swCHIP 1991 January

home *** CD-ROM | disk | FTP | other *** search

/ swCHIP 1991 January / swCHIP_95-1.bin / dos / crush18 / manual.doc < prev next >

Wrap

Text File | 1995-12-09 | 45KB | 998 lines

_________________________________ CRUSH v1.8 **** Shareware Manual **** __________________________________ File Compression Tool CRUSH (TM) Copyright Jeff Rollason, PocketWare (TM) 1994 All Rights Reserved All Trademarks Are Acknowledged No part of this publication may be reproduced, transmitted, transcribed, stored in a retrieval system, or translated into any other language in whole or in part, in any form or by any means, whether it be electronic, mechanical, magnetic, optical, manual or otherwise, without prior written consent of PocketWare. PocketWare PO Box 2369, Hendon, London NW4 1NR, England. CIS 100031,3537. Internet 100031.3537@compuserve.com __________________________________________________________________________ Contents: CONTENTS: Page Getting Started: . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 The CRUSH program: . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 Default use of PKZIP . . . . . . . . . . . . . . . . . . . . . . . 3 -j Use ARJ . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 -o Use ZOO . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 -l Use LHA . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 -h Use HA . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 -u Use UC2 . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 -U Create CRU files only . . . . . . . . . . . . . . . . . . . . 3 -m Moves file to archive . . . . . . . . . . . . . . . . . . . . 4 -c Compare with and without CRUSH . . . . . . . . . . . . . . . 4 -f Optimise until 5 failed attempts to improve compression . . . 4 -f<n> Optimise until "n" failed attempts to improve compression . . 4 -t Only create CRUSH archive if 5% improvement . . . . . . . . . 4 -t<n> Only create CRUSH archive if "n"% improvement . . . . . . . . 4 -r Recurse subdirectories . . . . . . . . . . . . . . . . . . . 4 -: Set range(s) of dates . . . . . . . . . . . . . . . . . . . . 5 -x Exclude pathspecs . . . . . . . . . . . . . . . . . . . . . . 6 -y Answer YES to all questions . . . . . . . . . . . . . . . . . 6 -z Add CRUSH archive comment . . . . . . . . . . . . . . . . . . 6 -v Show files for adding without CRUSHing . . . . . . . . . . . 6 -Ln Set limit for files per archive . . . . . . . . . . . . . . . 7 -n Store files without paths . . . . . . . . . . . . . . . . . . 7 -Q Query each file to be archived. . . . . . . . . . . . . . . . 7 -q QUIET mode, no output . . . . . . . . . . . . . . . . . . . . 7 -a Create an additional index . . . . . . . . . . . . . . . . . 7 -e Use the existing archive for comparison . . . . . . . . . . . 8 -C Colour off . . . . . . . . . . . . . . . . . . . . . . . . . 8 -C<n> Set colour . . . . . . . . . . . . . . . . . . . . . . . . . 8 -s Minimise temporary disk space (slow!) . . . . . . . . . . . . 8 -i<n> Control information type . . . . . . . . . . . . . . . . . . 8 -2 Set display of times to 24 hour clock . . . . . . . . . . . . 8 The UNCRUSH Program: . . . . . . . . . . . . . . . . . . . . . . . . . . 9 -v Display without uncompressing . . . . . . . . . . . . . . . .10 -V Display archive comment only . . . . . . . . . . . . . . . .10 -o Overwrite files without query . . . . . . . . . . . . . . . .10 -u Uncompress NEWER or NEW files . . . . . . . . . . . . . . . .10 -a Uncompress NEW files . . . . . . . . . . . . . . . . . . . .10 -i Ignore paths in CRUSH file . . . . . . . . . . . . . . . . .10 -I Only ignore drives in paths . . . . . . . . . . . . . . . . .10 -q Quiet mode: No output at all . . . . . . . . . . . . . . . .11 -z Only suppress output of archiver . . . . . . . . . . . . . .11 -e Extract even if not CRUSH archive . . . . . . . . . . . . . .11 -l Only display a simple file list . . . . . . . . . . . . . . .11 -2 Display times as 24-hour clock . . . . . . . . . . . . . . .11 Worked Examples: . . . . . . . . . . . . . . . . . . . . . . . . . . . .12 The UNDO and UNDO2 Programs: . . . . . . . . . . . . . . . . . . . . . .13 Configuration (including using other archivers): . . . . . . . . . . . .14 Technical Details: . . . . . . . . . . . . . . . . . . . . . . . . . . .15 _________________________________________________________________________ Getting Started Page 1 ____________________________________ Installation CRUSH installation is very simple. The user need only copy the contents of the release disk to the desired directory. A file INSTALL.BAT is provided to do this for you. Your chosen directory should be within your search path (See the PATH variable in your DOS documentation). In order to be able to use CRUSH, you must have one of the archivers PKZIP, ARJ, LHA, ZOO, HA and UC; or your own particular archiver. ____________________________________ What is CRUSH? CRUSH is a command-line tool that uses your existing archiver PKZIP, ARJ, LHA, ZOO, HA or UC to generate smaller archives than it could alone. With many applications this improvement may be less than 5%, but dramatic improvements of 10%-70% are much more typical. A block test of 128 archives totalling 2349 wordprocessing files (see README) averaged a 47% improvement (100k ZIP reduced to 53k CRUSH!). CRUSH is ideally suited to users who regularly archive material or who just need to save disk space. There is no other compression tool for DOS (as of June 1994) that can match the compression performance of CRUSH, and most do not come close. ____________________________________ CRUSH and other Archivers CRUSH is used from the command-line in much the same way as archivers such as PKZIP and ARJ. If you are already familiar with these, then you should be able to adapt to CRUSH with very little difficulty. If you are a ZOO, LHA or HA user, then you will observe some changes, but will find the basic operation of CRUSH very similar. ____________________________________ Simple use of CRUSH To create a CRUSH archive called BACKUP containing all the files in the current directory requires the following simple command: CRUSH BACKUP This will create the file BACKUP.ZIP (PKZIP must be available). The user can be more selective by adding filespecs, as follows: CRUSH BACKUP *.DOC README MAN*.* This will only save files with the specified filespecs. The user may also want to receive some feedback about how much space is being saved. The "-c" (compare) option will allow a comparison to be made: CRUSH -c BACKUP *.DOC README MAN*.* Once created, a CRUSH archive can be uncompressed using UNCRUSH in much the same way as PKUNZIP is used. e.g. UNCRUSH BACKUP *.DOC. This only extracts files matching *.DOC. The user is advised to read the "Worked Examples" section for more details of CRUSH usage, and also the README file. Simply type "CRUSH" to obtain on-line help. _________________________________________________________________________ The CRUSH program Page 2 This is the program that will replace the use of PKZIP, LHA, ARJ, ZOO, HA and UC. Whichever of these you already use will still be needed, but CRUSH will run it for you. The command-line syntax of CRUSH is, for most cases, identical to PKZIP usage. CRUSH preserves dates, times and attributes. USAGE: CRUSH [opts] archive [pathspecs] [@lists] [>outfile] Where: [opts] These are options that control CRUSH's behaviour. e.g. -r causes CRUSH to search subdirectories. These can be grouped together for brevity or specified one-by-one, e.g. -rca is equivalent to: -r -c -a The user can choose to use "/" in place of "-" for preference. e.g. -r and /r are equivalent. archive This is the name of the archive to be created e.g. BACKUP or BACKUP.001 etc. If no extension is specified, then the default extension for the archiver used will be adopted (default PKZIP). If the special name "[]" is given (e.g. CRUSH []), then CRUSH will generate a unique archive name based on the current date and time (assuming that archives are not created within 2 seconds of each other). [pathspecs] These are the files to be compressed (defaulting to all files in the current directory). e.g. CRUSH SAVE \WP\*.DOC *.TXT This will compress into SAVE.ZIP all files matching *.DOC from the directory \WP and all files matching *.TXT from the current directory. The path \WP\ will also be stored in the archive, to be re-created by UNCRUSH. UNCRUSH can optionally ignore stored pathnames. CRUSH uses proper rather than DOS wildcards. e.g. the DOS wildcard *FRED.* will match a file called JOE because *FRED.* and *.* are treated by DOS as equivalent. CRUSH interprets "*" as any sequence of characters, including zero length. See also the "-x" EXCLUSION filespecs. [@lists] CRUSH will accept response files containing lists of files to be processed. These should contain one pathspec per line and can contain comments starting with ";" e.g. *.EXE \WP\*.DOC ; save my docs ; Fred's stuff here onwards \BIN\*.EXE Unlike PKZIP, the response file can be of any length (PKZIP limits the user to 500 filespecs). [>outfile] CRUSH can generate a response file by re-directing the output to a file for later use e.g. CRUSH DOCS >SAVE __________________________________________________________________________ CRUSH Options Page 3 CRUSH will default to using PKZIP unless -j, -o, -l, -h or -u is specified. PKZIP is the world's best known archiver, written by Paul Katz and marketed as Shareware by PKWARE. CRUSH will work with any version of PKZIP, but the user is recommended to use PKZIP v2.04 or later, which achieves significantly better compression than v1.x. ____________________________________ -j Use ARJ This instructs CRUSH to use Robert K Jung's "ARJ" program instead of PKZIP. This can be specified as a general default using the CRUSH environment variable e.g. SET CRUSH=-j ____________________________________ -o Use ZOO This will use the ZOO program, creating ZOO files. ____________________________________ -l Use LHA This will use Haruyasu Yoshizaki's "LHA" program, creating LZH files. ____________________________________ -h Use HA This will use Harri Hirvola's "HA" program, creating HA files. HA produces very small archives, but is quite slow. ____________________________________ -u Use UC2 This uses Ad Infinitum Programs's "UC" program. This uses a great of DOS memory and will probably require the added program US.EXE to allow CRUSH access to the 80k it needs. e.g. US CRUSH -u SAVE ____________________________________ -U Create CRU file only This will only create a CRU file rather than go on to compress using PKZIP or similar. This is of benefit to Stacker, DoubleSpace and SuperStore users which will result the the best possible compression. ____________________________________ -m Moves file to archive This deletes the files after they have been successfully compressed. ____________________________________ -c Compare with and without CRUSH This allows CRUSH to show how much filespace has been saved over the archiver alone. This will make CRUSH slower as an additional compression is needed in order to make a comparison (unless -e is also used). The results are displayed as a bar graph, plotting archive and CRUSH sizes. The original file size can also be plotted by adding the option -i2. __________________________________________________________________________ CRUSH Options Page 4 ____________________________________ -f Optimise until 5 failed attempts to improve compression This will make CRUSH experiment with different orderings to improve the compression. CRUSH will end when it has failed 5 times to find an improvement. This will increase the run time of CRUSH by at least 5 times. Each time it provides an improvement, the time taken increases. CRUSH is unlikely to take more than 15 times the normal compress time. The results of this trial are displayed as two bar graphs, showing the size of the first and best CRUSH compression. (See option "-s"). ____________________________________ -f<n> Optimise until "n" failed attempts to improve compression As "-f", but allowing the user to specify how long CRUSH is to continue the trial before giving up e.g. CRUSH -f2 BACKUP will give up after only 2 failures. CRUSH -f100 BACKUP will give up after 100 failures. ____________________________________ -t Only create CRUSH archive if 5% improvement This creates a CRUSH archive only if it can improve the archive size by 5%, otherwise it creates a conventional archive. It invokes the -c option, which adds to the run time, as it needs to run the archiver without CRUSH for comparison. ____________________________________ -t<n> Only create CRUSH archive if "n"% improvement As "-t", but allowing the user to set the threshold e.g. CRUSH -t10 BACKUP CRUSH only if 10% better ____________________________________ -r Recurse subdirectories This will search subdirectories from the specified directories, also storing paths in the archive. Unlike PKZIP, CRUSH defaults to storing pathnames with the files. However this is done in a very efficient manner and so does not significantly increase the archive size (PKZIP uses a space-consuming format for this e.g. searching a subdirectory DOCUMENT\MYDOCS with 100 files would require an extra 1660 bytes, whereas CRUSH would require just 16). Example: CRUSH -r BACKUP \WP\DOCS\*.WP *.TXT This will search \WP\DOCS and all its subdirectories for the filespec *.WP, and the current directory and its subdirectories for *.TXT. CRUSH will store the full path e.g. \WP\DOCS\HOME\DOC.WP etc. See also the CRUSH option "-n" and UNCRUSH options "-i" and "-I" for ignoring paths and drives. __________________________________________________________________________ CRUSH Options Page 5 ____________________________________ -: Set range(s) of dates DATE OPTIONS. These are used to specify a date range for file matching. If a single date is specified then this will be the earliest date to be included. If a second date is specified, then this is taken as the as the end date. Dates can be mixed in "absolute" and "relative" format. The full syntax is: [!]:[=]<date> Where: ! inverts the meaning of the range set. e.g. "range 1st-2nd" becomes "outside 1st-2nd" = sets both the start and end date to the date following. and <date> can be: Example: DD/MM/YY absolute UK date 21/2/94 MM/DD/YY absolute US date (CRUSH determines which) 2/21/94 YY/MM/DD absolute JAPAN date ( " ) 94/2/21 DDMonYY absolute date in alternative format 21feb94 DDMon absolute date (current year) 21feb -DD number of days before current date, OR -3 m-MM number of months OR m-2 y-YY number of years (truncates to 83 years) y-1 .name where date is taken from file/dir "name" .LASTBAC The four date formats can be mixed, e.g. -:y-2:-5 (or /:y-2:-5) specifies files created/modified between 2 years and 5 days ago. -:y-99:5jan will select all files created on or before 5th Jan (current year). -!:=-1 selects all files except yesterday's. The ".name" option allows compression of files relative to the date and time of an existing file. The latter feature allows the user to issue commands such as "compress all files modified/created since last backup", using commands such as: CRUSH -:.BACKTAG BACKUP REM>BACKTAG The REM>BACKTAG will create an empty file with the current date and time, so that the next backup will start with the date and time of that file. General Examples: (all use the archive name "BACK") CRUSH -:11feb BACK Compressing files modified/created on or after February 11th of this year. CRUSH -:-3 BACK Files modified/created today or during the 3 previous days. CRUSH -:-0 BACK As above, but today only. CRUSH -!:y-3 BACK Not files created during the last 3 years. CRUSH -:m-4:m-3 BACK Will compress files modified/created between 4 and 3 months ago. __________________________________________________________________________ CRUSH Options Page 6 ____________________________________ -x Exclude pathspecs In programs such as PKZIP and ARJ, the exclusion pathspecs can appear in any position in the command-line. With CRUSH they MUST appear AFTER the archive name with the other pathspecs. CRUSH limits the user to 300 exclusion filespecs. Examples: (pathspecs are all case-insensitive) CRUSH BACKUP *.EXE -xTEST*.* *.COM This will compress all files matching *.EXE and *.COM, but not those matching TEST*.*. The order of these filespecs is irrelevant. CRUSH -r BACKUP -xDOC\*.BAK -x*.TMP This will recurse, searching for files to compress, but excluding files that match the path and name DOC\*.BAK, or the name *.TMP in any subdirectory. CRUSH -r BACKUP -x@NOTTHIS -x@ORTHIS This will recurse, searching for file to compress, but excluding any pathspecs in the response files NOTTHIS or ORTHIS. ____________________________________ -y Answer YES to all questions This will allow CRUSH to compress without pausing for queries, such as permission to overwrite an existing file. ____________________________________ -z Add CRUSH archive comment This will prompt the user for an archive comment to add to the CRUSH file. Ths will be displayed by UNCRUSH when uncrushing or verifying an archive. The user can automatically provide a CRUSH comment without the program pausing for input by using the DOS echo command. e.g. ECHO My Comment|CRUSH -z BACKUP which will give the archive the comment "My Comment". ____________________________________ -v Show files for adding without CRUSHing This will show which files CRUSH would compress without actually compressing, with a summary at the end. If the output is re-directed to a file, then the re-directed file will hold a simple list of files (equivalent to PKZIP option "-@") e.g. CRUSH -v BACK > LIST Creating a list of matching files that would be archived. The archive name specified will be ignored. __________________________________________________________________________ CRUSH Options Page 7 ____________________________________ -Ln Set limit for files per archive This will allow the user to limit the number of files per archive. By default CRUSH will limit each file to 1000 files, after which it will create multiple archives. The user can force CRUSH to set a lower limit. This has the advantage of reducing the temporary space required when compressing many files. It also creates smaller archives that can be more easily copied to floppy disk. ____________________________________ -n Store files without paths This allows the user to prevent pathnames from being stored e.g. CRUSH -n BACKUP MYDOC\*.DOC would store files matching *.DOC, but without the path MYDOC\. ____________________________________ -Q Query each file to be archived This allows the user to choose on-line which files are to be included in the archive e.g. CRUSH -Qr DOC *.WP will search recursively (-r) for all files matching *.WP and query whether each is to be added to the archive. The user has the option to accept or reject all the remaining files at any point, or to abort the compression. This option provides an easy way to construct a response file, e.g. the following expanded example: CRUSH -Qr DOC *.WP *.TXT > SAVE would allow the user to select files for compression while also creating a response file SAVE for future use, e.g. CRUSH DOC @SAVE. ____________________________________ -q QUIET mode, no output This will run CRUSH without any display output, unless CRUSH needs to prompt the user with a question (inhibited by -y). This is useful where CRUSH is called within BAT scripts or another program. ____________________________________ -a Create an additional index The price paid for creating CRUSH archives is that, although programs such as PocketD Plus can still browse the contents of the archives, searching for files inside CRUSH archives would be prohibitively slow if each CRUSH archive had to be temporarily decompressed whilst scanning. The -a option will force CRUSH to create an additional external index with the extension ".CRI". PocketD Plus will use this for searching (and also browsing) to allow files to be searched or selected very quickly. This will have an associated disk space overhead, but this would not be significant where the CRUSH saving is already substantial (i.e. more than 10%) __________________________________________________________________________ CRUSH Options Page 8 ____________________________________ -e Use the existing archive for comparison This will allow the user to test the compression improvement without re-creating the original archive for comparison. CRUSH will still need the files for compression to be available in uncompressed form. The option -c must also be used. Make sure that the archive to be compared is not already a CRUSH archive! ____________________________________ -C Colour off This will use whatever colour is currently in use rather than forcing white on blue. This normally means white on black. ____________________________________ -C<n> Set colour This will allow the user to override the colour of the bar graph displayed during compression. "n" is the normal colour attribute in the range 0 to 7. ____________________________________ -s Minimise temporary disk space (slow!) CRUSH will normally use about 25% extra temporary disk space when running with the -f optimise option. If you do not have sufficient space to run CRUSH with this option, then "-s" will reduce the space CRUSH uses, but at the cost of a longer compression time. This option is not recommended for routine use unless you are normally always short of disk space. CRUSH does not make use of extended or expanded memory. ____________________________________ -i<n> Control information type By default, CRUSH will display a bar graph. If -i1 is specified, a more comprehensive table will be shown instead, giving more information. If -i2 is specified, CRUSH will still show a bar graph, but when -c is selected it will simultaneously plot the file size, archive size and CRUSH size, rather than just archive and CRUSH. This will allow all three sizes to be compared, but will give limited resolution of the difference between just the archive and CRUSH. Note: CRUSH plots to a resolution of 1 in 114 (using the special half width character "221") ____________________________________ -2 Set display of times to 24 hour clock This will force the display of 24 hour clock times instead of 12 hour during browsing. __________________________________________________________________________ The UNCRUSH Program Page 9 The UNCRUSH program is used for uncompressing or viewing CRUSH archives. It requires access to the corresponding uncompressing programs used by the base archiver e.g. PKUNZIP for PKZIP files and ARJ for ARJ files. Unlike CRUSH, UNCRUSH is copyrighted FREEWARE. You may use and distribute freely providing that its intended use is not mis-represented and it is not modified in any way. USAGE: UNCRUSH [opts] archive [destination] [pathspecs] [>outfile] Where: [opts] These are options that control UNCRUSH's behaviour. e.g. -v causes CRUSH to display the names of the compressed files without uncompressing them. These can be grouped together for brevity or specified one-by-one, e.g. -ioq is equivalent to: -i -o -q The user can choose to use "/" in place of "-" for preference e.g. -i and /i are equivalent. archive This is the name of the archive to be uncompressed e.g. BACKUP or BACKUP.001 etc. If no extension is specified, then UNCRUSH will look for known extensions that use the specified name in the order CRU, CRI, ZIP, ARJ, LZH, ZOO, HA then UC2. UNCRUSH cannot uncompress files with non-standard extensions. Such files will need to be re-named before using UNCRUSH, or uncompressed first using PKUNZIP etc. [destination] This is the destination path where files are to be uncompressed. This must be terminated by ":" or "\" to allow it to be distinguished from normal pathspecs (same operation as PKUNZIP). Paths within the archive will be combined with any specified destination path, e.g. CRUSH BACKUP \WP\*.DOC followed by: UNCRUSH BACKUP MYDOC\HOME\ will uncompress to MYDOC\HOME\WP\*.DOC [pathspecs] These are the files to be uncompressed (defaulting to all files). These can include paths e.g. UNCRUSH SAVE \WP\*.DOC *.TXT This will uncompress pathspecs from SAVE matching \WP\*.DOC and any pathspecs with the filespec *.TXT. i.e. *.TXT will match *.TXT and \WP\*.TXT, but \WP\*.DOC would not match *.DOC. UNCRUSH can be set to optionally ignore stored pathnames. As with CRUSH, UNCRUSH uses proper rather than DOS wildcards (See CRUSH for details). [>outfile] UNCRUSH will allow the user to create response or summary files using re-directed output. See -v, -V and -l. __________________________________________________________________________ UNCRUSH Options: Page 10 ____________________________________ -v Display without uncompressing This will cause UNCRUSH to report the pathspecs inside the archive without uncompressing. This gives a sorted output reporting size, date and time of the file stored. Attributes are not displayed, even though stored. ____________________________________ -V Display archive comment only This will simply display the archive comment and the name of the archive. This is useful for cataloging purposes, allowing this information to be appended to a reference file e.g. UNCRUSH -V FRED >> REFFILE ____________________________________ -o Overwrite files without query This will force UNCRUSH to overwrite files regardless of whether the file already exists. UNCRUSH would normally inform the user if files already exist with an option to skip or overwrite the file. UNCRUSH helpfully also informs the user whether the file to be overwritten is NEWER, OLDER, IDENTICAL or SAME DATE, DIFFERENT SIZE. ____________________________________ -u Uncompress NEWER or NEW files This will only uncompress NEW or NEWER files from within the archive. This will still prompt for overwrites, unless -o is also specified. ____________________________________ -a Uncompress NEW files This will only uncompress NEW files. ____________________________________ -i Ignore paths in CRUSH file This will ignore all paths inside the archive, therefore possibly generating an overwrite warning if WP\FRED.DOC and \DOC2\FRED.DOC exist in the same CRUSH file. ____________________________________ -I Only ignore drives in paths As "-i", but only ignoring drive names stored in the CRUSH file. This accommodates situations where a CRUSH command such as: CRUSH BACKUP S:\WP\*.DOC is run, storing a network drive letter with the path. A user on a non-networked PC would then be unable to uncompress the CRUSH file without losing the path. The -I option allows this to be overridden. __________________________________________________________________________ UNCRUSH Options Page 11 ____________________________________ -q Quiet mode: No output at all This will inhibit all output from UNCRUSH, except where UNCRUSH queries an overwrite. ____________________________________ -z Only suppress output of archiver This will only suppress the output of the unarchiver run by UNCRUSH. e.g. UNCRUSH may call PKUNZIP which normally displays information while uncompressing. This option will inhibit this. This may not always be safe if the unarchiver has an unusual problem, in which case the error message the archiver displays will not be visible. ____________________________________ -e Extract even if not CRUSH archive This is convenient if the user wants to use UNCRUSH to process any archives, whether CRUSH or not. This will support any pathspecs given, but will not support specified destination directories. ____________________________________ -l Only display a simple file list This will replace the normal display by a simple list of files and pathnames. This is most usefully used with -v to generate a response file. ____________________________________ -2 Display times as 24-hour clock This is only a cosmetic detail, but many users are firmly committed to 12 or 24 hour clocks, in which case UNCRUSH will allow the default 12 hour clock file times to be replaced by a 24 hour clock times. __________________________________________________________________________ Worked Examples Page 12 There are many routine archiving tasks that CRUSH could be used for. The following selects a few that either take particular advantage of CRUSH facilities, or that may be particularly relevant when first using CRUSH. For simple examples see "Getting Started" at the beginning of this manual. ____________________________________ Generating a Response File for Future Backups Most archivers allow the user to use a "response" file containing file and path specifications to be used in addition to (or instead of) CRUSH command-line parameters. This is convenient when a list is very large. CRUSH allows the user to choose files for archiving on-line using the -Q option. e.g. CRUSH -rQv TEST *.C *.ASM >RESPLIST This will search all subdirectories (-r), querying each file found (-Q), but without actually creating an archive (-v). The output is directed to the file RESPLIST. The user can then use this for future CRUSH archiving in a command of the form: CRUSH BACKUP @RESPLIST ____________________________________ Backing up to multiple disks Because of the way that CRUSH works, the incremental creation of a multi-disk archive is not practical. However you can store a large CRUSH archive on multiple disks by using the facilities of your archiver, e.g. PKZIP -& -e0 A:MULTI CRUSHED.ZIP This will create a multi-part archive. ____________________________________ Performing Archival Backups CRUSH's option to backup files dated on or after a named file (not supported by PKZIP etc.) makes it possible to create backups of all files modified since the previous backup. This can be combined with CRUSH's ability to automatically generate a unique filename, as follows: BACK.BAT containing: CRUSH -r:.LASTBAC [] *.DOC *.C *.H IF ERRORLEVEL 1 ECHO Error occurred! IF ERRORLEVEL 1 GOTO EXIT REM>LASTBAC :EXIT The "[]" parameter replaces the normal archive name by a unique name generated using the current date and time (e.g. 1CD29E33.ZIP, which uses the hexadecimal value of the combined 32-bit date and time). CRUSH will then search all subdirectories for all files matching *.DOC *.C or *.H modified since the date and time of the file LASTBAC and add them to the archive. If the archive step fails (e.g. out of memory etc.), the error is reported and the command sequence ended. If successful, then "REM>LASTBAC" will create an empty file with the current date and time for reference by the next backup. On the next occasion that the command "BACK" is run, CRUSH will only search for files modified since the last backup. __________________________________________________________________________ Worked Examples Page 13 ____________________________________ Converting all Existing Archives to CRUSH This would be a potentially mammoth task that would benefit from some automation, unless you are willing to convert each file one-by-one. A good choice for this task would be PocketD Plus v4.1 (D.EXE), as follows: MD TMP CD TMP D .ZIP/W[PKUNZIP -d $w//CRUSH -ert $w//PAUSE//D /zd]Rq This example converts ZIP files by running a command sequence of the following type for each archive found: PKUNZIP -d \C\BATS.ZIP Uncompress with subdirectories into TMP. CRUSH -ert \C\BATS.ZIP (See below). PAUSE Allow inspection of result. D /zd Delete uncompressed files/dirs from TMP. The CRUSH command compares the new CRUSH archive with the existing archive (-e), but will not replace it unless the new archive is at least 5% smaller (-t). The files originally compressed may also include paths, and so CRUSH uses -r to search all uncompressed subdirectories created by the "PKUNZIP -d" command. For those with a cautious disposition, the D.EXE command-line includes the "q" option, allowing the user to confirm or select each archive to be processed. The PAUSE command allows the user to examine the output of CRUSH before the command "D /zd" deletes the files and subdirectories temporarily uncompressed. ____________________________________ The UNDO and UNDO2 Programs The registered version of CRUSH.EXE comes with two additional programs, UNDO.EXE and UNDO2.EXE, which provide tiny alternatives to UNCRUSH.EXE for unpacking an uncompressed CRUSH archive. USAGE: UNDO crushname UNDO2 crushname UNDO.EXE is 3.8k and ignores any path information. UNDO2.EXE is 4.3k and will also re-construct embedded paths. These programs can be included within distribution archives that hold CRUSH archives, providing the smallest package that will allow UNCRUSHing without additional programs. Ideally UNDO.EXE and a CRU archive could be embedded inside an SFX archive, e.g. using LHA: CRUSH -l DEMO ! *.DOC LHA A DEMO UNDO.EXE LHA S DEMO The "!" parameter above is the file "!" which contains a message displayed when the archive DEMO.EXE is run. This can contain details of how to uncompress the CRUSH archive, e.g. "Unpack the CRUSH archive by typing UNDO DEMO.CRU after the DEMO.EXE executable has finished decompression." The user may choose to rename UNDO2.EXE to UNDO.EXE or some other convenient name. Note that distribution of CRUSH archives requires a distribution licence (see REGISTER.DOC). __________________________________________________________________________ Configuration Page 14 __________________________________ Configuration of CRUSH and UNCRUSH CRUSH and UNCRUSH can be configured using environment variables, as follows: The CRUSH variable: This is particularly important if you decide not to use the default PKZIP used by CRUSH e.g. SET CRUSH=-j will force CRUSH to always use ARJ by default. Other options can also be added e.g. SET CRUSH=-jCc will also force the normal display colour and always force a comparison between ARJ and CRUSH when creating a CRUSH file. The UNCRUSH variable: This is less critical, but an example usage might be: SET UNCRUSH=-2e which would use the 24 hour clock and also allow UNCRUSH to be used for uncompressing ordinary PKZIP, ARJ, LHA, ZOO, HA and UC archives. Using CRUSH with other archivers: The user can allow CRUSH to use other archivers by configuring using environment variables, as follows: SET CRADD=pkzip -ex %s %s %s} nul SET CRMOV=pkzip -mex %s %s} nul SET CREXT=.ZIP These examples show how PKZIP would be added to CRUSH, if it were not already supported. This overrides the default archiver. The "}" symbol is replaced by ">" when running. The user can use this facility to change the default options used by the archiver they wish to use, e.g. to turn off full compression etc. UNCRUSH is configured in much the same way, as follows: SET CRUN=pkunzip %s Environment variables should be set in your AUTOEXEC.BAT, otherwise they will be lost next time the system is re-booted. __________________________________________________________________________ Technical Details Page 15 This information is only of interest to 3rd party developers who may want to provide support for CRUSH archives. __________________________________ Structure of CRUSH file (.CRU) Address Length in Bytes Description 000000 14 CRUSH v1.8 00000E 2 Chksum 000010 2 Number of paths (only 0-255 supported) 000012 2 Number of files 000014 2 Reserved 000016 4 Totalsize of files + header length(26) 00001A "totalsize" Data ; directory held at offset "Totalsize of files + header length(26)" 000000 1 Path number 000001 1 Attribute 000002 2 Time 000004 2 Date 000006 4 Size 00000A 13 Name 000017 1 Reserved (above repeated "Number of files" times) ; "n" Paths path1 Each terminated by NULL path2 .. ; Archive comment comment Terminated by NULL __________________________________ Structure of CRUSH fast Index (.CRI) As CRU but with Totalsize field set to 0 + Headerlen(26) 000016 4 Header length(26)