diff options
Diffstat (limited to 'doc/tarlz.texi')
| -rw-r--r-- | doc/tarlz.texi | 179 |
1 files changed, 95 insertions, 84 deletions
diff --git a/doc/tarlz.texi b/doc/tarlz.texi index 5bdd2af..2c757e8 100644 --- a/doc/tarlz.texi +++ b/doc/tarlz.texi @@ -6,8 +6,8 @@ @finalout @c %**end of header -@set UPDATED 23 September 2022 -@set VERSION 0.23 +@set UPDATED 20 September 2023 +@set VERSION 0.24 @dircategory Archiving @direntry @@ -50,7 +50,7 @@ This manual is for Tarlz (version @value{VERSION}, @value{UPDATED}). @end menu @sp 1 -Copyright @copyright{} 2013-2022 Antonio Diaz Diaz. +Copyright @copyright{} 2013-2023 Antonio Diaz Diaz. This manual is free documentation: you have unlimited permission to copy, distribute, and modify it. @@ -81,9 +81,9 @@ 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). It can also -create uncompressed tar archives. +per file (@option{--no-solid}), per block (@option{--bsolid}, default), per +directory (@option{--dsolid}), appendable solid (@option{--asolid}), and +solid (@option{--solid}). It can also create uncompressed tar archives. @noindent Of course, compressing each file (or each directory) individually can't @@ -104,7 +104,7 @@ 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 @samp{--keep-damaged} can be used +(uncompressed) tar. Moreover, the option @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. @@ -118,8 +118,8 @@ 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}. The command -@w{@samp{tarlz -tf archive.tar.lz > /dev/null}} can be used to verify that +@samp{star}, or @samp{v7}. The command +@w{@samp{tarlz -t -f archive.tar.lz > /dev/null}} can be used to check that the format of the archive is compatible with tarlz. @@ -137,9 +137,9 @@ tarlz @var{operation} [@var{options}] [@var{files}] @end example @noindent -All operations except @samp{--concatenate} and @samp{--compress} operate on -whole trees if any @var{file} is a directory. All operations except -@samp{--compress} overwrite output files without warning. If no archive is +All operations except @option{--concatenate} and @option{--compress} operate +on whole trees if any @var{file} is a directory. All operations except +@option{--compress} overwrite output files without warning. If no archive is specified, tarlz tries to read it from standard input or write it to standard output. Tarlz refuses to read archive data from a terminal or write archive data to a terminal. Tarlz detects when the archive being created or @@ -147,7 +147,7 @@ enlarged is among the files to be archived, appended, or concatenated, and skips it. Tarlz does not use absolute file names nor file names above the current -working directory (perhaps changed by option @samp{-C}). On archive creation +working directory (perhaps changed by option @option{-C}). 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 @samp{..} component. On extraction, leading and trailing slashes are also @@ -161,9 +161,9 @@ member names in the archive or given in the command line, so that @w{@samp{tarlz -xf foo ./bar baz}} extracts members @samp{bar} and @samp{./baz} from archive @samp{foo}. -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 @w{@samp{-1 --solid}}. +If several compression levels or @option{--*solid} options are given, the last +setting is used. For example @w{@option{-9 --solid --uncompressed -1}} is +equivalent to @w{@option{-1 --solid}}. tarlz supports the following operations: @@ -179,7 +179,7 @@ This version number should be included in all bug reports. @item -A @itemx --concatenate Append one or more archives to the end of an archive. If no archive is -specified with the option @samp{-f}, the input archives are concatenated to +specified with the option @option{-f}, concatenate the input archives to standard output. 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 @@ -202,23 +202,23 @@ Create a new archive from @var{files}. @item -d @itemx --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 +member in the archive, check 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 @samp{-C} may -be used in combination with @samp{--diff} when absolute file names were used +As tarlz removes leading slashes from member names, the option @option{-C} may +be used in combination with @option{--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. @item --delete 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 +compressed individually (@option{--no-solid} archives). Note that files of +about @option{--data-size} or larger are compressed individually even if +@option{--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 @@ -261,32 +261,36 @@ Extract files from an archive. If @var{files} are given, extract only the directory without extracting the files under it, use @w{@samp{tarlz -xf foo --exclude='dir/*' dir}}. Tarlz removes files and empty directories unconditionally before extracting over them. Other than -that, it will not make any special effort to extract a file over an +that, it does not make any special effort to extract a file over an incompatible type of file. For example, extracting a file over a non-empty -directory will usually fail. +directory usually fails. @item -z @itemx --compress Compress existing POSIX tar archives aligning the lzip members to the tar -members with choice of granularity (---bsolid by default, ---dsolid works -like ---asolid). The input archives are kept unchanged. Existing compressed -archives are not overwritten. A hyphen @samp{-} used as the name of an input -archive reads from standard input and writes to standard output (unless the -option @samp{--output} is used). Tarlz can be used as compressor for GNU tar -using a command like @w{@samp{tar -c -Hustar foo | tarlz -z -o foo.tar.lz}}. -Note that tarlz only works reliably on archives without global headers, or -with global headers whose content can be ignored. +members with choice of granularity (@option{--bsolid} by default, +@option{--dsolid} works like @option{--asolid}). Exit with error status 2 if +any input archive is an empty file. The input archives are kept unchanged. +Existing compressed archives are not overwritten. A hyphen @samp{-} used as +the name of an input archive reads from standard input and writes to +standard output (unless the option @option{--output} is used). Tarlz can be +used as compressor for GNU tar by using a command like +@w{@samp{tar -c -Hustar foo | tarlz -z -o foo.tar.lz}}. Tarlz can be used as +compressor for zupdate (zutils) by using a command like +@w{@samp{zupdate --lz="tarlz -z" foo.tar.gz}}. Note that tarlz only works +reliably on archives without global headers, or with global headers whose +content can be ignored. The compression is reversible, including any garbage present after the end-of-archive blocks. Tarlz stops parsing after the first end-of-archive block is found, and then compresses the rest of the archive. Unless solid compression is requested, the end-of-archive blocks are compressed in a lzip member separated from the preceding members and from any non-zero garbage -following the end-of-archive blocks. @samp{--compress} implies plzip +following the end-of-archive blocks. @option{--compress} implies plzip argument style, not tar style. Each input archive is compressed to a file -with the extension @samp{.lz} added unless the option @samp{--output} is -used. When @samp{--output} is used, only one input archive can be specified. -@samp{-f} can't be used with @samp{--compress}. +with the extension @samp{.lz} added unless the option @option{--output} is +used. When @option{--output} is used, only one input archive can be specified. +@option{-f} can't be used with @option{--compress}. @item --check-lib Compare the @@ -314,25 +318,25 @@ tarlz supports the following @anchor{--data-size} @item -B @var{bytes} @itemx --data-size=@var{bytes} -Set target size of input data blocks for the option @samp{--bsolid}. +Set target size of input data blocks for the option @option{--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 +value is two times the dictionary size, except for option @option{-0} where it defaults to @w{1 MiB}. @xref{Minimum archive sizes}. @item -C @var{dir} @itemx --directory=@var{dir} -Change to directory @var{dir}. When creating or appending, the position of -each @samp{-C} option in the command line is significant; it will change the -current working directory for the following @var{files} until a new -@samp{-C} option appears in the command line. When extracting or comparing, -all the @samp{-C} options are executed in sequence before reading the -archive. Listing ignores any @samp{-C} options specified. @var{dir} is -relative to the then current working directory, perhaps changed by a -previous @samp{-C} option. +Change to directory @var{dir}. When creating, appending, comparing, or +extracting, the position of each @option{-C} option in the command line is +significant; it changes the current working directory for the following +@var{files} until a new @option{-C} option appears in the command line. +@option{--list} and @option{--delete} ignore any @option{-C} options +specified. @var{dir} is relative to the then current working directory, +perhaps changed by a previous @option{-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 @samp{-C} -option appears after a relative file name in the command line. +Therefore multi-threading can't be used to create or decode an archive if a +@option{-C} option appears after a (relative) file name in the command line. +(All file names are made relative when decoding). @item -f @var{archive} @itemx --file=@var{archive} @@ -351,7 +355,7 @@ 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 option @samp{-C} above. +multi-threading in the option @option{-C} above. Note that the number of usable threads is limited during compression to @w{ceil( uncompressed_size / data_size )} (@pxref{Minimum archive sizes}), @@ -360,9 +364,9 @@ archive, which you can find by running @w{@samp{lzip -lv archive.tar.lz}}. @item -o @var{file} @itemx --output=@var{file} -Write the compressed output to @var{file}. @w{@samp{-o -}} writes the -compressed output to standard output. Currently @samp{--output} only works -with @samp{--compress}. +Write the compressed output to @var{file}. @w{@option{-o -}} writes the +compressed output to standard output. Currently @option{--output} only works +with @option{--compress}. @item -p @itemx --preserve-permissions @@ -381,8 +385,8 @@ Verbosely list files processed. Further -v's (up to 4) increase the verbosity level. @item -0 .. -9 -Set the compression level for @samp{--create}, @samp{--append}, and -@samp{--compress}. The default compression level is @samp{-6}. Like lzip, +Set the compression level for @option{--create}, @option{--append}, and +@option{--compress}. The default compression level is @option{-6}. Like lzip, tarlz also minimizes the dictionary size of the lzip members it creates, reducing the amount of memory required for decompression. @@ -401,10 +405,12 @@ reducing the amount of memory required for decompression. @end multitable @item --uncompressed -With @samp{--create}, don't compress the tar archive created. Create an -uncompressed tar archive instead. With @samp{--append}, don't compress the +With @option{--create}, don't compress the tar archive created. Create an +uncompressed tar archive instead. With @option{--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. +appended to an uncompressed archive, nor vice versa. @option{--uncompressed} +can be omitted if it can be deduced from the archive name. (An uncompressed +archive name lacks a @samp{.lz} or @samp{.tlz} extension). @item --asolid When creating or appending to a compressed archive, use appendable solid @@ -447,7 +453,7 @@ appendable. No more files can be later appended to the archive. Solid archives can't be created nor decoded in parallel. @item --anonymous -Equivalent to @w{@samp{--owner=root --group=root}}. +Equivalent to @w{@option{--owner=root --group=root}}. @item --owner=@var{owner} When creating or appending, use @var{owner} for files added to the archive. @@ -465,27 +471,28 @@ to match if any component of the file name matches. For example, @samp{*.o} matches @samp{foo.o}, @samp{foo.o/bar} and @samp{foo/bar.o}. If @var{pattern} contains a @samp{/}, it matches a corresponding @samp{/} in the file name. For example, @samp{foo/*.o} matches @samp{foo/bar.o}. -Multiple @samp{--exclude} options can be specified. +Multiple @option{--exclude} options can be specified. @item --ignore-ids -Make @samp{--diff} ignore differences in owner and group IDs. This option is -useful when comparing an @samp{--anonymous} archive. +Make @option{--diff} ignore differences in owner and group IDs. This option is +useful when comparing an @option{--anonymous} archive. @item --ignore-overflow -Make @samp{--diff} ignore differences in mtime caused by overflow on 32-bit +Make @option{--diff} ignore differences in mtime caused by overflow on 32-bit systems with a 32-bit time_t. @item --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. It is recommended -to run tarlz in single-threaded mode (---threads=0) when using this option. +to run tarlz in single-threaded mode (@option{--threads=0}) when using this +option. @item --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 @samp{GNU.crc32} -keyword, for example @samp{GNU.crc33}, is reported as a missing CRC instead +keyword, for example @samp{GNU.crc30}, is reported as a missing CRC instead of as a corrupt record. This misleading @w{@samp{Missing CRC}} message is the consequence of a flaw in the POSIX pax format; i.e., the lack of a mandatory check sequence of the extended records. @xref{crc32}. @@ -606,7 +613,7 @@ Zero or more blocks that contain the contents of the file. @end itemize 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 +parallel decoding operations like @option{--list} to work. If any tar member is split over two or more lzip members, the archive must be decoded sequentially. @xref{Multi-threaded decoding}. @@ -639,7 +646,7 @@ tar.lz @end verbatim @ignore -When @samp{--permissive} is used, the following violations of the +When @option{--permissive} is used, the following violations of the archive format are allowed:@* If several extended headers precede an ustar header, only the last extended header takes effect. The other extended headers are ignored. @@ -660,6 +667,10 @@ 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}. +Tarlz limits the size of the pax extended header data so that the whole +header set (extended header + extended data + ustar header) can be read and +decoded in a buffer of size INT_MAX. + The pax extended header data consists of one or more records, each of them constructed as follows:@* @w{@samp{"%d %s=%s\n", <length>, <keyword>, <value>}} @@ -689,11 +700,11 @@ greater than 2_097_151 (octal 7777777). @xref{ustar-uid-gid}. The file name of a link being created to another file, of any type, 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 symbolic link and the -linkpath value will be used as the contents of the symbolic link. The -linkpath record is created only for links with a link name that does not fit -in the space provided by the ustar header. +the type of link created. If typeflag of the following header block is 1, a +hard link is created. If typeflag is 2, a symbolic link is created and the +linkpath value is used as the contents of the symbolic link. The linkpath +record is created only for links with a link name that does not fit in the +space provided by the ustar header. @item mtime The signed decimal representation of the modification time of the following @@ -728,8 +739,8 @@ 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, @w{@samp{22 GNU.crc32=00000000\n}}. The keyword of the CRC record is -protected by the CRC to guarante that corruption is always detected when -using @samp{--missing-crc} (except in case of CRC collision). A CRC was +protected by the CRC to guarantee that corruption is always detected when +using @option{--missing-crc} (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. @@ -878,7 +889,7 @@ 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{--diff}, @samp{--list}, and @samp{--extract}. +it happens in parallel @option{--diff}, @option{--list}, and @option{--extract}. In order to achieve this goal and avoid some other flaws in the pax format, 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. @@ -919,11 +930,11 @@ 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 -@samp{name} and @samp{prefix} will be unable to contain the full file name. +@samp{name} and @samp{prefix} are unable to contain the full file name. (Some tar implementations store the truncated name in the field @samp{name} alone, truncating the name to only 100 bytes instead of 256). Therefore the files corresponding to both the extended header and the overridden ustar -header will be extracted using truncated file names, perhaps overwriting +header are 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. @@ -1117,9 +1128,9 @@ 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. -If the files in the archive are large, multi-threaded @samp{--list} on a +If the files in the archive are large, multi-threaded @option{--list} on a regular (seekable) tar.lz archive can be hundreds of times faster than -sequential @samp{--list} because, in addition to using several processors, +sequential @option{--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: @@ -1130,7 +1141,7 @@ 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 +On the other hand, multi-threaded @option{--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. This is another reason why the tar headers must provide their own integrity checking. @@ -1176,11 +1187,11 @@ 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 -(@pxref{--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 +(@pxref{--data-size}). Else some processors do not get any data to compress, +and compression is 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 14 processors at level -9. +or linux scales 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 |