The Business Master (3rd Edition)

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

/ The Business Master (3rd Edition) / The Business Master (3rd Edition).iso / files / utilfile / dos4pros / xmanual.exe / XMANUAL.DOC < prev

Wrap

Text File | 1992-05-11 | 62.4 KB | 1,824 lines

(manual.doc) Welcome to DOS Extensions for Professionals This is the Shareware Version of "DOS Extensions for Professionals", a Unified Package of file utilities bringing to the PC the power, functionality, and increased personal productivity usually associated only with minicomputer-based software. Designed to meet the processing needs of the corporate community as well as of demanding home users, these utilities feature designed-in robustness, flexibility, and generality leading to predictable results in the most trying and unusual situations. These programs are not crippled or hamstrung in any way. They are, in fact, identical to those offered in the full retail version. Use these programs and test them for THREE WEEKS to see if they meet your daily needs. If you don't like these programs - if they don't help you solve your problems, delete them from your disk. You are under no further obligation. If, however, you find these utilites useful and you want to make them a permanent part of your collection, you must register to obtain a full license for continued use. The "DOS Extensions for Professionals" registration fee is $19. For that fee you recieve a printed manual, a full-use license, a new disk insuring you have the latest version of "DOS Extensions for Professionals", and mail product support as needed. To register, fill out the order form found in the file "REGISTER.DOC" (copy register.doc prn), and send it with $19 plus $5 ($8 overseas) shipping and handling to Allison Software 166 Shady Lane Apollo, PA 15613 USA The latest version with printed manual and license will be shipped to you immediately. Thank you for helping make shareware work. John Allison DOS Extensions for Professionals Programs: Diff .................................................. 2 Notes binary output & synchronization Dump .................................................. 4 Notes viewing deleted directories ............ 7 boot tables ............................ 8 the boot sector the media descriptor byte directory format disk tutor ............................. 9 the boot sector the file allocation table directories the root directory Search ............................................... 14 Notes binary Split ................................................ 16 Sub .................................................. 18 Notes ................................... 21 file renaming convention order of execution Xdel ................................................. 23 Notes ................................... 24 XDEL and XDIR contrasted secure deletes Xdir ................................................. 26 Xhelp ................................................ 29 General Topics: Command Modifiers .................................... 30 Wildcards & Recursive Notation ....................... 30 Implied Wildcards .................................... 31 Examples of Notation ................................. 31 Bug Reporting ........................................ 32 Legalities ........................................... 33 2 DIFF Displays the differences between two similar text files line-by-line. Whether you've made minor or massive changes, DIFF displays only the differences while resyncing on identical text. usage: diff file_1 [file_2] [/Comand_Modifiers...] If no file_2 is specified, file_2 takes the name of file_1 but with the extension ".BAK". Wildcards and recursive notation are not supported by the diff command. Command Modifiers: /case /numbers /notabs /pause EXAMPLES: Compare the files RESUME.TXT and RESUME.BAK ignoring any differences due to changes in capitalization. DIFF resume.txt /nocase Note that because no second file was specified in the comparison, RESUME.BAK was assumed. Compare two files outputing any differences without including their respective line numbers. DIFF source.c source.old /nonumbers The second version of a file was edited with a text processor which replaces blanks with tabs. Because blanks and tabs are not the same character, the lines in which they fall would normally be displayed as different. In this case, flag only noticeable differences between the two versions of the file. DIFF smith.1 smith.2 /notabs NOTES: Binary - Diff is designed to work with text files only. Using it to compare binary files will give hard-to-interpret and unpredictable results. Comparing identical binary files will show them to be identical, but slightly unidentical binary files may also test okay. String comparison functions which don't scan past the first null in any given line are the immediate cause for this behavior. 3 The real problem, however, is how to report inequities between binary files and how to resynchronize after differences are found. The line, the normal unit of resynchronization, has no meaning in binary files. Resync could be done on a byte-by-byte basis, but the value of the whole effort seems questionable. Output - The diff command displays the differences between two text files on a line-by-line basis. Those lines which actually differ are displayed, first those from "file A", followed by those from "file B", a section at-a-time, until the entire files are compared. Diff does its best to resyncronize the two files, so if one has had lines altered, added, or deleted, diff will recognize when these sections of differences have past and will consider the files to be matching again. The last lines to be output for the two files are the first ones which match. Three matching lines are actually required for complete resyncronization. You may find in an unmatching section one or two identical lines surrounded by different lines. Both files are read and stored internally within an unmatched section. The rescanning effort increases as the square of the number of unmatched lines. If you compare two completely dissimilar files, the process may run out of memory before the end of file markers are detected. At the conclusion of a comparison, a summary is issued giving the total number of lines in each file, the total number of differing sections, and the total number of lines which differ. 4 DUMP Displays the contents of binary files, directories, disk sectors (including an interpreted boot sector), disk FAT's, and cluster chains. Addresses the internals of any file by sector or byte. Searches the FAT for bad clusters, shows what's really in your directories, and snoops through deleted files. Explains the details of the DOS file structure with the on-line -DISK_TUTOR and -BOOT_TABLE help sections. usage: dump [input_file] [/Comand_Modifiers...] Input File: The name of any disk, file, or directory to be viewed. Wildcards and recursive notation are not supported by the dump command. Command Modifiers: /boot /byte /cluster /contents /disk /extended /fat /first /interpret /last /numbering /pause /sector See also -BOOT_TABLE and -DISK_TUTOR available from within dump. EXAMPLES: Prepare output for a printer. DUMP test.1 /noextended /nopause > x.lis /Noextended suppresses control characters which might upset the printer. /Nopause should be used anytime the output is redirected to a file or to a printer to avoid extra characters in the output and to avoid having to strike a key after every page of data - the redirected process will hang for no apparent reason. View the entire contents of the binary file PRINT.COM. DUMP print.com If you wanted to view just the first 5 sectors, you can enter DUMP print.com /sector=1,5 or DUMP print.com /first=5 The range doesn't have to start with the first sector. It can start anywhere and can even be anchored to the end of the file. DUMP print.com /sector=-5,-1 5 or DUMP print.com /last=5 Of course /first or /last by itself means /first=1 or /last=1. Suppose you want to view the 1000th byte of a file. There are 512 bytes/sector, so the 1000th byte must fall close to the end of the second sector. But it is simpler to say DUMP print.com /byte=1000 and the line containing byte number 1000 is output. Remember, unlike sectors which are numbered in files (but not on the disk) from 1, the first byte is byte 0. To see the 1000th byte, you should actually ask for /byte=999. Again, a range of bytes is permitted. You could specify /bytes=6500,6600 , or you could anchor the search from the last byte of the file (-1) as in DUMP print.com /byte=-300,-200 to see the second-to-the-last 100 bytes. Of course the keywords /first and /last are also operative here as in DUMP print .com /byte /first=600 So far the output has been displayed in base 16 - hexadecimal. You can just as easily display it as good old- fashioned decimal or as base 8 (octal). DUMP print.com /contents=decimal or DUMP print.com /contents=octal Similar notation applies to numbering. Byte numbering defaults to hex, but you can force it to decimal or to octal. DUMP print.com /numbering=decimal or DUMP print.com /numbering=octal Closely related to files are directory and subdirectories. Technically these entities aren't files, but you can view them through a special directory-interpreted dump. DUMP . and DUMP \ dump the current and root directory. The latter especially may hold interesting hidden files as well as a volume id. In addition to viewing the contents of files and directories, DUMP 6 enables you to address disk sectors directly by using the /disk modifier. The first disk sector is 0. (In files, DUMP considers the first file sector to be 1.) Dump disk sectors 5 through 10 of the current disk. DUMP /disk=5,10 or DUMP /disk /sector=5,10 You can also address the last sector of the disk in different ways. DUMP /disk=-1 or DUMP /disk /last The first sector of a disk is special. It is the boot sector. DUMP /disk /first gives the normal uninterpreted dump. To get an interpreted dump of the boot sector, enter DUMP /boot Try this on different sized disks. Related to the boot sector is an area following it called the FAT (File Access Table. See the discussion under Notes (Disk Tutor).) DUMP allows you to view the FAT directly. DUMP /fat You can specify which clusters in the FAT to see by giving a range DUMP /fat=6,100 or DUMP /fat /first=30 The first cluster is 2. Each cluster holds a single value - the number of the next cluster in the chain. There are special values which are assigned meanings: free cluster, end of chain, and bad cluster. Normally you see the interpretations of these values. If you want to see the values themselves, use /nointerpret. DUMP /fat=-300,-200 /nointerpret A special kind of interpreted clusters are bad clusters set by the disk formatter to mark bad disk sectors. To view just these, enter DUMP /fat=bad The FAT is a collection of clusters - mappings of disk sectors. As 7 mentioned above, clusters can be chained. DUMP allows you to view these chains in toto or to pick up anywhere within a chain and trace to its end. DUMP print.com /cluster traces the entire chain associated with the file PRINT.COM telling you in which logical disk sectors the file resides. You can trace the cluster chains of subdirectories even though we know directories aren't files. DUMP . /cluster Chain tracing doesn't work on the root directory because it isn't in the FAT (it doesn't reside in the files area of the disk and is not extensible.) By entering DUMP /cluster=68 you pick up whatever chain, if any, includes cluster 68 tracing it to its end. NOTES: deleted directories - You can dump deleted directories in the standard directory format if you know the sectors where the directory lay and, of course, it has not yet been overwritten by new data. The command "dump \" causes the root directory of the current device to be dumped. In the upper left corner there is identifying information such as, "Sector 129 (1 of 32)". Using the /sector= Command Modifier, you can look at any range in those 32 sectors. By giving it a value greater than 32, you can look at any sector on the disk through directory-colored-glasses. Subtract one less than the root starting sector from the desired sector to calculate the proper offset. In our example if you want to view disk sector 661 as a directory, calculate the offset (661 - (129 - 1)) and enter DUMP \ /sector=533 8 The Boot Sector - Byte Size Contents Implemented 00H 3 E9 XX XX or EB XX 90 03H 8 OEM name and version 0BH 2 bytes per sector 0DH 1 sectors per allocation unit 0EH 2 reserved sectors starting at logical sector 0 10H 1 number of FAT's 11H 2 number of root directory entries 13H 2 total sectors in logical volume 15H 1 media descriptor byte 16H 2 number of sectors per FAT 18H 2 sectors per track DOS 3.0 1AH 2 number of heads 1CH 4 number of hidden sectors 2 bytes DOS 3 4 bytes DOS 4 20H 4 total sectors in logical volume DOS 4.0 & size > 32MB 24H 1 physical drive number 25H 1 reserved 26H 1 (29H) - extended boot signature 27H 4 32-bit binary volume ID 2BH 11 volume label 36H 8 reserved 3EH 193 bootstrap The Media Descriptor Byte - Code Size Description F0H 3.5 2-sided, 18 sector F8H fixed disk F9H 5.25 2-sided, 15 sector 3.5 2-sided, 9 sector FCH 5.25 1-sided, 9 sector FDH 5.25 2-sided, 9 sector 8 1-sided, single density FEH 5.25 1-sided, 8 sector 8 1-sided, single density 8 2-sided, double density FFH 5.25 2-sided, 8 sector 9 Directory Format - Byte Size Contents 00 8 file name 08 3 extension 11 1 file attribute bit 0 read only 1 hidden (excluded from normal searches) 2 system (excluded from normal searches) 3 volume label (root directory only) 4 directory (excluded from normal searches) 5 archive (set when modified) 6 reserved 7 reserved 12 10 reserved 22 2 time created or updated FEDCBA9876543210 hhhhhmmmmmmsssss h - hours (0-23) m - minutes (0-59) s - double seconds (0-29) 24 2 date created or updated FEDCBA9876543210 YYYYYYYMMMMDDDDD Y - year (0 - 199) add 1980 M - month (1-12) D - day (1-31) 26 2 starting cluster 28 4 file size (Least significant byte first.) DISK TUTOR: What follows is a brief introduction to the internals of the DOS disk structure. The smallest addressable portion of a disk is the sector. A sector is 512 bytes long. Even though a disk's capacity is measured in megabytes, the disk itself is accessed by the software in terms of logical sectors, numbered 0 through n. (Underlying logical sectors are the physical realities of heads, tracks, and sectors within each track. But these are too ugly to consider here.) The Boot Sector - The first sector on each disk, sector 0, is the boot sector. Even if the disk doesn't contain an operating system and is therefore not bootable, it contains a boot sector applied by the high level format program. 10 Two types of data are included in the boot sector. The physical characteristics of the disk (total number of sectors on the disk, sectors per cluster, etc) are found in the first part of the boot sector. You can view at this information with DUMP /boot. A bootstrap is stored in the second half of the boot sector. The bootstrap is a small program which, on the basis of information found in the first part of the sector, reaches out onto the disk, finds key operating system files (e.g. IBMBIO.COM and IBMDOS.COM), loads them into memory, and executes them thus bringing the system to life. If the required system files cannot be found on the disk, the familiar "Non-System disk or disk error" message is printed out. You can find this message near the end of sector 0 with the command DUMP /disk=0. The File Allocation Table - The File Allocation Table, affectionately known as the FAT, is the method DOS uses to keep track of which disk sectors are used and which are free. The FAT is really a big lookup table. Its starting sector and length are known from the boot sector. The entries, known as allocation units or clusters, chain from one to another throughout the FAT. Cluster 5 might contain the value 6, which chains it to the sixth cluster, which in turn might be chained to the tenth. A special code marks the end of each chain. In fact, many chains are only one link long. Each FAT chain provides the system with a map to the sectors allocated for each file. Given a cluster number, you can mathematically calculate its related sector's offset from the beginning of files area. This cluster-to-sector mapping, when combined with the cluster chaining property, provides a completely flexible sector methodology for file creation and access. Sectors can be accessed sequentially although their physical locations may be randomly distributed across the disk, and logical random access is likewise efficiently supported through the same chaining mechanism. Each cluster or allocation unit maps to one or more sectors depending on the size and type of disk. This ratio, known as cluster size, determines the minimum size of a file. Suppose you want to store one character. One character requires one byte, but the minimum addressable disk unit is the sector - 512 bytes. So whether you're storing one byte or 512, the disk requirements are the same, one sector. Suppose a disk's cluster size is 4. That means each cluster in the FAT points to a group of four sectors. Even though a one-byte file may require only one disk sector, four are effectively used (allocated) because their single common cluster is used by that first sector. This is why, in an XDIR /full listing, you may see 2048 bytes (4 sectors) allocated for even the smallest file. In view of the inherent waste in multiple sector clusters, you may well ask why larger clusters are used at all. There are several good 11 reasons. Multiple sector clusters raise disk I/O efficiency when large data transfers take place. Because all sectors of a cluster are contiguous, the underlying software can support multi-sector reads when loading executable files, for instance. Large clusters give rise to a smaller, more efficient FAT. A four sectors/cluster FAT takes only one quarter the disk and in-memory buffer space that a single sector/cluster arrangement requires. In addition, the greater-than-32MB-disks supported by DOS 4 and 5 are possible only because each cluster references more than one sector. The unsigned 16 bit cluster can address only 64k sectors - 32MB. Multiple sectors per cluster increase that top disk limit by the multiple factor. So you see, there are positive trade-offs in the sectors-per-cluster question. Depending on the size of a disk, DOS will use either 12 or 16 bit clusters. The smaller cluster size allows up to 4k entries while the larger, as previously mentioned, permits 64k. Some cluster values have special meanings. The first two entries of the FAT (0 and 1) are reserved. Cluster 2 is the first valid cluster. The special cluster values are: 16 bit cluster 12 bit cluster 0 0 - free cluster FFF0 - FFF6H FF0 - FF6H - reserved cluster FFF7H FF7H - bad cluster FFF8 - FFFFH FF8 - FFFH - last cluster in chain You can see that with the exception of the free cluster marker (0), they are all near the upper limit of the cluster address range. You should also note that if an area of the disk is bad, the formatter will mark the whole cluster as bad. Any good sectors within that cluster are unreachable by normal FAT-based I/O. This is not a significant problem in practice. Bad spots on disks tend to span several sectors and often several clusters. A bad FAT, for obvious reasons, is catastrophic. Some disks support multiple FAT's incase one becomes unreadable during use. Directories - The FAT keeps track of which sectors belong to which files and which sectors are free. Directories give names to files and impose the hierarchical file structure on the disk which, since the advent of the UNIX (TM) operating system in the early 1970's, has become standard in so many other operating systems. Directory entries are 32 bytes long, 16 to the sector. They contain the file name, extension, file attribute, reserved area, date and time of creation or last update, starting cluster, and file size. (See DUMP - BOOT_TABLES) Besides the file name, the most significant part of the directory entry 12 is the starting cluster. This points to the first cluster of a cluster chain in the FAT and to the physical location of the file on the disk. The file attribute gives special meaning to certain files. The attributes are: Hidden file System file Volume label Directory Archive (Pending) There can be only one Volume label on a disk, and it must be in the root directory. Its 11 characters are held in the file name and extension sections of the entry. Before they're first used, directory entries contain null (0) characters in the file name. As files are added, their names are entered starting at the top of the directory. Files are deleted by overwriting the first character of the name with the lower case sigma character (229). The FAT entries are free, but, as far as the directory is concerned, no other information is changed. This is why special utilities can "unerase" a file (except for recovering the first character of its name). The next time a file is added to a directory, it overwrites the first record it finds marked as deleted, or in the absence of any of these, it uses the first null entry. When a directory is searched, the search stops at the first null entry. There is no more data to be found. Thus a directory which once held many files takes noticeably longer to search than a nearly empty one, even after most of those files are deleted. Searches must still continue to the old high water mark. The Root Directory - There are two kinds of directories in DOS: the root directory and subdirectories. The root (top level) directory has its roots in the first version of DOS in which it was the only directory. The root directory is not allocated via the FAT, but is of a fixed length whose size is dependent upon the disk type. It is located after the FAT(s) and before the files area, where the FAT mapping begins. Its length and location is one of those constants specified in the boot sector. (Try DUMP /boot.) Once you fill up the root directory, it's filled up. Subdirectories, on the other hand, are mapped by the FAT, are extensible, and can reside anywhere. They never can be filled up although too-large subdirectories are awkward and slow. Subdirectories are distinguished from root directories in that the former have entries to themselves and to their parent directories, the familiar "." and ".." files. 13 The distinction between the root directory and subdirectories is hidden from both the user and from the programmer who uses BIOS calls. Only those foolish enough to grovel in the lower reaches of DOS ever need be concerned with those subtle differences. 14 SEARCH Quickly and silently locates text strings in any file anywhere on your whole disk. The simplicity of this favorite program belies its power and usefulness. usage: Search file_spec "text string" [/Comand_Modifiers...] file_spec: A wildcarded recursive file specification. "Text string" may be quoted or unquoted depending on whether it contains special characters which may affect command execution (eg: blanks or '/'). Command Modifiers: /before /case /numbers /pause /since EXAMPLES: Flag any C language source files in the program subdirectories on the current drive which call the binary_search() function. SEARCH \user\pgm...*.c binary_search( Normally searches are made independent of the case of either the search string or of the text in the files being searched. This case insensitive default would yield a match on a file containing the text "ABCDEF" against the requested string "AbCdEf". Make a case-sensitive search of all files in the letters subdirectory for the name Smith. This search would not find "blacksmith" because of case-sensitivity. SEARCH C:\admin\letters Smith /case Search all files in the current directory for the phrase "Wall Street", outputing the line number as well as the entire line in which the phrase occurs. SEARCH * "Wall Street" /numbers Note the use of quotes to include blanks and tabs in a search. Without the quotes, * Wall Street would have appeared to the program as three arguments instead of the require two. Locate all letters referencing airline travel in the current directory written between June 1st and July 15th. SEARCH *.ltr air /since=1-June /before=15-July Flag all references to Pittsburgh in your manuscript, output the line in which the reference occurs with the two following lines, and 15 redirect the output to a file. SEARCH chaptr*.doc Pittsburgh > pgh.ref /context=3 /nopause Note that /nopause should be used when output is redirected to a file or to the printer. Demonstrate a search of binary files in a directory holding system programs. SEARCH C:\DOS Copyright NOTES: Binary - The search command can scan both binary as well as ASCII text files while looking for a text string entered as a command line argument. The interpretation of exactly what consitutes a line is different in the two cases. A line in a text file is denoted by a carriage return/line feed sequence (newline character). When the text string is found, the entire line containing that string is displayed. In a binary file, the carriage return/line feed sequence has no special meaning and may be randomly distributed throughout the file. For this case, a line is defined as 512 bytes. When the text string is found, the text string and enough printable context to total approximately 70 characters is displayed. Null characters are taken as absolute delimiters before and beyond which text is not displayed and may limit the displayed text to substantially less than 70 characters. The remainder of the 512 byte line is scanned for additional matches and, if found, are displayed separately. Unlike text files, matching text is searched for across line boundaries in binary files. The search command attempts to determine whether a file is binary or text by reading the first 1024 bytes and scanning for included null characters. If found, these are interpreted as indicating a binary file. The alternate method of searching for the carriage return/line feed sequence would impose a maximum line length on text files. 16 SPLIT Breaks text files into pieces or segments by four methods. Specify the number of segments, the length of each segment, that breaks be made at certain lines, or that breaks be made only at lines containing specific search text. usage: split input_file [/Comand_Modifiers...] Input_file: A wildcarded recursive file specification. Output Files: Output files, if any, are generated using the names of the original files with extensions .001, .002, .003, etc. Command Modifiers: /after /at /before /every /into /pause EXAMPLES: Split all files ending in .BIG on the current drive into files no larger than 1000 lines each. SPLIT \...*.big /every=1000 The new split files will assume the names of their parents but will be given the extensions .001, .002, .003, etc. Split a file into five equal pieces. SPLIT work_c /into=5 The split is made by line count, not by file size. Each piece will have the same number of lines, if possible. Divide a file at lines 235, 609, and 964. SPLIT book.ms /at=235,609,964 /At makes the enumerated lines the first one in each new file. This example yields four files: source lines BOOK.001 1 - 234 BOOK.002 235 - 608 BOOK.003 609 - 963 BOOK.004 964 - end Files can also be divided base on text found in the file. If you want to split a file everywhere after the text " END" occurs in a concatenated collection of FORTRAN files, the command 17 SPLIT lib1.for /after=" END" creates a new file for each occurrence of the FORTRAN END. If only a single fine END statement is found, no split is made. Text-based splits can also be made before lines containing the target text is found. Split all files in the current directory beginning with a "T" and with extension .IGS before lines containing the string "SHEET NO." unless that string occurs only in the first line of the file. SPLIT t*.igs /before="SHEET NO." 18 SUB Substitutes characters in lines in text files. This program deletes characters, add characters, replaces tabs with spaces, and replaces spaces with any size tabs. It strips trailing blanks from lines or pads out each line to a fixed length. SUB can also limit its operations to lines containing search strings and anchor operations to that text. This utility is useful for dynamically creating complex batch and script files based on directory contents. This is one of those utilities for which you may see no need - until you need it, and then it's indispenable. usage: sub input_spec [output_file] [/Comand_Modifiers...] input_spec: A wildcarded recursive file specification. Output File: If no output_file is specified, the output_file takes the name of the input_file, and the input_file is given the extension ".BAK". Command Modifiers: /at /bak /before /copy /delete /for /from /insert /noexpand /pad /pause /silent /strip /tabs /truncate EXAMPLES: SUB commands normally are applied to each line of one or more files. There are roughly three categories of Command Modifiers which can be used with the SUB command: /pad, /strip, /tabs, and /truncate apply to whole lines; /copy, /delete, and /insert apply to specified columns within a line; and /at, /before (the same as /at) , /for, and /from indicate in which columns the action is to take place. Insure that all lines of a file are 120 character long (excluding the carriage return). SUB list.txt /pad=120 Insure that no line of a file exceeds 30 characters. SUB list.txt /truncate=30 Strip trailing blanks from those lines which contain them. SUB list.txt /strip Replace multiple spaces with tab characters where possible. 19 SUB list.txt /tab In this case, tab size is assumed to be (modulo) eight spaces. Replace all tabs with spaces in .C files edited with a 4-column-tab editor for compatibility with itself and the rest of the world. SUB *.c /notab=4 In all files ending in .C on the current drive, replace 4 column tabs with eight column tabs. SUB \...*.c /notab=4/tab=8 Now let's turn to operations requiring pointers to locate the columns in which the operations are to take place. Delete the first four columns of every line. SUB list.txt /delete=4 causes four characters to be deleted, but which four? Unless otherwise stated, there is an implied pointer to column 1, an implied /at=1 or /before=1. It is always better to state assumptions. SUB list.txt /delete=4 /at=1 To shift all text one inch to the right for printing and binding requires SUB report /insert=" " /before=1 Insertion text containing tabs or blanks must be quoted. This text, eight blanks, must therefore be quoted. Duplicate columns 10 through 13, inserting them before column 20 in each line of a file. SUB table.2 /copy=10,13 /before=20 All text after column 20 is shifted four columns to the right. Nothing is deleted or overwritten unless that action is specifically requested with /delete. You may not always know the exact columns in a file in which the operation should take place. Operations can be tied to specific text in each line with the /for modifier rather than to particular columns. In a file with the lines image1.dat imagetst.dat screen.iq 20 the command SUB x y /for=. /delete=4 creates the new file Y containing image1 imagetst screen /For=text is, in effect, a variable /at or /before. If the /for text occurs more than once in a line, you can restrict the search for that text to a specific range within the line. SUB filelist.2 /delete=4 /for=.bat /from=15,30 restricts the search for the string ".bat" to columns 15 through 30. A file containing duplicate text autoexec.bat autoexec.bat config.sys config.sys x.bat x.bat is changed to autoexec.bat autoexec config.sys config.sys x.bat x after processing. Lines missing the target text ".bat" are skipped. Illustrate the importance of the order of execution as multiple operations are applied to a file. In a file with the line "123456789abcdef", replace "567" with "abc". /Copy without /delete results in an insertion rather than a replacement. Modified by SUB x y /at=5 /delete=3 the line becomes "123489abcdef". Modifying the result by SUB y z /at=5 /copy=7,9 yields "1234abc89abcdef". These two operations to replace "567" by "abc" (columns 5 - 7 with 10 - 12) can be applied by the single command SUB x z /copy=7,9 /at=5 /delete=3 Notice that although the order in which Command Modifiers are placed on the command line is unimportant and has no effect on the order of execution, the order of execution is critical and is fixed by convention. (See Notes Order of Execution.) 21 Where ever the string "Copyright 1990" occurs in .HEL files, replace it with "Copyright 1991". SUB *.hel * /for="copyright 1990" /delete=14 /insert="Copyright 1991" SUB can be used to accomplish global changes across many files, but it is rather tricky. Since the /for command modifier is really a dynamic /at=column, you must specify /delete or "Copyright 1991" will simply be inserted before "copyright 1990", not exactly the replacement operation desired. The /for string is case insensitive. With the /insert string, however, you get the case you specify. All matching files are updated (copied) whether there are changes or not. Illustrate SUB's limited ability to rename files. In all files in the current directory beginning with a "T" and with extension .DAT, delete columns 15 - 21 of each line and store the results in similarly named files with the extension .S1. SUB t*.dat *.s1 /delete=6 /at=15 NOTES: File Renaming Conventions - There are three ways to specify the output filename[s] for the SUB command. Case 1: SUB file_1 file_1 is read and renamed file_1.BAK at the command's conclusion. file_1 is the name given to the newly created file. Case 2: SUB file_1 file_2 file_1 is read. file_2 is created. Only in Case 2 are files permitted to be in different directories or on different drives. Case 3: SUB *.ext1 *.ext2 .ext1 files are read. .ext2 files are created. 22 General wildcards are permitted in the file_1 name. A single all-inclusive wildcard is permitted in the file_2 name but not in the extension. SUB file_1 * and SUB file_1 *.* are the same as Case 1 "SUB file_1". Internal logic insures that the input and output filenames are not identical: that an output file cannot overwrite itself. An output file can, however, overwrite another file. Order of Execution - The order of application of multiple Command Modifiers to each line of the file is: /notab /delete /insert /copy /truncate /pad /strip /tab This order helps insure intended, predictable, and uniform results. 23 XDEL Recursively deletes files and removes directories operating in either the default normal mode or in a secure overwrite mode. XDEL deletes hidden, readonly, and system files or files by date. Verification is available on the command, directory, or file levels, or can be bypassed completely for batch operaions. The secure overwrite mode is flexible, allowing you to specify the number of overwirtes and whether they are with random data or with patterns of your own choosing. usage: xdel file-spec [/Command_Modifiers...] file-spec: A wildcarded recursive file specification. Command_Modifiers: /before /directories /hidden /list /modified /patterns /pause /readonly /secure /since /system /unmodified /verify EXAMPLES: If NAME is a file, delete it. If it is a directory, delete all files within it (not recursively), and then remove it if empty. XDEL name Delete the contents of the current directory but no subdirectories. XDEL *.* Delete the contents of the current directory and recursively the contents of all subdirectories. XDEL ... Delete all files in the current directory and recursively in all subdirectories below it verifying only at the directory level. XDEL ... /verify Delete all .BAK files from the C: drive. XDEL c:\...*.bak Delete all files from the A: drive leaving the directory structure intact. XDEL a:\... /nodirectories 24 Delete all files with the extension .LTR in the current directory created before yesterday appending their names to a file as they are deleted. XDEL *.ltr /before=yesterday /list=all /nopause >> deleted.log Delete the backup files on drive A: which are marked readonly. XDEL a:*.* /readonly Delete all files with the extension .LTR from the current directory overwriting them with random numbers two times, and a final time with zero's. XDEL *.ltr /secure=3/pattern=0 While requesting verification at the directory and file levels, delete all files under the \USER directory which are unmodified (already backed up). XDEL \user... /unmodified /verify=all Hint: When in doubt about the exact effect of an XDEL file spec syntax, try it first with the XDIR command which is similar in operation and a lot safer! NOTES: XDIR and XDEL Contrasted - In general, the xdel command parallels the xdir command quite closely in operation. If you want to delete a certain set of files, especially with recursive deletes, you may want to try the same syntax with the xdir command first. There are some significant differences, however. The attributes modifiers are exclusive with xdir and inclusive with xdel. The command "xdir *.* /readonly" will list ONLY readonly files. "xdel *.* /readonly" must be used to delete readonly files and will delete readonly files AS WELL AS normal files. To delete the standard system files, for example, requires a /system /hidden and /readonly combination. The xdir and xdel commands both do depth first descents on recursive operations, traveling down each successive subdirectory tree to the bottom before retracing upward to take alternate paths. The order of operation is different, however. Xdir lists each entire subdirectory as it is encountered while xdel travels the entire extent of the first branch before it begins to delete. Only when it encounters a subdirectory with no subdirectories within it does it begin file deletion. Then it backs out of the now empty subdirectory, removes it, and continues to process, trying to descend first, and then when at the bottom, deleting. The very last directory whose contents are deleted is the top level directory. The operation also takes place 25 in alphabetic order. If, horror of horrors, you discover that you're deleting the wrong tree and do a frantic <CTRL> C or hit the Big Red Switch, upon recovery you should be able to tell pretty easily where the delete stopped. (This is NOT recommeded procedure!) During a delete, if your current directory is in a path being deleted, the current directory and its parent directories are not removed. Files continue to be deleted as requested. (See xdel -help /directory for examples.) Volume labels (attribute V in the dump command) are not deletable by xdel. Secure Deletes - /Secure deletes remove all directory information for all deleted files within the delete path, independent of when those files were deleted. You can remove all directory information about all previous files on the entire disk by requesting a /secure delete on a non-existent file (e.g. xdel \...*.qqq /secure). Directory information would be removed, but there is no easy or complete way to overwrite sectors once they are deallocated. (Creating a hugh file to scavenge all free clusters and then doing a /secure delete on that file would help but would not be a complete answer. See the example below.) Demonstrate that overwriting a disk sector with a new file may not remove all the orignal data. (See /secure.) Create a small file with the sequence copy con x Hi! <Ctrl>Z and use dump to dump the disk sector used by file x: dump x /cluster dump /disk=n,m ; where n,m are the sectors used by x's cluster. If you are working with an old, much-used disk, you'll see previously deleted data which was not overwritten by the creation of x and will never have a chance of being overwritten until the little file x is first extended or itself deleted. No attempt is made during a secure delete to overwrite the FAT multiple times or with random data. It is just too dangerous, and there can be multiple FAT's. The most information anyone could get from your FAT with analysis of the once-overwritten information would be the chain of clusters comprising your securely deleted file. And that assumes he could get the starting cluster. Security ultimately rests in the secure deletion of the data itself, not of pointers to it from the FAT or from the directory entry. 26 XDIR Recursively lists files in directory trees giving subtotals and grandtotals of file sizes and disk allocations. Bottom line enthusiasts can specify totals or grandtotals only - without file detail. Directories are displayed in UPPER case and regular files in lower case for easy discrimination. The defualt sort order is alphabetic unless you request ordering by file size or date. Searches can specify hidden, readonly, system, or modified files only, and can limit the scope of a search before or after any time and date, or us the generic keywords "yesterday" and "today". usage: xdir [file-spec] [/Command_Modifiers...] file-spec: A wildcarded recursive file specification. If none is given, the current directory is assumed. Command Modifiers: /allocated /archive /attributes /before /bydate /bysize /columns /date /debug /directories /foot /full /grandtotal /heading /hidden /modified /pause /readonly /since /size /system /total EXAMPLES: List all files in the current directory and recursively in all subdirectories below it. XDIR ... Produce a full listing to reveal all that is known about all files in the current directory. XDIR /full Rather than ask for a /full listing, you can specify any combination of the components of a /full listing: /size /allocated /date /attributes Determine how much space you could save on your disk by deleting all the .BAK files from the current directory. XDIR *.bak /allocated 27 Find the total size by directory of all files on the C: drive. XDIR c:\... /total Find the total number, size, and allocation of all .DOC files on the current drive. XDIR \...*.doc /grandtotal List all files under the \USER directory modified since the last backup. XDIR \user... /modified Estimate how many diskettes will be needed for the next incremental backup. XDIR \... /modified /grandtotal /Modified gives the same results as /archive. List all the files which have been created or changed today on the current drive. XDIR \... /since=today List all the files in the DOCUMENT directory created or last updated in May. XDIR document /since=1-May /before=1-June Find the oldest program in your \DOS executable directory. XDIR \DOS /bydate The newest files sort to the top, the oldest to the bottom. List all files in the current directory sorted by size and which have been created or modification since the beginning of yesterday. XDIR /bysize /since=yesterday Check your whole disk for one or more copies of a misplaced document. XDIR \...northern.doc Check your disk for hidden or readonly files. XDIR \... /hidden /readonly Output a list of all the files on your C: disk along with their sizes for later sorting by the DOS sort utility. XDIR c:\... /size /noheading /nofoot /nodirectories /nopause>slist 28 /Nodirectories is added because you aren't interested in including directories in your list. /Nopause is necessary so that the output isn't hung up waiting for you to strike a key in response to invisible "more" messages. /Noheading and /nofoot remove separating headers and feet from between directory contents resulting in one long list. Note that only the names of the files, not the directories in which they reside will be included. Different files with the same name but residing in different directories will loose their unique identities. Find a file whose name you cannot remember and whose location you don't know. You do recall that the file was called sales-something, or was it something-sales? Use multiple wildcards and recursive notation to search the whole disk for it. XDIR \...*sales* You have a directory named FINAL containing 175 files whose names you want to use in a listing documenting a new product. Output the names in alphabetic order to a file for later editing. XDIR final /column=1 /noheading /nofoot /nopause > prod.lis Without /column=1 and in the absence of outputing any other file information (/size, /allocate, /date, or /attributes), the output would have defaulted to five columns wide, not the single-column listing format desired. Preserve an image of the directory structure of your C: drive. XDIR c:\... /directories /nopause > dir_skel /Directories causes only directories to be listed. You're not sure about the ins and outs of the syntax of the date/time modifiers of the /before /since Command Modifiers and want to experiment using /debug to test their use. XDIR /debug=date /before=6-13,9 Note that /debug=date must precede /before on the command line. The program's interpretation of your entry is displayed, but no directory listing is made. /Debug=path shows the full expansion of the file specification. 29 XHELP Supplements the help built into each command and becomes your first stop if you don't know what you're looking for. It contains a brief description of each program in "DOS Extensions for Professionals" as well as examples, notes, and discussions of features common to all commands. usage: xhelp [Command] [General Topics] [-ALL] [/nopause] Commands: diff dump makehelp search split sub xdel xdir xhelp General Topics: -BUG_REPORTING -LICENSE -MODIFIERS -WILDCARDS EXAMPLES: Type XHELP by itself to get a list of all the commands in "DOS Extensions for Professionals". XHELP If you want a high level view of what a particular command does or a list of its Command Modifiers, enter the command's name after XHELP. XHELP split Notes and examples are also included with the list of command modifiers. There are several general topics which, for your convenience, are also stored on-line in XHELP. Enter the general topic's name after XHELP XHELP -bug_reporting If you forget which general topics are treated, enter XHELP - for a listing of them all. 30 GENERAL TOPICS: Command Modifiers Command Modifiers are special command line arguments which begin with a "/" and alter the basic command's action in some way as in xdir /hidden Command Modifiers are delimited by a space or their "/" from the preceding command, command argument, or Command Modifier. The order of Command Modifiers among themselves or among other command arguments is unimportant. (The order of other command arguments is, in general, significant.) Only enough characters to unambiguously identify a Command Modifier must be given after the "/". Ambiguously matching entries are listed. A "/" by itself causes all possible Command Modifiers to be listed. A "-" by itself causes all possible Help Topics to be listed. Command Modifiers should be fully spelled out in batch files to avoid possible future ambiguities in case new Command Modifiers are ever added and for clearer documentation. Command Modifiers which require values are followed by an "=" and the value as in "/date=today". Multiple values are "," separated. Wildcards and Recursive Notation The wildcard and recursive notation supported by many of the DOS Extensions for Professionals' commands are a superset of DOS wildcards and can be applied to partial or complete DOS file specifications. C:\directory_path\filename.ext Wildcards apply exclusively to the filename portion of the specification and the recursive notation to the directory path. Supported notation: ? - matches a single character in a filename or extension. * - matches 0 or more characters in a filename or extension. ... - matches 0 or more directory names in a recursive descent down a directory tree. Note that leading and multiple wildcards are supported. 31 Implied Wildcards: Some commands support implied wildcards. Others do not. In general, those commands which involve read-only activity such as dir, xdir, and search use the broader implied wildcards. Commands such as sub, xdel, and copy, which alter files, do not support implied wildcards but enforce a stricter interpretation of the user's command. The ins and outs of implied wildcards are subtle. Many people lead full and meaningful lives never consciously learning the rules. They are meant to be natural. Implied wildcards are covered here so you won't be upset when some commands behave differently than others and so you'll understand why the details of the following examples don't hold for all commands. Notation Implied Wildcard Strict Interpretation xyz xyz.* xyz. xyz. xyz. xyz. * *.* *. C: C:*.* (XDEL) C: (DEL) The dir/xdir commands are peculiar in their approach to filenames which are directories. xdir ABC gives a listing of the CONTENTS of directory ABC. xdir A* ditto if A* is unique, otherwise it lists at the current level as in: ABC afile.dat Examples of Wildcard and Recursive Notation: ... or ...*.* Match all files in this or any subordinate directories. \...*.c or \...\*.c Match all files ending with a .c extension found in any directory in the current drive. C:\dos...*op*.??? or C:\dos\...*op*.??? or C:\dos\...\*op*.??? Match all files on the C: drive in the \DOS directory tree which contain the letters "op" and a three character extension. 32 ..\...xyz or ..\...xyz.* or ..\...\xyz.* Match all files recursively descending from one level up from the current directory with the name xyz and any or no extension. xyz. Match a file in the current directory with the name xyz and no extension. Note: The ellipsis (...) must be the final portion of the path specification. Split specifications (e.g. \abc\def\...\xyz\file.ext) are not supported. Bug Reporting: Every effort has been made to thoroughly exercise DOS Extensions for Professionals in a variety of circumstances. The fact remains, however, that no non-trivial work of software is completely bug free. If you should encounter an operational anomaly, there are several steps which you can take to make convergence on a solution as efficient and painless as possible. 1. Verify the repeatability of the occurrence that you haven't made a typo or silly mistake. 2. Review the documentation. Maybe the command is working properly, but not the way you think it should. (If an operation is, in your opinion, too difficult or counter- intuitive, suggest a better solution. 3. Programs can fail from resource exhaustion. In general, this mode of failure is not a bug. If an error message tells you that you are out of memory or disk space, you probably are. (The First Rule of Successful Computing: BELIEVE ALL ERROR MESSAGES. Surprisingly often they contain at least grain of truth, especially when interpreted from the program's myopic viewpoint.) 4. Try to discover the essence of the failure. Reproduce it in its simplest scenario. This step is the hardest part of tracing down a bug. If you can do this, we'll have a much better chance of correcting the underlying problem. 5. Suggest a work-around if possible or applicable. 6. Send a clear explanation of the bug and the work you've done to try to isolate it. Send pertinent information about your computer, its setup, and whether you've observed this failure on other machines. If the failure requires a lot of data or a complicated scenario, send a disk configured to demonstrate it. 33 7. All correspondence will be given careful and thorough consideration and, when practical, will be followed by prompt action. LEGALITIES Copyright: The collection of software described in this file and contained on this disk is the Shareware Version of "DOS Extensions for Professionals" (the Software) and is Copyright 1991 John B. Allison. Ownership: You have purchased or otherwise obtained media containing a copy of the Shareware Version of "DOS Extensions for Professionals". The Software contained on the media and all rights pertaining thereto remain the sole property of John B. Allison. By continuing to use the Software, you indicate your acceptance of the terms and conditions outlined below. Shareware License: You are permitted to use the Software for an evaluation period of THREE WEEKS. After the evaluation period, you must either remove all copies of the Software from your computer system(s) and refrain from all additional use, or register your copy of the Software with Allison Software 166 Shady Lane Apollo, PA 15613 USA (412) 727-2198 CompuServe 72600,1200 (See the file REGISTER.DOC on the distribution disk for registration instructions.) Shareware Distribution: You may give copies of the Software IN ITS UNMODIFIED ENTIRETY to friends or to other parties. You may distribute copies of the Software IN ITS UNMODIFIED ENTIRETY on a commercial basis for a fee not to exceed SIX DOLLARS (US), or you may post the Software on electronic bulletin boards if you package the Software so that it must be downloaded as a single unit IN ITS UNMODIFIED ENTIRETY. 34 ASP Member: This program is produced by a member of the Association of Shareware Professionals (ASP). ASP wants to make sure that the shareware principle works for you. If you are unable to resolve a shareware- related problem with an ASP member by contacting the member directly, ASP may be able to help. The ASP Ombudsman can help you resolve a dispute or problem with an ASP member, but does not provide technical support for members' products. Please write to the ASP Ombudsman at 545 Grover Road, Muskegon, MI 49442 or send a CompuServe message via CompuServe Mail to ASP Ombudsman 70007,3536. The OMB may be contacted by FAX by sending to the ASP FAX number: (616) 788-2765. In communi- cation with the OMB, please include a telephone number and/or FAX if available. Limited Warranty: THIS SOFTWARE IS PROVIDED "AS IS" WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESS OR IMPLIED, INCLUDING, BUT NOT LIMITED TO THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. THE ENTIRE RISK AS TO THE QUALITY AND PERFOMANCE OF THIS SOFTWARE IS YOURS. SHOULD THIS SOFTWARE PROVE DEFECTIVE, YOU (AND NOT JOHN B. ALLISON, ALLISON SOFTWARE, OR AN AGENT) ASSUME THE ENTIRE COST OF ALL NECESSARY SERVICING, REPAIR, OR CORRECTION. SOME STATES DO NOT ALLOW THE EXCLUSION OF IMPLIED WARRANTIES, SO THE ABOVE EXCLUSION MAY NOT APPLY TO YOU. THIS WARRANTY GIVES YOU SPECIFIC LEGAL RIGHTS, AND YOU MAY ALSO HAVE OTHER RIGHTS WHICH VARY FROM STATE TO STATE. Limitation of Remedies: IN NO EVENT SHALL JOHN B. ALLISON, ALLISON SOFTWARE, OR AN AGENT BE LIABLE FOR ANY DAMAGES WHATSOEVER (INCLUDING, WITHOUT LIMITATION, DAMAGES FOR LOSS OF BUSINESS PROFITS, BUSINESS INTERRUPTION, LOSS OF BUSINESS INFORMATION, OR OTHER PECUNIARY LOSS) ARISING OUT OF THE USE OF OR INABILITY TO US THIS SOFTWARE, EVEN IF JOHN B. ALLISON OR ALLISON SOFTWARE HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES. SOME STATES DO NOT ALLOW THE EXCLUSION OF LIABILITY FOR INCIDENTAL OR CONSEQUENTIAL DAMAGES, SO THE ABOVE LIMITATION OR EXCLUSION MAY NOT APPLY TO YOU. This agreement shall be governed by and construed in accordance with the laws of the Commonwealth of Pennsylvania and the applicable laws of the United States of America and shall inure to the benefit of John B. Allison dba Allison Software and any successors, administrators, heirs, and assigns. Any action or proceeding brought by either party against the other arising out of or related to this agreement shall be brought only in a Commonwealth or Federal Court of competent jurisdiction located in Allegheny County, Pennsylvania. The parties hereby consent to in personam jurisdiction of said courts.