NAME
ZIP - package and compress (archive) files
SYNOPSIS
CHAIN "zip zipfile [inpath] [-# -D -g -nodir -q -r -resume -X]
[-i@includefile] [-pp path] [-S suffixfile] [-t date] [-quit quitcmd]"
or *zip zipfile [inpath] [-# -D -g -nodir -q -r -resume -X]
[-i@includefile] [-pp path] [-S suffixfile] [-t date] [-quit quitcmd]
DESCRIPTION
zip is a compression and file packaging utility for Unix, VMS, MSDOS,
OS/2, CP/M, Windows NT, Minix, Atari, Macintosh, Amiga, Acorn RISC OS
and Acorn BBC.
A companion program (unzip(1)), unpacks zip archives. The zip and
unzip(1) programs can work with archives produced by PKZIP and
Archive(1), and PKZIP, PKUNZIP and Archive can work with archives
produced by zip. zip version 1.xx is compatible with PKZIP 2.04 and
later and Archive 2.16 and later. Note that PKUNZIP 1.10 cannot extract
files produced by PKZIP 2.04 or zip. You must use PKUNZIP 2.04g (or
later versions) to extract them.
The program is useful for packaging a set of files for distribution;
for archiving files; and for saving disk space by temporarily
compressing unused files or directories.
This describes the BBC version of zip which aims to replicate the
behaviour of InfoZip zip v3. It does not compress files, just archives
them, unless helped by SparkFS.
ARGUMENTS
zipfile
filename of the zipfile to create. The zipfile will be typed to
Archive (&DDC). On the Master and RISC OS the zipfile can have a
-fsname- or fsname: prefix to specify the filing system the file
is on.
inpath
Path to include. If inpath is a directory, then all files
within that directory will be included. If inpath is a
file, just that single file will be included. If inpath
contains wildcards then it will use only the first matching
object.
OPTIONS
Options supplied to the BBC version of zip must each be specified
singly. This is different to other platforms where multiple options can
be run together.
-? Display the zip command line syntax.
-D Do not create entries in the zip archive for directories.
Directory entries are created by default so that their
attributes can be saved in the zip archive.
-g Grow existing zipfile by appending to it. If the existing
zipfile is truncated it will be appended to after the last
valid entry.
-i@filename
Include the files listed in the file specified by filename.
If no filename is given, the filenames are read from the
input stream.
-nodir
Do not add a zip catalogue to the end of the zipfile. The
final file only contains the stored entries with their
inline metadata. This will shorten the zipfile by something
like 10%-15% depending on contents, but some unzip tools
may not be able to extract the contents. BBC unzip does
extract the contents as it ignores the zip catalogue and
uses the inline metadata.
-pp path
Prepends the specified path onto the stored filename in the
archive. This allows the entries within an archive to
appear to be within the specified directory. For example:
*zip outfile infile -pp DirName
will store infile in the archive, where it will be called
DirName.infile.
-q Quiet mode. All output messages are disabled.
-quit quitcmd
Command to execute on termination. This allows zip to be
called and then continue to a specified program. If quitcmd
starts with a '*' it is called as a *command, otherwise it
is CHAINed.
-r Travel the directory structure recursively, for example:
*zip outfile indir -r
In this case, all the files and directories in indir are
saved in a zip archive named outfile.
-resume
Restarts an aborted zip session by appending to the
zipfile, scanning through it and finding the last entry
successfully added and then continuing to add files from
that point onwards. This way an aborted zip session can be
restarted at the file that was being added when the
previous session terminated.
-S suffixfile
Specifies a text file containing a list of file extensions
to map to file types, commonly called a MimeMap file. If
this is specified then files added to the zip file which
have truncated file extentions will have those extensions
extended according to their filetype. For instance, the
file mainpage/h will be stored as mainpage/htm (translated
to mainpage.htm) if it is filetyped to &FAF and the
suffixfile contains the entry 'FAF,htm'.
-t date
Only include files that were modified on the specified date
or later. The date is specified as dd/mm/yyyy or yyyy/mm/dd.
-X Include extra file information. By default zip stores a
object's load and execution address, access permissions,
and the modification date and time. This option will
include all the additional metadata associated with objects
stored on SJ MDFS file servers and with filing systems such
as HADFS, this being the creation date and time, main
account and auxillary acount.
-# Specify the compression type used, where -0 is no
compression (store all files) and -8 is deflate. BBCZip
will compress files if used on RISC OS with SparkFS
running. Otherwise, any -# option is ignored and BBCZip
stores files. As a special option -255 causes zip to store
only objects' metadata and no data.
EXAMPLES
The simplest example:
*zip stuff things
creates the archive stuff and puts all the files in the directory
things in it. This will not include any subdirectories from the
directory. To do this, the command:
*zip stuff -r things
creates the archive stuff, containing all the files and directories in
directory things.
The use of the -pp option allows the entries within the archive to be
named as being in a specified directory. For example:
*zip BBCZip/zip BBCZip -r -pp Software.Archivers.BBCZip
will archive the entries within the directory BBCZip and within the
archive the entries will appear to be within a directory named
Software.Archivers.BBCZip. Using this small parts of a directory tree
can be archived without the whole directory tree having to be archived.
The -i@ option allows a list of files in a textfile to be archived.
For example:
*zip Update/zip :Docs.$ -i@uplist
will create the archive Update/zip containing only those files in
directory :Docs.$ that are listed in the text file uplist. This option
is useful when creating an update archive based on a comparison with
another archive.
The lines in the include file must be terminated with <cr>s. Note that
there is no space between the '-i@' option and the filename.
Using -@ with no filename will read the list of filenames from the
current input stream. This can allow you to have an *EXEC file zip up
objects from the EXEC file itself. For example, a file containing:
*zip ^.man/zip @ -@
man/src
man
*EXECing this file will start zip and include the files in the current
directory listed in the rest of the file.
Using the -X option will include all the extra file information that
systems such as the SJ MDFS fileserver have. Using:
*zip backup1 :SYSTEM.$ -r -X -pp :SYSTEM.$
will create an archive of everything on the disk SYSTEM including all
the additional file information, and including the name of the source
directory in the archive entry names. This archive can be unzipped to
completely restore all the files on disk SYSTEM along with all the
associated file information using:
*unzip backup1 -X
Note that to do this the appropriate user privileges will be required.
The -t option gives the ability to make incremental zip archives,
zipping up only files that have been modified after a certain date. As
an example, extending the previous example to:
*zip backup1 :SYSTEM.$ -r -X -t 01/12/1999 -pp :SYSTEM.$
would zip up only those files that were last modified on 1st December
1999 or later.
Note that you shouldn't try to create the zipfile within the directory
structure that you specify to include in the zipfile otherwise zip will
attempt to include the zipfile within itself. For example:
*zip stuff @
will create the zipfile stuff within the current directory, and try to
include within it all the files within the current directory, including
the zipfile.
TECHNICAL INFORMATION
ZIP files include an extras field that holds metadata specific to
particular systems. Acorn metadata is stored in the extras field in the
following format:
0 extra header id 2 bytes "AC"
2 extra header sublength 2 bytes ------------v
4 Acorn header id 4 bytes "ARC0" 4
8 load address 4 bytes 8
12 execution address 4 bytes 12
16 attributes 4 bytes 16
20 &00000000 4 bytes 20
24 creation time 2 bytes 22
26 creation date 2 bytes 24
28 main account number 2 bytes 26
30 auxilary account number 2 bytes 28
Only that metadata specified by the header sublength is present. Any
metadata that is omitted results in UnZip not setting that metadata on
extracted objects. If the execution address is missing, but the load
address is present, it is assumed to be the same as the load address.
Creation and modification times and dates are stored in MS-DOS format.
BBCZip bidirectionally swaps the following filename characters:
. <-> / # <-> ? $ <-> < ^ <-> > @ <-> =
When appending with the -resume option BBCZip puts a "PK",&0F,&10
'FileOpen' header at the start of the output file, followed by the
offset to the last correctly-added entry. This allows the -resume
option to quickly find the last entry. This header is replaced with a
"PK",&03,&04 header when the file is finalised.
NOTES
zip 1.xx does not compress files, just stores.
The -X switch is slightly non-symetrical between Zip and UnZip:
Zip stores just the common Acorn metadata, load/exec addresses,
attributes and modification date/time
Zip -X stores all available metadata on the source filing system
UnZip extracts all stored metadata except any account numbers
UnZip -X extracts all stored metadata and any account numbers
If running in an environment where SparkFS is available, then
compression is available by calling *ZipCompress.
zip attempts to find additional file metadata by using OSFILE &FD,
NetFS_Op 18,64 and NetFS_Op 18,65.
There is a bug in the CmdLine library used by zip where any hyphen is
seen as introducing an option, for example Zip -ramfs-out or Zip new-run
is seen as the -r option. This can be worked around by using upper
case for the file name, eg -RAMFS-out or a colon filing system
prefix, eg ramfs:out.
Need to do:
Some way of storing metadata on start directory, so zip out $ -r
stores infomation about $.
Store 'P' bit from MDFS
Is DefAccess info readable from MDFS?
SEE ALSO
Archive(1), UnZip(1), UnZip(9), Zip(1), Zip(5), Zip(9), ZipCat(1),
ZipDiff(1), ZipEx(1), ZipInfo(1).
AUTHOR
J.G.Harston <jgh@mdfs.net>