB} to @w{1 GiB}. Default value is two times the dictionary size, except for option @samp{-0} where it defaults to @w{1 MiB}. @xref{Minimum archive sizes}. @@ -210,7 +216,7 @@ standard output the differences found in type, mode (permissions), owner and group IDs, modification time, file size, file contents (of regular files), target (of symlinks) and device number (of block/character special files). -As tarlz removes leading slashes from member names, the @samp{-C} option may +As tarlz removes leading slashes from member names, the option @samp{-C} may be used in combination with @samp{--diff} when absolute file names were used on archive creation: @w{@samp{tarlz -C / -d}}. Alternatively, tarlz may be run from the root directory to perform the comparison. @@ -220,14 +226,18 @@ Make @samp{--diff} ignore differences in owner and group IDs. This option is useful when comparing an @samp{--anonymous} archive. @item --delete -Delete the specified files and directories from an archive in place. It -currently can delete only from uncompressed archives and from archives with -individually compressed files (@samp{--no-solid} archives). Note that files -of about @samp{--data-size} or larger are compressed individually even if +Delete files and directories from an archive in place. It currently can +delete only from uncompressed archives and from archives with files +compressed individually (@samp{--no-solid} archives). Note that files of +about @samp{--data-size} or larger are compressed individually even if @samp{--bsolid} is used, and can therefore be deleted. Tarlz takes care to not delete a tar member unless it is possible to do so. For example it won't -try to delete a tar member that is not individually compressed. To delete a -directory without deleting the files under it, use +try to delete a tar member that is not compressed individually. Even in the +case of finding a corrupt member after having deleted some member(s), tarlz +stops and copies the rest of the file as soon as corruption is found, +leaving it just as corrupt as it was, but not worse. + +To delete a directory without deleting the files under it, use @w{@samp{tarlz --delete -f foo --exclude='dir/*' dir}}. Deleting in place may be dangerous. A corrupt archive, a power cut, or an I/O error may cause data loss. @@ -241,14 +251,22 @@ the file name. For example, @samp{foo/*.o} matches @samp{foo/bar.o}. @item -f @var{archive} @itemx --file=@var{archive} -Use archive file @var{archive}. @samp{-} used as an @var{archive} argument -reads from standard input or writes to standard output. +Use archive file @var{archive}. A hyphen @samp{-} used as an @var{archive} +argument reads from standard input or writes to standard output. @item -h @itemx --dereference Follow symbolic links during archive creation, appending or comparison. Archive or compare the files they point to instead of the links themselves. +@item --mtime=@var{date} +When creating or appending, use @var{date} as the modification time for +files added to the archive instead of their actual modification times. The +value of @var{date} may be either @samp{@@} followed by the number of +seconds since the epoch, or a date in format @w{@samp{YYYY-MM-DD HH:MM:SS}}, +or the name of an existing file starting with @samp{.} or @samp{/}. In the +latter case, the modification time of that file is used. + @item -n @var{n} @itemx --threads=@var{n} Set the number of (de)compression threads, overriding the system's default. @@ -256,15 +274,22 @@ Valid values range from 0 to "as many as your system can support". A value of 0 disables threads entirely. If this option is not used, tarlz tries to detect the number of processors in the system and use it as default value. @w{@samp{tarlz --help}} shows the system's default value. See the note about -multi-threaded archive creation in the @samp{-C} option above. +multi-threaded archive creation in the option @samp{-C} above. Multi-threaded extraction of files from an archive is not yet implemented. -@xref{Multi-threaded tar}. +@xref{Multi-threaded decoding}. Note that the number of usable threads is limited during compression to @w{ceil( uncompressed_size / data_size )} (@pxref{Minimum archive sizes}), and during decompression to the number of lzip members in the tar.lz archive, which you can find by running @w{@samp{lzip -lv archive.tar.lz}}. +@item -p +@itemx --preserve-permissions +On extraction, set file permissions as they appear in the archive. This is +the default behavior when tarlz is run by the superuser. The default for +other users is to subtract the umask of the user running tarlz from the +permissions specified in the archive. + @item -q @itemx --quiet Quiet operation. Suppress all messages. @@ -298,7 +323,10 @@ Verbosely list files processed. Extract files from an archive. If @var{files} are given, extract only the @var{files} given. Else extract all the files in the archive. To extract a directory without extracting the files under it, use -@w{@samp{tarlz -xf foo --exclude='dir/*' dir}}. +@w{@samp{tarlz -xf foo --exclude='dir/*' dir}}. Tarlz will not make any +special effort to extract a file over an incompatible type of file. For +example, extracting a link over a directory will usually fail. (Principle of +least surprise). @item -0 .. -9 Set the compression level for @samp{--create} and @samp{--append}. The @@ -411,9 +439,9 @@ keyword appearing in the same block of extended records. @end table Exit status: 0 for a normal exit, 1 for environmental problems (file not -found, invalid flags, I/O errors, etc), 2 to indicate a corrupt or -invalid input file, 3 for an internal consistency error (eg, bug) which -caused tarlz to panic. +found, files differ, invalid flags, I/O errors, etc), 2 to indicate a +corrupt or invalid input file, 3 for an internal consistency error (eg, bug) +which caused tarlz to panic. @node Portable character set @@ -431,12 +459,16 @@ a b c d e f g h i j k l m n o p q r s t u v w x y z The last three characters are the period, underscore, and hyphen-minus characters, respectively. +File names are identifiers. Therefore, archiving works better when file +names use only the portable character set without spaces added. + @node File format @chapter File format @cindex file format In the diagram below, a box like this: + @verbatim +---+ | | <-- the vertical bars might be missing @@ -444,6 +476,7 @@ In the diagram below, a box like this: @end verbatim represents one byte; a box like this: + @verbatim +==============+ | | @@ -486,7 +519,7 @@ Zero or more blocks that contain the contents of the file. Each tar member must be contiguously stored in a lzip member for the parallel decoding operations like @samp{--list} to work. If any tar member is split over two or more lzip members, the archive must be decoded -sequentially. @xref{Multi-threaded tar}. +sequentially. @xref{Multi-threaded decoding}. At the end of the archive file there are two 512-byte blocks filled with binary zeros, interpreted as an end-of-archive indicator. These EOF @@ -530,28 +563,29 @@ An extended header just before the EOF blocks. @section Pax header block The pax header block is identical to the ustar header block described below -except that the typeflag has the value @samp{x} (extended). The size field -is the size of the extended header data in bytes. Most other fields in the -pax header block are zeroed on archive creation to prevent trouble if the -archive is read by an ustar tool, and are ignored by tarlz on archive -extraction. @xref{flawed-compat}. +except that the typeflag has the value @samp{x} (extended). The field +@samp{size} is the size of the extended header data in bytes. Most other +fields in the pax header block are zeroed on archive creation to prevent +trouble if the archive is read by an ustar tool, and are ignored by tarlz on +archive extraction. @xref{flawed-compat}. The pax extended header data consists of one or more records, each of them constructed as follows:@* @samp{"%d %s=%s\n", <length>, <keyword>, <value>} -The <length>, <blank>, <keyword>, <equals-sign>, and <newline> in the -record must be limited to the portable character set. The <length> field -contains the decimal length of the record in bytes, including the -trailing <newline>. The <value> field is stored as-is, without -conversion to UTF-8 nor any other transformation. +The fields <length> and <keyword> in the record must be limited to the +portable character set (@pxref{Portable character set}). The field <length> +contains the decimal length of the record in bytes, including the trailing +newline. The field <value> is stored as-is, without conversion to UTF-8 nor +any other transformation. The fields are separated by the ASCII characters +space, equal-sign, and newline. -These are the <keyword> fields currently supported by tarlz: +These are the <keyword> values currently supported by tarlz: @table @code @item linkpath The pathname of a link being created to another file, of any type, -previously archived. This record overrides the linkname field in the +previously archived. This record overrides the field @samp{linkname} in the following ustar header block. The following ustar header block determines the type of link created. If typeflag of the following header block is 1, it will be a hard link. If typeflag is 2, it will be a @@ -559,8 +593,8 @@ symbolic link and the linkpath value will be used as the contents of the symbolic link. @item path -The pathname of the following file. This record overrides the name and -prefix fields in the following ustar header block. +The pathname of the following file. This record overrides the fields +@samp{name} and @samp{prefix} in the following ustar header block. @item size The size of the file in bytes, expressed as a decimal number using @@ -610,31 +644,30 @@ shown in the following table. All lengths and offsets are in decimal. All characters in the header block are coded using the ISO/IEC 646:1991 (ASCII) standard, except in fields storing names for files, users, and groups. For maximum portability between implementations, names should only -contain characters from the portable character set. But if an implementation -supports the use of characters outside of @samp{/} and the portable -character set in names for files, users, and groups, tarlz will use the byte -values in these names unmodified. +contain characters from the portable character set (@pxref{Portable +character set}), but if an implementation supports the use of characters +outside of @samp{/} and the portable character set in names for files, +users, and groups, tarlz will use the byte values in these names unmodified. -The fields name, linkname, and prefix are null-terminated character -strings except when all characters in the array contain non-null -characters including the last character. +The fields @samp{name}, @samp{linkname}, and @samp{prefix} are +null-terminated character strings except when all characters in the array +contain non-null characters including the last character. -The name and the prefix fields produce the pathname of the file. A new -pathname is formed, if prefix is not an empty string (its first +The fields @samp{prefix} and @samp{name} produce the pathname of the file. A +new pathname is formed, if prefix is not an empty string (its first character is not null), by concatenating prefix (up to the first null -character), a <slash> character, and name; otherwise, name is used -alone. In either case, name is terminated at the first null character. -If prefix begins with a null character, it is ignored. In this manner, -pathnames of at most 256 characters can be supported. If a pathname does -not fit in the space provided, an extended record is used to store the -pathname. - -The linkname field does not use the prefix to produce a pathname. If the -linkname does not fit in the 100 characters provided, an extended record +character), a slash character, and name; otherwise, name is used alone. In +either case, name is terminated at the first null character. If prefix +begins with a null character, it is ignored. In this manner, pathnames of at +most 256 characters can be supported. If a pathname does not fit in the +space provided, an extended record is used to store the pathname. + +The field @samp{linkname} does not use the prefix to produce a pathname. If +the linkname does not fit in the 100 characters provided, an extended record is used to store the linkname. -The mode field provides 12 access permission bits. The following table -shows the symbolic name of each bit and its octal value: +The field @samp{mode} provides 12 access permission bits. The following +table shows the symbolic name of each bit and its octal value: @multitable {Bit Name} {Value} {Bit Name} {Value} {Bit Name} {Value} @headitem Bit Name @tab Value @tab Bit Name @tab Value @tab Bit Name @tab Value @@ -644,29 +677,28 @@ shows the symbolic name of each bit and its octal value: @item S_IROTH @tab 00004 @tab S_IWOTH @tab 00002 @tab S_IXOTH @tab 00001 @end multitable -The uid and gid fields are the user and group ID of the owner and group -of the file, respectively. +The fields @samp{uid} and @samp{gid} are the user and group IDs of the owner +and group of the file, respectively. -The size field contains the octal representation of the size of the file -in bytes. If the typeflag field specifies a file of type '0' (regular -file) or '7' (high performance regular file), the number of logical +The field @samp{size} contains the octal representation of the size of the +file in bytes. If the field @samp{typeflag} specifies a file of type '0' +(regular file) or '7' (high performance regular file), the number of logical records following the header is @w{(size / 512)} rounded to the next -integer. For all other values of typeflag, tarlz either sets the size -field to 0 or ignores it, and does not store or expect any logical -records following the header. If the file size is larger than -8_589_934_591 bytes @w{(octal 77777777777)}, an extended record is used -to store the file size. - -The mtime field contains the octal representation of the modification -time of the file at the time it was archived, obtained from the stat() -function. - -The chksum field contains the octal representation of the value of the -simple sum of all bytes in the header logical record. Each byte in the -header is treated as an unsigned value. When calculating the checksum, -the chksum field is treated as if it were all <space> characters. - -The typeflag field contains a single character specifying the type of +integer. For all other values of typeflag, tarlz either sets the size field +to 0 or ignores it, and does not store or expect any logical records +following the header. If the file size is larger than 8_589_934_591 bytes +@w{(octal 77777777777)}, an extended record is used to store the file size. + +The field @samp{mtime} contains the octal representation of the modification +time of the file at the time it was archived, obtained from the function +@samp{stat}. + +The field @samp{chksum} contains the octal representation of the value of +the simple sum of all bytes in the header logical record. Each byte in the +header is treated as an unsigned value. When calculating the checksum, the +chksum field is treated as if it were all space characters. + +The field @samp{typeflag} contains a single character specifying the type of file archived: @table @code @@ -680,8 +712,8 @@ Hard link to another file, of any type, previously archived. Symbolic link. @item '3', '4' -Character special file and block special file respectively. In this case -the devmajor and devminor fields contain information defining the +Character special file and block special file respectively. In this case the +fields @samp{devmajor} and @samp{devminor} contain information defining the device in unspecified format. @item '5' @@ -697,14 +729,15 @@ regular file (type 0). @end table -The magic field contains the ASCII null-terminated string "ustar". The -version field contains the characters "00" (0x30,0x30). The fields uname, -and gname are null-terminated character strings except when all characters -in the array contain non-null characters including the last character. Each -numeric field contains a leading space- or zero-filled, optionally -null-terminated octal number using digits from the ISO/IEC 646:1991 (ASCII) -standard. Tarlz is able to decode numeric fields 1 byte longer than standard -ustar by not requiring a terminating null character. +The field @samp{magic} contains the ASCII null-terminated string "ustar". +The field @samp{version} contains the characters "00" (0x30,0x30). The +fields @samp{uname} and @samp{gname} are null-terminated character strings +except when all characters in the array contain non-null characters +including the last character. Each numeric field contains a leading space- +or zero-filled, optionally null-terminated octal number using digits from +the ISO/IEC 646:1991 (ASCII) standard. Tarlz is able to decode numeric +fields 1 byte longer than standard ustar by not requiring a terminating null +character. @node Amendments to pax format @@ -714,10 +747,10 @@ ustar by not requiring a terminating null character. Tarlz creates safe archives that allow the reliable detection of invalid or corrupt metadata during decoding even when the integrity checking of lzip can't be used because the lzip members are only decompressed partially, as -it happens in parallel @samp{--list} and @samp{--extract}. In order to -achieve this goal, tarlz makes some changes to the variant of the pax format -that it uses. This chapter describes these changes and the concrete reasons -to implement them. +it happens in parallel @samp{--diff}, @samp{--list}, and @samp{--extract}. +In order to achieve this goal, tarlz makes some changes to the variant of +the pax format that it uses. This chapter describes these changes and the +concrete reasons to implement them. @sp 1 @anchor{crc32} @@ -735,7 +768,7 @@ Metadata like file name and file size must be always protected in an archive format because of the adverse effects of undetected corruption in them, potentially much worse that undetected corruption in the data. Even more so in the case of pax because the amount of metadata it stores is potentially -large, making undetected corruption more probable. +large, making undetected corruption and archiver misbehavior more probable. Headers and metadata must be protected separately from data because the integrity checking of lzip may not be able to detect the corruption before @@ -753,12 +786,12 @@ In order to allow the extraction of pax archives by a tar utility conforming to the POSIX-2:1993 standard, POSIX.1-2008 recommends selecting extended header field values that allow such tar to create a regular file containing the extended header records as data. This approach is broken because if the -extended header is needed because of a long file name, the name and prefix -fields will be unable to contain the full pathname of the file. Therefore -the files corresponding to both the extended header and the overridden ustar -header will be extracted using truncated file names, perhaps overwriting -existing files or directories. It may be a security risk to extract a file -with a truncated file name. +extended header is needed because of a long file name, the fields +@samp{prefix} and @samp{name} will be unable to contain the full pathname of +the file. Therefore the files corresponding to both the extended header and +the overridden ustar header will be extracted using truncated file names, +perhaps overwriting existing files or directories. It may be a security risk +to extract a file with a truncated file name. To avoid this problem, tarlz writes extended headers with all fields zeroed except size, chksum, typeflag, magic and version. This prevents old tar @@ -778,10 +811,10 @@ between the extended header and the ustar header. The tarlz format is mainly ustar. Extended pax headers are used only when needed because the length of a file name or link name, or the size of a file -exceed the limits of the ustar format. Adding extended headers to each -member just to record subsecond timestamps seems wasteful for a backup -format. Moreover, minimizing the overhead may help recovering the archive -with lziprecover in case of corruption. +exceed the limits of the ustar format. Adding @w{1 KiB} of extended headers +to each member just to record subsecond timestamps seems wasteful for a +backup format. Moreover, minimizing the overhead may help recovering the +archive with lziprecover in case of corruption. Global pax headers are tolerated, but not supported; they are parsed and ignored. Some operations may not behave as expected if the archive contains @@ -797,7 +830,88 @@ accidental double UTF-8 conversions. If the need arises this behavior will be adjusted with a command line option in the future. -@node Multi-threaded tar +@node Program design +@chapter Internal structure of tarlz +@cindex program design + +The parts of tarlz related to sequential processing of the archive are more +or less similar to any other tar and won't be described here. The interesting +parts described here are those related to Multi-threaded processing. + +The structure of the part of tarlz performing Multi-threaded archive +creation is somewhat similar to that of plzip with the added complication of +the solidity levels. A grouper thread and several worker threads are +created, acting the main thread as muxer (multiplexer) thread. A "packet +courier" takes care of data transfers among threads and limits the maximum +number of data blocks (packets) being processed simultaneously. + +The grouper traverses the directory tree, groups together the metadata of +the files to be archived in each lzip member, and distributes them to the +workers. The workers compress the metadata received from the grouper along +with the file data read from the file system. The muxer collects processed +packets from the workers, and writes them to the archive. + +@verbatim +,--------, +| data|---> to each worker below +| | ,------------, +| file | ,-->| worker 0 |--, +| system | | `------------' | +| | ,---------, | ,------------, | ,-------, ,---------, +|metadata|--->| grouper |-+-->| worker 1 |--+-->| muxer |-->| archive | +`--------' `---------' | `------------' | `-------' `---------' + | ... | + | ,------------, | + `-->| worker N-1 |--' + `------------' +@end verbatim + +Decoding an archive is somewhat similar to how plzip decompresses a regular +file to standard output, with the differences that it is not the data but +only messages what is written to stdout/stderr, and that each worker may +access files in the file system either to read them (diff) or write them +(extract). As in plzip, each worker reads members directly from the archive. + +@verbatim +,--------, +| file |<---> data to/from each worker below +| system | +`--------' + ,------------, + ,-->| worker 0 |--, + | `------------' | +,---------, | ,------------, | ,-------, ,--------, +| archive |-+-->| worker 1 |--+-->| muxer |-->| stdout | +`---------' | `------------' | `-------' | stderr | + | ... | `--------' + | ,------------, | + `-->| worker N-1 |--' + `------------' +@end verbatim + +As misaligned tar.lz archives can't be decoded in parallel, and the +misalignment can't be detected until after decoding has started, a +"mastership request" mechanism has been designed that allows the decoding to +continue instead of signalling an error. + +During parallel decoding, if a worker finds a misalignment, it requests +mastership to decode the rest of the archive. When mastership is requested, +an error_member_id is set, and all subsequently received packets with +member_id > error_member_id are rejected. All workers requesting mastership +are blocked at the request_mastership call until mastership is granted. +Mastership is granted to the delivering worker when its queue is empty to +make sure that all preceding packets have been processed. When mastership is +granted, all packets are deleted and all subsequently received packets not +coming from the master are rejected. + +If a worker can't continue decoding for any cause (for example lack of +memory or finding a split tar member at the beginning of a lzip member), it +requests mastership to print an error and terminate the program. Only if +some other worker requests mastership in a previous lzip member can this +error be avoided. + + +@node Multi-threaded decoding @chapter Limitations of parallel tar decoding @cindex parallel tar decoding @@ -827,8 +941,8 @@ decoding it safely in parallel. Tarlz is able to automatically decode aligned and unaligned multimember tar.lz archives, keeping backwards compatibility. If tarlz finds a member misalignment during multi-threaded decoding, it switches to single-threaded -mode and continues decoding the archive. Currently only the @samp{--list} -option is able to do multi-threaded decoding. +mode and continues decoding the archive. Currently only the options +@samp{--diff} and @samp{--list} are able to do multi-threaded decoding. If the files in the archive are large, multi-threaded @samp{--list} on a regular (seekable) tar.lz archive can be hundreds of times faster than @@ -843,6 +957,10 @@ time plzip -cd silesia.tar.lz | tar -tf - (3.256s) time tarlz -tf silesia.tar.lz (0.020s) @end example +On the other hand, multi-threaded @samp{--list} won't detect corruption in +the tar member data because it only decodes the part of each lzip member +corresponding to the tar member header. + @node Minimum archive sizes @chapter Minimum archive sizes required for multi-threaded block compression @@ -860,7 +978,7 @@ least as large as the number of worker threads times the block size compress, and compression will be proportionally slower. The maximum speed increase achievable on a given archive is limited by the ratio @w{(uncompressed_size / data_size)}. For example, a tarball the size of gcc -or linux will scale up to 10 or 12 processors at level -9. +or linux will scale up to 10 or 14 processors at level -9. The following table shows the minimum uncompressed archive size needed for full use of N processors at a given compression level, using the default @@ -1,18 +1,18 @@ -/* Tarlz - Archiver with multimember lzip compression - Copyright (C) 2013-2019 Antonio Diaz Diaz. +/* Tarlz - Archiver with multimember lzip compression + Copyright (C) 2013-2020 Antonio Diaz Diaz. - 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 of the License, or - (at your option) any later version. + 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 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. + 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. - You should have received a copy of the GNU General Public License - along with this program. If not, see <http://www.gnu.org/licenses/>. + You should have received a copy of the GNU General Public License + along with this program. If not, see <http://www.gnu.org/licenses/>. */ #define _FILE_OFFSET_BITS 64 diff --git a/extended.cc b/extended.cc index 0e647c6..d03494f 100644 --- a/extended.cc +++ b/extended.cc @@ -1,18 +1,18 @@ -/* Tarlz - Archiver with multimember lzip compression - Copyright (C) 2013-2019 Antonio Diaz Diaz. +/* Tarlz - Archiver with multimember lzip compression + Copyright (C) 2013-2020 Antonio Diaz Diaz. - 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 of the License, or - (at your option) any later version. + 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 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. + 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. - You should have received a copy of the GNU General Public License - along with this program. If not, see <http://www.gnu.org/licenses/>. + You should have received a copy of the GNU General Public License + along with this program. If not, see <http://www.gnu.org/licenses/>. */ #define _FILE_OFFSET_BITS 64 @@ -32,6 +32,8 @@ namespace { +const CRC32 crc32c( true ); + unsigned decimal_digits( unsigned long long value ) { unsigned digits = 1; @@ -280,7 +282,8 @@ void Extended::fill_from_ustar( const Tar_header header ) /* Returns file size from record or from ustar header, and resets file_size_. - Used for fast parsing of headers in uncompressed archives. */ + Used for fast parsing of headers in uncompressed archives. +*/ long long Extended::get_file_size_and_reset( const Tar_header header ) { const long long tmp = file_size_; @@ -1,18 +1,18 @@ -/* Tarlz - Archiver with multimember lzip compression - Copyright (C) 2013-2019 Antonio Diaz Diaz. +/* Tarlz - Archiver with multimember lzip compression + Copyright (C) 2013-2020 Antonio Diaz Diaz. - 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 of the License, or - (at your option) any later version. + 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 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. + 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. - You should have received a copy of the GNU General Public License - along with this program. If not, see <http://www.gnu.org/licenses/>. + You should have received a copy of the GNU General Public License + along with this program. If not, see <http://www.gnu.org/licenses/>. */ #define _FILE_OFFSET_BITS 64 @@ -32,19 +32,21 @@ #include <utime.h> #include <sys/stat.h> #include <sys/types.h> +#if !defined __FreeBSD__ && !defined __OpenBSD__ && !defined __NetBSD__ && \ + !defined __DragonFly__ && !defined __APPLE__ #include <sys/sysmacros.h> // for major, minor, makedev +#endif #include <lzlib.h> #include "arg_parser.h" -#include "lzip_index.h" #include "tarlz.h" +#include "lzip_index.h" +#include "archive_reader.h" namespace { Resizable_buffer grbuf; -bool archive_is_uncompressed_seekable = false; -bool archive_has_lz_ext; // local var for archive_read bool skip_warn( const bool reset = false ) // avoid duplicate warnings { @@ -57,6 +59,13 @@ bool skip_warn( const bool reset = false ) // avoid duplicate warnings } +void read_error( const Archive_reader & ar ) + { + show_file_error( ar.ad.namep, ar.e_msg(), ar.e_code() ); + if( ar.e_skip() ) skip_warn(); + } + + bool make_path( const std::string & name ) { const mode_t mode = S_IRWXU | S_IRGRP | S_IXGRP | S_IROTH | S_IXOTH; @@ -86,255 +95,47 @@ bool make_path( const std::string & name ) } -// Return value: 0 = OK, 1 = damaged member, 2 = fatal error. -// If sizep and error, return in *sizep the number of bytes read. -// The first 6 bytes of the archive must be intact for islz to be meaningful. -int archive_read( const char * const archive_namep, const int infd, - uint8_t * const buf, const int size, int * const sizep = 0 ) - { - static LZ_Decoder * decoder = 0; - static bool at_eof = false; - static bool fatal = false; - static bool first_call = true; - - if( sizep ) *sizep = 0; - if( fatal ) return 2; - if( first_call ) // check format - { - first_call = false; - if( size != header_size ) - internal_error( "size != header_size on first call." ); - const int rd = readblock( infd, buf, size ); - if( sizep ) *sizep = rd; - if( rd != size && errno ) - { show_file_error( archive_namep, "Error reading archive", errno ); - fatal = true; return 2; } - const Lzip_header & header = (*(const Lzip_header *)buf); - bool islz = ( rd >= min_member_size && header.verify_magic() && - header.verify_version() && - isvalid_ds( header.dictionary_size() ) ); - const bool istar = ( rd == size && verify_ustar_chksum( buf ) ); - const bool iseof = - ( !islz && !istar && rd == size && block_is_zero( buf, size ) ); - if( !islz && !istar && !iseof ) // corrupt or invalid format - { - show_file_error( archive_namep, posix_msg ); - if( archive_has_lz_ext && rd >= min_member_size ) islz = true; - if( !islz ) return 1; - } - if( !islz ) // uncompressed - { if( rd == size ) return 0; fatal = true; return 2; } - archive_is_uncompressed_seekable = false; // compressed - decoder = LZ_decompress_open(); - if( !decoder || LZ_decompress_errno( decoder ) != LZ_ok ) - { show_error( mem_msg ); - LZ_decompress_close( decoder ); fatal = true; return 2; } - if( LZ_decompress_write( decoder, buf, rd ) != rd ) - internal_error( "library error (LZ_decompress_write)." ); - const int res = archive_read( archive_namep, infd, buf, size, sizep ); - if( res != 0 ) { if( res == 2 ) fatal = true; return res; } - if( verify_ustar_chksum( buf ) || block_is_zero( buf, size ) ) return 0; - show_file_error( archive_namep, posix_lz_msg ); - fatal = true; return 2; - } - - if( !decoder ) // uncompressed - { - const int rd = readblock( infd, buf, size ); if( rd == size ) return 0; - if( sizep ) *sizep = rd; - show_file_error( archive_namep, end_msg ); fatal = true; return 2; - } - const int ibuf_size = 16384; - uint8_t ibuf[ibuf_size]; - int sz = 0; - while( sz < size ) - { - const int rd = LZ_decompress_read( decoder, buf + sz, size - sz ); - if( rd < 0 ) - { - if( LZ_decompress_sync_to_member( decoder ) < 0 ) - internal_error( "library error (LZ_decompress_sync_to_member)." ); - skip_warn(); set_error_status( 2 ); return 1; - } - if( rd == 0 && LZ_decompress_finished( decoder ) == 1 ) - { LZ_decompress_close( decoder ); - show_file_error( archive_namep, end_msg ); fatal = true; return 2; } - sz += rd; if( sizep ) *sizep = sz; - if( sz < size && !at_eof && LZ_decompress_write_size( decoder ) > 0 ) - { - const int rsize = std::min( ibuf_size, LZ_decompress_write_size( decoder ) ); - const int rd = readblock( infd, ibuf, rsize ); - if( LZ_decompress_write( decoder, ibuf, rd ) != rd ) - internal_error( "library error (LZ_decompress_write)." ); - if( rd < rsize ) - { - at_eof = true; LZ_decompress_finish( decoder ); - if( errno ) - { show_file_error( archive_namep, "Error reading archive", errno ); - fatal = true; return 2; } - } - } - } - return 0; - } - - -int skip_member( const char * const archive_namep, const int infd, - const Extended & extended ) +int skip_member( Archive_reader & ar, const Extended & extended ) { - long long rest = extended.file_size(); - const int rem = rest % header_size; - if( rem ) rest += header_size - rem; // padding - if( archive_is_uncompressed_seekable && lseek( infd, rest, SEEK_CUR ) > 0 ) - return 0; - const int bufsize = 32 * header_size; - uint8_t buf[bufsize]; - while( rest > 0 ) - { - const int rsize = ( rest >= bufsize ) ? bufsize : rest; - const int ret = archive_read( archive_namep, infd, buf, rsize ); - if( ret != 0 ) { if( ret == 2 ) return 2; else break; } - rest -= rsize; - } + const int ret = ar.skip_member( extended ); + if( ret != 0 ) { read_error( ar ); if( ret == 2 ) return 2; } return 0; } -void show_file_diff( const char * const filename, const char * const msg ) - { - if( verbosity >= 0 ) - { std::printf( "%s: %s\n", filename, msg ); std::fflush( stdout ); } - } - - -int compare_member( const char * const archive_namep, const int infd1, - const Extended & extended, const Tar_header header, - const bool ignore_ids ) +int compare_member( const Cl_options & cl_opts, Archive_reader & ar, + const Extended & extended, const Tar_header header ) { if( !show_member_name( extended, header, 1, grbuf ) ) return 1; - long long rest = extended.file_size(); - const char * const filename = extended.path().c_str(); + std::string estr, ostr; + const bool stat_differs = + !compare_file_type( estr, ostr, cl_opts, extended, header ); + if( estr.size() ) std::fputs( estr.c_str(), stderr ); + if( ostr.size() ) { std::fputs( ostr.c_str(), stdout ); std::fflush( stdout ); } + if( extended.file_size() <= 0 ) return 0; const Typeflag typeflag = (Typeflag)header[typeflag_o]; - bool diff = false, size_differs = false, type_differs = true; - struct stat st; - if( hstat( filename, &st ) != 0 ) - show_file_error( filename, "Warning: Can't stat", errno ); - else if( ( typeflag == tf_regular || typeflag == tf_hiperf ) && - !S_ISREG( st.st_mode ) ) - show_file_diff( filename, "Is not a regular file" ); - else if( typeflag == tf_symlink && !S_ISLNK( st.st_mode ) ) - show_file_diff( filename, "Is not a symlink" ); - else if( typeflag == tf_chardev && !S_ISCHR( st.st_mode ) ) - show_file_diff( filename, "Is not a character device" ); - else if( typeflag == tf_blockdev && !S_ISBLK( st.st_mode ) ) - show_file_diff( filename, "Is not a block device" ); - else if( typeflag == tf_directory && !S_ISDIR( st.st_mode ) ) - show_file_diff( filename, "Is not a directory" ); - else if( typeflag == tf_fifo && !S_ISFIFO( st.st_mode ) ) - show_file_diff( filename, "Is not a FIFO" ); - else - { - type_differs = false; - if( typeflag != tf_symlink ) - { - const mode_t mode = parse_octal( header + mode_o, mode_l ); // 12 bits - if( mode != ( st.st_mode & ( S_ISUID | S_ISGID | S_ISVTX | - S_IRWXU | S_IRWXG | S_IRWXO ) ) ) - { show_file_diff( filename, "Mode differs" ); diff = true; } - } - if( !ignore_ids ) - { - if( (uid_t)parse_octal( header + uid_o, uid_l ) != st.st_uid ) - { show_file_diff( filename, "Uid differs" ); diff = true; } - if( (gid_t)parse_octal( header + gid_o, gid_l ) != st.st_gid ) - { show_file_diff( filename, "Gid differs" ); diff = true; } - } - if( typeflag != tf_symlink ) - { - if( typeflag != tf_directory ) - { - const time_t mtime = parse_octal( header + mtime_o, mtime_l ); // 33 bits - if( mtime != st.st_mtime ) - { show_file_diff( filename, "Mod time differs" ); diff = true; } - } - if( ( typeflag == tf_regular || typeflag == tf_hiperf ) && - rest != st.st_size ) // don't compare contents - { show_file_diff( filename, "Size differs" ); size_differs = true; } - if( ( typeflag == tf_chardev || typeflag == tf_blockdev ) && - ( parse_octal( header + devmajor_o, devmajor_l ) != - (unsigned)major( st.st_rdev ) || - parse_octal( header + devminor_o, devminor_l ) != - (unsigned)minor( st.st_rdev ) ) ) - { show_file_diff( filename, "Device number differs" ); diff = true; } - } - else - { - char * const buf = new char[st.st_size+1]; - long len = readlink( filename, buf, st.st_size ); - bool e = ( len != st.st_size ); - if( !e ) - { - while( len > 1 && buf[len-1] == '/' ) --len; // trailing '/' - buf[len] = 0; - if( extended.linkpath() != buf ) e = true; - } - delete[] buf; - if( e ) { show_file_diff( filename, "Symlink differs" ); diff = true; } - } - } - if( diff || size_differs || type_differs ) - { diff = false; set_error_status( 1 ); } - if( rest <= 0 ) return 0; - if( ( typeflag != tf_regular && typeflag != tf_hiperf ) || - size_differs || type_differs ) - return skip_member( archive_namep, infd1, extended ); + if( ( typeflag != tf_regular && typeflag != tf_hiperf ) || stat_differs ) + return skip_member( ar, extended ); // else compare file contents - const int rem = rest % header_size; - const int padding = rem ? header_size - rem : 0; - const int bufsize = 32 * header_size; - uint8_t buf1[bufsize]; - uint8_t buf2[bufsize]; + const char * const filename = extended.path().c_str(); const int infd2 = open_instream( filename ); if( infd2 < 0 ) - { set_error_status( 1 ); - return skip_member( archive_namep, infd1, extended ); } - int retval = 0; - while( rest > 0 ) - { - const int rsize1 = ( rest >= bufsize ) ? bufsize : rest + padding; - const int rsize2 = ( rest >= bufsize ) ? bufsize : rest; - const int ret = archive_read( archive_namep, infd1, buf1, rsize1 ); - if( ret != 0 ) { if( ret == 2 ) retval = 2; diff = true; break; } - if( !diff ) - { - const int rd = readblock( infd2, buf2, rsize2 ); - if( rd != rsize2 ) - { - if( errno ) show_file_error( filename, "Read error", errno ); - else show_file_diff( filename, "EOF found in file" ); - diff = true; - } - else - { - int i = 0; while( i < rsize2 && buf1[i] == buf2[i] ) ++i; - if( i < rsize2 ) - { show_file_diff( filename, "Contents differ" ); diff = true; } - } - } - if( rest < bufsize ) break; - rest -= rsize1; - } - if( diff ) set_error_status( 1 ); - close( infd2 ); + { set_error_status( 1 ); return skip_member( ar, extended ); } + int retval = compare_file_contents( estr, ostr, ar, extended.file_size(), + filename, infd2 ); + if( retval ) { read_error( ar ); if( retval != 2 ) retval = 0; } + else { if( estr.size() ) std::fputs( estr.c_str(), stderr ); + if( ostr.size() ) + { std::fputs( ostr.c_str(), stdout ); std::fflush( stdout ); } } return retval; } -int list_member( const char * const archive_namep, const int infd, +int list_member( Archive_reader & ar, const Extended & extended, const Tar_header header ) { if( !show_member_name( extended, header, 0, grbuf ) ) return 1; - return skip_member( archive_namep, infd, extended ); + return skip_member( ar, extended ); } @@ -346,17 +147,27 @@ bool contains_dotdot( const char * const filename ) } -int extract_member( const char * const archive_namep, const int infd, - const Extended & extended, const Tar_header header, - const bool keep_damaged ) +mode_t get_umask() + { + static mode_t mask = 0; // read once, cache the result + static bool first_call = true; + if( first_call ) { first_call = false; mask = umask( 0 ); umask( mask ); + mask &= S_IRWXU | S_IRWXG | S_IRWXO; } + return mask; + } + + +int extract_member( const Cl_options & cl_opts, Archive_reader & ar, + const Extended & extended, const Tar_header header ) { const char * const filename = extended.path().c_str(); if( contains_dotdot( filename ) ) { show_file_error( filename, "Contains a '..' component, skipping." ); - return skip_member( archive_namep, infd, extended ); + return skip_member( ar, extended ); } - const mode_t mode = parse_octal( header + mode_o, mode_l ); // 12 bits + mode_t mode = parse_octal( header + mode_o, mode_l ); // 12 bits + if( geteuid() != 0 && !cl_opts.preserve_permissions ) mode &= ~get_umask(); const time_t mtime = parse_octal( header + mtime_o, mtime_l ); // 33 bits const Typeflag typeflag = (Typeflag)header[typeflag_o]; const bool islink = ( typeflag == tf_link || typeflag == tf_symlink ); @@ -381,7 +192,7 @@ int extract_member( const char * const archive_namep, const int infd, ( !hard && symlink( linkname, filename ) != 0 ) ) { if( verbosity >= 0 ) - std::fprintf( stderr, "Can't %slink file '%s' to '%s': %s.\n", + std::fprintf( stderr, "Can't %slink '%s' to '%s': %s.\n", hard ? "" : "sym", linkname, filename, std::strerror( errno ) ); return 2; @@ -442,14 +253,14 @@ int extract_member( const char * const archive_namep, const int infd, while( rest > 0 ) { const int rsize = ( rest >= bufsize ) ? bufsize : rest + padding; - int rd; - const int ret = archive_read( archive_namep, infd, buf, rsize, &rd ); + const int ret = ar.read( buf, rsize ); if( ret != 0 ) { + read_error( ar ); if( outfd >= 0 ) { - if( keep_damaged ) - { writeblock( outfd, buf, std::min( rest, (long long)rd ) ); + if( cl_opts.keep_damaged ) + { writeblock( outfd, buf, std::min( rest, (long long)ar.e_size() ) ); close( outfd ); } else { close( outfd ); std::remove( filename ); } } @@ -473,85 +284,188 @@ int extract_member( const char * const archive_namep, const int infd, } -bool parse_records( const char * const archive_namep, const int infd, - Extended & extended, const Tar_header header, - Resizable_buffer & rbuf, const bool permissive ) +void format_file_diff( std::string & ostr, const char * const filename, + const char * const msg ) { - const long long edsize = parse_octal( header + size_o, size_l ); - const long long bufsize = round_up( edsize ); - if( edsize <= 0 || edsize >= 1LL << 33 || bufsize >= INT_MAX ) - return false; // overflow or no extended data - if( !rbuf.resize( bufsize ) ) return false; // extended records buffer - return ( archive_read( archive_namep, infd, (uint8_t *)rbuf(), bufsize ) == 0 && - extended.parse( rbuf(), edsize, permissive ) ); + if( verbosity < 0 ) return; + { ostr += filename; ostr += ": "; ostr += msg; ostr += '\n'; } } } // end namespace -int decode( const std::string & archive_name, const Arg_parser & parser, - const int filenames, const int num_workers, const int debug_level, - const Program_mode program_mode, const bool ignore_ids, - const bool keep_damaged, const bool missing_crc, - const bool permissive ) +bool compare_file_type( std::string & estr, std::string & ostr, + const Cl_options & cl_opts, + const Extended & extended, const Tar_header header ) + { + const char * const filename = extended.path().c_str(); + const Typeflag typeflag = (Typeflag)header[typeflag_o]; + struct stat st; + bool diff = false, size_differs = false, type_differs = true; + if( hstat( filename, &st, cl_opts.dereference ) != 0 ) + format_file_error( estr, filename, "Warning: Can't stat", errno ); + else if( ( typeflag == tf_regular || typeflag == tf_hiperf ) && + !S_ISREG( st.st_mode ) ) + format_file_diff( ostr, filename, "Is not a regular file" ); + else if( typeflag == tf_symlink && !S_ISLNK( st.st_mode ) ) + format_file_diff( ostr, filename, "Is not a symlink" ); + else if( typeflag == tf_chardev && !S_ISCHR( st.st_mode ) ) + format_file_diff( ostr, filename, "Is not a character device" ); + else if( typeflag == tf_blockdev && !S_ISBLK( st.st_mode ) ) + format_file_diff( ostr, filename, "Is not a block device" ); + else if( typeflag == tf_directory && !S_ISDIR( st.st_mode ) ) + format_file_diff( ostr, filename, "Is not a directory" ); + else if( typeflag == tf_fifo && !S_ISFIFO( st.st_mode ) ) + format_file_diff( ostr, filename, "Is not a FIFO" ); + else + { + type_differs = false; + if( typeflag != tf_symlink ) + { + const mode_t mode = parse_octal( header + mode_o, mode_l ); // 12 bits + if( mode != ( st.st_mode & ( S_ISUID | S_ISGID | S_ISVTX | + S_IRWXU | S_IRWXG | S_IRWXO ) ) ) + { format_file_diff( ostr, filename, "Mode differs" ); diff = true; } + } + if( !cl_opts.ignore_ids ) + { + if( (uid_t)parse_octal( header + uid_o, uid_l ) != st.st_uid ) + { format_file_diff( ostr, filename, "Uid differs" ); diff = true; } + if( (gid_t)parse_octal( header + gid_o, gid_l ) != st.st_gid ) + { format_file_diff( ostr, filename, "Gid differs" ); diff = true; } + } + if( typeflag != tf_symlink ) + { + if( typeflag != tf_directory ) + { + const time_t mtime = parse_octal( header + mtime_o, mtime_l ); // 33 bits + if( mtime != st.st_mtime ) + { format_file_diff( ostr, filename, "Mod time differs" ); diff = true; } + } + if( ( typeflag == tf_regular || typeflag == tf_hiperf ) && + extended.file_size() != st.st_size ) // don't compare contents + { format_file_diff( ostr, filename, "Size differs" ); size_differs = true; } + if( ( typeflag == tf_chardev || typeflag == tf_blockdev ) && + ( parse_octal( header + devmajor_o, devmajor_l ) != + (unsigned)major( st.st_rdev ) || + parse_octal( header + devminor_o, devminor_l ) != + (unsigned)minor( st.st_rdev ) ) ) + { format_file_diff( ostr, filename, "Device number differs" ); diff = true; } + } + else + { + char * const buf = new char[st.st_size+1]; + long len = readlink( filename, buf, st.st_size ); + bool e = ( len != st.st_size ); + if( !e ) + { + while( len > 1 && buf[len-1] == '/' ) --len; // trailing '/' + buf[len] = 0; + if( extended.linkpath() != buf ) e = true; + } + delete[] buf; + if( e ) { format_file_diff( ostr, filename, "Symlink differs" ); diff = true; } + } + } + if( diff || size_differs || type_differs ) set_error_status( 1 ); + return !( size_differs || type_differs ); + } + + +bool compare_file_contents( std::string & estr, std::string & ostr, + Archive_reader_base & ar, const long long file_size, + const char * const filename, const int infd2 ) + { + long long rest = file_size; + const int rem = rest % header_size; + const int padding = rem ? header_size - rem : 0; + const int bufsize = 32 * header_size; + uint8_t buf1[bufsize]; + uint8_t buf2[bufsize]; + int retval = 0; + bool diff = false; + estr.clear(); ostr.clear(); + while( rest > 0 ) + { + const int rsize1 = ( rest >= bufsize ) ? bufsize : rest + padding; + const int rsize2 = ( rest >= bufsize ) ? bufsize : rest; + if( ( retval = ar.read( buf1, rsize1 ) ) != 0 ) { diff = true; break; } + if( !diff ) + { + const int rd = readblock( infd2, buf2, rsize2 ); + if( rd != rsize2 ) + { + if( errno ) format_file_error( estr, filename, "Read error", errno ); + else format_file_diff( ostr, filename, "EOF found in file" ); + diff = true; + } + else + { + int i = 0; while( i < rsize2 && buf1[i] == buf2[i] ) ++i; + if( i < rsize2 ) + { format_file_diff( ostr, filename, "Contents differ" ); diff = true; } + } + } + if( rest < bufsize ) break; + rest -= rsize1; + } + close( infd2 ); + if( diff ) set_error_status( 1 ); + return retval; + } + + +int decode( const Cl_options & cl_opts ) { - const bool from_stdin = archive_name.empty(); - const char * const archive_namep = - from_stdin ? "(stdin)" : archive_name.c_str(); - const int infd = from_stdin ? STDIN_FILENO : open_instream( archive_name ); - if( infd < 0 ) return 1; + // open archive before changing working directory + const Archive_descriptor ad( cl_opts.archive_name ); + if( ad.infd < 0 ) return 1; // Execute -C options and mark filenames to be compared, extracted or listed. // name_pending is of type char instead of bool to allow concurrent update. - std::vector< char > name_pending( parser.arguments(), false ); - for( int i = 0; i < parser.arguments(); ++i ) + std::vector< char > name_pending( cl_opts.parser.arguments(), false ); + for( int i = 0; i < cl_opts.parser.arguments(); ++i ) { - const int code = parser.code( i ); - if( code == 'C' && program_mode != m_list ) + const int code = cl_opts.parser.code( i ); + if( code == 'C' && cl_opts.program_mode != m_list ) { - const char * const dir = parser.argument( i ).c_str(); + const char * const dir = cl_opts.parser.argument( i ).c_str(); if( chdir( dir ) != 0 ) { show_file_error( dir, "Error changing working directory", errno ); return 1; } } - if( !code && parser.argument( i ).size() && - !Exclude::excluded( parser.argument( i ).c_str() ) ) + if( !code && cl_opts.parser.argument( i ).size() && + !Exclude::excluded( cl_opts.parser.argument( i ).c_str() ) ) name_pending[i] = true; } // multi-threaded --list is faster even with 1 thread and 1 file in archive - if( program_mode == m_list && num_workers > 0 ) + // (but multi-threaded --diff probably needs at least 2 of each) + if( ( cl_opts.program_mode == m_diff || cl_opts.program_mode == m_list ) && + cl_opts.num_workers > 0 ) { - const Lzip_index lzip_index( infd, true, false ); // only regular files - const long members = lzip_index.members(); - if( lzip_index.retval() == 0 && members >= 2 ) // one file + eof + if( ad.indexed && ad.lzip_index.members() >= 2 ) // one file + eof { - // show_file_error( archive_namep, "Is compressed seekable" ); - return list_lz( archive_namep, parser, name_pending, lzip_index, - filenames, debug_level, infd, - std::min( (long)num_workers, members ), - missing_crc, permissive ); + // show_file_error( ad.namep, "Is compressed seekable" ); + return decode_lz( cl_opts, ad, name_pending ); } - if( lseek( infd, 0, SEEK_SET ) == 0 && lzip_index.retval() != 0 && - lzip_index.file_size() > 3 * header_size ) - archive_is_uncompressed_seekable = true; // unless compressed corrupt } - archive_has_lz_ext = has_lz_ext( archive_name ); // var for archive_read + Archive_reader ar( ad ); Extended extended; // metadata from extended records int retval = 0; bool prev_extended = false; // prev header was extended while( true ) // process one tar header per iteration { Tar_header header; - const int ret = archive_read( archive_namep, infd, header, header_size ); - if( ret == 2 ) { retval = 2; break; } + const int ret = ar.read( header, header_size ); + if( ret != 0 ) { read_error( ar ); if( ret == 2 ) { retval = 2; break; } } if( ret != 0 || !verify_ustar_chksum( header ) ) { if( ret == 0 && block_is_zero( header, header_size ) ) { - if( !prev_extended || permissive ) break; // EOF - show_file_error( archive_namep, fv_msg1 ); + if( !prev_extended || cl_opts.permissive ) break; // EOF + show_file_error( ad.namep, fv_msg1 ); retval = 2; break; } if( skip_warn() && verbosity >= 2 ) @@ -563,26 +477,24 @@ int decode( const std::string & archive_name, const Arg_parser & parser, const Typeflag typeflag = (Typeflag)header[typeflag_o]; if( typeflag == tf_global ) { - if( prev_extended && !permissive ) - { show_file_error( archive_namep, fv_msg2 ); retval = 2; break; } + if( prev_extended && !cl_opts.permissive ) + { show_file_error( ad.namep, fv_msg2 ); retval = 2; break; } Extended dummy; // global headers are parsed and ignored - if( !parse_records( archive_namep, infd, dummy, header, grbuf, true ) ) - { show_file_error( archive_namep, - "Error in global extended records. Skipping to next header." ); - set_error_status( 2 ); } + const int ret = ar.parse_records( dummy, header, grbuf, true ); + if( ret != 0 ) { show_file_error( ad.namep, gblrec_msg ); skip_warn(); + set_error_status( ret ); } continue; } if( typeflag == tf_extended ) { - if( prev_extended && !permissive ) - { show_file_error( archive_namep, fv_msg3 ); retval = 2; break; } - if( !parse_records( archive_namep, infd, extended, header, grbuf, - permissive ) ) - { show_file_error( archive_namep, - "Error in extended records. Skipping to next header." ); - extended.reset(); set_error_status( 2 ); } - else if( !extended.crc_present() && missing_crc ) - { show_file_error( archive_namep, mcrc_msg ); retval = 2; break; } + if( prev_extended && !cl_opts.permissive ) + { show_file_error( ad.namep, fv_msg3 ); retval = 2; break; } + const int ret = ar.parse_records( extended, header, grbuf, + cl_opts.permissive ); + if( ret != 0 ) { show_file_error( ad.namep, extrec_msg ); skip_warn(); + extended.reset(); set_error_status( ret ); } + else if( !extended.crc_present() && cl_opts.missing_crc ) + { show_file_error( ad.namep, mcrc_msg ); retval = 2; break; } prev_extended = true; continue; } @@ -590,29 +502,26 @@ int decode( const std::string & archive_name, const Arg_parser & parser, extended.fill_from_ustar( header ); // copy metadata from header - if( check_skip_filename( parser, name_pending, extended.path().c_str(), - filenames ) ) - retval = skip_member( archive_namep, infd, extended ); - else if( program_mode == m_list ) - retval = list_member( archive_namep, infd, extended, header ); - else if( program_mode == m_diff ) - retval = compare_member( archive_namep, infd, extended, header, ignore_ids ); - else retval = extract_member( archive_namep, infd, extended, header, - keep_damaged ); + if( check_skip_filename( cl_opts, name_pending, extended.path().c_str() ) ) + retval = skip_member( ar, extended ); + else if( cl_opts.program_mode == m_list ) + retval = list_member( ar, extended, header ); + else if( cl_opts.program_mode == m_diff ) + retval = compare_member( cl_opts, ar, extended, header ); + else retval = extract_member( cl_opts, ar, extended, header ); extended.reset(); if( retval ) { show_error( "Error is not recoverable: exiting now." ); break; } } - if( close( infd ) != 0 && !retval ) - { show_file_error( archive_namep, "Error closing archive", errno ); + if( close( ad.infd ) != 0 && !retval ) + { show_file_error( ad.namep, "Error closing archive", errno ); retval = 1; } - if( retval == 0 ) for( int i = 0; i < parser.arguments(); ++i ) - if( !parser.code( i ) && parser.argument( i ).size() && name_pending[i] ) - { - show_file_error( parser.argument( i ).c_str(), "Not found in archive." ); - retval = 1; - } - return final_exit_status( retval, program_mode != m_diff ); + if( retval == 0 ) + for( int i = 0; i < cl_opts.parser.arguments(); ++i ) + if( nonempty_arg( cl_opts.parser, i ) && name_pending[i] ) + { show_file_error( cl_opts.parser.argument( i ).c_str(), + "Not found in archive." ); retval = 1; } + return final_exit_status( retval, cl_opts.program_mode != m_diff ); } diff --git a/list_lz.cc b/list_lz.cc deleted file mode 100644 index bdfad36..0000000 --- a/list_lz.cc +++ /dev/null @@ -1,586 +0,0 @@ -/* Tarlz - Archiver with multimember lzip compression - Copyright (C) 2013-2019 Antonio Diaz Diaz. - - 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 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. - - You should have received a copy of the GNU General Public License - along with this program. If not, see <http://www.gnu.org/licenses/>. -*/ - -#define _FILE_OFFSET_BITS 64 - -#include <algorithm> -#include <cerrno> -#include <climits> -#include <cstdio> -#include <cstdlib> -#include <cstring> -#include <queue> -#include <string> -#include <vector> -#include <pthread.h> -#include <stdint.h> -#include <unistd.h> -#include <lzlib.h> - -#include "arg_parser.h" -#include "lzip_index.h" -#include "tarlz.h" - - -namespace { - -// Returns the number of bytes really read. -// If (returned value < size) and (errno == 0), means EOF was reached. -// -int preadblock( const int fd, uint8_t * const buf, const int size, - const long long pos ) - { - int sz = 0; - errno = 0; - while( sz < size ) - { - const int n = pread( fd, buf + sz, size - sz, pos + sz ); - if( n > 0 ) sz += n; - else if( n == 0 ) break; // EOF - else if( errno != EINTR ) break; - errno = 0; - } - return sz; - } - -/* -// Returns the number of bytes really written. -// If (returned value < size), it is always an error. -// -int pwriteblock( const int fd, const uint8_t * const buf, const int size, - const long long pos ) - { - int sz = 0; - errno = 0; - while( sz < size ) - { - const int n = pwrite( fd, buf + sz, size - sz, pos + sz ); - if( n > 0 ) sz += n; - else if( n < 0 && errno != EINTR ) break; - errno = 0; - } - return sz; - } -*/ - -struct Packet // member name and metadata or error message - { - enum Status { ok, member_done, error }; - long member_id; // lzip member containing the header of this tar member - std::string line; // member name and metadata ready to print, if any - Status status; - Packet( const long i, const char * const msg, const Status s = ok ) - : member_id( i ), line( msg ), status( s ) {} - }; - - -class Packet_courier // moves packets around - { -public: - unsigned ocheck_counter; - unsigned owait_counter; -private: - long error_member_id; // first lzip member with error/misalign/eof - int deliver_worker_id; // worker queue currently delivering packets - int master_worker_id; // worker in charge if error/misalignment/eof - std::vector< std::queue< const Packet * > > opacket_queues; - int num_working; // number of workers still running - const int num_workers; // number of workers - const unsigned out_slots; // max output packets per queue - pthread_mutex_t omutex; - pthread_cond_t oav_or_exit; // output packet available or all workers exited - std::vector< pthread_cond_t > slot_av; // output slot available - pthread_cond_t check_master; - bool eof_found_; - - Packet_courier( const Packet_courier & ); // declared as private - void operator=( const Packet_courier & ); // declared as private - -public: - Packet_courier( const int workers, const int slots ) - : ocheck_counter( 0 ), owait_counter( 0 ), - error_member_id( -1 ), deliver_worker_id( 0 ), master_worker_id( -1 ), - opacket_queues( workers ), num_working( workers ), - num_workers( workers ), out_slots( slots ), slot_av( workers ), - eof_found_( false ) - { - xinit_mutex( &omutex ); xinit_cond( &oav_or_exit ); - for( unsigned i = 0; i < slot_av.size(); ++i ) xinit_cond( &slot_av[i] ); - xinit_cond( &check_master ); - } - - ~Packet_courier() - { - xdestroy_cond( &check_master ); - for( unsigned i = 0; i < slot_av.size(); ++i ) xdestroy_cond( &slot_av[i] ); - xdestroy_cond( &oav_or_exit ); xdestroy_mutex( &omutex ); - } - - bool eof_found() const { return eof_found_; } - void report_eof() { eof_found_ = true; } - - bool mastership_granted() const { return master_worker_id >= 0; } - - bool request_mastership( const long member_id, const int worker_id ) - { - xlock( &omutex ); - if( mastership_granted() ) // already granted - { xunlock( &omutex ); return ( master_worker_id == worker_id ); } - if( error_member_id < 0 || error_member_id > member_id ) - error_member_id = member_id; - while( !mastership_granted() && ( worker_id != deliver_worker_id || - !opacket_queues[deliver_worker_id].empty() ) ) - xwait( &check_master, &omutex ); - if( !mastership_granted() && worker_id == deliver_worker_id && - opacket_queues[deliver_worker_id].empty() ) - { - master_worker_id = worker_id; // grant mastership - for( int i = 0; i < num_workers; ++i ) // delete all packets - while( !opacket_queues[i].empty() ) - opacket_queues[i].pop(); - xbroadcast( &check_master ); - xunlock( &omutex ); - return true; - } - xunlock( &omutex ); - return false; // mastership granted to another worker - } - - void worker_finished() - { - // notify muxer when last worker exits - xlock( &omutex ); - if( --num_working == 0 ) xsignal( &oav_or_exit ); - xunlock( &omutex ); - } - - /* Collect a packet from a worker. - If a packet is rejected, the worker must terminate. */ - bool collect_packet( const int worker_id, const long member_id, - const char * const msg, - const Packet::Status status = Packet::ok ) - { - const Packet * const opacket = new Packet( member_id, msg, status ); - xlock( &omutex ); - if( ( mastership_granted() && master_worker_id != worker_id ) || - ( error_member_id >= 0 && error_member_id < opacket->member_id ) ) - { xunlock( &omutex ); delete opacket; return false; } // reject packet - while( opacket_queues[worker_id].size() >= out_slots ) - xwait( &slot_av[worker_id], &omutex ); - opacket_queues[worker_id].push( opacket ); - if( worker_id == deliver_worker_id ) xsignal( &oav_or_exit ); - xunlock( &omutex ); - return true; - } - - /* Deliver a packet to muxer. - If packet.status == Packet::member_done, move to next queue. - If packet.line.empty(), wait again (empty lzip member). */ - const Packet * deliver_packet() - { - const Packet * opacket = 0; - xlock( &omutex ); - ++ocheck_counter; - while( true ) - { - while( opacket_queues[deliver_worker_id].empty() && num_working > 0 ) - { - ++owait_counter; - if( !mastership_granted() && error_member_id >= 0 ) - xbroadcast( &check_master ); // mastership requested not yet granted - xwait( &oav_or_exit, &omutex ); - } - if( opacket_queues[deliver_worker_id].empty() ) break; - opacket = opacket_queues[deliver_worker_id].front(); - opacket_queues[deliver_worker_id].pop(); - if( opacket_queues[deliver_worker_id].size() + 1 == out_slots ) - xsignal( &slot_av[deliver_worker_id] ); - if( opacket->status == Packet::member_done && !mastership_granted() ) - { if( ++deliver_worker_id >= num_workers ) deliver_worker_id = 0; } - if( !opacket->line.empty() ) break; - delete opacket; opacket = 0; - } - xunlock( &omutex ); - return opacket; - } - - bool finished() // all packets delivered to muxer - { - if( num_working != 0 ) return false; - for( int i = 0; i < num_workers; ++i ) - if( !opacket_queues[i].empty() ) return false; - return true; - } - }; - - -int list_member_lz( LZ_Decoder * const decoder, const int infd, - long long & file_pos, const long long member_end, - const long long cdata_size, long long & data_pos, - const long long mdata_end, Packet_courier & courier, - const Extended & extended, const Tar_header header, - Resizable_buffer & rbuf, const long member_id, - const int worker_id, const char ** msg, const bool skip ) - { - long long rest = extended.file_size(); - const int rem = rest % header_size; - if( rem ) rest += header_size - rem; // padding - const long long data_rest = mdata_end - ( data_pos + rest ); - - if( verbosity < 0 || skip ) rbuf()[0] = 0; - else if( !format_member_name( extended, header, rbuf, verbosity > 0 ) ) - { *msg = mem_msg; return 1; } - if( !courier.collect_packet( worker_id, member_id, rbuf(), - data_rest ? Packet::ok : Packet::member_done ) ) - { *msg = "other worker found an error"; return 1; } - if( data_rest ) - return skip_member_lz( decoder, infd, file_pos, member_end, cdata_size, - data_pos, rest, msg ); - data_pos = mdata_end; - return 0; - } - - -struct Worker_arg - { - const Lzip_index * lzip_index; - Packet_courier * courier; - const Arg_parser * parser; - std::vector< char > * name_pending; - int worker_id; - int num_workers; - int infd; - int filenames; - bool missing_crc; - bool permissive; - }; - - -/* Read lzip members from archive, list their tar members, and give the - packets produced to courier. */ -extern "C" void * tworker( void * arg ) - { - const Worker_arg & tmp = *(const Worker_arg *)arg; - const Lzip_index & lzip_index = *tmp.lzip_index; - Packet_courier & courier = *tmp.courier; - const Arg_parser & parser = *tmp.parser; - std::vector< char > & name_pending = *tmp.name_pending; - const int worker_id = tmp.worker_id; - const int num_workers = tmp.num_workers; - const int infd = tmp.infd; - const int filenames = tmp.filenames; - const int missing_crc = tmp.missing_crc; - const bool permissive = tmp.permissive; - - Resizable_buffer rbuf; - LZ_Decoder * const decoder = LZ_decompress_open(); - if( !rbuf.size() || !decoder || LZ_decompress_errno( decoder ) != LZ_ok ) - { show_error( mem_msg ); cleanup_and_fail(); } - - const long long cdata_size = lzip_index.cdata_size(); - bool master = false; - for( long i = worker_id; !master && i < lzip_index.members(); i += num_workers ) - { - long long data_pos = lzip_index.dblock( i ).pos(); - const long long mdata_end = lzip_index.dblock( i ).end(); - long long data_end = mdata_end; - long long file_pos = lzip_index.mblock( i ).pos(); - const long long member_end = lzip_index.mblock( i ).end(); - if( data_pos >= data_end ) // empty lzip member - { - if( courier.collect_packet( worker_id, i, "", Packet::member_done ) ) - continue; else break; - } - - Extended extended; // metadata from extended records - bool prev_extended = false; // prev header was extended - LZ_decompress_reset( decoder ); // prepare for new member - while( true ) // process one tar header per iteration - { - if( data_pos >= data_end ) - { - if( data_pos == data_end && !prev_extended ) break; - // member end exceeded or ends in extended, process rest of file - if( !courier.request_mastership( i, worker_id ) ) goto done; - master = true; - if( data_end < lzip_index.udata_size() ) - data_end = lzip_index.udata_size(); - else - { courier.collect_packet( worker_id, i, end_msg, Packet::error ); - goto done; } - } - Tar_header header; - const char * msg = 0; - const int ret = archive_read_lz( decoder, infd, file_pos, member_end, - cdata_size, header, header_size, &msg ); - if( ret != 0 ) - { - if( !courier.request_mastership( i, worker_id ) ) goto done; - master = true; - courier.collect_packet( worker_id, i, msg, Packet::error ); - goto done; - } - data_pos += header_size; - if( !verify_ustar_chksum( header ) ) - { - if( !courier.request_mastership( i, worker_id ) ) goto done; - master = true; - if( block_is_zero( header, header_size ) ) // EOF - { - if( !prev_extended || permissive ) courier.report_eof(); - else courier.collect_packet( worker_id, i, fv_msg1, Packet::error ); - goto done; - } - courier.collect_packet( worker_id, i, ( data_pos > header_size ) ? - bad_hdr_msg : posix_lz_msg, Packet::error ); - goto done; - } - - const Typeflag typeflag = (Typeflag)header[typeflag_o]; - if( typeflag == tf_global ) - { - if( prev_extended && !permissive ) - { courier.collect_packet( worker_id, i, fv_msg2, Packet::error ); - goto done; } - Extended dummy; // global headers are parsed and ignored - if( parse_records_lz( decoder, infd, file_pos, member_end, cdata_size, - data_pos, dummy, header, rbuf, &msg, true ) == 0 ) - { - if( data_pos == data_end && // end of lzip member - !courier.collect_packet( worker_id, i, "", Packet::member_done ) ) - goto done; - continue; - } - if( courier.request_mastership( i, worker_id ) ) - courier.collect_packet( worker_id, i, msg ? msg : gblrec_msg, - Packet::error ); - goto done; - } - if( typeflag == tf_extended ) - { - int ret = 0; - if( prev_extended && !permissive ) { msg = fv_msg3; ret = 2; } - else ret = parse_records_lz( decoder, infd, file_pos, member_end, - cdata_size, data_pos, extended, header, - rbuf, &msg, permissive ); - if( ret == 0 && !extended.crc_present() && missing_crc ) - { msg = mcrc_msg; ret = 2; } - if( ret == 0 ) { prev_extended = true; continue; } - if( courier.request_mastership( i, worker_id ) ) - courier.collect_packet( worker_id, i, msg ? msg : extrec_msg, - Packet::error ); - goto done; - } - prev_extended = false; - - extended.fill_from_ustar( header ); // copy metadata from header - - const bool skip = check_skip_filename( parser, name_pending, - extended.path().c_str(), filenames ); - - if( list_member_lz( decoder, infd, file_pos, member_end, cdata_size, - data_pos, mdata_end, courier, extended, - header, rbuf, i, worker_id, &msg, skip ) != 0 ) - { courier.collect_packet( worker_id, i, msg, Packet::error ); - goto done; } - extended.reset(); - } - } -done: - if( LZ_decompress_close( decoder ) < 0 ) - courier.collect_packet( worker_id, lzip_index.members(), - "LZ_decompress_close failed.", Packet::error ); - courier.worker_finished(); - return 0; - } - - -/* Get from courier the processed and sorted packets, and print - the member lines on stdout or the diagnostics on stderr. */ -void muxer( const char * const archive_namep, Packet_courier & courier ) - { - while( true ) - { - const Packet * const opacket = courier.deliver_packet(); - if( !opacket ) break; // queue is empty. all workers exited - - if( opacket->status == Packet::error ) - { show_file_error( archive_namep, opacket->line.c_str() ); - cleanup_and_fail( 2 ); } - if( opacket->line.size() ) - { std::fputs( opacket->line.c_str(), stdout ); std::fflush( stdout ); } - delete opacket; - } - if( !courier.eof_found() ) // no worker found EOF blocks - { show_file_error( archive_namep, end_msg ); cleanup_and_fail( 2 ); } - } - -} // end namespace - - -/* Read 'size' decompressed bytes from the archive. - Return value: 0 = OK, 1 = damaged member, 2 = fatal error. */ -int archive_read_lz( LZ_Decoder * const decoder, const int infd, - long long & file_pos, const long long member_end, - const long long cdata_size, uint8_t * const buf, - const int size, const char ** msg ) - { - int sz = 0; - - while( sz < size ) - { - const int rd = LZ_decompress_read( decoder, buf + sz, size - sz ); - if( rd < 0 ) - { *msg = LZ_strerror( LZ_decompress_errno( decoder ) ); return 1; } - if( rd == 0 && LZ_decompress_finished( decoder ) == 1 ) - { *msg = end_msg; return 2; } - sz += rd; - if( sz < size && LZ_decompress_write_size( decoder ) > 0 ) - { - const long long ibuf_size = 16384; // try 65536 - uint8_t ibuf[ibuf_size]; - const long long rest = ( file_pos < member_end ) ? - member_end - file_pos : cdata_size - file_pos; - const int rsize = std::min( LZ_decompress_write_size( decoder ), - (int)std::min( ibuf_size, rest ) ); - if( rsize <= 0 ) LZ_decompress_finish( decoder ); - else - { - const int rd = preadblock( infd, ibuf, rsize, file_pos ); - if( LZ_decompress_write( decoder, ibuf, rd ) != rd ) - internal_error( "library error (LZ_decompress_write)." ); - file_pos += rd; - if( rd < rsize ) - { - LZ_decompress_finish( decoder ); - if( errno ) { *msg = "Error reading archive"; return 2; } - } - } - } - } - return 0; - } - - -int parse_records_lz( LZ_Decoder * const decoder, const int infd, - long long & file_pos, const long long member_end, - const long long cdata_size, long long & data_pos, - Extended & extended, const Tar_header header, - Resizable_buffer & rbuf, const char ** msg, - const bool permissive ) - { - const long long edsize = parse_octal( header + size_o, size_l ); - const long long bufsize = round_up( edsize ); - if( edsize <= 0 || edsize >= 1LL << 33 || bufsize >= INT_MAX ) - return 1; // overflow or no extended data - if( !rbuf.resize( bufsize ) ) return 1; // extended records buffer - int retval = archive_read_lz( decoder, infd, file_pos, member_end, - cdata_size, (uint8_t *)rbuf(), bufsize, msg ); - if( retval == 0 ) - { if( extended.parse( rbuf(), edsize, permissive ) ) data_pos += bufsize; - else retval = 1; } - return retval; - } - - -int skip_member_lz( LZ_Decoder * const decoder, const int infd, - long long & file_pos, const long long member_end, - const long long cdata_size, long long & data_pos, - long long rest, const char ** msg ) - { - const int bufsize = 32 * header_size; - uint8_t buf[bufsize]; - while( rest > 0 ) // skip tar member - { - const int rsize = ( rest >= bufsize ) ? bufsize : rest; - const int ret = archive_read_lz( decoder, infd, file_pos, member_end, - cdata_size, buf, rsize, msg ); - if( ret != 0 ) return ret; - data_pos += rsize; - rest -= rsize; - } - return 0; - } - - - // init the courier, then start the workers and call the muxer. -int list_lz( const char * const archive_namep, const Arg_parser & parser, - std::vector< char > & name_pending, const Lzip_index & lzip_index, - const int filenames, const int debug_level, const int infd, - const int num_workers, const bool missing_crc, - const bool permissive ) - { - const int out_slots = 65536; // max small files (<=512B) in 64 MiB - - /* If an error happens after any threads have been started, exit must be - called before courier goes out of scope. */ - Packet_courier courier( num_workers, out_slots ); - - Worker_arg * worker_args = new( std::nothrow ) Worker_arg[num_workers]; - pthread_t * worker_threads = new( std::nothrow ) pthread_t[num_workers]; - if( !worker_args || !worker_threads ) { show_error( mem_msg ); return 1; } - for( int i = 0; i < num_workers; ++i ) - { - worker_args[i].lzip_index = &lzip_index; - worker_args[i].courier = &courier; - worker_args[i].parser = &parser; - worker_args[i].name_pending = &name_pending; - worker_args[i].worker_id = i; - worker_args[i].num_workers = num_workers; - worker_args[i].infd = infd; - worker_args[i].filenames = filenames; - worker_args[i].missing_crc = missing_crc; - worker_args[i].permissive = permissive; - const int errcode = - pthread_create( &worker_threads[i], 0, tworker, &worker_args[i] ); - if( errcode ) - { show_error( "Can't create worker threads", errcode ); cleanup_and_fail(); } - } - - muxer( archive_namep, courier ); - - for( int i = num_workers - 1; i >= 0; --i ) - { - const int errcode = pthread_join( worker_threads[i], 0 ); - if( errcode ) - { show_error( "Can't join worker threads", errcode ); cleanup_and_fail(); } - } - delete[] worker_threads; - delete[] worker_args; - - int retval = 0; - if( close( infd ) != 0 ) - { show_file_error( archive_namep, "Error closing archive", errno ); - retval = 1; } - - if( retval == 0 ) for( int i = 0; i < parser.arguments(); ++i ) - if( !parser.code( i ) && parser.argument( i ).size() && name_pending[i] ) - { - show_file_error( parser.argument( i ).c_str(), "Not found in archive." ); - retval = 1; - } - - if( debug_level & 1 ) - std::fprintf( stderr, - "muxer tried to consume from workers %8u times\n" - "muxer had to wait %8u times\n", - courier.ocheck_counter, - courier.owait_counter ); - - if( !courier.finished() ) internal_error( "courier not finished." ); - return retval; - } diff --git a/lzip_index.cc b/lzip_index.cc index 6c599a2..015bef5 100644 --- a/lzip_index.cc +++ b/lzip_index.cc @@ -1,25 +1,25 @@ -/* Tarlz - Archiver with multimember lzip compression - Copyright (C) 2013-2019 Antonio Diaz Diaz. +/* Tarlz - Archiver with multimember lzip compression + Copyright (C) 2013-2020 Antonio Diaz Diaz. - 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 of the License, or - (at your option) any later version. + 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 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. + 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. - You should have received a copy of the GNU General Public License - along with this program. If not, see <http://www.gnu.org/licenses/>. + You should have received a copy of the GNU General Public License + along with this program. If not, see <http://www.gnu.org/licenses/>. */ #define _FILE_OFFSET_BITS 64 #include <algorithm> #include <cerrno> -#include <climits> +#include <climits> // for tarlz.h #include <cstdio> #include <cstring> #include <string> @@ -28,8 +28,8 @@ #include <stdint.h> #include <unistd.h> -#include "lzip_index.h" #include "tarlz.h" +#include "lzip_index.h" int seek_read( const int fd, uint8_t * const buf, const int size, @@ -53,6 +53,19 @@ const char * bad_version( const unsigned version ) } // end namespace +bool Lzip_index::check_header_error( const Lzip_header & header, + const bool first ) + { + if( !header.verify_magic() ) + { error_ = bad_magic_msg; retval_ = 2; if( first ) bad_magic_ = true; + return true; } + if( !header.verify_version() ) + { error_ = bad_version( header.version() ); retval_ = 2; return true; } + if( !isvalid_ds( header.dictionary_size() ) ) + { error_ = bad_dict_msg; retval_ = 2; return true; } + return false; + } + void Lzip_index::set_errno_error( const char * const msg ) { error_ = msg; error_ += std::strerror( errno ); @@ -68,14 +81,24 @@ void Lzip_index::set_num_error( const char * const msg, unsigned long long num ) } +bool Lzip_index::read_header( const int fd, Lzip_header & header, + const long long pos ) + { + if( seek_read( fd, header.data, Lzip_header::size, pos ) != Lzip_header::size ) + { set_errno_error( "Error reading member header: " ); return false; } + return true; + } + + // If successful, push last member and set pos to member header. bool Lzip_index::skip_trailing_data( const int fd, unsigned long long & pos, - const bool ignore_trailing, const bool loose_trailing ) + const bool ignore_trailing, + const bool loose_trailing ) { + if( pos < min_member_size ) return false; enum { block_size = 16384, buffer_size = block_size + Lzip_trailer::size - 1 + Lzip_header::size }; uint8_t buffer[buffer_size]; - if( pos < min_member_size ) return false; int bsize = pos % block_size; // total bytes in buffer if( bsize <= buffer_size - block_size ) bsize += block_size; int search_size = bsize; // bytes to search for trailer @@ -98,26 +121,30 @@ bool Lzip_index::skip_trailing_data( const int fd, unsigned long long & pos, if( member_size > ipos + i || !trailer.verify_consistency() ) continue; Lzip_header header; - if( seek_read( fd, header.data, Lzip_header::size, - ipos + i - member_size ) != Lzip_header::size ) - { set_errno_error( "Error reading member header: " ); return false; } - const unsigned dictionary_size = header.dictionary_size(); - if( !header.verify_magic() || !header.verify_version() || - !isvalid_ds( dictionary_size ) ) continue; - if( (*(const Lzip_header *)( buffer + i )).verify_prefix( bsize - i ) ) - { error_ = "Last member in input file is truncated or corrupt."; - retval_ = 2; return false; } - if( !loose_trailing && bsize - i >= Lzip_header::size && - (*(const Lzip_header *)( buffer + i )).verify_corrupt() ) + if( !read_header( fd, header, ipos + i - member_size ) ) return false; + if( !header.verify() ) continue; + const Lzip_header & header2 = *(const Lzip_header *)( buffer + i ); + const bool full_h2 = bsize - i >= Lzip_header::size; + if( header2.verify_prefix( bsize - i ) ) // last member + { + if( !full_h2 ) error_ = "Last member in input file is truncated."; + else if( !check_header_error( header2, false ) ) + error_ = "Last member in input file is truncated or corrupt."; + retval_ = 2; return false; + } + if( !loose_trailing && full_h2 && header2.verify_corrupt() ) { error_ = corrupt_mm_msg; retval_ = 2; return false; } if( !ignore_trailing ) { error_ = trailing_msg; retval_ = 2; return false; } pos = ipos + i - member_size; + const unsigned dictionary_size = header.dictionary_size(); member_vector.push_back( Member( 0, trailer.data_size(), pos, member_size, dictionary_size ) ); + if( dictionary_size_ < dictionary_size ) + dictionary_size_ = dictionary_size; return true; } - if( ipos <= 0 ) + if( ipos == 0 ) { set_num_error( "Bad trailer at pos ", pos - Lzip_trailer::size ); return false; } bsize = buffer_size; @@ -131,7 +158,8 @@ bool Lzip_index::skip_trailing_data( const int fd, unsigned long long & pos, Lzip_index::Lzip_index( const int infd, const bool ignore_trailing, const bool loose_trailing ) - : insize( lseek( infd, 0, SEEK_END ) ), retval_( 0 ) + : insize( lseek( infd, 0, SEEK_END ) ), retval_( 0 ), dictionary_size_( 0 ), + bad_magic_( false ) { if( insize < 0 ) { set_errno_error( "Input file is not seekable: " ); return; } @@ -142,14 +170,8 @@ Lzip_index::Lzip_index( const int infd, const bool ignore_trailing, retval_ = 2; return; } Lzip_header header; - if( seek_read( infd, header.data, Lzip_header::size, 0 ) != Lzip_header::size ) - { set_errno_error( "Error reading member header: " ); return; } - if( !header.verify_magic() ) - { error_ = bad_magic_msg; retval_ = 2; return; } - if( !header.verify_version() ) - { error_ = bad_version( header.version() ); retval_ = 2; return; } - if( !isvalid_ds( header.dictionary_size() ) ) - { error_ = bad_dict_msg; retval_ = 2; return; } + if( !read_header( infd, header, 0 ) ) return; + if( check_header_error( header, true ) ) return; unsigned long long pos = insize; // always points to a header or to EOF while( pos >= min_member_size ) @@ -159,7 +181,7 @@ Lzip_index::Lzip_index( const int infd, const bool ignore_trailing, pos - Lzip_trailer::size ) != Lzip_trailer::size ) { set_errno_error( "Error reading member trailer: " ); break; } const unsigned long long member_size = trailer.member_size(); - if( member_size > pos || !trailer.verify_consistency() ) + if( member_size > pos || !trailer.verify_consistency() ) // bad trailer { if( member_vector.empty() ) { if( skip_trailing_data( infd, pos, ignore_trailing, loose_trailing ) ) @@ -167,12 +189,8 @@ Lzip_index::Lzip_index( const int infd, const bool ignore_trailing, set_num_error( "Bad trailer at pos ", pos - Lzip_trailer::size ); break; } - if( seek_read( infd, header.data, Lzip_header::size, - pos - member_size ) != Lzip_header::size ) - { set_errno_error( "Error reading member header: " ); break; } - const unsigned dictionary_size = header.dictionary_size(); - if( !header.verify_magic() || !header.verify_version() || - !isvalid_ds( dictionary_size ) ) + if( !read_header( infd, header, pos - member_size ) ) break; + if( !header.verify() ) // bad header { if( member_vector.empty() ) { if( skip_trailing_data( infd, pos, ignore_trailing, loose_trailing ) ) @@ -181,8 +199,11 @@ Lzip_index::Lzip_index( const int infd, const bool ignore_trailing, break; } pos -= member_size; + const unsigned dictionary_size = header.dictionary_size(); member_vector.push_back( Member( 0, trailer.data_size(), pos, member_size, dictionary_size ) ); + if( dictionary_size_ < dictionary_size ) + dictionary_size_ = dictionary_size; } if( pos != 0 || member_vector.empty() ) { diff --git a/lzip_index.h b/lzip_index.h index 08758da..77d2a00 100644 --- a/lzip_index.h +++ b/lzip_index.h @@ -1,18 +1,18 @@ -/* Tarlz - Archiver with multimember lzip compression - Copyright (C) 2013-2019 Antonio Diaz Diaz. +/* Tarlz - Archiver with multimember lzip compression + Copyright (C) 2013-2020 Antonio Diaz Diaz. - 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 of the License, or - (at your option) any later version. + 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 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. + 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. - You should have received a copy of the GNU General Public License - along with this program. If not, see <http://www.gnu.org/licenses/>. + You should have received a copy of the GNU General Public License + along with this program. If not, see <http://www.gnu.org/licenses/>. */ #ifndef INT64_MAX @@ -52,9 +52,13 @@ class Lzip_index std::string error_; const long long insize; int retval_; + unsigned dictionary_size_; // largest dictionary size in the file + bool bad_magic_; // bad magic in first header + bool check_header_error( const Lzip_header & header, const bool first ); void set_errno_error( const char * const msg ); void set_num_error( const char * const msg, unsigned long long num ); + bool read_header( const int fd, Lzip_header & header, const long long pos ); bool skip_trailing_data( const int fd, unsigned long long & pos, const bool ignore_trailing, const bool loose_trailing ); @@ -65,6 +69,8 @@ public: long members() const { return member_vector.size(); } const std::string & error() const { return error_; } int retval() const { return retval_; } + unsigned dictionary_size() const { return dictionary_size_; } + bool bad_magic() const { return bad_magic_; } long long udata_size() const { if( member_vector.empty() ) return 0; @@ -1,24 +1,24 @@ -/* Tarlz - Archiver with multimember lzip compression - Copyright (C) 2013-2019 Antonio Diaz Diaz. +/* Tarlz - Archiver with multimember lzip compression + Copyright (C) 2013-2020 Antonio Diaz Diaz. - 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 of the License, or - (at your option) any later version. + 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 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. + 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. - You should have received a copy of the GNU General Public License - along with this program. If not, see <http://www.gnu.org/licenses/>. + You should have received a copy of the GNU General Public License + along with this program. If not, see <http://www.gnu.org/licenses/>. */ /* - Exit status: 0 for a normal exit, 1 for environmental problems - (file not found, invalid flags, I/O errors, etc), 2 to indicate a - corrupt or invalid input file, 3 for an internal consistency error - (eg, bug) which caused tarlz to panic. + Exit status: 0 for a normal exit, 1 for environmental problems (file not + found, files differ, invalid flags, I/O errors, etc), 2 to indicate a + corrupt or invalid input file, 3 for an internal consistency error + (eg, bug) which caused tarlz to panic. */ #define _FILE_OFFSET_BITS 64 @@ -29,6 +29,7 @@ #include <cstdio> #include <cstdlib> #include <cstring> +#include <ctime> #include <string> #include <vector> #include <fcntl.h> @@ -59,9 +60,8 @@ int verbosity = 0; namespace { const char * const program_name = "tarlz"; -const char * const program_year = "2019"; +const char * const program_year = "2020"; const char * invocation_name = program_name; // default value -bool dereference = false; void show_help( const long num_online ) @@ -86,47 +86,49 @@ void show_help( const long num_online ) "can be used to recover some of the damaged members.\n" "\nUsage: %s [options] [files]\n", invocation_name ); std::printf( "\nOptions:\n" - " --help display this help and exit\n" - " -V, --version output version information and exit\n" - " -A, --concatenate append archives to the end of an archive\n" - " -B, --data-size=<bytes> set target size of input data blocks [2x8=16 MiB]\n" - " -c, --create create a new archive\n" - " -C, --directory=<dir> change to directory <dir>\n" - " -d, --diff find differences between archive and file system\n" - " --ignore-ids ignore differences in owner and group IDs\n" - " --delete delete files/directories from an archive\n" - " --exclude=<pattern> exclude files matching a shell pattern\n" - " -f, --file=<archive> use archive file <archive>\n" - " -h, --dereference follow symlinks; archive the files they point to\n" - " -n, --threads=<n> set number of (de)compression threads [%ld]\n" - " -q, --quiet suppress all messages\n" - " -r, --append append files to the end of an archive\n" - " -t, --list list the contents of an archive\n" - " -v, --verbose verbosely list files processed\n" - " -x, --extract extract files/directories from an archive\n" - " -0 .. -9 set compression level [default 6]\n" - " --uncompressed don't compress the archive created\n" - " --asolid create solidly compressed appendable archive\n" - " --bsolid create per block compressed archive (default)\n" - " --dsolid create per directory compressed archive\n" - " --no-solid create per file compressed archive\n" - " --solid create solidly compressed archive\n" - " --anonymous equivalent to '--owner=root --group=root'\n" - " --owner=<owner> use <owner> name/ID for files added\n" - " --group=<group> use <group> name/ID for files added\n" - " --keep-damaged don't delete partially extracted files\n" - " --missing-crc exit with error status if missing extended CRC\n" - " --out-slots=<n> number of 1 MiB output packets buffered [64]\n" -/* " --permissive allow repeated extended headers and records\n"*/, + " --help display this help and exit\n" + " -V, --version output version information and exit\n" + " -A, --concatenate append archives to the end of an archive\n" + " -B, --data-size=<bytes> set target size of input data blocks [2x8=16 MiB]\n" + " -c, --create create a new archive\n" + " -C, --directory=<dir> change to directory <dir>\n" + " -d, --diff find differences between archive and file system\n" + " --ignore-ids ignore differences in owner and group IDs\n" + " --delete delete files/directories from an archive\n" + " --exclude=<pattern> exclude files matching a shell pattern\n" + " -f, --file=<archive> use archive file <archive>\n" + " -h, --dereference follow symlinks; archive the files they point to\n" + " --mtime=<date> use <date> as mtime for files added to archive\n" + " -n, --threads=<n> set number of (de)compression threads [%ld]\n" + " -p, --preserve-permissions don't subtract the umask on extraction\n" + " -q, --quiet suppress all messages\n" + " -r, --append append files to the end of an archive\n" + " -t, --list list the contents of an archive\n" + " -v, --verbose verbosely list files processed\n" + " -x, --extract extract files/directories from an archive\n" + " -0 .. -9 set compression level [default 6]\n" + " --uncompressed don't compress the archive created\n" + " --asolid create solidly compressed appendable archive\n" + " --bsolid create per block compressed archive (default)\n" + " --dsolid create per directory compressed archive\n" + " --no-solid create per file compressed archive\n" + " --solid create solidly compressed archive\n" + " --anonymous equivalent to '--owner=root --group=root'\n" + " --owner=<owner> use <owner> name/ID for files added to archive\n" + " --group=<group> use <group> name/ID for files added to archive\n" + " --keep-damaged don't delete partially extracted files\n" + " --missing-crc exit with error status if missing extended CRC\n" + " --out-slots=<n> number of 1 MiB output packets buffered [64]\n" +/* " --permissive allow repeated extended headers and records\n"*/, num_online ); if( verbosity >= 1 ) { - std::printf( " --debug=<level> (0-1) print debug statistics to stderr\n" ); + std::printf( " --debug=<level> (0-1) print debug statistics to stderr\n" ); } - std::printf( "\nExit status: 0 for a normal exit, 1 for environmental problems (file\n" - "not found, invalid flags, I/O errors, etc), 2 to indicate a corrupt or\n" - "invalid input file, 3 for an internal consistency error (eg, bug) which\n" - "caused tarlz to panic.\n" + std::printf( "\nExit status: 0 for a normal exit, 1 for environmental problems (file not\n" + "found, files differ, invalid flags, I/O errors, etc), 2 to indicate a\n" + "corrupt or invalid input file, 3 for an internal consistency error (eg, bug)\n" + "which caused tarlz to panic.\n" "\nReport bugs to lzip-bug@nongnu.org\n" "Tarlz home page: http://www.nongnu.org/lzip/tarlz.html\n" ); } @@ -193,6 +195,17 @@ unsigned long long getnum( const char * const ptr, } +void set_archive_name( std::string & archive_name, const std::string & new_name ) + { + static bool first_call = true; + + if( first_call ) { if( new_name != "-" ) archive_name = new_name; + first_call = false; return; } + show_error( "Only one archive can be specified.", 0, true ); + std::exit( 1 ); + } + + void set_mode( Program_mode & program_mode, const Program_mode new_mode ) { if( program_mode != m_none && program_mode != new_mode ) @@ -204,28 +217,58 @@ void set_mode( Program_mode & program_mode, const Program_mode new_mode ) } -void set_owner( const char * const arg ) +void set_mtime( long long & mtime, const char * arg ) + { + if( *arg == '@' ) + { mtime = getnum( arg + 1, 0, ( 1ULL << 33 ) - 1 ); return; } + else if( *arg == '.' || *arg == '/' ) + { + struct stat st; + if( stat( arg, &st ) == 0 ) { mtime = st.st_mtime; return; } + show_file_error( arg, "Can't stat", errno ); std::exit( 1 ); + } + else // format 'YYYY-MM-DD HH:MM:SS' + { + unsigned y, mo, d, h, m, s; + const int n = std::sscanf( arg, "%u-%u-%u %u:%u:%u", + &y, &mo, &d, &h, &m, &s ); + if( n == 6 && y >= 1970 && mo >= 1 ) + { + struct tm t; + t.tm_year = y - 1900; t.tm_mon = mo - 1; t.tm_mday = d; + t.tm_hour = h; t.tm_min = m; t.tm_sec = s; t.tm_isdst = -1; + mtime = std::mktime( &t ); if( mtime >= 0 ) return; + } + } + show_error( "Invalid mtime.", 0, true ); std::exit( 1 ); + } + + +void set_owner( int & owner, const char * const arg ) { const struct passwd * const pw = getpwnam( arg ); - if( pw ) cl_owner = pw->pw_uid; + if( pw ) owner = pw->pw_uid; else if( std::isdigit( (unsigned char)arg[0] ) ) - cl_owner = getnum( arg, 0, INT_MAX ); + owner = getnum( arg, 0, INT_MAX ); + else if( std::strcmp( arg, "root" ) == 0 ) owner = 0; else { show_file_error( arg, "Invalid owner" ); std::exit( 1 ); } } -void set_group( const char * const arg ) +void set_group( int & group, const char * const arg ) { const struct group * const gr = getgrnam( arg ); - if( gr ) cl_group = gr->gr_gid; + if( gr ) group = gr->gr_gid; else if( std::isdigit( (unsigned char)arg[0] ) ) - cl_group = getnum( arg, 0, INT_MAX ); + group = getnum( arg, 0, INT_MAX ); + else if( std::strcmp( arg, "root" ) == 0 ) group = 0; else { show_file_error( arg, "Invalid group" ); std::exit( 1 ); } } } // end namespace -int hstat( const char * const filename, struct stat * const st ) +int hstat( const char * const filename, struct stat * const st, + const bool dereference ) { return dereference ? stat( filename, st ) : lstat( filename, st ); } @@ -251,10 +294,10 @@ int open_outstream( const std::string & name, const bool create ) } -// This can be called from any thread, main thread or sub-threads alike, -// since they all call common helper functions that call cleanup_and_fail() -// in case of an error. -// +/* This can be called from any thread, main thread or sub-threads alike, + since they all call common helper functions that call cleanup_and_fail() + in case of an error. +*/ void cleanup_and_fail( const int retval ) { // calling 'exit' more than once results in undefined behavior @@ -278,10 +321,21 @@ void show_error( const char * const msg, const int errcode, const bool help ) } +void format_file_error( std::string & estr, const char * const filename, + const char * const msg, const int errcode ) + { + if( verbosity < 0 ) return; + estr += program_name; estr += ": "; estr += filename; estr += ": "; + estr += msg; + if( errcode > 0 ) { estr += ": "; estr += std::strerror( errcode ); } + estr += '\n'; + } + + void show_file_error( const char * const filename, const char * const msg, const int errcode ) { - if( verbosity >= 0 ) + if( verbosity >= 0 && msg && msg[0] ) std::fprintf( stderr, "%s: %s: %s%s%s\n", program_name, filename, msg, ( errcode > 0 ) ? ": " : "", ( errcode > 0 ) ? std::strerror( errcode ) : "" ); @@ -298,128 +352,131 @@ void internal_error( const char * const msg ) int main( const int argc, const char * const argv[] ) { - std::string archive_name; - int debug_level = 0; - int level = 6; // compression level, < 0 means uncompressed - int num_workers = -1; // start this many worker threads - int out_slots = 64; - Program_mode program_mode = m_none; - bool ignore_ids = false; - bool keep_damaged = false; - bool missing_crc = false; - bool permissive = false; if( argc > 0 ) invocation_name = argv[0]; - if( LZ_version()[0] < '1' ) - { show_error( "Bad library version. At least lzlib 1.0 is required." ); +#if !defined LZ_API_VERSION || LZ_API_VERSION < 1 // compile-time test +#error "lzlib 1.8 or newer needed." +#elif LZ_API_VERSION >= 2 + if( LZ_api_version() < 1 ) // runtime test + { show_error( "Wrong library version. At least lzlib 1.8 is required." ); return 1; } +#endif enum { opt_ano = 256, opt_aso, opt_bso, opt_crc, opt_dbg, opt_del, opt_dso, - opt_exc, opt_grp, opt_hlp, opt_id, opt_kd, opt_nso, opt_out, opt_own, - opt_per, opt_sol, opt_un }; + opt_exc, opt_grp, opt_hlp, opt_id, opt_kd, opt_mti, opt_nso, opt_out, + opt_own, opt_per, opt_sol, opt_un }; const Arg_parser::Option options[] = { - { '0', 0, Arg_parser::no }, - { '1', 0, Arg_parser::no }, - { '2', 0, Arg_parser::no }, - { '3', 0, Arg_parser::no }, - { '4', 0, Arg_parser::no }, - { '5', 0, Arg_parser::no }, - { '6', 0, Arg_parser::no }, - { '7', 0, Arg_parser::no }, - { '8', 0, Arg_parser::no }, - { '9', 0, Arg_parser::no }, - { 'A', "concatenate", Arg_parser::no }, - { 'B', "data-size", Arg_parser::yes }, - { 'c', "create", Arg_parser::no }, - { 'C', "directory", Arg_parser::yes }, - { 'd', "diff", Arg_parser::no }, - { 'f', "file", Arg_parser::yes }, - { 'h', "dereference", Arg_parser::no }, - { 'H', "format", Arg_parser::yes }, - { 'n', "threads", Arg_parser::yes }, - { 'q', "quiet", Arg_parser::no }, - { 'r', "append", Arg_parser::no }, - { 't', "list", Arg_parser::no }, - { 'v', "verbose", Arg_parser::no }, - { 'V', "version", Arg_parser::no }, - { 'x', "extract", Arg_parser::no }, - { opt_ano, "anonymous", Arg_parser::no }, - { opt_aso, "asolid", Arg_parser::no }, - { opt_bso, "bsolid", Arg_parser::no }, - { opt_dbg, "debug", Arg_parser::yes }, - { opt_del, "delete", Arg_parser::no }, - { opt_dso, "dsolid", Arg_parser::no }, - { opt_exc, "exclude", Arg_parser::yes }, - { opt_grp, "group", Arg_parser::yes }, - { opt_hlp, "help", Arg_parser::no }, - { opt_id, "ignore-ids", Arg_parser::no }, - { opt_kd, "keep-damaged", Arg_parser::no }, - { opt_crc, "missing-crc", Arg_parser::no }, - { opt_nso, "no-solid", Arg_parser::no }, - { opt_out, "out-slots", Arg_parser::yes }, - { opt_own, "owner", Arg_parser::yes }, - { opt_per, "permissive", Arg_parser::no }, - { opt_sol, "solid", Arg_parser::no }, - { opt_un, "uncompressed", Arg_parser::no }, - { 0 , 0, Arg_parser::no } }; - - const Arg_parser parser( argc, argv, options, true ); + { '0', 0, Arg_parser::no }, + { '1', 0, Arg_parser::no }, + { '2', 0, Arg_parser::no }, + { '3', 0, Arg_parser::no }, + { '4', 0, Arg_parser::no }, + { '5', 0, Arg_parser::no }, + { '6', 0, Arg_parser::no }, + { '7', 0, Arg_parser::no }, + { '8', 0, Arg_parser::no }, + { '9', 0, Arg_parser::no }, + { 'A', "concatenate", Arg_parser::no }, + { 'B', "data-size", Arg_parser::yes }, + { 'c', "create", Arg_parser::no }, + { 'C', "directory", Arg_parser::yes }, + { 'd', "diff", Arg_parser::no }, + { 'f', "file", Arg_parser::yes }, + { 'h', "dereference", Arg_parser::no }, + { 'H', "format", Arg_parser::yes }, + { 'n', "threads", Arg_parser::yes }, + { 'p', "preserve-permissions", Arg_parser::no }, + { 'q', "quiet", Arg_parser::no }, + { 'r', "append", Arg_parser::no }, + { 't', "list", Arg_parser::no }, + { 'v', "verbose", Arg_parser::no }, + { 'V', "version", Arg_parser::no }, + { 'x', "extract", Arg_parser::no }, + { opt_ano, "anonymous", Arg_parser::no }, + { opt_aso, "asolid", Arg_parser::no }, + { opt_bso, "bsolid", Arg_parser::no }, + { opt_dbg, "debug", Arg_parser::yes }, + { opt_del, "delete", Arg_parser::no }, + { opt_dso, "dsolid", Arg_parser::no }, + { opt_exc, "exclude", Arg_parser::yes }, + { opt_grp, "group", Arg_parser::yes }, + { opt_hlp, "help", Arg_parser::no }, + { opt_id, "ignore-ids", Arg_parser::no }, + { opt_kd, "keep-damaged", Arg_parser::no }, + { opt_crc, "missing-crc", Arg_parser::no }, + { opt_mti, "mtime", Arg_parser::yes }, + { opt_nso, "no-solid", Arg_parser::no }, + { opt_out, "out-slots", Arg_parser::yes }, + { opt_own, "owner", Arg_parser::yes }, + { opt_per, "permissive", Arg_parser::no }, + { opt_sol, "solid", Arg_parser::no }, + { opt_un, "uncompressed", Arg_parser::no }, + { 0, 0, Arg_parser::no } }; + + const Arg_parser parser( argc, argv, options, true ); // in_order if( parser.error().size() ) // bad option { show_error( parser.error().c_str(), 0, true ); return 1; } + Cl_options cl_opts( parser ); const long num_online = std::max( 1L, sysconf( _SC_NPROCESSORS_ONLN ) ); long max_workers = sysconf( _SC_THREAD_THREADS_MAX ); if( max_workers < 1 || max_workers > INT_MAX / (int)sizeof (pthread_t) ) max_workers = INT_MAX / sizeof (pthread_t); - int filenames = 0; for( int argind = 0; argind < parser.arguments(); ++argind ) { const int code = parser.code( argind ); if( !code ) // skip non-options - { if( parser.argument( argind ).size() ) ++filenames; continue; } + { + if( parser.argument( argind ).empty() ) + { show_error( "Empty non-option argument." ); return 1; } + ++cl_opts.filenames; continue; + } const std::string & sarg = parser.argument( argind ); const char * const arg = sarg.c_str(); switch( code ) { case '0': case '1': case '2': case '3': case '4': case '5': case '6': case '7': case '8': case '9': - level = code - '0'; break; - case 'A': set_mode( program_mode, m_concatenate ); break; - case 'B': cl_data_size = getnum( arg, min_data_size, max_data_size ); + cl_opts.level = code - '0'; break; + case 'A': set_mode( cl_opts.program_mode, m_concatenate ); break; + case 'B': cl_opts.data_size = getnum( arg, min_data_size, max_data_size ); break; - case 'c': set_mode( program_mode, m_create ); break; + case 'c': set_mode( cl_opts.program_mode, m_create ); break; case 'C': break; // skip chdir - case 'd': set_mode( program_mode, m_diff ); break; - case 'f': if( sarg != "-" ) archive_name = sarg; break; - case 'h': dereference = true; break; + case 'd': set_mode( cl_opts.program_mode, m_diff ); break; + case 'f': set_archive_name( cl_opts.archive_name, sarg ); break; + case 'h': cl_opts.dereference = true; break; case 'H': break; // ignore format - case 'n': num_workers = getnum( arg, 0, max_workers ); break; + case 'n': cl_opts.num_workers = getnum( arg, 0, max_workers ); break; + case 'p': cl_opts.preserve_permissions = true; break; case 'q': verbosity = -1; break; - case 'r': set_mode( program_mode, m_append ); break; - case 't': set_mode( program_mode, m_list ); break; + case 'r': set_mode( cl_opts.program_mode, m_append ); break; + case 't': set_mode( cl_opts.program_mode, m_list ); break; case 'v': if( verbosity < 4 ) ++verbosity; break; case 'V': show_version(); return 0; - case 'x': set_mode( program_mode, m_extract ); break; - case opt_ano: set_owner( "root" ); set_group( "root" ); break; - case opt_aso: solidity = asolid; break; - case opt_bso: solidity = bsolid; break; - case opt_crc: missing_crc = true; break; - case opt_dbg: debug_level = getnum( arg, 0, 3 ); break; - case opt_del: set_mode( program_mode, m_delete ); break; - case opt_dso: solidity = dsolid; break; + case 'x': set_mode( cl_opts.program_mode, m_extract ); break; + case opt_ano: set_owner( cl_opts.owner, "root" ); + set_group( cl_opts.group, "root" ); break; + case opt_aso: cl_opts.solidity = asolid; break; + case opt_bso: cl_opts.solidity = bsolid; break; + case opt_crc: cl_opts.missing_crc = true; break; + case opt_dbg: cl_opts.debug_level = getnum( arg, 0, 3 ); break; + case opt_del: set_mode( cl_opts.program_mode, m_delete ); break; + case opt_dso: cl_opts.solidity = dsolid; break; case opt_exc: Exclude::add_pattern( sarg ); break; - case opt_grp: set_group( arg ); break; + case opt_grp: set_group( cl_opts.group, arg ); break; case opt_hlp: show_help( num_online ); return 0; - case opt_id: ignore_ids = true; break; - case opt_kd: keep_damaged = true; break; - case opt_nso: solidity = no_solid; break; - case opt_out: out_slots = getnum( arg, 1, 1024 ); break; - case opt_own: set_owner( arg ); break; - case opt_per: permissive = true; break; - case opt_sol: solidity = solid; break; - case opt_un: level = -1; break; + case opt_id: cl_opts.ignore_ids = true; break; + case opt_kd: cl_opts.keep_damaged = true; break; + case opt_mti: set_mtime( cl_opts.mtime, arg ); break; + case opt_nso: cl_opts.solidity = no_solid; break; + case opt_out: cl_opts.out_slots = getnum( arg, 1, 1024 ); break; + case opt_own: set_owner( cl_opts.owner, arg ); break; + case opt_per: cl_opts.permissive = true; break; + case opt_sol: cl_opts.solidity = solid; break; + case opt_un: cl_opts.level = -1; break; default : internal_error( "uncaught option" ); } } // end process options @@ -429,22 +486,18 @@ int main( const int argc, const char * const argv[] ) setmode( STDOUT_FILENO, O_BINARY ); #endif - if( num_workers < 0 ) num_workers = std::min( num_online, max_workers ); + if( cl_opts.num_workers < 0 ) // 0 disables multi-threading + cl_opts.num_workers = std::min( num_online, max_workers ); - switch( program_mode ) + switch( cl_opts.program_mode ) { - case m_none: show_error( "Missing operation.", 0, true ); return 2; + case m_none: show_error( "Missing operation.", 0, true ); return 1; case m_append: - case m_create: return encode( archive_name, parser, filenames, level, - num_workers, out_slots, debug_level, - program_mode == m_append, dereference ); - case m_concatenate: return concatenate( archive_name, parser, filenames ); - case m_delete: return delete_members( archive_name, parser, filenames, - missing_crc, permissive ); + case m_create: return encode( cl_opts ); + case m_concatenate: return concatenate( cl_opts ); + case m_delete: return delete_members( cl_opts ); case m_diff: case m_extract: - case m_list: return decode( archive_name, parser, filenames, - num_workers, debug_level, program_mode, - ignore_ids, keep_damaged, missing_crc, permissive ); + case m_list: return decode( cl_opts ); } } @@ -1,18 +1,18 @@ -/* Tarlz - Archiver with multimember lzip compression - Copyright (C) 2013-2019 Antonio Diaz Diaz. +/* Tarlz - Archiver with multimember lzip compression + Copyright (C) 2013-2020 Antonio Diaz Diaz. - 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 of the License, or - (at your option) any later version. + 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 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. + 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. - You should have received a copy of the GNU General Public License - along with this program. If not, see <http://www.gnu.org/licenses/>. + You should have received a copy of the GNU General Public License + along with this program. If not, see <http://www.gnu.org/licenses/>. */ #define max_file_size ( LLONG_MAX - header_size ) @@ -157,6 +157,50 @@ public: }; +class CRC32 + { + uint32_t data[256]; // Table of CRCs of all 8-bit messages. + +public: + CRC32( const bool castagnoli = false ) + { + const unsigned cpol = 0x82F63B78U; // CRC32-C Castagnoli polynomial. + const unsigned ipol = 0xEDB88320U; // IEEE 802.3 Ethernet polynomial. + const unsigned poly = castagnoli ? cpol : ipol; + + for( unsigned n = 0; n < 256; ++n ) + { + unsigned c = n; + for( int k = 0; k < 8; ++k ) + { if( c & 1 ) c = poly ^ ( c >> 1 ); else c >>= 1; } + data[n] = c; + } + } + + void update_byte( uint32_t & crc, const uint8_t byte ) const + { crc = data[(crc^byte)&0xFF] ^ ( crc >> 8 ); } + + void update_buf( uint32_t & crc, const uint8_t * const buffer, + const int size ) const + { + uint32_t c = crc; + for( int i = 0; i < size; ++i ) + c = data[(c^buffer[i])&0xFF] ^ ( c >> 8 ); + crc = c; + } + + // Calculates the crc of size bytes except a window of 8 bytes at pos + uint32_t windowed_crc( const uint8_t * const buffer, const int pos, + const int size ) const + { + uint32_t crc = 0xFFFFFFFFU; + update_buf( crc, buffer, pos ); + update_buf( crc, buffer + pos + 8, size - pos - 8 ); + return crc ^ 0xFFFFFFFFU; + } + }; + + enum { min_dictionary_bits = 12, min_dictionary_size = 1 << min_dictionary_bits, @@ -178,7 +222,7 @@ struct Lzip_header { uint8_t data[6]; // 0-3 magic bytes // 4 version - // 5 coded_dict_size + // 5 coded dictionary size enum { size = 6 }; bool verify_magic() const @@ -208,6 +252,10 @@ struct Lzip_header sz -= ( sz / 16 ) * ( ( data[5] >> 5 ) & 7 ); return sz; } + + bool verify() const + { return verify_magic() && verify_version() && + isvalid_ds( dictionary_size() ); } }; @@ -255,50 +303,41 @@ struct Lzip_trailer }; -class CRC32 +enum Program_mode { m_none, m_append, m_concatenate, m_create, m_delete, + m_diff, m_extract, m_list }; +enum Solidity { no_solid, bsolid, dsolid, asolid, solid }; +class Arg_parser; +struct Cl_options // command line options { - uint32_t data[256]; // Table of CRCs of all 8-bit messages. - -public: - CRC32( const bool castagnoli = false ) - { - const unsigned cpol = 0x82F63B78U; // CRC32-C Castagnoli polynomial. - const unsigned ipol = 0xEDB88320U; // IEEE 802.3 Ethernet polynomial. - const unsigned poly = castagnoli ? cpol : ipol; - - for( unsigned n = 0; n < 256; ++n ) - { - unsigned c = n; - for( int k = 0; k < 8; ++k ) - { if( c & 1 ) c = poly ^ ( c >> 1 ); else c >>= 1; } - data[n] = c; - } - } - - void update_byte( uint32_t & crc, const uint8_t byte ) const - { crc = data[(crc^byte)&0xFF] ^ ( crc >> 8 ); } - - void update_buf( uint32_t & crc, const uint8_t * const buffer, - const int size ) const - { - uint32_t c = crc; - for( int i = 0; i < size; ++i ) - c = data[(c^buffer[i])&0xFF] ^ ( c >> 8 ); - crc = c; - } - - // Calculates the crc of size bytes except a window of 8 bytes at pos - uint32_t windowed_crc( const uint8_t * const buffer, const int pos, - const int size ) const - { - uint32_t crc = 0xFFFFFFFFU; - update_buf( crc, buffer, pos ); - update_buf( crc, buffer + pos + 8, size - pos - 8 ); - return crc ^ 0xFFFFFFFFU; - } + const Arg_parser & parser; + std::string archive_name; + long long mtime; + Program_mode program_mode; + Solidity solidity; + int data_size; + int debug_level; + int filenames; + int level; // compression level, < 0 means uncompressed + int num_workers; // start this many worker threads + int out_slots; + int owner; + int group; + bool dereference; + bool ignore_ids; + bool keep_damaged; + bool missing_crc; + bool permissive; + bool preserve_permissions; + + Cl_options( const Arg_parser & ap ) + : parser( ap ), mtime( -1 ), program_mode( m_none ), solidity( bsolid ), + data_size( 0 ), debug_level( 0 ), filenames( 0 ), level( 6 ), + num_workers( -1 ), out_slots( 64 ), owner( -1 ), group( -1 ), + dereference( false ), ignore_ids( false ), keep_damaged( false ), + missing_crc( false ), permissive( false ), preserve_permissions( false ) + {} }; -extern const CRC32 crc32c; const char * const bad_magic_msg = "Bad magic number (file not in lzip format)."; const char * const bad_dict_msg = "Invalid dictionary size in member header."; @@ -330,24 +369,19 @@ void xbroadcast( pthread_cond_t * const cond ); unsigned long long parse_octal( const uint8_t * const ptr, const int size ); int readblock( const int fd, uint8_t * const buf, const int size ); int writeblock( const int fd, const uint8_t * const buf, const int size ); +bool nonempty_arg( const Arg_parser & parser, const int i ); // defined in common_decode.cc -class Arg_parser; bool block_is_zero( const uint8_t * const buf, const int size ); bool format_member_name( const Extended & extended, const Tar_header header, Resizable_buffer & rbuf, const bool long_format ); bool show_member_name( const Extended & extended, const Tar_header header, const int vlevel, Resizable_buffer & rbuf ); -bool check_skip_filename( const Arg_parser & parser, +bool check_skip_filename( const Cl_options & cl_opts, std::vector< char > & name_pending, - const char * const filename, const int filenames ); + const char * const filename ); // defined in create.cc -enum Solidity { no_solid, bsolid, dsolid, asolid, solid }; -extern int cl_owner; -extern int cl_group; -extern int cl_data_size; -extern Solidity solidity; bool copy_file( const int infd, const int outfd, const long long max_size = -1 ); bool writeblock_wrapper( const int outfd, const uint8_t * const buffer, const int size ); @@ -364,18 +398,13 @@ int final_exit_status( int retval, const bool show_msg = true ); unsigned ustar_chksum( const uint8_t * const header ); bool verify_ustar_chksum( const uint8_t * const header ); bool has_lz_ext( const std::string & name ); -int concatenate( const std::string & archive_name, const Arg_parser & parser, - const int filenames ); -int encode( const std::string & archive_name, const Arg_parser & parser, - const int filenames, const int level, const int num_workers, - const int out_slots, const int debug_level, const bool append, - const bool dereference ); +int concatenate( const Cl_options & cl_opts ); +int encode( Cl_options & cl_opts ); // defined in create_lz.cc -int encode_lz( const char * const archive_namep, const Arg_parser & parser, +int encode_lz( const Cl_options & cl_opts, const char * const archive_namep, const int dictionary_size, const int match_len_limit, - const int num_workers, const int outfd, const int out_slots, - const int debug_level, const bool dereference ); + const int outfd ); // defined in delete.cc class Lzip_index; @@ -384,17 +413,14 @@ int tail_copy( const char * const archive_namep, const Arg_parser & parser, std::vector< char > & name_pending, const Lzip_index & lzip_index, const long long istream_pos, const int infd, const int outfd, int retval ); -int delete_members( const std::string & archive_name, const Arg_parser & parser, - const int filenames, const bool missing_crc, - const bool permissive ); +int delete_members( const Cl_options & cl_opts ); // defined in delete_lz.cc -int delete_members_lz( const char * const archive_namep, - const Arg_parser & parser, +int delete_members_lz( const Cl_options & cl_opts, + const char * const archive_namep, std::vector< char > & name_pending, const Lzip_index & lzip_index, - const int filenames, const int infd, const int outfd, - const bool missing_crc, const bool permissive ); + const int infd, const int outfd ); // defined in exclude.cc namespace Exclude { @@ -403,35 +429,21 @@ bool excluded( const char * const filename ); } // end namespace Exclude // defined in extract.cc -enum Program_mode { m_none, m_append, m_concatenate, m_create, m_delete, - m_diff, m_extract, m_list }; -int decode( const std::string & archive_name, const Arg_parser & parser, - const int filenames, const int num_workers, const int debug_level, - const Program_mode program_mode, const bool ignore_ids, - const bool keep_damaged, const bool missing_crc, - const bool permissive ); - -// defined in list_lz.cc -struct LZ_Decoder; -int archive_read_lz( LZ_Decoder * const decoder, const int infd, - long long & file_pos, const long long member_end, - const long long cdata_size, uint8_t * const buf, - const int size, const char ** msg ); -int parse_records_lz( LZ_Decoder * const decoder, const int infd, - long long & file_pos, const long long member_end, - const long long cdata_size, long long & data_pos, - Extended & extended, const Tar_header header, - Resizable_buffer & rbuf, const char ** msg, - const bool permissive ); -int skip_member_lz( LZ_Decoder * const decoder, const int infd, - long long & file_pos, const long long member_end, - const long long cdata_size, long long & data_pos, - long long rest, const char ** msg ); -int list_lz( const char * const archive_namep, const Arg_parser & parser, - std::vector< char > & name_pending, const Lzip_index & lzip_index, - const int filenames, const int debug_level, const int infd, - const int num_workers, const bool missing_crc, - const bool permissive ); +bool compare_file_type( std::string & estr, std::string & ostr, + const Cl_options & cl_opts, + const Extended & extended, const Tar_header header ); +class Archive_reader_base; +bool compare_file_contents( std::string & estr, std::string & ostr, + Archive_reader_base & ar, const long long file_size, + const char * const filename, const int infd2 ); +int decode( const Cl_options & cl_opts ); + +// defined in decode_lz.cc +int preadblock( const int fd, uint8_t * const buf, const int size, + const long long pos ); +struct Archive_descriptor; +int decode_lz( const Cl_options & cl_opts, const Archive_descriptor & ad, + std::vector< char > & name_pending ); // defined in lzip_index.cc int seek_read( const int fd, uint8_t * const buf, const int size, @@ -440,12 +452,15 @@ int seek_read( const int fd, uint8_t * const buf, const int size, // defined in main.cc extern int verbosity; struct stat; -int hstat( const char * const filename, struct stat * const st ); +int hstat( const char * const filename, struct stat * const st, + const bool dereference ); int open_instream( const std::string & name ); int open_outstream( const std::string & name, const bool create = true ); void cleanup_and_fail( const int retval = 1 ); // terminate the program void show_error( const char * const msg, const int errcode = 0, const bool help = false ); +void format_file_error( std::string & estr, const char * const filename, + const char * const msg, const int errcode = 0 ); void show_file_error( const char * const filename, const char * const msg, const int errcode = 0 ); void internal_error( const char * const msg ); diff --git a/testsuite/check.sh b/testsuite/check.sh index 1b6f3f7..74e8549 100755 --- a/testsuite/check.sh +++ b/testsuite/check.sh @@ -1,9 +1,9 @@ #! /bin/sh # check script for Tarlz - Archiver with multimember lzip compression -# Copyright (C) 2013-2019 Antonio Diaz Diaz. +# Copyright (C) 2013-2020 Antonio Diaz Diaz. # # This script is free software: you have unlimited permission -# to copy, distribute and modify it. +# to copy, distribute, and modify it. LC_ALL=C export LC_ALL @@ -60,20 +60,24 @@ eof="${testdir}"/eof.tar eof_lz="${testdir}"/eof.tar.lz fail=0 lwarn=0 +lwarnc=0 test_failed() { fail=1 ; printf " $1" ; [ -z "$2" ] || printf "($2)" ; } lzlib_1_11() { [ ${lwarn} = 0 ] && printf "\nwarning: testing --keep-damaged requires lzlib-1.11 or newer\n$1" lwarn=1 ; } +cyg_symlink() { [ ${lwarnc} = 0 ] && + printf "\nwarning: your OS follows symbolic links to directories even when tarlz asks it not to\n$1" + lwarnc=1 ; } # Description of test files for tarlz: # test.txt.tar.lz: 1 member (test.txt). # t155.tar[.lz]: directory + links + file + eof, all with 155 char names # t155_fv?.tar[.lz]: like t155.tar but with 3 kinds of format violations -# tar_in_tlz1.tar.lz 2 members (test.txt.tar test3.tar) 3 lzip members -# tar_in_tlz2.tar.lz 2 members (test.txt.tar test3.tar) 5 lzip members +# tar_in_tlz1.tar.lz: 2 members (test.txt.tar test3.tar) 3 lzip members +# tar_in_tlz2.tar.lz: 2 members (test.txt.tar test3.tar) 5 lzip members # ts_in_link.tar.lz: 4 symbolic links (link[1-4]) to / /dir/ dir/ dir(107/) -# test_bad1.tar.lz: truncated at offset 6000 (of 7495) -# test_bad2.tar.lz: byte at offset 6000 changed from 0x56 to 0x46 +# test_bad1.txt.tar.lz: truncated at offset 6000 (of 7495) +# test_bad2.txt.tar.lz: byte at offset 6000 changed from 0x56 to 0x46 # test3.tar[.lz]: 3 members (foo bar baz) + 2 zeroed 512-byte blocks # test3_dir.tar[.lz] like test3.tar but members /dir/foo /dir/bar /dir/baz # test3_dot.tar.lz: 3 times 3 members ./foo ././bar ./././baz @@ -108,12 +112,17 @@ lzlib_1_11() { [ ${lwarn} = 0 ] && # ug32chars.tar.lz: 1 member (foo) with 32-character owner and group names # ug32767.tar.lz: 1 member (foo) with numerical-only owner and group +# Note that multi-threaded --list succeeds with test_bad2.txt.tar.lz and +# test3_bad3.tar.lz because their headers are intact. + printf "testing tarlz-%s..." "$2" "${TARLZ}" -q -tf "${in}" [ $? = 2 ] || test_failed $LINENO "${TARLZ}" -q -tf "${in_lz}" [ $? = 2 ] || test_failed $LINENO +"${TARLZ}" -q -tf "${in_tar_lz}" -f "${in_tar_lz}" +[ $? = 1 ] || test_failed $LINENO "${TARLZ}" -q -tf nx_file [ $? = 1 ] || test_failed $LINENO "${TARLZ}" -tf 2> /dev/null @@ -159,6 +168,8 @@ rm -f empty.tar.lz empty.tlz || framework_failure [ $? = 1 ] || test_failed $LINENO "${TARLZ}" -q -ctx [ $? = 1 ] || test_failed $LINENO +"${TARLZ}" -q -tf "${in_tar_lz}" "" +[ $? = 1 ] || test_failed $LINENO "${TARLZ}" --help > /dev/null || test_failed $LINENO "${TARLZ}" -V > /dev/null || test_failed $LINENO "${TARLZ}" --bad_option -tf "${test3_lz}" 2> /dev/null @@ -382,16 +393,16 @@ for i in "${tarint1_lz}" "${tarint2_lz}" ; do "${TARLZ}" -tvf "$i" -n$j > outv$j || test_failed $LINENO "$i $j" done - diff -u out0 out2 || test_failed $LINENO $i - diff -u out0 out6 || test_failed $LINENO $i - diff -u out2 out6 || test_failed $LINENO $i - diff -u outv0 outv2 || test_failed $LINENO $i - diff -u outv0 outv6 || test_failed $LINENO $i - diff -u outv2 outv6 || test_failed $LINENO $i + diff -u out0 out2 || test_failed $LINENO "$i" + diff -u out0 out6 || test_failed $LINENO "$i" + diff -u out2 out6 || test_failed $LINENO "$i" + diff -u outv0 outv2 || test_failed $LINENO "$i" + diff -u outv0 outv6 || test_failed $LINENO "$i" + diff -u outv2 outv6 || test_failed $LINENO "$i" rm -f out0 out2 out6 outv0 outv2 outv6 || framework_failure - "${TARLZ}" -xf "$i" || test_failed $LINENO $i - cmp "${in_tar}" test.txt.tar || test_failed $LINENO $i - cmp "${test3}" test3.tar || test_failed $LINENO $i + "${TARLZ}" -xf "$i" || test_failed $LINENO "$i" + cmp "${in_tar}" test.txt.tar || test_failed $LINENO "$i" + cmp "${test3}" test3.tar || test_failed $LINENO "$i" rm -f test.txt.tar test3.tar || framework_failure done @@ -414,16 +425,16 @@ done # extended tar members split among lzip members for i in em1 em2 em3 em4 em5 em6 gh1 gh2 gh3 gh4 gh5 gh6 sm1 sm2 sm3 sm4 ; do for j in 0 2 6 ; do - "${TARLZ}" -tf "${testdir}"/test3_${i}.tar.lz -n$j > out || + "${TARLZ}" -n$j -tf "${testdir}"/test3_${i}.tar.lz > out || test_failed $LINENO "$i $j" diff -u list3 out || test_failed $LINENO "$i $j" - "${TARLZ}" -tvf "${testdir}"/test3_${i}.tar.lz -n$j > out || + "${TARLZ}" -n$j -tvf "${testdir}"/test3_${i}.tar.lz > out || test_failed $LINENO "$i $j" diff -u vlist3 out || test_failed $LINENO "$i $j" done rm -f out || framework_failure for j in 0 2 6 ; do - "${TARLZ}" -xf "${testdir}"/test3_${i}.tar.lz -n$j || + "${TARLZ}" -n$j -xf "${testdir}"/test3_${i}.tar.lz || test_failed $LINENO "$i $j" cmp cfoo foo || test_failed $LINENO "$i $j" cmp cbar bar || test_failed $LINENO "$i $j" @@ -431,6 +442,7 @@ for i in em1 em2 em3 em4 em5 em6 gh1 gh2 gh3 gh4 gh5 gh6 sm1 sm2 sm3 sm4 ; do rm -f foo bar baz || framework_failure done done +rm -f list3 vlist3 || framework_failure # test --concatenate compressed cat "${in}" > out.tar.lz || framework_failure # invalid tar.lz @@ -556,7 +568,7 @@ cat cbar > bar || framework_failure cat cbaz > baz || framework_failure "${TARLZ}" -0 -cf out.tar.lz foo bar baz --out-slots=1 || test_failed $LINENO "${TARLZ}" -0 -q -cf aout.tar.lz foo bar aout.tar.lz baz || test_failed $LINENO -cmp out.tar.lz aout.tar.lz || test_failed $LINENO +cmp out.tar.lz aout.tar.lz || test_failed $LINENO # test reproducible rm -f aout.tar.lz || framework_failure # "${TARLZ}" -0 -cf aout.tar.lz foo bar baz -C / || test_failed $LINENO @@ -638,25 +650,33 @@ rm -f out.tar foo bar baz || framework_failure "${TARLZ}" --uncompressed -cf out.tar foo || test_failed $LINENO "${TARLZ}" --uncompressed -cf aout.tar foo --anonymous || test_failed $LINENO if cmp out.tar aout.tar > /dev/null ; then - printf "\nwarning: '--diff' test can't be run as root." + printf "\nwarning: '--diff' test can't be run as root." else - "${TARLZ}" -df "${test3_lz}" > /dev/null - [ $? = 1 ] || test_failed $LINENO - "${TARLZ}" -df "${test3_lz}" --ignore-ids || test_failed $LINENO - "${TARLZ}" -df "${test3_lz}" --exclude '*' || test_failed $LINENO - "${TARLZ}" -df "${in_tar_lz}" --exclude '*' || test_failed $LINENO - rm -f bar || framework_failure - "${TARLZ}" -df "${test3_lz}" foo baz --ignore-ids || test_failed $LINENO - "${TARLZ}" -df "${test3_lz}" --exclude bar --ignore-ids || - test_failed $LINENO - rm -f foo baz || framework_failure - "${TARLZ}" -q -xf "${test3dir_lz}" || test_failed $LINENO - "${TARLZ}" -q -df "${test3dir_lz}" --ignore-ids || test_failed $LINENO - "${TARLZ}" -q -df "${test3dir_lz}" dir --ignore-ids || test_failed $LINENO - "${TARLZ}" -df "${test3_lz}" --ignore-ids -C dir || test_failed $LINENO - rm -rf dir || framework_failure + for i in 0 2 ; do + "${TARLZ}" -n$i -xf "${test3_lz}" || test_failed $LINENO + "${TARLZ}" -n$i -df "${test3_lz}" > out$i + [ $? = 1 ] || test_failed $LINENO $i + "${TARLZ}" -n$i -df "${test3_lz}" --ignore-ids || test_failed $LINENO $i + "${TARLZ}" -n$i -df "${test3_lz}" --exclude '*' || test_failed $LINENO $i + "${TARLZ}" -n$i -df "${in_tar_lz}" --exclude '*' || test_failed $LINENO $i + rm -f bar || framework_failure + "${TARLZ}" -n$i -df "${test3_lz}" foo baz --ignore-ids || + test_failed $LINENO $i + "${TARLZ}" -n$i -df "${test3_lz}" --exclude bar --ignore-ids || + test_failed $LINENO $i + rm -f foo baz || framework_failure + "${TARLZ}" -q -n$i -xf "${test3dir_lz}" || test_failed $LINENO $i + "${TARLZ}" -q -n$i -df "${test3dir_lz}" --ignore-ids || + test_failed $LINENO $i + "${TARLZ}" -q -n$i -df "${test3dir_lz}" dir --ignore-ids || + test_failed $LINENO $i + "${TARLZ}" -n$i -df "${test3_lz}" --ignore-ids -C dir || + test_failed $LINENO $i + rm -rf dir || framework_failure + done fi -rm -f out.tar aout.tar foo bar baz || framework_failure +cmp out0 out2 || test_failed $LINENO +rm -f out0 out2 out.tar aout.tar foo bar baz || framework_failure # test --delete for e in "" .lz ; do @@ -778,11 +798,12 @@ if [ "${ln_works}" = yes ] ; then "${TARLZ}" -0 -n0 -h -cf hout3 dir_link || test_failed $LINENO rm -rf dir dir_link || framework_failure for i in 1 2 3 ; do - "${TARLZ}" -xf out$i || test_failed $LINENO $i + "${TARLZ}" -xf out$i --exclude='dir_link/*' dir_link || + test_failed $LINENO $i # Cygwin stores dir_link/* [ -h dir_link ] || test_failed $LINENO $i - [ ! -e dir_link/foo ] || test_failed $LINENO $i - [ ! -e dir_link/bar ] || test_failed $LINENO $i - [ ! -e dir_link/baz ] || test_failed $LINENO $i + "${TARLZ}" -q -tf out$i dir_link/foo && cyg_symlink $LINENO $i + "${TARLZ}" -q -tf out$i dir_link/bar && cyg_symlink $LINENO $i + "${TARLZ}" -q -tf out$i dir_link/baz && cyg_symlink $LINENO $i rm -rf dir_link out$i || framework_failure "${TARLZ}" -xf hout$i || test_failed $LINENO $i [ -d dir_link ] || test_failed $LINENO $i |