DIX[/qualifiers] filename [/qualifiers]
This is a program to dump/display/edit records in all types of files.
-You can select records by key value and number (indexed files) or
record number (sequential files or relative files).
-You can display the record-data in raw format (like OpenVMS DUMP) or
interpreted format. For the interpreted mode you need a description.
See the help about [DIX/HELP] record format for descriptions.
-You can use views to combine multiple records from multiple files
to one "record" (See the help about "views")
-You can dump/display/edit the record data in four modes.
1. Screen oriented (using SMG) (See help about "modes screen_mode").
2. Screen oriented (using DECwindows) (See help about "modes DECW_mode").
3. Dump records to a file (See help about "modes file_mode").
4. Interactive mode with a powerful scripting language.
(See help about "modes interactive_mode").
The program supports INDEXED, RELATIVE, DIRECT ACCESS and SEQUENTIAL files.
The program can open multiple files and multiple descriptions per file.
A fieldname in a record can contain a pointer to (a key in) another file,
and you can follow that link to the other file.
Files are opened READ-ONLY, unless you specify /MODIFY or /WRITE.
Files can also be opened bypassing RMS (via the /FAST qualifier).
You may even open modules in a .TLB or .MLB library
If the first character of the command line is not a /, DIX checks if the
first parameter (so the start of the line) is (part of) a valid DIX verb.
If so line is executed as a DIX command. If there is no match, DIX
assumes it is a filename to be displayed/editted in screen mode
If you define the symbol DIX_NO_AUTO_COMMAND to Y, DIX will assume the
parameter is a filename.
The first parameter is filename to be "DIX"ed.
The filename can contain a list of wildcards. If
you specify /MULTI_FILE DIX will open all files.
If you do not specify /MULTI_FILE and more than one files matches
the filename(s), you will be prompted to select one file.
Parameter 2..8 are optional and contain the optional
record-search parameters. See the help about "Record_searches"
The filename can also be specified as a file-id.
Format: [device:](n1,n2,n3).
The default for Device: is SYS$DISK:
So $ DIX (10,20,0) Will open the file with fileid (10,20,0) on the
disk specified by SYS$DISK
$ DIX mydisk:(10,20,0) Will open the file with fileid (10,20,0) on
disk "mydisk"
DIX will translate the disk/fileid to a filename, and that filename is used
for the determination of the description.
If you specify the filename as module@[library] or as [library](module),
DIX will read the module "module" from the (.TLB or .MLB) file "library".
Both module and library may contain wildcards.
If no library is specified, DIX will search SYS$LIBRARY:STARLET.MLB
and SYS$LIBRARY:FORSYSDEF.TLB (if present) by default.
There are 2 special names for the library name
DIX_SYS : The DIX_DES.TLB file
DIX_USER : The DIX_DES_USER.TLB file (if present)
Parameter can also be a dix interactive command.
If the symbol dix_no_auto_command is not set to Y, and the command
line does not start with a /, DIX will insert a /command just before
the command line. This setting makes the /screen necessary if you want
to display a file in smg mode
The following symbols may be handy. 'arch' must be VAX, Alpha or IA64
and dev:[dir] the place where the image of DIX can be found
DIX:==$dev:[dir]dix_'arch' !default DIX startup in SMG mode
DXW:==$dev:[dir]dix_'arch'/decw !dix startup in DECwindows mode
DXI:==$dev:[dir]dix_'arch'/inter !dix startup in interactive mode
DXC:==$dev:[dir]dix_'arch'/command !dix startup for one line command
DXP:==$dev:[dir]dix_'arch'/plot !dix startup for T4 plotting
DXD:==$dev:[dir]dix_'arch'/command dir/fast !dix startup fast directory search (like DFU)
DIX also supports a number of symbols, that are checked on startup
DIX_DEBUG : Y : All debug classes enabled at level 0 to the file DIX_DEBUG.DBG
0..9 : The same as above, but the level is set to 0..9
Any other value will be parsed, and debug classes,
level and output will be set.
See the help about INTERACTIVE SET DEBUG
DIX_STATISTICS: If defined as Y, DIX will print a line on exit with the
elapsed, CPU, bufio, dirio, faults (like ^T)
DIX_ENABLE_P2 : If defined as Y, DIX will make use of P2 data (Alpha, IA64).
This is still experimental, so you must enable it.
DIX_FAKE_ARCH : Override the current architecture (VAX/Alpha/IA64)
DIX supports a number of logical names
DIX_HELP : May point to the DIX_HELP.HLB file
The default place is the same directory as the image
DIX_PLOT_SETUP : May point to the dix_plot_setup file.
The default name is DIX_PLOT_SETUP and the default is
SYS$LOGIN:.SETUP
DIX_DES : May point to the system-wide description file.
The default name is DIX_DES and the default is
same_dir_as_image:.TLB
DIX_DES_USER : May point to the user defined description file.
The default name is DIX_DES_USER and the default is
SYS$LOGIN:.TLB
DIX_INI : Set the name for the initialisation file. By default
this file is SYS$LOGIN:DIX_INI.INI. You can also
specify the file via the /startup qualifier.
DIX_DECENT : Set the name of the decent setup file,
By default same_dir_as_image:.setup
DIX_NO_AUTO_COMMAND : Set this symbol to Y to disable the automatic
insertion of /command for command lines starting
with a valid dix verb
DIX can operate in 9 modes, 3 are interactive, they let you change the
records (if you specified /modify)
The interactive modes
/SCREEN_ORIENTED : DIX will display data full screen via the SMG routines
and lets you look at the record and possibly modify it.
/DECW_ORIENTED : DIX will display data full screen via the DECwindows routines
and lets you look at the record and possibly modify it.
/INTERACTIVE : DIX enter interactive (line) mode. This mode has a powerful
scripting language, and can also be used in BATCH.
The non-interactive commands
/FILE_ORIENTED : DIX will behave like DUMP, no interaction possible
/INFO : DIX will display information about the file.
/VERIFY : DIX will verify an indexed file
/ANALYZE : DIX will analyze an indexed file with lots of info
/COMMAND : DIX will take the argument and process it
/PLOT : DIX will take the argument and process it as PLOT/T4 'arg'
The qualifiers /SCREEN, /DECWINDOW, /FILE, /INTERACTIVE, /INFO
/PLOT /VERIFY, /ANALYZE, /COMMAND
must be entered just after the DIX verb.
Three more major-qualifiers qualifiers can be given
/HELP : Display help
/DEMO : Display demo pages
/VERSION : Display the version, link date of DIX (/ALL for all modules)
DIX/ANALYZE [/DISPLAY=(option,option..)] [/VERBOSE=level]
option can be
KEY=n : Analyze only key
LEVEL=n : Analyze only key level 'n'
LEVEL="+n" : Analyze only the top 'n' key levels
LEVEL=-n : Analyze only the lowest '-n' levels
[NO]DATA : Analyze the data buckets too. Since this can take a lot
of time, this is not the default.
[NO]SIDR : Display the SIDR records, this is the default
[NO]KEYBUCKETS : Display the buckets in the key tree, this is the default
OUTPUT=filename : If verbose is >0, the extra output can be redirected
to another file. The normal output will be in the file
specified with /OUTPUT=filename
START=n : Start analysis at bucket 'n' in chain, default=0
(n may be negative, last but 'n' bucket of the chain)
END=n : End analysis at bucket 'n' in chain, default is whole chain
(n may be negative, last but 'n' bucket of the chain)
/VERBOSE=level
The verbose qualifier defines the amount of output
VERBOSE=0 : Just the normal output, see below
VERBOSE=1 : Print out the buckets read
VERBOSE=2 : Detail 1 and the bucket header info
VERBOSE=3 : Detail 2 and SIDR records and the data records
VERBOSE=4 : Detail 3 and the SIDR pointers
If VERBOSE>0, the output of the detail option can be redirected to another
file. See the OUTPUT= option
Examples:
$ DIX/ANA/DISP=DATA DUPKEY.IDX
Analysis of key 1 !analysis info for key 1
Type STG, flags %X05:Duplicate,Nullkey=%X30
Idx area 2, Level1 area 2, bucketsize 1, Fill 512, Level 1, rootvbn 18
Data area 2, bucketsize 1, Fill 512, VBN 17
Keysize 4, number of keysegments 1, minimum recordsize 9
Segment 1, Position 5, size 4
Level 1 !level 1 (toplevel) for key 1
Start vbn : 18, #records = 1, #buckets 1(1 blocks), 100% contiguous
Bytes in : Overhead: 17, Keyvalues: 4, Pointers: 2, Bucket fill: 5%
Bytes in keys : 4(compressed), 4(uncompressed) = 100%
Analysis of SIDR buckets
Start vbn : 17, #buckets 1(1 blocks), 100% contiguous
Bytes in : Overhead: 14, Keyvalues: 12, Pointers: 43, Bucket fill: 16%
Bytes in keys : 12(compressed), 12(uncompressed) = 100%
Records : Keyvalues: 3, RFA's: 7, Deleted: 0
Maximum SIDR length : 3, average 2, max SIDR length for key value:2222
Average key length : 4, Average data length 14
Analysis of key 0 !analysis for key 0
Type STG, flags %X81:Duplicate,Data compress
Idx area 1, Level1 area 1, bucketsize 1, Fill 512, Level 1, rootvbn 13
Data area 0, bucketsize 1, Fill 512, VBN 4
Keysize 4, number of keysegments 2, minimum recordsize 4
Segment 1, Position 2, size 2
Segment 2, Position 0, size 2
Level 1 !level 1 (toplevel) for key 0
Start vbn : 13, #records = 1, #buckets 1(1 blocks), 100% contiguous
Bytes in : Overhead: 17, Keyvalues: 4, Pointers: 2, Bucket fill: 5%
Bytes in keys : 4(compressed), 4(uncompressed) = 100%
Analysis of DATA buckets
Start vbn : 4, #buckets 1(1 blocks), 100% contiguous
Bytes in : Overhead: 14, Keyvalues: 44, Data: 178, Bucket fill: 70%
Bytes in keys : 44(compressed), 44(uncompressed) = 100%
Bytes in datarecords: compressed 178, Uncompressed 145, = 123% (without primary key)
Datarecords : Valid 11, #Deleted 0, #RRV 0
Max data length : 18, Average = 17
Maximum DUP keylen : 1
The contiguousness of the areas are not very significant except for the
SIDR and data area. They are determined in the following way.
If the next bucket of a bucket is adjacent to the current buffer, 5 is added
If the next bucket lies within a bucketsize next to the current buffer, 4 is added
If the next bucket lies within 2 bucketsizes next to the current buffer, 3 is added
If the next bucket lies within 3 bucketsizes next to the current buffer, 2 is added
If the next bucket lies within 4 bucketsizes next to the current buffer, 1 is added
If the next bucket lies before the current bucket, 0 is added
The total is multiplied by 100 and divided by the 5*(bucketcount-1). This gives a
number between 0 and 100.
If all buckets are adjacent in ascending order we get 100%.
DIX/COMMAND dix_command Take the command line and execute it. This is a one-line interactive mode Command may be a list of commands via the ;command;command construct Each ; must be preceded by a space, or it could be interpreted as part of a filename
If you specify /DECWINDOW, the program will work full screen
using DECwindows, and you can look at the data, and possibly
modify it (see /MODIFY qualifier).
If you have a description present
1. You can modify fields (if in modify mode) by entering a right arrow,
Enter, or any printable character. You are in field modifying mode until
you enter an Enter. After you type Enter DIX checks if the typed data
is valid for that field(-type) and restores the original value if not.
The record will not be updated until you type the DO key.
2. For fields that contain field-names, you can also enter a key_getfields
key (default PF1-G) and modify/enter the field value.
3. If you type a key_contdis (default F8) DIX will open a popup-display
containing information about the description of the current line, and
update it when you move between lines. If you type key_contdis again
DIX will remove the display.
And lots more, see the help (Default PF2 or Help)
on the command line you may set some settings for the DECwindows displays
See the help in /attributes.
Syntax
DIX/DECWINDOW filespec[/qualifiers] [searchparameter] [searchparameter] ..
The searchparameter is optional, see the help about [DIX/HELP] record_search
Up to 7 (search) parameters are supported. DCL allows 8 parameters, and
the first one is the filename.
DIX will enter demo mode, You will be show (some of) the capabilities of DIX.
If you specify /FILE the program will dump the data to a file or the terminal (see /OUTPUT qualifier) in either interpreted or raw dump mode. Syntax DIX/FILE filespec[/qualifiers] [searchparameter] [searchparameter] The searchparameter is optional, see the help about [DIX/HELP] record_search Up to 7 (search) parameters are supported. DCL allows 8 parameters, and the first one is the filename.
Display the DIX help, you are doing it now
You can specify arguments for the help. If the first argument is a
qualifier you must use "'s around the help text, otherwise DCL will get
confused.
The following qualifiers are supported for the HELP command, and must
be entered directly after the HELP command
/DCL : Display the help in SYS$HELP:HELPLIB.HLB instead of the DIX help
/DECWINDOW : Display the help in a DECwindow window. This requires the
availability of DECwindows and a freeware package FSHELP
(see below)
Examples:
$ DIX/HELP modes ! Give help about the mode
$ DIX/HELP modes info ! Give help about the modes info subtopic
$ DIX/HELP "/DISPLAY" ! Give help about the /DISPLAY qualifier. You need
the quotes here
$ DIX/HELP record_sear/exist ! Give help about the /exist subtopic in record_sear
$ DIX/HELP/DCL copy ! Give help about the COPY command in SYS$HELP:HELPLIB.HLB
$ DIX/HELP/DECW modes ! Give the help about modes in a DECwindows display
(if FSHELP is installed)
DIX will use the LBR$OUTPUT_HELP routine. If you also have the
FSHELP freeware package installed (available on the same website),
DIX will use the FSHELP_SHR inclusive a DECwindows interface.
The /INFO qualifier will display information about the file(s), including
information about the keys of an indexed file.
This information includes name, position, size and type of keys.
The /INFO qualifier must be the first on the command line.
If you specify the /FDL qualifier also, DIX will output the info in
an FDL file. This should give the same output as ANALYZE/RMS/FDL,
but is does not analyze the data, so it is a lot faster for large files
(but cannot be used for optimizing)
Example:
$ DIX/INFO SYSUAF
Information about file KXSYS_DISK:[VMS_COMMON.SYSEXE]SYSUAF.DAT;2
File
Opened for : Normal I/O,Fast I/O(Default)
Fast Blocksize : 1024
Fast Multibuffer: 4
Organization : Indexed
Lock mode : Normal
Record format : Variable
Creation date : 25-NOV-2003 17:01:26.28
Expiration date : 5-FEB-2007 08:15:44.15
Backup date : <No backup recorded>
Revision date : 9-AUG-2006 08:15:44.15
File id : (66004,27,0)
File owner : [SYSTEM] = [1,4]
File protection : S:RWED, O:RWED, G:RE, W:
Globalbuffercount: 0
Record
Record attribute : NONE
Max record size : 1412
Longest record : 0
Allocation
EOF block : 72 (FFB = 0)
Allocated blocks : 72
Bucketsize : 3
Areas
Area Allocation bucketsize extension
0 72 3 10
Keyinformation
D C I K D N ..INDEX. ..DATA..
Nr Type u h c c c u Lvl Pos Siz Area Bkt Area Bkt Field
p g p p p l Nr siz Nr siz
0 STG N N Y Y Y NO 1 4 32 0 3 0 3 UAF$T_USERNAME
1 BIN4 Y Y N N N NO 1 36 4 0 3 0 3 UAF$W_MEM,UAF$W_GRP,UAF
$L_UIC
2 BIN8 Y Y N N N NO 1 36 8 0 3 0 3 UAF$W_MEM,UAF$W_GRP,UAF
$L_UIC,UAF$L_SUB_ID
3 BIN8 Y Y N N N NO 1 44 8 0 3 0 3 UAF$Q_PARENT_ID(1),UAF$
Q_PARENT_ID(2)
* means the current file
The tags in the keyinformation are
Nr : The key number
Type: The key type (Descending keys have D in front)
Dup : This key allows duplicate
Chg : This key is changeable
Icp : Key values are compressed in index buckets
Kcp : Key values are compressed in data buckets
Dcp : Data values are compressed (in data buckets)
Nul : The null-key value (in hexadecimal) for this key or NO
Lvl : The tree depth for index buckets
Pos : The starting position of the key (multiple lines if more segments)
Siz : The size of the key (multiple lines if more segments)
Index:
Area_nr : The area for the index buckets
Bktsize : The bucket size for the index buckets
Data:
Area_nr : The area for the data buckets
Bktsize : The bucket size for the data buckets
If DIX has a description for this file (as above) DIX will also
display the fields that are part of this key
An example of an FDL command
$ DIX/info/fdl/out=test.fdl test.dat
If you specify /INTERACTIVE, DIX will enter interactive mode
and lets you display/modify fields. This mode works in BATCH too,
so you can create command-files and use DIX to display/modify fields.
See also the /INTERACTIVE qualifier.
![Interactive mode]DIX /INTERACTIVE
Syntax
DIX/INTER [@file] [filespec] [/qualifiers] [searchparameter] ..
@file maybe a file containing DIX commands. If specified it must
be the first parameter. The default filename is [].DIX
This file can also be specified via the /script qualifier.
Filespec is optional. If you do not specify a filespec DIX will start
without any file open. You can then later open/close files with the
interactive commands OPEN and CLOSE.
Example:
$ DIX/INTER @mycommands SYSUAF/MODIFY
$ DIX/INTER SYSUAF/MODIFY /script=mycommands
Will both execute the commands in mycommands[.DIX] and exit.
The searchparameter(s) is/are optional, see the help about
[DIX/HELP] record_search
Since DCL allows 8 parameters, up to 7, or 6 if @file is present,
searchparameters are supported.
If you specify DIX/RESTORE, DIX will continue from the point you exited DIX the last time (if you had auto mode enabled then). It will open all files you had open, and select the record(s) you were processing at that time. It will also restore your recall buffer, key definitions for the interactive mode, display settings and so on. See the SET AUTO, SHOW AUTO, EXIT/[NO]AUTO command in interactive mode.
pipe command [;command [;command]] Make a sequence of commands. You can use loops in this statement i.e. DIX> pipe for k=1,10 ; show symb k ; end for DIX> pipe k=1 ; lp:show symb k ; k=k+1 ; if k <= 10 then goto lp the ; must be preceded by a space or it could be part of a filename This is handled as a new command procedure level. Any symbols defined in the pipe command will be deleted after the pipe is complete. If you want to preserve the symbol, declare it with /level=-1 so DIX> ; int a ; a=10 ; say a !will print 10, and a is deleted and DIX> ; int a/level=-1 ; a=10 ; say a !will print 10, but a is still there You can replace the PIPE verb by a ; so DIX> pipe command ; command ; command will be the same as DIX> ; command ; command ; command
DIX/PLOT filemask/qualifiers DIX will execute the PLOT/T4 'filemask/qualifiers'. For more help see the INTERACTIVE_MODE PLOT command. You can also use the DIX/PLOT/CSV and the DIX/PLOT/PIE command.
If you specify /SCREEN , DIX will work in full screen mode
using the SMG routines, and you can look at the data, and possibly
modify it (see /MODIFY qualifier).
If you have a description present
1. You can modify fields (if in modify mode) by entering a right arrow,
Enter, or any printable character. You are in field modifying mode until
you enter an Enter. After you type Enter DIX checks if the typed data
is valid for that field(-type) and restores the original value if not.
The record will not be updated until you type the DO key.
2. For fields that contain field-names, you can also enter a key_getfields
key (default PF1-G) and modify/enter the field value through a SMG-menu.
3. If you type a key_contdis (default F8) DIX will open a popup-display
containing information about the description of the current line, and
update it when you move between lines. If you type key_contdis again
DIX will remove the display.
And lots more, see the help (Default PF2 or Help)
Syntax
DIX[/SCREEN] filespec[/qualifiers] [searchparameter] [searchparameter] ..
The searchparameter is optional, see the help about [DIX/HELP] record_search
Up to 7 (search) parameters are supported. DCL allows 8 parameters, and
the first one is the filename.
DIX/verify indexedfile [qualifiers] Check all buckets of an indexed file for corruption. Supported qualifiers are /verbose=n : n=0..4 Output more info during the check /output=filename : Print errors to file DIX can be used to detect and repair corrupt indexed files See the deeper help BE VERY CAREFUL. =============== Always make a copy of the corrupt file and edit the copied file.
If a file has a corrupt bucket, You may use DIX to repair the file
or get as many records out of it as possible.
If the bucket is not a data bucket, DIX using Fast I/O can read all records
and copy them to a new file.
Example:
We have a file with a corrupt index bucket (it was patched)
$ DIX/info/verify sysu.dat/verbose=3
Information about file USER50:[STUBBF.PROGRAMS.DIX.TEST]SYSU.DAT;2
...
Reading Prologue block 1
Checking area 0
Reading area admin offset 0 in block 3
Checking area fields 0
Checking key 0
Reading key admin block offset 0 in block 1
Checking key fields for key 0
Checking key tree for key 0
%DIX_RMS-W,Bucket 7 level 1 Check bytes different, first 0A, last 0E, previous bucket 0
Checking key 1
Reading key admin block offset 0 in block 2
Checking key fields for key 1
Checking key tree for key 1
Checking key 2
Reading key admin block offset 102 in block 2
Checking key fields for key 2
Checking key tree for key 2
Checking key 3
Reading key admin block offset 204 in block 2
Checking key fields for key 3
Checking key tree for key 3
This tells us that a bucket is corrupt in the primary key chain at level 1.
Since this is NOT a data bucket we can use DIX to get all records from this
file and copy it to a new file using
$ DIX/interactive sysu.dat
%RMS-F-CHK, bucket format check failed for VBN = 7
%DIX-I-USINGFIL, Using file USER50:[STUBBF.PROGRAMS.DIX.TEST]SYSU.DAT;2/NOMOD
DIX (RMS) is telling us that the file has a corrupt bucket (the root bucket
for the primary key. RMS cannot process this file, so we try again using
Fast I/O
$ DIX/interactive sysu.dat/fastio
%DIX-I-USINGFIL, Using file USER50:[STUBBF.PROGRAMS.DIX]SYSU.DAT;2/NOMOD
DIX> WRITE/ALL/LOG new_file.dat
%DIX-I-RECAPP, 62 records written to NEW_FILE.DAT
So now we have a copy of the file with all record. Now use convert to
make the file indexed again, and we have all records back
If the corrupt bucket IS a data bucket you have two options.
1. Use DIX to repair the corrupt bucket.
2. Use DIX to skip the data bucket.
Now suppose the corrupt bucket is the in one of the data buckets.
RMS as well as Fast I/O will follow the chain of the data-buckets and
will find the corrupt bucket, so we cannot use the previous method.
Below is the technique to repair a DATA-bucket (if not too heavily damaged)
$ DIX/interactive [.test]sysu.dat
%DIX-I-USINGFIL, Using file USER50:[STUBBF.PROGRAMS.DIX]SYSU.DAT;2/NOMOD
DIX> ne
DIX> ne
%RMS-F-CHK, bucket format check failed for VBN = 55
%RMS-F-CHK, bucket format check failed for VBN = 0
This tells us we have a corrupt bucket at bucket 55. We need to know the
bucketsize of the data-buckets
$ DIX/info sysu.dat !show key info of the file
Keyinformation
.....
D C I K D N ..INDEX. ..DATA..
Nr Type u h c c c u Lvl Pos Siz Area Bkt Area Bkt
p g p p p l Nr siz Nr siz
0 STG N N Y Y Y NO 1 4 32 0 3 0 3 <== this 3 is the
data bucketsize
...
Enter DIX in block mode for this bucket (using the built-in $BUCKET
description)
$ DIX/INT/MODIFY /BLOCK=3/record=55/descr=$bucket SYSU.DAT
%DIX-I-USINGFIL, Using file USER50:[STUBBF.PROGRAMS.DIX]SYSU.DAT;2/MOD
%DIX-I-MESSAGE, "Data bucket"
%DIX-I-MESSAGE, "Check and Check1 not equal, corrupt bucket"
%DIX-I-USINGDES, Using description DSA50:[STUBBF.PROGRAMS.DIX]DIX_DES.TLB;135($BUCKET)
DIX> !DIX tells us the bucket is corrupt (the check and check1 are unequal)
DIX> examine *CHECK*
0.0|CHECK |19
2.0|CHECK_VBN|55
1535.0|CHECK1 |17
And indeed : check and check1 are not equal. We can patch them to be equal
DIX> DEPOSIT CHECK=17
%DIX-I-MESSAGE, "Data bucket" !DIX tells us it is data bucket
DIX> UPDATE
If this was the only problem, the bucket is now correct.
You may test to see if the rest of the buckets looks OK
DIX> EXAMINE
If this looks fine, all is done. If not we need to skip this data bucket
from the data-bucket chain, so see the other help (skipping bucket)
Now the bucket is very corrupt and does not contain any valid data.
We can use DIX to skip the bucket from the data-chain. This means
we use a bucketful of records, not perfect but better than nothing at all.
$ DIX/info/verify sysu.dat
Information about file USER50:[STUBBF.PROGRAMS.DIX.TEST]SYSU.DAT;2
%DIX-E-VERIFY, Bucket 55, level 0 Check bytes different, first 13, last 11,
previous bucket 4
%DIX-E-VERIFY, Data bucket 55 points to bucket 33,but the upper key bucket
points to 34
%DIX-S-VERIFY, Using pointer from index bucket 7 to bucket 34
%DIX-E-ERRSEEN, Error seen during verify
So bucket 55 has two problems
1. The check bytes do not match
2. The next_bucket pointer (33) is probably incorrect, since the above
lying index buckets points to bucket 34, and this index bucket seems
to be valid.
The previous bucket of 55 is 4, so now we need to patch bucket 4 to point
to bucket 34, effectively skipping bucket 55.
We still know the bucket size for data=3
$ DIX/INT/MODIFY /BLOCK=3/record=4/descr=$bucket SYSU.DAT
%DIX-I-USINGFIL, Using file USER50:[STUBBF.PROGRAMS.DIX.TEST]SYSU.DAT;2/MOD
%DIX-I-MESSAGE, "Data bucket"
%DIX-I-USINGDES, Using description DSA50:[STUBBF.PROGRAMS.DIX]DIX_DES.TLB;135($BUCKET)
DIX> examine next_bucket
8.0|NEXT_BUCKET>|55
Indeed bucket 4 points to bucket 55, we now set its forward to bucket 34
DIX> deposit next_bucket=34
DIX> UPDATE
DIX> REWIND
DIX> WRITE/ALL/LOG x.x
%DIX-I-RECAPP, 58 records written to x.x
Now we have copied all (58 and not 62) records to x.x, and recovered most of
the records
DIX/VERSION [/ALL][/SYMBOL] [/OUTPUT=file]
Display the version and link date of DIX.
If /ALL is specified, versions and creation dates of all DIX modules
will be displayed.
If /SYMBOL is specified DIX will define the following symbols
DIX_VERSION : Containing the version e.g. "7.1"
DIX_LINK_DATE : Containing the link date e.g. "23-AUG-2008 10:12.19.45"
If /ALL is also specified, DIX will also define the following symbols
DIX_MODULE_COUNT : Count of modules
DIX_MODULE_NAME_'n' : The name of the module 'n'
DIX_MODULE_DATE_'n' : The creation date of the module 'n'
with n=1..dix_module_count
The following file types are supported.
For SEQUENTIAL or RELATIVE files the record selection is based on the /RECORD qualifier. If the start record is nn, the first nn-1 records will be skipped, this may take some time. If the record format is fixed, DIX can read the record directly without skipping the previous records.
/RECORD=nr You can specify the starting record number. For indexed files you can also use the /key to go to the 'n'th record on key 'key'. If not specified the program starts at the first record.
You can specify a record number. DIX will rewind the file and skip 'recordnumber'-1 records. You can also specify a cell number with the /EQ, /GE, /GT In this case the access will be direct.
/RECORD=nr You can specify the starting record number.
/EQ=value Look for a record with a cell number EQ the specified value. The values will be converted by DCL to uppercase, unless enclosed in quotes. See also the help about keyvalues. If DIX is in case_sensitive mode, DIX will not uppercase the value. See the help about "set case "
/GE=value Look for a record with a cell number GE the specified value. The values will be converted by DCL to uppercase, unless enclosed in quotes. See also the help about keyvalues. If DIX is in case_sensitive mode, DIX will not uppercase the value. See the help about "set case "
/GT=value Look for a record with a cell number GT the specified value. The values will be converted by DCL to uppercase, unless enclosed in quotes. See also the help about keyvalues. If DIX is in case_sensitive mode, DIX will not uppercase the value. See the help about "set case "
When specifying a record number or a cell number with the /record,/eq,/ge,/gt DIX will convert the text to a number.
For INDEXED files the record selection is based on the /EQ or /GE or /GT or /LE or /LT and the /KEY qualifier. If the key has Ascending attributes, the /EQ, /GE or /GT qualifiers must be used, and if the key is descending, the /EQ, /LE or /LT qualifiers. Information about the key types can be obtained via the /INFO qualifier.
/RECORD=nr You can specify the starting record number. You can also use the /key to go to the 'n'th record on key 'key'.
/KEY=nr Look for the keyvalue on the specified key. The default is the primary key (0). If none of the LT, LE, EQ, GE, GT options is specified, the program will start with the first record on the specified key.
/LT=value Look for a record with keyvalue LT the specified value. Information about the key types can be obtained via the /INFO qualifier. The values will be converted by DCL to uppercase, unless enclosed in quotes. See also the help about keyvalues. If DIX is in case_sensitive mode, DIX will not uppercase the value. See the help about "set case "
/LE=value Look for a record with keyvalue LE the specified value. Information about the key types can be obtained via the /INFO qualifier. The values will be converted by DCL to uppercase, unless enclosed in quotes. See also the help about keyvalues. If DIX is in case_sensitive mode, DIX will not uppercase the value. See the help about "set case "
/EQ=value Look for a record with keyvalue EQ the specified value. The values will be converted by DCL to uppercase, unless enclosed in quotes. See also the help about keyvalues. If DIX is in case_sensitive mode, DIX will not uppercase the value. See the help about "set case "
/GE=value Look for a record with keyvalue GE the specified value. Information about the key types can be obtained via the /INFO qualifier. The values will be converted by DCL to uppercase, unless enclosed in quotes. See also the help about keyvalues. If DIX is in case_sensitive mode, DIX will not uppercase the value. See the help about "set case "
/GT=value Look for a record with keyvalue GT the specified value. Information about the key types can be obtained via the /INFO qualifier. The values will be converted by DCL to uppercase, unless enclosed in quotes. See also the help about keyvalues. If DIX is in case_sensitive mode, DIX will not uppercase the value. See the help about "set case "
When specifying a keyvalue with /lt, /le, /eq, /ge, /gt the program will check for the key type (in the file). For binary key types the program will convert the data from ASCII to binary. For string types you can enter binary values via the %DD syntax where DD are hexadecimal digits (%% will enter one %). See the conversion options below
INT2 values can only be specified as integers.
INT4 values can be specified in 3 ways
1. A normal integer.
2. A number with another radix in two formats
'DDDD'F with F : X (hexadecimal), O(octal), D(decimal), B(binary) or Rn[n]
or %Fddddd with F : X (hexadecimal), O(octal), D(decimal) or B(binary)
3. A Date which will be converted to DATE*4 format.
4. A OpenVMS Identifier
5. A UIC e.g. [SYSTEM]
6. An ASCII string "ABCD"
INT8 keyvalues can be specified in 3 ways
1. One big integer up to 18 characters
2. Two INT4 values separated by a . (e.g. 10.20, see INT4 conversion)
(e.g. SYSTEM.1-jan-1991)
3. A date string (e.g. 1-apr-1991).
A VFC file is a sequential file with a fixed prefix. You will normally see this type of file from the DCL OPEN statement. This will generate a VFC file with a 2 byte header, that is used for carriage control. But you can also create your own VFC file with other sizes (up to 255) for other meanings. DIX will display the data from the normal record in one window, and the VFC data in another (in screen mode), or lets you display the data via EXAMINE/VFC (in interactive mode). The file mode can also display the data when you specify /DISPLAY=(all) or /DISPLAY=VFC (see the help about [DIX/HELP] "/DISPLAY") The DEPOSIT/VFC will change data in VFC buffer (in interactive mode) and the PF1-Y command switches from normal data to VFC data in screen mode.
DIX can work either in RAW dump format or INTERPRETED (see the /DESCRIPTION qualifier).
In the RAW dump mode the dump is equal to the OpenVMS DUMP format. It is used if no description file can be found, or if the /RAW qualifier is specified. The display is split into two parts. The left part is the data (byte, word, longword) in decimal or hexadecimal. The right part is the data in ASCII (if printable). In RAW dump mode the following qualifiers can be used.
The data will be dumped in bytes. See also the /WORD and /LONG and /UNSIGNED qualifiers.
The data will be dumped in words. See also the /BYTE and /LONG and /UNSIGNED qualifiers.
The data will be dumped in longwords. See also the /BYTE and /WORD and /UNSIGNED qualifiers. This is the default.
The data (if not in /HEX mode) will be interpreted as unsigned.
Display the data in hexadecimal (Default). See also the /BINARY,/OCTAL,/DECIMAL qualifiers.
Display the data in decimal. See also the /HEX,/BINARY,/OCTAL qualifiers.
Display the data in octal. See also the /HEX,/BINARY,/DECIMAL qualifiers.
Display the data in binary. See also the /HEX,/OCTAL,/DECIMAL qualifiers.
In the interpreted mode, the program will try to display the fields
of a record. This mode requires a description file (see the /DESCRIPTION)
This is the default if a description can be found.
Descriptions can be in files (default 'filename'.DES) or a text library.
DIX will look for a description file in the same directory as the data file
DIX will look for a file in the current directory.
DIX will look for a module in the DIX_DES.TLB file (system_library)
(see the help for the [DIX/HELP] "/SYSTEM_LIBRARY" qualifier)
DIX will look for a module in the DIX_DES_USER.TLB file (user_library)
(see the help for the [DIX/HELP] "/USER_LIBRARY" qualifier)
For details see /DESCRIPTION.
If no description can be found (and you did not specify /NODES),
DIX will load a default description
union
map
bytearr*(32765) $bytes
end map
map
character*32765 $line
end map
map
character $chars(32765)
end map
end union
You can reference (part of) these fields, but DIX will not display
the fields with the examine command or screen mode display
DIX> say Examine ! Will display in raw mode
DIX> Say $line ! Will display the field $line
DIX> say $line{10:20} ! Will display elements 10:20 of $line
If the /ALL is given, all matching descriptions will be used. If /ALL is not given, only the first match will be used. The search order is 1. The .DES file in the same directory as the datafile 2. The .DES file in the current directory 3. Entries in user_library (if present) 4. Entries in system_library (if present)
/DESCRIPTION[=name[,name]]
Give the name of a record description file(s) or module.
The format of description file/module is described in the RECORD_FORMAT help.
If the /DESCRIPTION qualifier is given without the filename, the program will
look for a file with a filename as the input file, but with the extension
.DES. DIX also searches the Text library DIX_DES (with default file
"IMAGE_DIRECTORY".TLB), so the text library will be searched
in the same directory as the DIX image (but you can redefine the
logical name DIX_DES to any text library) for a module with the name that
matches the filename of the datafile. A match can be on device, directory,
name and extension. This key may define multiple fields separated by a |.
If there are multiple field, any field may match.
DIX also supports a user defined library. DIX searches for the file
DIX_DES_USER with default SYS$LOGIN:.TLB.
Normally DIX will stop as soon as a match is found, but if /ALL is given
all matching descriptions will be used.
Example:
DIX MYDEV:[MYDIR]MYDATA.MYEXTENSION [/DES]
1. DIX searches in the current directory for the file MYDATA.DES
2. DIX searches for a description file MYDEV:[MYDIR]MYDATA.DES
3. DIX searches DIX_DES_USER.TLB for a module that matches the name
MYDEV:[MYDIR]MYDATE.MYEXTENSION. Empty parts of the module name match.
So a module with the name *MY*, .MYEXTENSION, *:[*]*.* will all match.
4. DIX searches DIX_DES.TLB for a module that matches the name
MYDEV:[MYDIR]MYDATE.MYEXTENSION. Empty parts of the module name match.
So a module with the name *MY*, .MYEXTENSION, *:[*]*.* will all match.
The first or all (/ALL specified) matches will be loaded and can be used.
Example for multiple fields:
Suppose the key is ".BCK|.SAV". This description will match files
*.BCK and *.SAV.
To insert a module in one of the text libraries you can i.e. use :
$ LIBRARY/REPLACE DIX_DES.TLB backupdes.txt/MODULE=".BCK|.SAV"
DIX can use Views. A View is a way to connect multiple records in
multiple files to a single entity. With DIX you can display
and modify the accumulated data as one record.
A view is a module in the DIX_DES.TLB file (or a text file) that contains
instructions to build the record from the various records in various files.
The name of the module is "VIEW#view_name". The syntax of "view_name" is
the same as that of the description records. In the examples is a view
for a directory entry. The description has the module name ".DIR" and the
view has the module name "VIEW#.DIR"
If you use a file, the default file type is ".VIEW".
You can open a file with a view. Is this view you can select fields from
this (first) file (FIELD statement), and you can ask DIX to follow links
to other records in the same or other files (the FOLLOW/BACK statement).
In order to use a file in a view, it should have a description, since views
reference fieldnames via the FIELD statement, but if you do not have
a description you can still use the RANGE command.
If you start DIX, you may specify the /VIEW[=namemask] after the file you
want to "DIX". You can also use the /follow qualifier to create a
simple "follow all" view. See the help about /follow
DIX does support modification of views, but only in a limited way.
The fields that may be modified must have all of the following
a. Be fixed AND
b. Not change the link to another file/record
This restriction may be lifted in a future release.
Below are some examples of views.
The following example is the view for a directory entry, and all of the
file headers up to the [000000] directory. The view will only display the
name fields in the file headers.
field verslim !examine the version limit
field name !examine the name
follow v(1).fileid !follow the link to the file header
!of the file
loop !loop
field *_filename !examine the name in ODS-5 part
field fnam !and in ODS-2 part
test (file_id <> backl_fid) !stop if fid=backl_fid
follow backl_fid !follow link to parent
end loop !end of loop
The file header for the 000000.DIR directory has a fid of (4,4,0) and
a back_link (to its parent directory) of (4,4,0) too.
We assume the VIEW#.DIR module in the DIX_DES.TLB file exists and
contains the above lines.
$ DIX/FILE/COUNT=1 ALPHA.dir/view
%Info:Recordnumber = 1 , Recordsize = 84 , RFA=(1,0)
0|VERSLIM|none !from the directory entry
2|NAME |DIX_AUTOSAVE_LIBRARY.OBJ !from the directory entry
27|FNAM |DIX_AUTOSAVE_LIBRARY !file header of the .obj file
47|FNAM |ALPHA.DIR;1 !file header of its directory
67|FNAM |DIX.DIR;1 !" " " "
87|FNAM |PROGRAMS.DIR;1 !" " " "
107|FNAM |STUBBF.DIR;1 !" " " "
127|FNAM |000000.DIR;1 !File header of the 000000.dir file
An example for a set of three files that make a simple cross reference
system:
The first file (CROSS_REF.CRF_FILE_NAMES) has the following description
(.CRF_FILE_NAMES in the system or user text library)
integer*2 file_nr !primary key
character*60 file_name
The second file (CROSS_REF.CRF_MOD_NAMES) has the following description
(.CRF_MOD_NAMES in the system or user text library)
integer*2 mod_nr !primary key
character*32 mod_name
integer*2 file_nr/file=.CRF_FILE_NAMES !link to the filename
The third file (CROSS_REF.CRF_CROSS) has the following description
(.CRF_CROSS in the system or user text library)
integer*2 caller_nr /file=.CRF_MOD_NAMES !link to module name
integer*2 called_nr /file=.CRF_MOD_NAMES !link to module name
We can define a view for the .CRF_CROSS file (either in i.e. the
cross_ref.view file or in the VIEW#.CRF_CROSS module in DIX_DEF.TLB)
(We assume that above named descriptions are in the DIX_DES.TLB file)
$ DIX/file/count=1 cross_rec.crf_cross/view
field caller_nr !examine the value of caller_nr
follow caller_nr !follow the link to .CRF_MOD_NAMES
gosub do_common !common code for caller_nr and called_nr
field called_nr !examine called_nr
follow called_nr !follow the link
gosub do_common !common code
exit !all done
do_common:
field mod_name !examine the name of the module
follow file_nr !follow the link to the filename in .crf_file
field file_name !examine the filename
back !pop back to the .CRF_MOD_NAMES
back !pop back to the .CRF_CROSS
return
The result will be something like
%Info:Recordnumber = 1 , Recordsize = 4 , RFA=(4,1)
0|CALLER_NR|1 !from the CROSS.CRF_CROSS file
2|MOD_NAME |DIX !from the .MOD_NAMES file
34|FILE_NAME|DIX_MAIN.FOR !from the .FILE_NAMES file
94|CALLED_NR|2 !from the CROSS.CRF_CROSS file
96|MOD_NAME |DIX_MAIN !from the .MOD_NAMES file
128|FILE_NAME|DIX_MAIN.FOR !from the .FILE_NAMES file
Another example for the command line /follow qualifier
$ DIX/file/count=1 cross_rec.crf_cross/follow=(end,skip)
%Info:Recordnumber = 1 , Recordsize = 4 , RFA=(4,1)
0|CALLER_NR|1
2|CALLED_NR|2
4|MOD_NAME |DIX
36|FILE_NR |1
38|FILE_NAME|DIX_MAIN.FOR
98|MOD_NAME_|DIX_MAIN
130|FILE_NR_1|1
132|FILE_NAME|DIX_MAIN.FOR
the /follow=(end,skip) means :
First display all fields of the record and then follow all
links in a record and skip key fields
So the display is created as:
1: display the fields in .CRF_CROSS (CALLER_NR and CALLED_NR)
2: Follow the link in CALLER_NR to the .CRF_MOD_NAMES file
3: Display the fields in .crf_mod_name record with mod_nr=1
(we do not display the MOD_NR field since the skipkey is given)
4: Follow the link in FILE_NR to the .CRF_FILE_NAMES
5: Display the fields in .CRF_FILE_NAMES record with FILE_NR=1
(we do not display the FILE_NR field since the skipkey is given)
7: Follow the link in CALLED_NR to the .CRF_MOD_NAMES file
8: Display the fields in .crf_mod_name record with mod_nr=1
(we do not display the MOD_NR field since the skipkey is given)
9: Follow the link in FILE_NR to the .CRF_FILE_NAMES
10: Display the fields in .CRF_FILE_NAMES record with FILE_NR=1
(we do not display the FILE_NR field since the skipkey is given)
A view file/module is a text file.
It may contain the following statements
- All the control command from interactive mode
(if/loop/while/case...) including the gosub/return
but not the @ or CALL statements.
- FOLLOW Statement to follow the link to another file
- BACK Statement to return to the previous file (level)
- FIELD To include fields.
- RANGE To include a range of bytes (can be useful if you have no description)
- OPEN To open files used to follow the fields. The open is only
needed if you need to specify file open options others than
the default options from the follow command (i.e. the description)
- ECHO To display comments to the terminal while expanding
- EXIT Statement to exit
- symbol assignments
For all the statements (except the FIELD and RANGE) the help can be
found in the help about interactive mode.
See also the examples
If you want to use the FIELD command, the file must have a description
FIELD[/follow=follow_type][/skipkey][/readonly][/new_name=name] -
[/quiet] fieldnamemask
The qualifiers have the following meaning
/readonly : Declares the field readonly (if the field is not yet readonly).
/follow=type : Let DIX automatically follow all link fields
type can be
begin : First display the followed fields (and follow them)
and then the normal fields
end : First display the normal fields and then
the followed fields (and follow them)
field : Display the followed fields and follow the
link immediately.
/skipkey : If /follow is used, this will suppress the keyfields
of the followed field, since they will already be
displayed in the current field
/new_name=name : You may override the original name.
If the new_name ends with a . ,the original name is
appended to the new_name
/quiet : If mask does not match any field, DIX will signal a message
If you specify /quiet, DIX will not complain
For an example see the help for the /follow qualifier
The qualifiers must be specified before the fieldnamemask.
Fieldnamemask is a wildcard mask to select fields of the current description
The RANGE command may be used whether or not the file has a description
RANGE[/READONLY][/quiet][/name=name] bpos:[epos] !epos is default equal to bpos
/readonly : Set the field to readonly (if the file is not yet readonly).
/quiet : If mask does not match any field, DIX will signal a message
If you specify /quiet, DIX will not complain
/name=name : Set the name (default=B_'BPOS'_'EPOS')
/type=fieldtype : Set the fieldtype (default is BYTEARR)
Include bytes bpos up to and including epos
If you do not specify epos, DIX will take the default size for the
selected fieldtype. If you do specify epos, the size (epos-bpos+1)
must be valid for the selected type
so "RANGE /type=integer 0:9" will be invalid, since INTEGER*10 is not
supported.
and "RANGE /TYPE=INTEGER 12" will insert an INTEGER*4 (the default)
Originally DIX_DES.TLB and DIX_DES_USER.TLB were designed to store
record definitions (descriptions). In a later phase VIEWS were added.
To discriminate between views and normal definitions the module names of the
views (in the .TLB) files are preceded by VIEW#. So a view with a name
.DIR is stored in the .TLB file with a key VIEW#.DIR and the record definition
for the same file with key .DIR.
Later the same technique was used to store DEMO files (DEMO#...)
Since DIX 8.3 DIX can look up a lot of things in .TLB files so this concept
has been enhanced. Now the following types of data can be stored in .TLBs
Suppose the name of the object is MYNAME
Prefix Type
None The normal record definitions
VIEW# The view descriptions to be used via the /VIEW=MYNAME
DEMO# The DEMO files to be called in the DIX/DEMO mode
FUNC# Function definitions to be used via the LOAD/FUNC MYNAME
PARM# Parameter definitions to be used via LOAD /MODULE=MYNAME
PROC# Command scripts to be called like @MYNAME@
STRU# Structure definitions to be used via load/structure MYNAME
You can store the definitions yourself via the (example for parameter file)
$ LIBRARY/INSERT DIX_DES_USER.TLB myparam.txt/module="PARM#MYPARAMS"
and then use if via the
DIX> LOAD /modu=MYPARAMS !DIX will search SYS$LIBRARY:STARLET.MLB
!and SYS$LIBRARY:FORSYSDEF.TLB (if is exists)
! for the module MYNAME
!and DIX_DES_USER (the user des file, if it exists)
!and DIX_DES, the normal description file
! for the module PARM#MYNAME
The description records are in normal ASCII, and can be edited with a normal editor. Lines starting with C, c or * are skipped (Fortran comment lines) and data after the ! (comment) is skipped. Continuation lines are not like the normal Fortran syntax but use the DCL syntax (- at the end of the line) Spaces are ignored, except within names and types and between the type and its qualifier. Single lines can be up to 255 chars, and the total line length (including continuation lines) can be up to 4096 chars. There are 6 types of lines -Parameter, value &type definitions -Conditional statements -Extra statements -Field Declarations. -Control Statements -Macro definitions During the reading of the description, DIX may use symbol substitution. See the "INTERACT SET SUBST" help topic.
DIX supports macros. A macro starts with the
#MACRO NAME [PARAM1 [PARAM2] ...] statement and ends with a
#ENDMACRO statement
Macros can be used to repeat a number of lines.
Inside the macro you specify param1 and param2 by
'param1' and 'param2'.
Up to 16 parameters are supported, and macro calls can
be nested to a depth of 64
Example:
A definition of an ACMS file
....
integer*2 menu_offs
ubyte menu_leng
range (menu_offs: menu_offs+menu_leng)
character*(menu_leng) menu
end range
integer*2 mdb_offs
ubyte mdb_leng
range (mdb_offs: mdb_offs+mdb_leng)
character*(mdb_leng) mdb
end range
integer*2 init_appl_offs
ubyte init_appl_leng
range (init_appl_offs: init_appl_offs+init_appl_leng)
character*(init_appl_leng) init_appl
end range
.....
Above description contains a lot of similar statements.
They can be replaced by
....
#macro item name !define a macro first
integer*2 'name'_offs !it contains one parameter
ubyte 'name'_leng !name. During expansion
range ('name'_offs: 'name'_offs+'name'_leng)
character*('name'_leng) 'name' !'name' will be replaced by
end range !the actual argument
#endmacro !(it may be empty)
item menu !call the macro and replace 'name' by menu
item mdb !call the macro and replace 'name' by mdb
item init_appl !call the macro and replace 'name' by init_appl
....
With the parameter definition, you can define parameter values,
that can be used later in other parameter definitions, or in
field declarations.
Syntax:
PARAMETER [(]name=expression[,name=expression] [)]
VALUE [(]name=expression[,name=expression] [)]
TYPE name=expression
PARAMETERs may only be defined once in a description,
VALUEs may occur multiple times in a description
TYPEs may be used for fieldnames. See below
All can be referenced, but none can be updated in interactive mode.
The expression is first evaluated at read time. If this fails
(i.e. because of dependencies on field data), DIX will retry the
evaluation at expand time (when the record data is present)
You may not redefine a parameter within a description, but you
may redefine a value
Examples:
PARAMETER VALUE1 = 121 ! Will be evaluated at read time
PARAMETER VALUE2 = (VALUE1*4+12) ! Will be evaluated at read time
INTEGER*2 Test_field
PARAMETER (p1=10) ! Will be evaluated at read time
PARAMETER (p2=test_field*10) ! Will be evaluated at expand time
PARAMETER (p3=F$FILE("RECORDL")) ! Will be evaluated at read time
The value is evaluated as an expression.
in the help about "interactive_commands expressions"
![Expressions]DIX INTERACTIVE_COMMAND expressions
Parameters may also be defined globally. See the interactive command
ADD PARAMETER. ADD VALUE, ADD TYPE and LOAD
Example for types:
Suppose you want to have a text name for two specific values
You could specify
INTEGER*4 MYNUMBER [0=zero,5=five]
But if the string is used more often you can also use
TYPE nice_name ="0=zero,5=five" !declare a type
INTEGER*4 MYNUMBER [#nice_name] !use a name starting with a #
These types can be refenced in an expression by
#[\filetag[\descriptiontag]]typename"fieldname"
Suppose the definition is in a description with tag MYDES and a file MYFILE
All of the below will return the value (integer) 5
#nice_name"five"
#myfile\nice_name"five"
#myfile\mydes\nice_name"five"
You can use the symbol table to conditionally use parts of the
description. The #IF(DEF), #ELSE, #ELSEIF and #ENDIF are available to
dis/enable parts of the description.
The syntax is
#IF Logical-statement
#IFDEF symbolname
#ELSEIF Logical-statement
#ELSE
#ENDIF
The Logical-statement is evaluated using the available DIX-symbols and
field variables.
You can define the symbols in interactive mode via the normal
assignments and in other modes via the /DEFINE=symbol qualifier
Example (the indentation is just for visibility):
#IF F$EXISTS("LANGUAGE")
#IF LANGUAGE="DUTCH"
CHARACTER*20 DUTCH_NAME
#MESSAGE "The Dutch version"
#ELSEIF LANGUAGE="GERMAN"
CHARACTER*20 GERMAN_NAME
#MESSAGE "The German version"
#ELSE
CHARACTER*20 DEFAULT_NAME
#MESSAGE "The default version"
#ENDIF
#ELSE
CHARACTER*20 DEFAULT_NAME
#MESSAGE "The default version"
#ENDIF
If you change a symbol that would affect the #IF/#ELSEIF clauses,
you need to force DIX to re-evaluate the description.
You can do that via EXAMINE/EXPAND
#MESSAGE text Display message during expansion
#COMMENT text Display text in the fields display
INCLUDE 'filename[(module)]'
LOAD file[/module]
#INSERT NAME [expression]
Add fieldnames with values computed from other fields. This fields
will be displayed with offset "blank", and cannot be modified.
The expression can contain fields, parameters, and values from the description,
but not symbols, global values/parameters.
Example:
In the INDEXF description, the ODS-2 filename is split in two parts.
at offset 82 are the first 20 character of the filename (field FNAM)
and at offset 134 the last 66 characters (field REST_FNAM)
You could specify:
#INSERT TOTAL_FILENAME [FNAM+REST_FNAM]
And DIX would add a field with the name TOTAL_FILENAME and as contents
the total filename (all 86 characters). This fieldname can be used
and displayed in expressions.
If you i.e. specified #INSERT TOTAL_FILENAME [F$TRIM(FNAM+REST_FNAM)]
DIX would trim the trailing blanks
This allows you to include a file (or a module in a file)
Since this is a typical Fortran statement, Fortran defaults are used
formats
INCLUDE 'filename' Include a file, default filename = SYS$DISK:[].FOR
INCLUDE '(module)' Include a module from a file, default library
is SYS$LIBRARY:FORSYSDEF.TLB
INCLUDE 'library(module)' Include a module from a file, in .TLB file library
Form this file /module ONLY the parameters will be loaded
See also the help about interactive mode LOAD
LOAD file Include a file, default filename = SYS$DISK:[].TXT
LOAD library/module=modnam Include module from a library. the library
will be searched with a default name of
SYS$LIBRARY:STARLET.MLB and if that fails
SYS$LIBRARY:FORSYSDEF.TLB
Form this file /module ONLY the parameters will be loaded
See also the help about interactive mode LOAD
You can include the #MESSAGE statement in the description
#MESSAGE text
The text will be displayed via a message when the expander finds
this statement. This can be useful with the #IF constructs
see example below.
declaration statement
declaration statement
...
#IF F$EXISTS("_DEBUG")
#MESSAGE Debug variant
integer*4 temp
#ELSE
integer*4 temp /nodisplay
#ENDIF
If the symbol debug is defined the value of temp will be displayed
and during the expansion of the descriptions the message "Debug variant"
will be displayed on the terminal
If the symbol debug is not defined, the value of temp will NOT be
displayed and no message will occur.
#COMMENT text
This statement can be used in a description or a view
The text will be displayed with the examine command
You can use this to separate parts of the description or
in a view parts from different files
Example (for a view):
field verslim !examine the version limit
field name !examine the name
#comment ============To file header:=============
follow v(1).fileid !follow the link to the file header
loop !loop
field *_filename !examine the name in ODS-5 part
field fnam !and in ODS-2 part
test (file_id <> backl_fid) !stop if fid=backl_fid
#comment ============Backlink:=============
follow backl_fid !follow link to parent
end loop !end of loop
You can use any normal Fortran declaration, with some extensions
General format
TYPE*(SIZE)[/qualifiers] fieldname[(dim1[,dim2[,dim3]])]
The qualifiers must be specified BEFORE the fieldname.
The following types are supported
The {*n} means an implied size. You may specify the size, but if you
do not, this size is assumed. You cannot specify a different size.
REAL_F{*4},REAL_D{*8},REAL_G{*8},REAL_H{*16}
The data types are the VAX realtypes
REAL_S{*4}, REAL_T{*8},REAL_X{*16}
The data types are the IEEE realtypes
REAL*4,REAL*8,REAL*16
These are aliases, the actual datatype depends on
the size and the architecture. The default size = 4.
Replacement according to the following table
Size
Arch 4 8 16
VAX REAL_F REAL_D REAL_H
Alpha REAL_F REAL_G REAL_X
IA64 REAL_S REAL_T REAL_X
COMPLEX*8, COMPLEX*16
The complex datatype (real,imaginary).
the complex*8 contains 2 real*4 values
the complex*16 contains 2 real*8 values
LOGICAL*nn (nn=1..128 in bitfieldmode, 1..16 in normal mode)
The date types are the logical data types
Default size = 4(bytes) in normal mode,
8(bits) in BITFIELDMODE
RLOGICAL*nn (nn=1..128 in bitfieldmode, 1..16 in normal mode)
The date types are the logical data types, only the
value is reversed
Default size = 4(bytes) in normal mode,
8(bits) in BITFIELDMODE
CHARACTER*NN or CHARACTER*(NN)
The data types are the character datatypes
Default size = 1 (byte)
[U]INTEGER*nn (nn=1..64 in bitfieldmode, 1..8 in normal mode)
The data type is [un]signed integer
Default size = 4(bytes) in normal mode,
8(bits) in BITFIELDMODE
[U]BYTE is accepted as alias for [U]INTEGER*1
[U]WORD is accepted as alias for [U]INTEGER*2
[U]LONGWORD is accepted as alias for [U]INTEGER*4
[U]QUADWORD is accepted as alias for [U]INTEGER*8
BITS*NN (nn=1..128 in bitfieldmode, 1..16 in normal mode)
The data structure is a bit stream, nn<=8 Bytes
Default size = 4(bytes) in normal mode,
8(bits) in BITFIELDMODE
RBITS*NN (nn=1..128 in bitfieldmode, 1..16 in normal mode)
The data structure is a bit stream, nn<=8 Bytes
This is the same as BITS, but the meaning of each
bit is reversed (0 means active).
Default size = 4(bytes) in normal mode,
8(bits) in BITFIELDMODE
For the integer and bits fields, you may add a list of names to make
the display more readable, see the name_syntax help below
DATE*4 The date structure is an internal date format, the number of
minutes since 1857 e.g. VMSdate/(10*1000*1000*60)
No seconds or ticks (hundredths) are allowed
DATE*8 The date structure is a OpenVMS date format. This is
the default date data type.
UDATE{*4} The date structure is a UNIX time
(#seconds since 1-JAN-1970).
No ticks (hundredths) are allowed
UIC{*4} The data structure is a UIC
CPUTIME{*4} The data is a 32 bits integer containing CPU ticks.
This will be displayed as a delta time.
RAD50{*2} The data is a 16 bit integer contains 3 rad50 chars
IDENTIFIER{*4} The data structure is an OpenVMS-identifier
RFA{*6} The normal RFA structure. It contains a 2-byte
record-id/byte offset and a 4 byte blocknumber.
The display is (block,recid)
BKTRFA*NN (nn=4,5,6) The data structure is an RFA as found in the indexed
file buckets. It contains a 2-byte record-id/byte offset
and a 2-4 byte blocknumber. The display is (block,recid)
PROTECTION{*2} The data structure is a protection mask.
RINTEGER{*4} The data is a INTEGER*4 with low and high 16 bits
swapped (PDP-11 compatible format)
FILEID{*6} The data is a Fileid (filenr,revnr,volnr)
PRIVILEGE*nn The data is a privilege mask
The size may be 4 or 8, the default=8
If the size is 8, the value is a normal OpenVMS privilege mask
If the size is 4, the value is a privilege mask used for
functions like $CHKPRO (the $CHPDEF privilege mask)
This type has a built-in fieldname list.
DISKMAP The data is an ODS-2/5 storage description found in file headers
There are 4 formats.
P:dd dd placement map
T:1 begin-end 8 bits count and 22 bits blocknumber (4 bytes)
T:2 begin-end 14 bits count and 32 bits blocknumber (6 bytes)
T:3 begin-end 30 bits count and 32 bits blocknumber (8 bytes)
If you do not specify the t:n when converting from text to
binary DIX will take the shortest format
ACE The data is the binary ACE description
The text must include the leading and trailing ()
ACL The data is a list of ACEs
VFC{*1} The data is interpreted as a carriage control byte
in a DCL VFC file. See the OpenVMS docs about the layout.
BYTEARRAY*(SIZE) The data is a list of bytes, and the display is a
line with the byte values separated by a space
String types
STRING [*NN] The data is a counted ASCII string (count=byte)
WSTRING[*NN] The data is a counted ASCII string (count=word)
LSTRING[*NN] The data is a counted ASCII string (count=longword)
ZSTRING[*NN] The data is zero terminated string
HSTRING[*NN] The data is terminated by any char >127 (bit 7 = 1)
If *NN is specified, the strings are variable in a fixed allocation size.
The total allocation is NN (HSTRING)
NN+1 (STRING,ZSTRING)
NN+2 (WSTRING)
NN+4 (LSTRING)
If *NN is not specified the allocation is variable also.
DIX can handle display and modification of variable fields, but only if
the record length can be changed (indexed or relative files and not fixed).
DECIMAL strings
[U]DECIMAL The data is an (unsigned) decimal string. Depending on
some qualifiers the allowed data is 0..9,".",E,+,-
0..9 are always allowed.
If the type is UDECIMAL, the + and - are not allowed.
If the type specifies /fraction, the . is allowed
If the type specifies /exponent, the E is allowed
For this datatype there are some extra qualifiers defined
/ZERO_FILL The field is filled with leading zeros. All spaces
will be replaced by 0.
/FRACTION Allow the string to contain a decimal fraction.
This qualifier adds the "." character to the allowed
character set.
/EXPONENT Allow the string to contain an exponent.
This qualifier adds the E/e and the +/- for the
exponent (E+12 or e-12) characters to the allowed
character set.
/LEFT_JUST If the /ZERO_FILL is not specified, this qualifier
decides if a short string must be left or right
justified.
Fieldnames are strings up to 32 characters. They must start with a
letter (a-z), and the following character may contain a letter (a-z),
a digit (0-9), a $ or an _.
Field names starting with a % and PAD fields are not displayed (fillers).
These fields cannot be used for field-dependent computation.
The display can also be disabled with the /NODISPLAY option, and these
fields CAN be used.
Qualifiers can be
/HEXADECIMAL Always display data in hexadecimal format
/BINARY Always display data in binary format
/OCTAL Always display data in octal format
/RADIX=nn Always display the data in radix nn format
This can only be used for [u]integer fields
/RANGE=min_val:max_val For integer values, limit the range of the value
/NODISPLAY Do not display data (can be referenced)
/USER The type is a user defined type. See the help about
[DIX/HELP] record_for field usertypes.
/UPPERCASE Force conversion to uppercase on field modification
/LOWERCASE Force conversion to lowercase on field modification
/LIMIT=value If the value is reached, the repeat loop is terminated
/LIMIT=>value If the field has a repeat, this will terminate, else
/LIMIT=<value DIX will terminate the next structure repeat
See example
/VFC This field is in the VFC area (if present) of the record
/RELATIVE Qualifier for POSITION (see there)
/READONLY This field is read-only. DIX cannot modify it.
/EBCDIC This field contains EBCDIC data.
DIX will convert to and from ASCII. This will only
work correctly for characters that are in
both character sets.
You can also open the file in EBCDIC mode. This
will automatically append a /EBCDIC to all fields
This conversion only takes place in char/string
and decimal fields.
/DEPENDENCY If this field changes, DIX must re-expand the description
since there is a dependency somewhere on this value.
DIX allows you to mark a field as an index into another file
To do this you need to specify at least one of the following qualifiers
The value of the file will be used as a value for a
key/rfa/recnr/byteoffset in the target file.
/KEY, /RECORD, /RFA, and /BYTEOFFSET are mutually exclusive.
/FILE=filename Specifies that this field is a key into file
or 'filename'. The default is the same file.
The filename can also be a fileid. See the help
about [DIX/HELP] record_format parameters.
/FILE="expression" If you specify the name in "", it will be
interpreted as a string to be evaluated. In this
way you can compute the name.
/DESCRIPTION="descr" The explicit description used for this file.
/KEY=keynumber Specifies the keynumber (default=0)
/RECORD Specifies that the value is a record number into
'filename'. If used on an indexed file, it will
read record 'computestring' on the primary key.
/RFA Specifies that the value is an RFA into the target
file.
/BYTEOFFSET Specifies that the value is a byte-index into the
target file (value 0 = byte 0 in block 1)
/MATCH=LT,LE,EQ,GE,GT Keymatch, the default is EQ search.
/COMPUTE="compstring" Allows you to do computation on the field contents
the compute string must be a valid expression that
delivers an integer or character value to be used
as key (for indexed files) or record number.
/IF="IFstring" If specified the IFstring expression must be TRUE
to take this link.
During the evaluation of the IF, COMPUTE and FILE string the normal
symbol substitution is active, and also the special symbols
$FILE (the current file)
$DESCRIPTION (The current description)
$FIELD (The contents of the current field converted to text)
$FLD (the current field, may be any type)
can be used (as any other symbol/function/parameter)
See the examples for this
See also the radix_override help below
Any variable can be followed by a dimension description up to 3 dimensions.
Dimension are specified as
high (low=1, incr=1) or
Low:high (incr=1) or
low:high:incr incr may be negative, but then low must be >= high
The sizes and dimensions can be expressions with numbers, parameters or
other fieldnames in the description. If fieldnames are used, they must
appear before the line in which they are referenced.
see the help about expressions
![expressions]DIX INTERACT expressions
With the (R)BITS*(*),(U)INTEGER*(*),LOGICAL*(*) and the (u)DATE*(*)
declarations you can append a list of names to the description to
give a readable interpretation of the values
There are two formats
1. [field,field,,field...] A list of fields numbered from 0 up to nn
Fields may be omitted, but commas are
significant. Mostly used for bits. A #
means do never display this bit even if set.
2. [idx=name,idx=name,...] A list of numbered names, the indexes
do not have to be in order. Mostly used
for integers. The idx's are integer*4 only,
if the values are 8 bytes wide, DIX will check
if the high longword is a sign extension of the
low longword.
The text can also be stored in a parameter.
if you specify [#type_name], the actual text will be searched
in type "type_name". It must be defined earlier via the ADD TYPE command
This can be useful if the same strings happens more than once
The names (field or name) can have aliases. They are specified as
name1;name2;name3...
All alias names can be used for text to binary translation, but only the
first alias will be used for binary to text translation. See example below
Example for case 1 (the privileges in SYSUAF):
BITS*8 UAF$Q_DEFPRIV -
[CMKRNL,CMEXEC,SYSNAM,GRPNAM,ALLSPOOL,IMPERSONATE,DIAGNOSE,-
LOG_IO,GROUP,NOACNT,PRMCEB,PRMMBX,PSWAPM,SETPRI,SETPRV,TMPMBX,-
WORLD,MOUNT,OPER,EXQUOTA,NETMBX,VOLPRO,-
PHY_IO,BUGCHK,PRMGBL,SYSGBL,PFNMAP,SHMEM,SYSPRV,BYPASS,-
SYSLCK,SHARE,UPGRADE,DOWNGRADE,GRPPRV,READALL,-
IMPORT,AUDIT,SECURITY]
bit(0) = CMKRNL, bit(1)= CMEXEC ...
Bits are named in order, but bits may be skipped.
If the non-named bits are in the mask they will be displayed with
the text "BITnn". If any bit has the name # it will not be
displayed (see below)
A more complicated example for the bits in the file characteristics
in the file header in the INDEXF.SYS
The longword has a lot of bits defined, but also a 2-bit field that
contains the setting for the caching options
First we define all normal bits, and the 2-bit field is defined
with #, so they will not be displayed (and cannot be modified using
the bitnames). Then we skip 24 bits back and define a 2-bit
integer with 4 possible values, and then we skip 22 bits forward again
Also notice the ,, after the #, this signifies an unused bit.
BITS*4 FILE_CHAR -
[Wascontig,Nobackup,Writeback,Readcheck,Writecheck,-
Contigb,Locked,Contig,#,#,,Badacl,-
Spool,Directory,Badblock,Markdel,Nocharge,Erase,alm_aip,-
shelved,scratch,nomove,noshelvable,shelv_res]
bitfield !start in bits mode, the next 2 bits define the caching
position/relative (-24) !skip back 24 bits, the /relative
!is optional since the value is negative
integer*2 caching [writethrough,writeback,flush_on_clos,no_cache]
position /relative (22) !skip the next 22 bits(in total 3 bytes)
end bitfield !back to non-fieldmode again
Example for case 2:
INTEGER*2 type [1=special,4=normal,10=fatal]
Field values 1,4,10 will be by displayed as a name, all others in
numeric
Example for specification via a TYPE
TYPE (user_type = "1=special,4=normal,10=fatal") !must be a string
.... !scalar type
INTEGER*2 type [#user_type] !reference to parameter user_type.
Example for name aliases:
In the SYSUAF description is the UAF$L_FLAGS field. One of the bits
defines if the user is allowed to login. This bit has two different names
1. In AUTHORIZE this bit is referenced by the /[NO]DISUSER name.
2. In the $UAIDEF include file this bit is called UAI$M_DISACNT
So the names for UAF$L_FLAGS contains an alias for this bit
BITS*4 UAF$L_FLAGS -
[DISCTLY,DEFCLI,LOCKPWD,RESTRICTED,DISUSER;DISACNT,DISWELCOME,DISMAIL,NOMAIL,
GENPWD,PWD_EXPIRED,PWD2_EXPIRED,AUDIT,DISREPORT,DISRECONNECT,AUTOLOGIN,-
DISFORCE_PWD_CHANGE,CAPTIVE,DISIMAGE,DISPWDDIC,DISPWDHIS,DEFCLSVAL,EXTAUTH,
MIGRATEPWD,VMSAUTH,DISPWDSYNCH,PWDMIX]
The bit has the name DISUSER;DISACNT.
On text to binary you may use either name (DISUSER or DISACNT).
On binary to text DIX will use the first name (DISUSER)
For all fieldtypes or normal values containing multiple data,
RBITS,BITS,CHARACTER,xSTRING,(U)INTEGER,BYTEARRAY,COMPLEX,
UIC,(U)DATE,DELTATIME,CPU,PROTECTION,FILEID,RFA,BKTRFA,RAD50,ACL
DIX will normally display all data on one line, and
lets you examine/modify the whole structure in one time.
Sometimes it may be useful to specify a specific subfield of the whole
structure. Below is a table of transformation
Subfields can be used with the
a. examine command (fieldnames)
This will display a multiple line display with each subfield
on a separate line.
b. Evaluation of fieldnames/symbols.
This will deliver a table containing the subfield values
For the following type you can only specify a bit/bytenumber
CHARACTER,xSTRING byte-size character 1..len(string)
BYTEARRAY byte-size byte 1..len(bytearray)
(U)INTEGER one-bit integer 0..#bits-1
RAD50 byte-size character 1..3
ACL The ACE in the ACL
For the following types you can specify a list of bit values via the
fieldnames construct [name1,name2,name1]
RBITS LOGICAL
BITS LOGICAL
The following fields have predefined subfields
U(DATE) Integer YEAR,MONTH,DAY,HOUR,MINUTE,SECOND,HUNDREDTH
DELTATIME,CPU Integer DAY,HOUR,MINUTE,SECOND,HUNDREDTH
PROTECTION Integer SYSTEM,GROUP,OWNER,WORLD
FILEID Integer FILENR,REVISION,VOLUME
RFA Integer BLOCK,OFFSET
BKTRFA Integer RECID,BUCKET
PRIVILEGE[*8] Logical The normal 39 privilege names
PRIVILEGE[*4] Logical The 12 privileges in the audit log
COMPLEX Real REAL,IMAG
You may specify a list of either
a. subfield name masks (between "")
If the name mask does not contain a wildcard (* or %), DIX will
match the beginning of the keyword, this may match more than one name.
If the name mask does contain a wildcard, DIX will select all names
matching the wildcard.
or
a. element numbers
b. ranges like 1:4
c. ranges with stride like 1:5:2 or 5:1:-2
Suppose you have define the following fields
RBITS*7 UAF$B_PRIMEDAYS - !value=[t,t,t,t,t,f,f]
[MONDAY,TUESDAY,WEDNEDSAY,THURSDAY,FRIDAY,SATURDAY,SUNDAY]
CHARACTER*32 UAF$T_ACCOUNT !value=TACV
You can now specify
UAF$B_PRIMEDAYS{"*T*"} to reference the three one-rlog values
514.1|UAF$B_PRIMEDAYS{TUESDAY} |True
514.3|UAF$B_PRIMEDAYS{THURSDAY}|True
514.5|UAF$B_PRIMEDAYS{SATURDAY}|False
UAF$T_ACCOUNT{3:5,1}
54.0|UAF$T_ACCOUNT{3}|C
55.0|UAF$T_ACCOUNT{4}|V
56.0|UAF$T_ACCOUNT{5}|
54.0|UAF$T_ACCOUNT{1}|T
But also "TEST"{*} will deliver a table of 4 characters
and 10{*} will deliver a table of 32 bits (if integersize=4)
The syntax between the {} is
"*XYZ*": a mask. This only works on fields with fieldnames defined
"AB" : Match all names beginning with AB
"*" : All elements with a fieldname
* : All elements
4-5 : Element 4..5
4-* : element 4..end_of_field
*-4 : element begin_of_field..4
12 : Element 12 only
1:5:2 : Element 1,3 and 5.
5:1:-2 : Element 5,3 and 1
Subfields can be used on fields (you may use the field value names)
or normal symbol values
a byte wide integer
DIX> exa UAF$B_RTYPE
0.0|UAF$B_RTYPE|1
DIX> exa UAF$B_RTYPE{*}
0.0|UAF$B_RTYPE{0}|1
0.1|UAF$B_RTYPE{1}|0
0.2|UAF$B_RTYPE{2}|0
0.3|UAF$B_RTYPE{3}|0
0.4|UAF$B_RTYPE{4}|0
0.5|UAF$B_RTYPE{5}|0
0.6|UAF$B_RTYPE{6}|0
0.7|UAF$B_RTYPE{7}|0
A UIC
DIX> exa UAF$L_UIC
36.0|UAF$L_UIC|[100,1515]
DIX> exa UAF$L_UIC{"*"} !display all fields
36.0|UAF$L_UIC{GROUP} |00000001515
38.0|UAF$L_UIC{MEMBER}|00000000100
DIX> exa UAF$L_UIC{"MEM*"} !display fields matching MEM*
38.0|UAF$L_UIC{MEMBER}|00000000100
A data field
DIX> exa UAF$Q_LASTLOGIN_N !the normal way
404.0|UAF$Q_LASTLOGIN_N|10-JUL-2008 17:20:01.30
DIX> exa UAF$Q_LASTLOGIN_N{*} !examine all subfields
404.0|UAF$Q_LASTLOGIN_N{YEAR} |2008
404.0|UAF$Q_LASTLOGIN_N{MONTH} |7
404.0|UAF$Q_LASTLOGIN_N{DAY} |10
404.0|UAF$Q_LASTLOGIN_N{HOUR} |17
404.0|UAF$Q_LASTLOGIN_N{MINUTE} |20
404.0|UAF$Q_LASTLOGIN_N{SECOND} |1
404.0|UAF$Q_LASTLOGIN_N{HUNDREDTH}|30
DIX> exa UAF$Q_LASTLOGIN_N{"*"} !examine all fields matching *
404.0|UAF$Q_LASTLOGIN_N{YEAR} |2008
404.0|UAF$Q_LASTLOGIN_N{MONTH} |7
404.0|UAF$Q_LASTLOGIN_N{DAY} |10
404.0|UAF$Q_LASTLOGIN_N{HOUR} |17
404.0|UAF$Q_LASTLOGIN_N{MINUTE} |20
404.0|UAF$Q_LASTLOGIN_N{SECOND} |1
404.0|UAF$Q_LASTLOGIN_N{HUNDREDTH}|30
DIX> exa UAF$Q_LASTLOGIN_N{"*U*"} !examine all fields matching *U*
404.0|UAF$Q_LASTLOGIN_N{HOUR} |17
404.0|UAF$Q_LASTLOGIN_N{MINUTE} |20
404.0|UAF$Q_LASTLOGIN_N{HUNDREDTH}|30
Fieldname and symbols may be used in evaluations with subnames.
DIX> examine UAF$Q_LASTLOGIN_N
404.0|UAF$Q_LASTLOGIN_N|10-JUL-2008 17:20:01.30
DIX> eval UAF$Q_LASTLOGIN_N
10-JUL-2008 17:20:01.30
DIX> eval UAF$Q_LASTLOGIN_N{*}
[2008,7,10,17,20,1,30]
DIX> eval UAF$Q_LASTLOGIN_N{"*U*"}
[17,20,30]
Now for a symbol
DIX> a=UAF$Q_LASTLOGIN_N !contains a date value
DIX> eval a
10-JUL-2008 17:20:01.30
DIX> eval a{*}
[2008,7,10,17,20,1,30]
DIX> eval a{"*U*"}
[17,20,30]
DIX> exa UAF$Q_LASTLOGIN_N
404.0|UAF$Q_LASTLOGIN_N|10-JUL-2008 17:20:01.30
DIX> dep UAF$Q_LASTLOGIN_N{"DAY"}=23
DIX> exa UAF$Q_LASTLOGIN_N
404.0|UAF$Q_LASTLOGIN_N|23-JUL-2008 17:20:01.30
DIX> exa UAF$Q_LASTLOGIN_N
404.0|UAF$Q_LASTLOGIN_N|10-JUL-2008 17:20:01.30
DIX> UAF$Q_LASTLOGIN_N{"DAY"}=23
DIX> exa UAF$Q_LASTLOGIN_N
404.0|UAF$Q_LASTLOGIN_N|23-JUL-2008 17:20:01.30
DIX> a=UAF$Q_LASTLOGIN_N
a{"DAY"}=25
DIX> say a
23-JUL-2008 17:20:01.30
Or with much more fun
DIX> a{2:7}=1 !change all subfields except the year
DIX> say a
1-JAN-2008 01:01:01.01
DIX> A{4:7}=[1,2,3,4] !table shape must be equal (both have 4 elements)
DIX> say A
1-JAN-2008 01:02:03.04
DIX> UAF$Q_LASTLOGIN_N{4:7}=[1,2,3,4] !table shape must match
DIX> exa UAF$Q_LASTLOGIN_N
404.0|UAF$Q_LASTLOGIN_N|23-JUL-2008 01:02:03.04
DIX> UAF$Q_LASTLOGIN_N{"*U*"}=[10,11,12] !table shape must match
DIX> exa UAF$Q_LASTLOGIN_N{"*U*"}
404.0|UAF$Q_LASTLOGIN_N{HOUR} |10
404.0|UAF$Q_LASTLOGIN_N{MINUTE} |11
404.0|UAF$Q_LASTLOGIN_N{HUNDREDTH}|12
DIX> exa UAF$Q_LASTLOGIN_N
404.0|UAF$Q_LASTLOGIN_N|23-JUL-2008 10:11:03.12
The user can define his own types. The translations between ASCII and
binary must be programmed in the DIX interactive language.
This translation script can be defined in the following ways.
1. In the .DES file
Somewhere in the .DES file there must be a section the contains the
script code. This section must begin with #TYPE name and end with #ENDTYPE
2. In a module in the DIX_DES.TLB (either user or system) library.
The module name must be USERTYPE_'name'. In this case the
#TYPE and #ENDTYPE tags are not necessary.
This script is called with 4 symbols defined
1. ACTION : a string symbol that defines the action to take, values can be
SIZE : Return the size (in bits) of the binary data.
(only used in variable length data)
ASCBIN : Convert from ASCII to binary
BINASC : Convert from binary to ASCII
2. DATA : A string variable containing the source data.
In case "SIZE" and "BINASC" DATA contains the binary data
In case "ASCBIN" DATA contains the ASCII text.
In all cases the length of the string is the size in bytes.
3. BITSIZE: An integer value that contains the size in bits. Normally this
is 8*length(data), but not for bitfields.
4. FIELDS : The fields as defined in the declaration statement.
The script must set the following symbols
Case "SIZE" : SIZE, an integer containing the size in BITS.
case "ASCBIN" : RESULT, a string variable containing the binary data.
SIZE, an integer value containing the size in BITS.
case "BINASC" : RESULT, a string variable containing the ASCII text.
Also the script must finish with an EXIT statement that returns the status.
For BINASC 0 means the text will be suppressed by the /compress qualifier.
1 means is not regarded as "suppressible"
For others 1=ok, 2=error.
Example:
You have a line in the description record:
mytype*8/user fieldname !user type mytype, size=fixed to 8.
Now we must have either a section in the description record enclosed
in the following statements:
#type usertype
...
#endtype
Or a module in the DIX_DES or DIX_DES_USER file with the name USERYTPE_MYTYPE
In this case the #type and #endtype are not needed.
Below an example of a script that will work. It will convert a 8 byte
string to a sequence like 12,34,56,43,123,0,56,9
#type mytype
if (action = "SIZE") then !return symbol SIZE (size in bits)
size = 8*8 !will not be called since the size is
stat = 1 !fixed at 8 bytes (is 8*8 bits)
endif
if (action = "BINASC") then !return symbol RESULT
nk = f$length(data)
stat = 0 !assume all bytes are ZERO
result = ""
for k=1,nk
kar = f$extract(k-1,1,data)
byt = f$extzv(0,8,kar)
if (byt <> 0) stat = 1 !not all bytes ZERO
result = result + "," + string(byt)
endfor
result = f$extract(1,-1,result)
endif
if (action = "ASCBIN") then !return symbol RESULT and SIZE
set noon
result = ""
k = 0
stat = 2 !assume error in conversion
lp2: part = f$element(k,",",data)
if (part <> ",") then
byt = int(part)
if (even($STATUS)) goto err
if (byt < 0 | byt > 255) goto err
result = result + f$char(byt)
k = k + 1
goto lp2
endif
size = f$length(result)*8 !return length in bits
stat = 1 !successful convert
err:
endif
exit 'stat' !return conversion result
#endtype
Normally all fields are displayed in decimal or text. If the fieldtype is followed by the a radix qualifier the display (and also the input is in the requested radix) The following radices are supported /HEXADECIMAL /OCTAL /BINARY Example: INTEGER*4 FIELD Will be displayed and modified in decimal INTEGER*2/BIN DATA Will be displayed and modified in binary INTEGER*2/OCT DATA Will be displayed and modified in octal CHARACTER*3/HEX TMP Will be displayed and modified in hexadecimal
DIX also supports a number of control statements
STRUCTURE,ENDSTRUCTURE After the structure statement follows a name
optionally followed by a dimension (dim1,dim2,dim3)
If the name is specified the fieldname is
extended by .structurename. if no name is specified,
the fieldname is not extended
UNION,ENDUNION
MAP,ENDMAP See MAP_SYNTAX below.
RANGE,ENDRANGE See RANGE_SYNTAX below
BLOCK,ENDBLOCK See RANGE_SYNTAX below
POSITION[/RELATIVE] (value) Set the bit/byte offset to the value.
The () must be specified.
This can be useful in some overlay structures
See also the example in RANGE_SYNTAX. The value
is the new byte position, unless in a bitfield,
where the value is in bits.
If the /relative is given (or the value is negative)
the value is added/subtracted to/from the current
offset, else it new offset=value.
POINTER Temporary override the bit offset for the
following structure. See pointer_syntax below
PAD[*NN] Round up bit/byte-offset to NN fold. nn=2,4,8
The default in bitmode 8 (align to next byte)
else 2 (align to next word)
Rounding takes 0..NN-1 bytes/bits.
The verb ALIGN is a synonym for PAD
The /VFC or not decides where the pad is done
PAD PACKED|NATURAL Set a policy for padding/aligning.
The verb ALIGN is a synonym for PAD
PACKED : Align from now on fields on bit/byte boundary
NATURAL: Align from now on fields on its natural size
i.e. INTEGER*4 will be aligned on a 4 byte boundary
The PAD*nn overrules this setting for the next
field only.
EXIT [condition] If condition is not specified, it will be evaluated
true. If condition is true, skip the repeat of the
current structure (there must be one)
See exit_syntax below
ABORT reason [condition] If the abort statement is seen, DIX checks if
condition evaluates to TRUE (or is not present).
If this is the case, the description is regarded
as not valid. This statement can be used to do
some sanity checking. See the examples help
for an example.
DEFAULT INTEGER size Set the default integer size.
Size can be 4,8,32 or 64
DEFAULT REAL size Set the default real size
Size can be 4,8,16,32,64 or 128
You can also define your own types. See the help about
[DIX/HELP] record_form field usertypes.
BITFIELD
ENDBITFIELD
Normally the number after the * will be interpreted as bytes, but
between a BITFIELD and an ENDBITFIELD the numbers will be interpreted
as bits. Only (U)INTEGER, (R)BITS, LOGICAL and PAD fields are allowed between
BITFIELD and ENDBITFIELD. After an ENDBITFIELD the record pointer will
be padded to the next byte.
Case construct. Select one of the case constructs.
SWITCH/CASE/ENDSWITCH See the help about the switch_case_endswitch
The record descriptions can contain one or more RANGES or BLOCKS
RANGE[/RELATIVE] (min_offset:max_offset)
field
field
...
END RANGE
BLOCK[/RELATIVE] (start_pos:end_pos)
field
field
...
END BLOCK
The () are required syntax.
The fields start at offset "min_offset" and continue until the offset
"max_offset". If the byte_offset is >= max_offset, the rest of
the RANGE is ignored. RANGE and BLOCK are very much the same, but
after the ENDRANGE is found, the bit offset is restored to the value
before the RANGE, and the BLOCK sets the bit offset to end_pos.
So if the range/block is in another part of the record you should use
RANGE, and if the range/block is inline you should use BLOCK
If /relative is specified, the start_pos and end_pos are added to the
current bit_offset, otherwise they will set the bit_offset
An example is in the ODS-2 File headers:
byte id_offset ! Offset to Ident area
byte map_offset ! Offset mapping area
byte acl_offset ! Offset to ACL area
byte res_offset ! Offset to ACL area
integer*2 seg_num ! Extension segment number
....
range (map_offset*2:acl_offset*2-1)
diskmap maps(256) !enough
end range
.....
position (510)
integer*2/hex checksum
The DISKMAP entries start at byte offset MAP_OFFSET*2 (MAP_OFFSET is in
words) and continues until ACL_OFFSET*2-1.
The checksum is at fixed location 510
An example for the BLOCK statement is in the $BUCKET description for
a bucket in an indexed file:
...
uinteger*2 first_free_byte
...
block (14:first_free_byte) !this part extends from byte 14 ..the
!value in first_free_byte
structure sidrvalues(*) !we don't know how many sidr values there are
!but if we cannot exceed the block
integer*2 total_len !the total length of the sidrvalue
... !display the keyvalue (compressed or not)
block/relative (0:total_len+%loc("TOTAL_LEN")-%location+2)
!the sidrvalue record
!starts at offset 0 (relative) and continues
!until the length is exceeded
structure rrvs(*) !the rrvs are limited to the block
...
end structure !end of the rrvs
end block !end of the rrv block
end structure !end of the sidr values structure
end block !end of the total block
...
The MAP command has a special syntax. It is included in a UNION statement.
UNION
MAP [selectionstring]
END MAP
MAP [selectionstring]
END MAP
..
END UNION
If there is no selectionstring specified, DIX will show the mapping.
If there is a selection string, DIX will show the map only if the
selection string matches. The format for the selection string is:
Expression=sel_value[,sel_value...]
And sel_value is
1. expression Just a single (integer) expression
2. expression:expression A range of (integer) expressions
At this moment only integer expressions are allowed.
You can specify a dependency on a previously defined variable.
SYNTAX:
integer*4 sel !any field will do (dependency field)
UNION
MAP SEL=value !Display this map if SEL has value "value"
END MAP
MAP SEL=VALUE1 !Display this map if SEL has value "value1"
END MAP
..
END UNION
If all maps have a dependency and none of them is used, DIX will always
use the last one, unless one of the MAP entries has a dependency fieldname
of *
Example:
integer*4 sel !any field will do (dependency field)
UNION
MAP SEL = 0,10:12 !case for sel=0 and 10:12
character*4 valuec
END MAP
MAP SEL = 1 !case for sel=1
integer*4 value4
END MAP
MAP * !case for all other values of sel
logical*4 valuel
END MAP
MAP SEL = 2 !case for sel=2
integer*2 value2(2)
END MAP
END UNION
If SEL has value 0 or is between 10 and 12 the character VALUEC is displayed
If SEL has value 1 the INTEGER*4 VALUE4 is displayed
If SEL has value 2 the INTEGER*2 VALUE2(2) is displayed
If SEL has any other value the LOGICAL*4 VALUEL is displayed
If the MAP * had not existed, and SEL would not have been 0 or 10-12 or
1 or 2 the last map (SEL=2 : = VALUE2(2)) would have been used.
The pointer statement allows you to decode a piece of data in an other
place in the record.
Syntax POINTER[*n] [/RANGE=min:max] fieldname [pointee]
The next structure block (there must be a structure block directly
after this statement) will contain the fields that are part of the
pointed data.
The *n is optional, default *4
The /range is optional, if allows you to mark certain values of the
pointervalue as not valid, so POINTER/RANG=1: name [pointee] will
not be taken is the contents of the pointer is <1.
The [pointee] is necessary. If contains the fieldname, of which the
address will be added to the pointer value to get the correct offset
An example is the PENDING structure (pending.dat) for Allin1/Office server:
character*64 pending_key !MAIL + the name of the user
character*1 record_code !continuation record index
character*1 cont_flag ! flag for continuation record
integer*2 total_attr_size !total size of the pointers+data
! of all records together
integer*2 this_rec_attr !size of pointer+data in this record
character*2 usage_count !how often is this record used
character*5 new_mail_count !new mail count for the user
character*3 padding
structure pointers(100) [1] !start of pointer/data
integer*2 type [7=mail] !type of field
integer*2 %dummy
pointer*4 offs [pointers] [2] !offset to the "pointer" field
structure work [3] !the data section
integer*2 length !length of string
byte %unknown(6)
character*(length-6) test !the string
end structure [4] !end of the data definition
exit [offs+length+2>=this_rec_attr] [5] !exit statement
end structure
The record is built up as follows
a. 80 bytes of fixed data
b. a pointer section containing an array of
1. Type
2. Offset to a position in the record after the pointers
c. the text area pointed to by the offset
The statement pointer*4 offset[pointers] at [2] must be followed by a
structure statement [3].
DIX will take the value from the record at the position of the pointer
statement [2] (the length contains the size of the field), adds the
location of "POINTERS" [1], and uses that offset to expand the
statements in the structure "work" [3]. When the end structure [4]
is found, DIX will resume processing the next statement [5]
with the saved offset.
At the end [5] is the exit test. If the offs > cur rec length : exit
This means to leave the "pointers" structure. Since there is no more
data available, the display is stopped.
See the display below
0|PENDING_KEY |MAIL JOHN.DOE
64|RECORD_CODE |
65|CONT_FLAG |.
66|TOTAL_ATTR_SIZE |86
68|THIS_REC_ATTR |82
70|USAGE_COUNT |
72|NEW_MAIL_COUNT |3
77|PADDING |
80|POINTERS(1).TYPE |MAIL
84|POINTERS(1).OFFS |16
96|POINTERS(1).WERK.LENGTH|31
104|POINTERS(1).WERK.TEST |OA$SHARA7718:IFQ3FORG.TXT
88|POINTERS(2).TYPE |MAIL
92|POINTERS(2).OFFS |49
129|POINTERS(2).WERK.LENGTH|31
137|POINTERS(2).WERK.TEST |OA$SHARE2664:IFR1XUY1.TXT
96|POINTERS(3).TYPE |31
You could suppress the display of pointers(1).offs, by changing the
definition of pointer to :
pointer*4/nodisplay offs [pointers]
Another example is from the records of the accounting file:
BITFIELD
INTEGER*1 FLAG
INTEGER*7 TOTTYPE [,PRCDEL,PRCPUR,IMGDEL,IMGPUR,SYSINIT,SETTIME,LOGFAIL,PRINT,USER,ENABLE,DISABLE,ALTACM,FILE_FL,FILE_BL]
INTEGER*4 SUBTYPE [,INTERACTIVE,SUBPROCESS,DETACHED,BATCH,NETWORK]
INTEGER*3 VERSION
INTEGER*1 CUSTUMER
ENDBITFIELD
INTEGER*2 LENGTH
DATE SYSTIME
STRUCTURE PACKET(10) [1]
BITFIELD
INTEGER*1 PACKETFLAG
INTEGER*7 PACKETTYPE [,ID,RESOURCE,IMAGENAME,FILENAME,USER_DATA,,,PRINT]
INTEGER*4 SUBTYPE
INTEGER*3 VERSION
INTEGER*1 CUSTUMER
ENDBITFIELD
INTEGER*2 SUBLENGTH
UNION
MAP PACKETTYPE=1
INTEGER*4/HEX PID
INTEGER*4/HEX OWNER
UIC UIC
INTEGER*4 PRIVS(2)
BYTE PRIO
BYTE FLAGS
POINTER*2/RANGE=1: USERNAMEOFFSET [PACKET] [2]
STRUCTURE [3]
STRING USERNAME
END STRUCTURE [4]
POINTER*2/RANGE=1: ACCOUNTOFFSET [PACKET] [5]
STRUCTURE
STRING ACCOUNTNAME
END STRUCTURE
.... and some more pointers for other fields
END MAP
... for other packettypes
END UNION
ENDSTRUCTURE
The pointer statement at [2]
POINTER*2/RANGE=1: USERNAMEOFFSET [PACKET] [2]
will be processed as follows.
Take the value from the record, 2 bytes The (*2).
See if this value is at least 1 (the /rang=1:)
If not, skip the following structure [3], and continue with the next [5]
If it is >=1, add the value to the offset of PACKET [1]
and decode the data at that offset as the structure [3]
So we expect a STRING datatype there
(the structure contains only one field)
After this sidestep (we found the ENDSTRUCTURE [4]),
return to processing the data at [5]
The pointer statement allows you to terminate a structure repeat count
character*64 test
structure pointers(100) [1]
integer*4 type [7=mail]
integer*4 value
exit [type=10] [2]
end structure
character*10 rest
The record is built up as follows
a. 64 bytes of fixed data (test)
b. a structure section containing an array of
1. Type
2. value
c. 10 bytes trailer (rest)
The statement "exit [type=10]" at [2] will exit the *100 repeat for [1]
if the contents for type = 10.
0|TEST |MAIL 01CURSIST
80|POINTERS(1).TYPE |MAIL
96|POINTERS(1).VALUE |20
88|POINTERS(2).TYPE |MAIL
92|POINTERS(2).VALUE |25
96|POINTERS(3).TYPE |10 <- this terminates with the exit
100|POINTERS(3).VALUE |25
104|REST |The rest
If the description had been
character*64 test
structure pointers(100) [1]
integer*4 type [7=mail]
exit [type=10] [2]
integer*4 value
end structure
character*10 rest
The result would have been (POINTERS(3).VALUE at 100 is not displayed,
and so the offset for REST is 4 lower
0|TEST |MAIL 01CURSIST
80|POINTERS(1).TYPE |MAIL
96|POINTERS(1).VALUE |20
88|POINTERS(2).TYPE |MAIL
92|POINTERS(2).VALUE |25
96|POINTERS(3).TYPE |10 <- this terminates with the exit
100|REST |The rest
INTEGER*4 value[_expression]
SWITCH value
CASE 1 !case taken if value=1
integer*4 value1
character*20 text1
CASE 4 !case taken if value=4
integer*4 val2
logical*4 ext
CASE !default case if nothing matches
character*(20) default
CASE 1 !will not be found, since we had a
integer*4 not_used !match before
CASE !will not be found, since we had a
integer*4 not_used_def !match before
ENDSELECT
INTEGER*4 value value contains 4 bytes = 32 bits
BITFIELD
integer*4 first_nibble value contains 4 bits.
logical*1 first_bit[0=no,1=yes]
value contains 1 bit
PAD 2 Align to even bit (so this takes 1 bit)
integer*1 second_bit value contains 1 bit
ENDBITFIELD aligns on byte boundary (so 1 pad bits)
LOGICAL*1 temp value contains 1 byte.
CHARACTER*12/FILE=SYSUAF Username
You can follow this link to the file SYSUAF
keyvalue the contents of "USERNAME", KEY=0
INTEGER*4/FILE=DATAFILE/RECORD/COMPUTE="$field+10" index
You can follow this link to the file DATAFILE
with the record number the contents of
the field "index" + 10
INTEGER*4/FILE=SYSUAF/KEY=1/IF="$field>0" identifier
Follow the identifier to the SYSUAF file, but
only if the identifier is >0 (UIC identifier)
Example for not displayable fields:
The file_char bits in the INDEXF.SYS contain 2 bits that define the
caching attributes (bit 8 and 9). Since the names for these bits are
a #, DIX will not display these bits.
The POSITION (-24) : Reposition the byte_offset by 24 (bits).
The INTEGER*2 : Define a 2 bits integer with 4 possible values.
The POSITION/RELATIVE (22) : Sets the byte_offset back to where we were.
bits*4 file_char -
[Wascontig,Nobackup,Writeback,Readcheck,Writecheck,-
Contigb,Locked,Contig,#,#,,Badacl,-
Spool,Directory,Badblock,Markdel,Nocharge,Erase,alm_aip,-
shelved,scratch,nomove,noshelvable,shelv_res]
bitfield !start in bits mode, the next 2 bits define the caching
position/relative (-24) !skip back 24 bits, the /relative
!is optional since the value is negative
integer*2 caching [writethrough,writeback,flush_on_clos,no_cache]
position /relative (22) !skip the next 22 bits(in total 3 bytes)
end bitfield !back to non field mode again
This could also have been achieved by
structure !no name attached
union
map
bits*4 file_char -
[Wascontig,Nobackup,Writeback,Readcheck,Writecheck,-
Contigb,Locked,Contig,#,#,,Badacl,-
Spool,Directory,Badblock,Markdel,Nocharge,Erase,alm_aip,-
shelved,scratch,nomove,noshelvable,shelv_res]
end map
map
byte %temp1
bitfield
integer*2 caching [writethrough,writeback,flush_on_clos,no_cache]
end bitfield !will align back to the next byte
end map
end union
end structure
The following description is from the SYSUAF description. The UAF$B_RTYPE is expected to be 1. BYTE UAF$B_RTYPE ABORT wrongtype [UAF$B_RTYPE <> 1] .... rest of description If the UAF$B_RTYPE is <> 1, DIX will assume that the current record cannot be described by this description. DIX will issue a warning "description is aborted wrongtype. The description will remain open for the file, so if another record does have the right UAF$B_RTYPE this record will be displayed using this description.
PARAMETER REP_BITS=2
INTEGER*4 RECNR
STRUCTURE DATUSR
DATE*4 DATTIM !internal date format #minutes since 1857
UIC*4 UIC !standard OpenVMS UIC format (group,member)
END STRUCTURE
INTEGER*2 type - !1,4,10 will be by name, all others in
[1=special,4=normal,10=fatal] !decimal format
INTEGER*2 COUNT
INTEGER*4 RESCHED
STRUCTURE BITS(REP_BITS) !2 structures, with both bits*1 flag=1 byte long
BITS*1 FLAGS [mon,tue,wed,thu,fri,sat,sun] !the total is 2 bytes
END STRUCTURE
UNION
MAP TYPE=Special !this case is valid if type=1 (special)
INTEGER*2 time(2)
END MAP
MAP * !this case is valid for any other field value
CHARACTER*2 dunnow !
ENDMAP
MAP type=normal !this case is value if type=4 (normal)
INTEGER*4 time
ENDMAP
MAP Type=fatal !this case is valid if type=10 (fatal)
ENDMAP !No allocation is this type
ENDUNION !total structure length is dependent on type
WSTRING*60 MESSAGE !word counted ASCII string with 60 bytes room
INTEGER*4 FIELD(1:COUNT)
!field has COUNT members
An example about field_following to another file:
The example is about 3 RMS indexed files that form a simple source module
cross_reference system
The first file (CROSS_REF.CRF_FILE_NAMES) has the following description
(.CRF_FILE_NAMES in the system or user text library)
integer*2 file_nr !primary key
character*60 file_name
integer*2 %filler
The second file (CROSS_REF.CRF_MOD_NAMES) has the following description
(.CRF_MOD_NAMES in the system or user text library)
integer*2 mod_nr !primary key
character*32 mod_name
integer*2 file_nr/file=.CRF_FILE_NAMES !link to the filename
The third file (CROSS_REF.CRF_CROSS) has the following description
(.CRF_CROSS in the system or user text library)
integer*2 caller_nr /file=.CRF_MOD_NAMES !link to module name
integer*2 called_nr /file=.CRF_MOD_NAMES !link to module name
Now if you open a record in the third file
$ DIX CROSS_REF.CRF_CROSS
%DIX-I-USINGFIL, Using file DSA40:[DIR]CROSS_REC.CRF_CROSS
%DIX-I-USINGDES, Using description DIX_SYS(.CRF_CROSS)
DIX> EXA *
0|CALLER_NR>|738 !the > tells us there is a link present
2|CALLED_NR>|-262
DIX> EXA/DES *
0|CALLER_NR>|738
Type :INTEGER*2
Linenumber:0
Variable :False
Dependency:False
Link file :.CRF_MOD_NAMES !and yes there is a link defined
Link field:0
Match :EQ
2|CALLED_NR>|-262
Type :INTEGER*2
Linenumber:1
Variable :False
Dependency:False
Link file :.CRF_MOD_NAMES
Link field:0
Match :EQ
DIX> Follow CALLER_NR !try to follow this link
File .CRF_MOD_NAMES not (yes) opened, open it (y/[n]):Y !do you want to open the file
%DIX-I-USINGFIL, Using file DSA40:[DIR]CROSS_REF.CRF_MOD_NAMES
%DIX-I-USINGDES, Using description DIX_SYS(.CRF_MOD_NAMES)
DIX> Exa *
0|MOD_NR |738
2|MOD_NAME|CHECK_ALLOWED_USER
34|FILE_NR>|66 !and this field has a link defined
DIX> Follow file_nr
File .CRF_FILE_NAMES not (yes) opened, open it (y/[n]):y
%DIX-I-USINGFIL, Using file DSA40:[DIR]CROSS_REF.CRF_FILE_NAMES
%DIX-I-USINGDES, Using description DIX_SYS(.CRF_FILE_NAMES)
DIX> Exa *
0|FILE_NR |66
2|FILE_NAME|REM_SERVER_CHECK_ACCESS
DIX> Back !now back trace
%DIX-I-USINGFIL, Using file DSA40:[DIR]CROSS_REF.CRF_MOD_NAMES
%DIX-I-USINGDES, Using description DIX_SYS(.CRF_MOD_NAMES)
DIX> Exa *
0|MOD_NR |738
2|MOD_NAME|CHECK_ALLOWED_USER
34|FILE_NR>|66
DIX> Back !and again a back trace
%DIX-I-USINGFIL, Using file DSA40:[DIR]CROSS_REC.CRF_CROSS
%DIX-I-USINGDES, Using description DIX_SYS(.CRF_CROSS)
DIX> EXA *
0|CALLER_NR>|738 !the > tells us the is a link present
2|CALLED_NR>|-262
As you see, you can jump around an look at the record structures and links
Another example is the .DIR description:
integer*2 verslim [32767=None] !version limit
bits*1 flags !flag byte
string name !bytecounted string (variable)
pad*2 !align to word boundary
structure v(64) !n*version,fileid
integer*2 version
fileid fileid/file=[000000]indexf.sys/record-
/comp="f$getd($file,""blnr"")+int(f$extr(1,-1,f$elem(0,"","",$FIELD)))"
end structure
The fileid has a link to the corresponding INDEXF.SYS with a record number
equal to the first number in the fileid (layout = (filenr,revnr,volume))
+ the record number in INDEXF.SYS of the file header with fileid 0.
You need read access to the index file for this
$ DIX somedir.dir/int
%DIX-I-USINGFIL, Using file USER50:[STUBBF.PROGRAMS.DIX]SOMEDIR.DIR;1/NOMOD
%DIX-I-USINGDES, Using description DSA50:[STUBBF.PROGRAMS.DIX]DIX_DES.TLB;54(.DIR)
DIX> exa *
0|VERSLIM |NONE
2|FLAGS |
3|NAME |Q.Q
8|V(1).VERSION|1
10|V(1).FILEID>|(63474,3,0)
DIX> foll v(1).fileid !follow link to file header
File [000000]INDEXF.SYS not (yet) opened, open it (y/[n]):y
%DIX-I-USINGFIL, Using file USER50:[000000]INDEXF.SYS;1/NOMOD
%DIX-I-USINGDES, Using description DSA50:[STUBBF.PROGRAMS.DIX]DIX_DES.TLB;54(INDEXF)
DIX> exa *
0|ID_OFFSET |40
...
8|FILE_ID |(63474,3,0)
...
66|BACKL_FID >|(62674,1,0)
80|FNAM |Q.Q;1
...
DIX> Follow backl_fid !follow link to parent directory
0|ID_OFFSET |40
...
8|FILE_ID |(62674,1,0)
...
66|BACKL_FID >|(44986,1,0)
...
80|FNAM |SOMEDIR.DIR;1
...
DIX> back !back to the file header
DIX> exa *
0|ID_OFFSET |40
...
8|FILE_ID |(63474,3,0)
...
66|BACKL_FID >|(62674,1,0)
80|FNAM |Q.Q;1
...
DIX> back !back to the entry in the somedir.dir file
%DIX-I-USINGFIL, Using file USER50:[STUBBF.PROGRAMS.DIX]SOMEDIR.DIR;1/NOMOD
%DIX-I-USINGDES, Using description DSA50:[STUBBF.PROGRAMS.DIX]DIX_DES.TLB;54(.DIR)
DIX> exa *
0|VERSLIM |NONE
2|FLAGS |
3|NAME |Q.Q
8|V(1).VERSION|1
10|V(1).FILEID>|(63474,3,0)
Example for the /LIMIT qualifier:
Description
CHARACTER*10 TEXT
STRUCTURE data(100)
INTEGER*4 flag/limit=>100
CHARACTER*20 what
END STRUCTURE
If the field limit contains a value > 100, the repeat for the
structure data(100) will be terminated.
CHARACTER*10 TEXT
STRUCTURE data(100)
INTEGER*4 flag(5)/limit=>100
CHARACTER*20 what
END STRUCTURE
If the field limit contains a value > 100, the repeat for the
structure flags (5) will be terminated, but the repeat for
the structure will be continued
You can search records in three ways.
a. Via the key record selection (/ge/gt/eq/le/lt). This will only
work for indexed or relative files
b. Via the record selection (/record). This will work for all types of files
c. Via a search on record contents. This help page describes this method.
DIX has a powerful record search function. You can look for data in
records, but also in fields (if you have a description).
Parameters 2..8 can be used for search input.
If the fields are not text-fields but binary data, DIX will do a convert
of the search string to binary and then do the compare.
DIX also allows you to search data in more than one record, so a query
like "search a record that contains A and the next record contains B" is
possible, and this is not limited to just 2 records.
The search can be a simple locate, or a wildcard match, or a compare
(le,le,eq,ge,gt), and the compares can be done in any known type.
The searchsyntax is
Parameter [parameter2] [parameter3...] Up to 7 parameters.
If more than one parameter is given, parameter2 must match data in the
following record, parameter3 in the next record and so on.
Each parameter has the syntax (a list of values and qualifiers)
string1[/qual],string2[/qual],..
Search can be used in 4 places
Interactive mode : Entered on the command line or in the interactive commands:
Will set the record pointer to the first match
Screen mode : Will set the record pointer to the first match
DECwindows mode : Will set the record pointer to the first match
File mode : Will display all records matching the search
(unless /count is given)
All these examples are for the DIX/FILE mode. They can also be used
for the screen and interactive mode. The file mode will display all matches,
the interactive and screen will select the first record that matches.
1. $ DIX/FILE data.dat TEST
Display all records that contain TEST (/MATCH=LOCATE). The search is
NOT case sensitive.
2. $ DIX/FILE data.dat TEST,12/FIELD=FOR*MAT/MATCH=GE
Display all records that contain TEST (/MATCH=LOCATE),
OR have the value in any field that matches "FOR*MAT" >=12
3. $ DIX/FILE data.dat TEST/LOGIC=AND,12/FIELD=FORMAT/MATCH=GE
Display all records that contain TEST (/MATCH=LOCATE),
AND have the value of field "FORMAT" >=12
4. $ DIX/FILE data.dat TEST/WINDOW=(position=10,size=40)
Display all records that contain TEST (/MATCH=LOCATE) in the
byte 10..49 of the record. The search is NOT case sensitive.
5. $ DIX/FILE data.dat *XYZ/MATCH=match
Display all records that match *XYZ (so the XYZ must be at
the end of the record.
6. $ DIX/FILE data.dat "TesT"/case_sensitive
Display all records that contain TesT is this case.
7. $ DIX/FILE data.dat string1,string2/next_record/wild=extended
Display the records that contains string1 and the next record
matches string2 (with extended wild characters).
Only the first record is displayed.
8. $ DIX/FILE data.dat 123:456/match=range/field=myfield
Displays the records that contain a value between 123 and 456
(including) in field myfield
9. $ DIX/FILE data.dat 123/field=myfield/type=integer*2/mat=eq
Displays the records that contain an value of 123 in the
first two bytes (converted as integer*2) of the field myfield
Field myfield does not have to be of type integer.
10. $ DIX/FILE data.dat 123/field=myfield/type=integer*2/wind=(pos=10)/mat=eq
Displays the records that contain an value of 123 in the
bytes 10,11 (converted as integer*2) of the field myfield.
If bytes 10-11 do not exist in the field, the match returns false.
11. $ DIX/FILE data.dat 123:456/field=myfield/wind=(pos=10,siz=4)/mat=Range
Displays the records that contain an value between 123 and 456 in the
bytes 10-13 (converted as integer*4 (size=4)) of the field myfield.
If bytes 10-13 do not exist in the field, the match returns false.
12. $ DIX/file data.dat *str*ing1*/mat=match/wild 1234/matc=eq/field=myfield
search for a record that match *str*ing1*
AND the next record contains the value 1234 in field myfield.
13. $ DIX/FILE data.dat string1 string2 string3
Will search for a record that contains "string1"
AND the next record contains "string2"
AND the third record contains "string3".
Record 1 will be displayed, record 2 and 3 not.
14. $ DIX/FILE data.dat string1 string2/show string3/show
Will search for a record that contains "string1"
AND the next record containing "string2"
AND the third record containing "string3".
Record 1 will not be displayed, record 2 and 3 will.
15. $ DIX/FILE data.dat string1 10/mat=ge/loca=recl string3
Will search for a record that contains "string1"
AND the next record has a record length of 10 or more,
AND the 3 record contains string3
16. $ DIX/FILE SYSUAF RESTRICTED/field=*flag*/mat=eq
Will search all SYSUAF records that have ONLY the restricted flag
17. $ DIX/FILE SYSUAF DISACNT/field=*flag*/mat=ge
Will search all SYSUAF records that have AT LEAST the DISACNT flag
18. $ DIX/FILE filename "1"/win=(pos=0,size=1)/log=and,1/loca=recl/match=eq
Will search all records with a ! in the first position
and no more on the line (record length=1)
19. $ DIX/FILE filename 10:12/loca=recl/match=range
Will search all records with a record length between 10 and 12
Search the file in the backward direction. Beware that this may take a long time, especially for sequential files (backspace is implemented as a rewind/read n-1 records).
The text in STRING is to be matched case sensitive. DCL will change all parameters to uppercase unless within quotes. So if you use /case_sensitive, make sure STRING is within quotes. This qualifier is positional, so each search item can have its own value.
If the /field is specified, DIX checks if the field is present. If the /WINDOW is also used, DIX checks if the range specified is in the record (no /FIELD) or in the field. The search string will not be used in this case.
/FIELD=NAMEMASK
namemask is fieldnamemask
descriptionmask\fieldnamemask
filenamemask\descriptionmask\fieldnamemask
If filename mask is present, the pattern is matched against the filename
if no match is found, this entry is skipped.
If descriptionname mask is present, the pattern is matched against the
description. If no match is found, this entry is skipped.
Normally DIX will search the string in the whole record. If you
enter /FIELD only the field(s) that match the "fieldnamemask"
are searched. This match uses the normal OpenVMS wildcard * and %.
See the examples, see also the /WINDOW qualifier.
The type of the field (integer, character, etc.) determines the way the
compares are done. For example, for INTEGERS the compare is done in
binary etc.
This qualifier is positional, so each search item can have its own value.
/LOCATION=location_value
Where the data should be searched.
location_value can be
data : The normal data (Default)
VFC : The VFC data
both : In the VFC as well in the normal data
recl : Special value for comparing record_length. In the case only the
/match=lt,le,eq,ge,gt and range can be used. The default will be eq
/LOGIC=AND /LOGIC=OR !Default If you specify /LOGIC=AND, this string must match as well as the next. This qualifier is positional, so each search item can have its own value. Example: STRING1/LOGIC=AND,STRING2 Matches only if both STRING1 and STRING2 are found. STRING1/LOGIC=OR,STRING2 (the /logic=or is default) Matches only if either STRING1 or STRING2 are found. STRING/LOGIC=AND,STRING2,STRING3 Matches only if either (STRING1 and STRING2) or STRING3 is found.
/MATCH=LOCATE (Default)
/MATCH=MATCH
/MATCH=LT
/MATCH=LE
/MATCH=EQ
/MATCH=GE
/MATCH=GT
/MATCH=RANGE
Defines the match.
This qualifier is positional, so each search item can have its own value.
See also the /wildcard qualifier
LOCATE : The string must occur somewhere in the (part of the) record
MATCH : The string must match the whole record
LE,LE,EQ,GE,GT :
The value is compared with the STRING. This can be used with a /FIELD
If the field value is a numeric value, compares are done in the normal
numeric way. If the field value is a STRING, compares are done as
strings. See also the /TYPE command to overrule the default type.
For the [R]BIT datatype DIX has the following rule:
LE : All specified bits may be present, but not more
GE : All specified bits must be present, but there may be more
LT : All specified bits may not be present
EQ : All specified bits must be present, but not more
RANGE : The search string is LOW:HIGH. A match will be found it the value
is between LOW and HIGH (including the boundaries).
Example:
123/FIELD=UAF$W_BYTLM/MATCH=GE (binary (integer*2) field)
Will display all records with the UAF$W_BYTLM value >=123. (binary compare)
XYZ/FIELD=UAF$T_USERNAME/MATCH=LT (ASCII (character*12) field)
Will display all records with the UAF$T_USERNAME value <XYZ (ASCII compare)
123:345/field=UAF$L_BYTLM/MATCH=RANGE (binary (integer*4) field)
Will display all records with BYTLM between 124 and 345
With the /NEXT_RECORD you can specify a search over more than one record. If you specify string1,string2/next_record, DIX will look for records that contain string1 in this record and string2 in the next record. This should normally be done by using more than one parameter, and the above example is equivalent to : string1 string2, but if you need more than 6 records to match, you need this option (DCL allows only 8 parameters). This qualifier is positional, so each search item can have its own value.
The result of the test will be reversed. So a match with /LOCATE will return true if the string is not found. This qualifier is positional, so each search item can have its own value.
/RECORD=N Start search at record N. If /record is not specified, DIX will start the search from the first (or current) record N can be negative, in that case DIX will start from N record before the end_of_file. See also the /LIMIT qualifier. Example: (assume a file of 50 records) /RECORD=5/LIMIT=10 ! Search record 4..14 /RECORD=-10/LIMIT=4 ! Search record 41..44
If no /show is given on a multi record search, the first record is displayed You can give a /show for any of for all records. See the examples
/TYPE=datatype[*size] This allows you to overrule the default datatype (the type of the field, or CHARACTER if no field is given). Datatype can be any valid declaration i.e. INTEGER, CHARACTER, UIC, BITS, etc. The size has its default for each datatype (see the description help) "help record_format_for_descriptions field_decl"
/WINDOW=(position=n,size=m) Do not use the whole record, but take a subset. See also the /FIELD qualifier. Position starts at byte 0. Size defaults to 1. If the /TYPE is used, the size defaults to the size of the datatype. This qualifier is positional, so each search item can have its own value.
/WILDCARD=NONE
/WILDCARD=STANDARD (Default)
/WILDCARD=EXTENDED
/WILDCARD=NONE No wildcard matching is used.
/WILDCARD=STANDARD The following wildcard characters can be used
* : matches all substrings (0 or more chars)
% : matches exactly 1 char
/WILDCARD=EXTENDED The Following extended wildcard characters can be used
[abc] : Matches a "a" or a "b" or a "c"
[-abc]: Matches anything except a,b,c
[a-z] : Matches all letters (a-z)
[-a-z]: Matches anything except letters
'a : Char "a" is no longer a special char ([*%' etc.)
~ : If in front of the search string, the search string must be in the
beginning of the line, if at the end of the search string,
the search string must be at the end of the line
! : Matches one or more whitespace chars
The search string must be converted to a specific type. If this conversion succeeds, this match succeeds This qualifier can be useful if you have specified a /FIELD with a wildcard somewhere, and the matching fields have different types.
DIX has a number of descriptions. They are stored in the DIX_DES.TLB file. If you want to add descriptions you may add them to this .TLB file, or if you do not have write access to this file, you may add them to the DIX_DES_USER.TLB file which can be placed in your own directory. See the help about [DIX/HELP] "/user_library". The standard descriptions are (besides a number of entries for my own use)
A description for a disk block containing one or more indexed-files
areas. You must open the indexed-file in block-mode.
DIX[/int] indexedfile/block=1/record=n/descr=$area
If the record=n points to a block that contains areas, this
description will show the contents of the area(s).
A description for a disk block(s) containing one or more indexed-files
buckets. You must open the indexed-file in block-mode.
DIX[/int] indexedfile/block=k/record=n/descr=$bucket
DIX> exa *
DIX> FULL_BUCKET=1
DIX> exami/expand
in the /block=k, the k is the bucket_size of this bucket.
If the record=n points to a block that contains a bucket, this
description will show the contents of the bucket.
All variants of the bucket (INDEX, SIDR and DATA) are supported
If you want to see the contents of the record data in the
data-buckets, you must define a symbol FULL_BUCKET.
A default description containing
BYTEARR $BYTES(record_length)
CHARACTER*(record_length) $LINE
CHARACTER $CHARS(record_length)
A description for a disk block(s) containing one or more indexed-files
key blocks. You must open the indexed-file in block-mode.
DIX[/int] indexedfile/block=1/record=n/descr=$key
If the record=n points to a block that contains a key block, this
description will show the contents of the key.
A description for a disk block(s) containing and indexed/relative-file
prologue block. You must open the indexed-file in block-mode.
The prologue block is the first block in the file.
DIX[/int] indexedfile/block=1/record=1/descr=$prologue
The whole record will be display as a character line.
Unprintables will be displayed depending on the "SET FORMAT" command
The layout of the All-In-1/office-server DAF file records.
The layout of the SECURITY.AUDIT$JOURNAL file.
The layout of a backup save set This description will match all files with names *.BCK and *.SAV
The layout of directory file records
The layout of ACCOUNTNG.DAT file records
The layout of the All-In-1/office-server DOCDB file records.
The layout of the All-In-1/office-server FILECAB file records.
The layout of the INDEXF.SYS file records, but only those records
that contain file headers. The first few hundred blocks contain
a bitmap of used file headers
The layout of the LMF$LURT file. This file contains the information about
how many license points are needed for all computer types for
the various types of licences
Files containing monitor recording files
The contents for the NETPROXY file
The contents for the NET$PROXY file
The layout of the All-In-1/office-server NETWORK file records.
The layout of the All-In-1/office-server PARTITION file records.
The layout of the All-In-1/office-server PENDING file records.
The layout of the All-In-1/office-server PROFILE file records.
The layout of disk quota files, disk:[000000]QUOTA.SYS
The layout of the All-In-1/office-server RESERVATIONS file records.
The layout of the RIGHTSLIST file records.
The layout of the SYSUAF file records.
The layout of the first (2) block(s) of the SYSDUMP.DMP file To be used with $ DIX/block=2 SYS$SYSTEM:SYSDUMP.DMP
The layout of the VMS$PASSWORD_HISTORY file records.
The layout of the VMSMAIL_PROFILE file records.
By default, DIX is not case sensitive, all commands/data will
be converted to upper case (unless within quotes).
You can enable case_sensitivity via the interactive command
DIX> enable case or disable case or set process/case_lookup
You can also enable case_sensitivity on the command line via
the /flag=case_sensitive, but you also need to tell DCL that
you want case_sensitive mode via the command:
$ SET PROCESS/PARSE_STYLE=EXTENDED
I am sorry, but that will not work on VAX (and older alpha versions)
If you work in case_sensitive mode, all symbol/field/parameter
names are case sensitive, and also values from commands are not
uppercased. So i.e. if you want to read a record from an indexed
file (with the /eq=xyz qualifier), you need to specify xyz in
the correct case (and you do not need the ""'s)
For more info see the help about "interac set case"
Examples:
DIX> exa sym1 !in case sensitive mode the symbol sym1 is printed
!in not case_sensitive mode SYM1
DIX> exa "sym1" !in both cases the sym1 is printed
Version Author Date Comment
1.0 F.S. 19-jan-1991 Initial release
1.1 F.S. 1-mar-1991 Support for sequential files
2.0 F.S. 1-jan-1992 Use of SMG routines
2.1 F.S. 26-feb-1992 Support block I/O
2.2 F.S. 27-feb-1992 Support fieldnames for bits values
2.3 F.S. 18-mar-1992 Support for multiple files
3.0 F.S. 1-jul-1998 Multiple changes
support for parameters
support for run-time evaluation of expressions
3.1 F.S. 10-mar-2003 Support for scripting in interactive mode
3.2 F.S. 16-Jun-2003 Added user defined keys
4.0 F.S. 7-may-2004 Added user defined types
5.0 F.S. 19-apr-2006 Added a lot of new features, see the release_notes
6.0 F.S. 1-dec-2006 Added DECwindow interface
7.0 F.S. 1-feb-2008 Added views and a lot more
7.1 F.S. 14-apr-2009 Added a lot of DCL compatibility
7.2 F.S. 1-dec-2009 Added compiled scripts/indexed symbols
8.0 F.S. 1-oct-2011 Added plot command
8.1 F.S. 22-mar-2012 Added plotting of CSV files
8.2 F.S. 31-oct-2012 plotting pie charts.
8.3 F.S. 1-oct-2013 Smaller enhancements, access to library modules
Added description and structured type symbols.
8.4 F.S. 1-jun-2014 A lot of smaller enhancements, Added complex datatype
Added auto command completion (set term/auto)
8.5 F.S. 1-mar-2016 Smaller enhancements', added pdf output for plots
8.6 F.S. 1-sep-2016 Added checksums for sha, md5
8.7 F.S. 1-jan-2019 Added the DECENT command
c
For interactive, screen and DECwindows mode only. Set the autosave flag to enabled. If you later exit DIX, the context will be saved, and you can restore with DIX/RESTORE See also the "modes restore", "interactive_com set auto" help
In DECwindows mode only. You can set some attributes of the DECwindows
displays. You can later change them via the interactive command
SET DECW[/qualifiers]
Syntax /ATTRIBUTES=(attr,attr...)
Attr can be
HEIGHT=n : The height of the main display in pixels
The default is 80 characters, so it depends on the font
WIDTH=n : The width of the main display in pixels
The default is 30 lines, so it depends on the font
FONT=name : Name can be Small, Medium, Large, name=fontname
The default is medium on display with more than 1024
pixels, and small for smaller ones.
COLOUR=display=(fore=colour,back=colour,High=colour)
: Change the colours for the various displays
The default colours are
background=gray50, foreground=white, high=red
Display can be
main The main display
offset The offset part of the main display
field The field part of the main display
value The value part of the main display
bar The scroll bars
menu The menu
message The message display
alldisplay The all-format display
desdisplay The description display
If the /BINARY qualifier is given, the program will display the data in binary. The default is /HEX in raw dump mode. The default is /DECIMAL in interpreted mode. See also the /HEX, /BINARY and /OCTAL qualifier. /BINARY is a shortcut for /radix=2
/BLOCKED[=blocksize] /BLOCKED[=blocksize.bytesize] If the /BLOCK qualifier is specified, the file is accessed via block I/O. You can either specify a number, this will interpreted as a number of disk blocks, or a block.byte where you can specify part of one or more blocks. The total size cannot exceed 65535. The data transfer is done via a blocksize 512 bytes, and no record structure is assumed. This method can destroy file/record integrity and should only be used if you know what you are doing. Blocksize is default 1 (512 bytes if you specify /BLOCKED). It can be useful for files with undefined record type (as the DUMP file). In blockmode you can use the /record=n qualifier to select the 'n'th record of the /blnr=n to start the read at virtual blocknumber 'n' Examples: $ DIX/INT filename/block=4 !use a block size of 4*512 bytes $ DIX/INT filename/block=0.32 !Use a blocksize of 32 bytes $ DIX/INT filename/block=2.32 !Use a lblocksize of 1056 (2*512+32) $ DIX/INT filename/block=0.1056 !the same as above $ DIX/INT filename/block=0.128/rec=6 !read blocknumber 2, bytes 128..255 $ DIX/INT filename/block=0.128/blnr=2 !read blocknumber 2, bytes 0..127 $ DIX/INT filename/block=2/rec=4 !read blocknumber 7 and 8 $ DIX/INT filename/block=2/blnr=4 !read blocknumber 4 and 5
/BLNR=nn If block mode I/O, you can select a virtual blocknumber to start the read $ DIX/INT filename/block=0.128/rec=6 !read blocknumber 2, bytes 128..255 $ DIX/INT filename/block=0.128/blnr=2 !read blocknumber 2, bytes 0..127 $ DIX/INT filename/block=2/rec=4 !read blocknumber 7 and 8 $ DIX/INT filename/block=2/blnr=4 !read blocknumber 4 and 5
/COMMAND dix_command(s) In command mode
/COMMAND=dix_command(s) In interactive mode
The command mode will start DIX and execute dix_command with all
functions available in interactive mode. You may specify multiple
commands separated by ;'s
command;command;command..
This mode does not allow parameters.
In interactive mode (DIX/inter/command=(command[,command)) [parameters]
The will startup DIX in interactive mode, inclusive the parameters
and execute the commands.
COmmand lines not starting with a /, will have the /command qualifier
inserted before the line, unless the symbol DIX_NO_AUTO_COMMAND is set to Y
Examples:
DIX/COMMAND dir *.*/sort=(organization,date=creat)
Will generate a list of files sorted by
1. Organization (idx,rel,seq)
2. The creating date
DIX/INTER/COMMAND=("while true","exam uaf$t_user*","next rec","end while") -
sysuaf
Display a list of users defined in the SYSUAF file.
By default DIX will display all fields. If you specified /COMPRESS, "empty" fields will not be displayed. For text data "empty" means all blank (spaces, binary %X20) For binary data "empty" means all zeros (binary %X00)
/COUNT[=number]
If the /COUNT qualifier is given, the program will dump "number"
records. If number is omitted, DIX will default to /COUNT=1
if a record is specified via the /LT,../GT or the /RECORD qualifier and
to /COUNT=0 (all records) otherwise.
This will only work in FILE_ORIENTED mode.
The /LIMIT qualifier limits the number of records searched, the
/COUNT the number of results
/CSV
/CSV=(Csv_option[,csv_option...])
Csv_option can be
HEADER : Print a header line before all data
NOQUOTES : Do not 'quote' fields (see below)
ALL_QUOTES: User quotes around all fields
SEPARATOR=COMMA : Use COMMA as separator (Default)
SEPARATOR=TAB : Use TAB char as separator
SEPARATOR=COLON : Use COLON (:) as separator
SEPARATOR=BAR : Use | as separator
SEPARATOR=SEMICOLON: Use SEMICOLON (;) as separator
SEPARATOR=CHARACTER=%Xdd : Use %Xdd as separator
QUOTE=DOUBLEQUOTE : Use the " character for quotes around fields (Default)
QUOTE=QUOTE : Use the ' character for quotes around fields
QUOTE=CHARACTER=%Xdd: Use the %Xdd character for quotes around field
If you want quoting (the default), DIX will insert the "quotecharacter"
around text fields and field that contain the "separatorcharacter" or space
If you specify ALL_QUOTES, DIX will insert the "quotecharacter"
around all fields.
If you specify NOQUOTES, DIX will never insert the "quotecharacter"
The header fields are regarded as "text" fields.
See also the /SELECT qualifier for help about selection of fields
Example:
DIX/FILE/CSV=(header,all_quotes,quote=quote,separator=semicolon) file
Output is a file with all fields quoted, using ' as quote and ; as
separator with a header line.
DIX/FILE/CSV=(header,noquote,separator=char=32) file
or
DIX/FILE/CSV=(noquote,separator=char=%X20) file
Output is a file with no fields quoted, and SPACE as separator
If the /DECIMAL qualifier is given, the program will display the data in decimal. The default is /HEX in raw dump mode. The default is /DECIMAL in interpreted mode. See also the /HEX, /BINARY, /OCTAL and /RADIX qualifier. /DECIMAL is a shortcut for /radix=10
/DECWINDOW If the /DECWINDOW qualifier is given, the program will display a record, (either interpreted or raw) and lets you specify editing (only in /MODIFY mode) or display commands. You can 1. Alter display formats (raw/interpreted) 2. Modify/insert/delete records (only in modify mode) Press the PF2/HELP key for info in screen mode. The /DECWINDOW qualifier must be the first on the command line. This mode will only work if you have DECwindows installed.
/DEFINE=(definition[,definition...])
A definition can be
1. name : This will deliver an integer symbol with the value 1
2. name=value : DIX will try to convert value to integer. If this
succeeds an integer symbol will be defined. If this
fails a character symbol will be defined.
See also the /script qualifier
DIX/DEMO [demoname] The [demoname] is a mask for the selection of demos. Start a demonstration. If you do not specify the demoname, or the mask matches more than one demo DIX will display a screen with a list of demos. You can then select one and a demo slide-show will begin.
/DISPLAY=(displayitem[,displayitem...])
/DISPLAY=ALL Default
Valid with /FILE only. This qualifier defines which parts of the
record are printed.
Valid displayitems are
[NO]RECNR : Print the record number
[NO]RECSIZE : Print the record size
[NO]VFC : Print the VFC part of the record (If the file contains VFC data)
[NO]DATA : Print the data part of the record
[NO]RFA : Print the RFA of the record
ALL : Print all parts
WHOLE_RECORD : Do not try to format, just dump the whole record. In this case
all other flags are cleared. You may use /FORMAT to select the
display of binary data, the default is passall. You may also
specify /display=(whole,[recnr],[recsize],[rfa]) to add an
additional record with this info.
[NO]WRAP : If set, DIX will wrap output to terminal width. The default is WRAP
[NO]BIT_OFFSET: Force offsets to be displayed with the bit offset part.
Normally DIX displays bit_offsets only if somewhere in the
record a data field is not byte aligned.
SEPARATOR=sep : Set the separator between offsets/fieldnames/data
Allowed options are
SEPARATOR=COMMA : Use COMMA as separator
SEPARATOR=TAB : Use TAB char as separator
SEPARATOR=COLON : Use COLON (:) a separator
SEPARATOR=SEMICOLON: Use SEMICOLON (;) as separator
SEPARATOR=BAR : Use | as separator (Default)
SEPARATOR=CHARACTER=%Xdd: Use %Xdd as separator
DESCRIPTION : Display the description of a field after the field data.
The file contains EBCDIC data. DIX assumes the whole record is EBCDIC. You can also define fields with EBCDIC contents. If you have a description present, DIX will add a /EBCDIC to all fields. If you have no description the raw display will display the data after conversion to ASCII
/FAST=number
/FAST=1024 (Default)
If you specify /FAST, DIX will read the file bypassing RMS, using
SYS$QIOW to access the data blocks. This may speed up searches for
indexed files, skipping the overhead for RMS-locking. It also allows
you to read (not modify) files that are locked. Since DIX bypasses
RMS locking there is no guarantee that you will get the correct data,
you are completely on your own.
If number is positive, it specifies the size of the buffers
(in blocks), DIX will use for file I/O.
If you specify number as -1, DIX will
map the file in memory via a SYS$CRMPSC function and all I/O
will be done via paging. On Alpha and IA64 DIX will choose
between p0 and p2 space depending on the size of the file.
Specifying number=-2, will force p2 space.
For searches though files it may speed up things, but not always.
It depends on how the buckets are distributed over the file.
If the buckets are mostly sequential in the file,
DIX can use large buffers, and read many buckets in one I/O. If
buckets are scattered over the file, these large reads will not be
beneficial, and a normal search with the SYS$GET may be faster.
The qualifier is supported for all types of files, but the gain for
non-indexed files may be small.
DIX will dump the record(s) like OpenVMS dump, or if a description is given in description format. If you do not specify /COUNT DIX will output all records of the file. The /FILE qualifier must be the first on the command line.
/FLAG=(name[,name...]) Set initial mode settings for interactive, DECwindows and screen mode name can be DCL_COMPATIBLE : For scripts DCL_FALLBACK : Dcl_compat + spawning of unknown verbs USE_MOUSE : Enable mouse usage LOW_FIRST : For multidimensional tables CASE_SENSITIVE : Enable case sensitive mode For more info see the help in interactive_mode set ...
/FOLLOW=(begin,[skipkey])
/FOLLOW=(end[,skipkey])
/FOLLOW=(field[,skipkey])
This is an easy way to generate a one line view
The generated source for the view file is
"field/foll=begin|end|field[/skipkey] *"
See the documentation about VIEWs for more info
If you have a file with a description that contains a link to another file
DIX can automatically include the fields from that file.
If you specify /FOLLOW=BEGIN, DIX will first follow all link fields, and
then display all fields of current description
If you specify /FOLLOW=END, DIX will first display all fields of the
current description, and then the follow fields.
If you specify /FOLLOW=FIELD, DIX will follow the link directly after
the follow field is found.
If you ADD the skipkeys keyword, DIX will not print the keyfields of
the followed file, since that value will probably also be a field
in the follow field.
Examples:
This example is a set of the files that make a (simple) cross reference
system
File 1. name.CRF_CROSS
Contains records with 2 numbers.
First the module number of the caller
And then the module number of the callee
integer*2 /file=.CRF_MOD_NAMES caller_nr !link to mod_names
integer*2 /file=.CRF_MOD_NAMES called_nr !link to mod_names
File 2. name.CRF_MOD_NAMES
Contains the module number and the name and a link to the filename
There is a key on the mod_nr
integer*2 mod_nr
character*32 mod_name
integer*2 /file=.CRF_FILE_NAMES file_nr !link to file_names
File 3. name.CRF_FILE_NAMES
Contains the file number and the file name
There is a key on the file_nr
integer*2 file_nr
character*60 file_name
$ DIX/FILE/COUNT=1/CSV=HEADER name.CRF_CROSS !DIX will not follow the links
"CALLER_NR","CALLED_NR"
738,-262
$ DIX/FILE/COUNT=1/CSV=HEADER name.CRF_CROSS/FOLLOW=END
"CALLER_NR","CALLED_NR",
"MOD_NR","MOD_NAME","FILE_NR",
"FILE_NR","FILE_NAME",
"MOD_NR","MOD_NAME","FILE_NR",
"FILE_NR","FILE_NAME"
738,-262, !both caller and callee
738,"CHECK_ALLOWED_USER",66, !info about caller number
66,"REM_SERVER_CHECK_ACCESS", !and its file
-262,"SYS$VERIFY_PROXY",69, !info about callee
69,"SYSTEM" !and its file
The indentation is only to show the actions.
$ DIX/FILE/COUNT=1/CSV=HEADER name.CRF_CROSS/FOLLOW=FIELD
"CALLER_NR",
"MOD_NR","MOD_NAME","FILE_NR",
"FILE_NR","FILE_NAME",
"CALLED_NR",
"MOD_NR","MOD_NAME","FILE_NR",
"FILE_NR","FILE_NAME"
738, !caller number
738,"CHECK_ALLOWED_USER",66, !its name
66,"REM_SERVER_CHECK_ACCESS", !and its file
-262, !callee number
-262,"SYS$VERIFY_PROXY",69, !its name
69,"SYSTEM" !and its file
Again, the indentation is only to show the actions.
You will see that the mod_nr and file_nr are specified twice
If you had added a SKIPKEYS, the display would have been
$ DIX/FILE/COUNT=1/CSV=HEADER name.CRF_CROSS/FOLLOW=(FIELD,skip)
"CALLER_NR",
"MOD_NAME","FILE_NR",
"FILE_NAME",
"CALLED_NR",
"MOD_NAME","FILE_NR",
"FILE_NAME"
738, !caller number
"CHECK_ALLOWED_USER",66, !its name
"REM_SERVER_CHECK_ACCESS", !and its file
-262, !callee number
"SYS$VERIFY_PROXY",69, !its name
"SYSTEM" !and its file
The display of the MOD_NR/FILE_NR for the second and third
level files are now gone, since they are the key values
/FORMAT=(option,option...)
/FORMAT=(DOT,ANSI) (Default)
Normally all "character"-like fields will not contain unprintable data.
These field types are CHARACTER,xSTRING.
If the data does contain unprintables (hex 0:1f,7f,80:9f or ff, this
includes the TAB character) the /FORMAT qualifier determines how this
data should be displayed.
Option can be
DOT :All unprintables are replaced by a ".". This encoding is not
reversible, after an unprintable value is replaced by a ".",
DIX cannot know what the unprintable value was. This may
be a problem in the screen mode.
HEX :All unprintable data is displayed as %Xdd, a hexadecimal display.
A % in the data will be displayed as %%. This display is reversible,
so DIX can reconstruct the original unprintable value.
If you are in this mode, on input a % must be entered as %%
PASSALL:DIX will not change the unprintable data. The data is displayed as
it is. This mode will not work in screen mode, and maybe poorly
when the output device is a terminal.
DUMP :Unprintable bytes are represented by a 2 or 3 letter mnemonic
like <DEL> or <CR>. A < in the text will be displayed as <<.
This display is reversible, so DIX can reconstruct the original
unprintable value.
If you are in this mode, on input a < must be entered as <<
ANSI :In combination with one of the first 4, allow ansi sequences to
be printed
SPECIAL:In combination with one of the first 4, allow BEEP (char(7)) and
TAB (char(9)) to be printed
If the /HEX qualifier is given, the program will display the data in hexadecimal. In interpreted mode the CHARACTER*(*) fields will be dumped in ascending order separated by spaces. All other fields will be dumped in descending byte order (the OpenVMS standard). The default is /HEX in raw dump mode. The default is /DECIMAL in interpreted mode. See also the /HEX, /BINARY, /OCTAL and /RADIX qualifier. /HEX is a shortcut for /radix=16
Enters INTERACTIVE mode. This is a command line mode that allows you to enter commands at the prompt. This mode is useful for DCL procedures. See the INTERACTIVE_COMMANDS help for possible commands. This mode also supports scripting commands (IF, GOTO etc.). The /INTERACTIVE qualifier must be the first on the command line. Example: $ DIX/INTERACTIVE SYSUAF/EQ=SYSTEM Processing file: SYS$COMMON:[SYSEXE]SYSUAF.DAT;1/NOMOD Using description: DSA0:[PROGRAMS.DIX]DIX_DES.TLB;9(SYSUAF) DIX> disp *pwd* 340.0|UAF$Q_PWD(1) |22B619B5 344.0|UAF$Q_PWD(2) |E19FAE22 348.0|UAF$Q_PWD2(1) |00000000 352.0|UAF$Q_PWD2(2) |00000000 362.0|UAF$B_PWD_LENGTH |8 372.0|UAF$Q_PWD_LIFETIME|30 00:00:00.00 380.0|UAF$Q_PWD_DATE |17-MAY-2002 18:34:12.00 388.0|UAF$Q_PWD2_DATE |17-NOV-1858 00:00:00.00 DIX> EXIT You can type HELP at the prompt to obtain online help.
/KEYBOARD=value
Define the input keyboard. Value can be
Normal : The LK45X keyboard layout
PC : The normal PC-type keyboard. Since the PC-keyboard
misses some keys, those keys are redefined. This is
visible in the help.
f13..f20 are displayed as Shift-F3..Shift-f10
Find Home
Select End
Next_screen PageDown
Prev_screen Pageup
LAPTOP : The normal LAPTOP misses also the PF1-PF4 keys, they
are redefined to ^F, ^G, ^K and ^L as well as the PC
keyboard modifications.
EXTENDED: Use extend input method for command input
See the set terminal/exten for more info
Set limits to the records DIX will process from this file
For indexed files
/LIMIT=(value=end,key=nn) !limit to key value <=end
/LIMIT=(value=(begin,end),key=nn) !limit to key values >=beg and <=end
/LIMIT=(number,value=(begin,end),key=nn) !limit to record numbers
!>=beg and <=end for key nn
For non-indexed files or if you do not want to use key values but record numbers
/LIMIT=(number,value=end) !limit to record number <=end
/LIMIT=(number,value=(begin,end)) !limit to record number >=beg and <=end
Limit the number of records searched. If /limit is not specified
DIX will process all records of the file.
If /LIMIT=VALUE=nnnnn the nnnnn is the last record number searched
If /LIMIT=VALUE=(xyx,key) the xyz is the last key value processed
Together with the /RECORD, you can search a part of the file
The /COUNT qualifier limits the number of records printed.
If the begin is set, DIX will position the file on the first record
that satisfies the record/key value.
If the end is set, DIX will issue a warning when the
record number/keyvalue exceeds the value
Example:
$ DIX/FILE myfile.dat/limit=value=10
Will process records 1..10 of the file
$ DIX/FILE myfile.dat/limit=value=(4,10)
Will process records 4..10 of the file
$ DIX/FILE myfile.dat/limit=(key=0,value=("A","B") )
Will process records with keyvalues >="A" and <="B" of the file
/LOCKING=(val,val...)
Val can be
RRL : Read all record with the RAB$M_RRL option. This means that even
if the record is locked by another stream, DIX can read it.
The usage is at your own risk, since any other user can
modify the data you are looking at (and possibly change).
OPTIMISTIC: DIX will read the record with the RAB$M_NLK option. The record
is read, but not locked. Other streams can read the data.
When you want to update a modified record, DIX will check if
the record still contains the original data, and if not
warn you, and give you the option to not modify the record.
If the file is opened with /nomodify (Default) records will
never be locked (the file is readonly).
If you leave the program DIX will define a symbol DIXRFA containing
the filename and RFA of the current record. If you start DIX at a
later moment with the same file and the /RFA qualifier, DIX will return
to the same record as before.
Example:
$ DIX/INTER SYSUAF/EQ=SYSTEM
DIX> EXA * ! Will display all data of the SYSTEM record
DIX> READ ! Will read the next record from the SYSUAF
DIX> Exit ! Leave DIX
$ DIX/RFA SYSUAF ! Will return you to the record in SYSUAF following
the SYSTEM record.
/MAX_BYTES=n
Limit the read from the file to maximum 'n' bytes
Example:
DIX/FILE filename/max_bytes=10
Dump the first 10 bytes of all records in the file.
If the /MODIFY qualifier is specified, you are allowed to modify /delete/insert records, Either in interactive or screen mode, and in raw (DEP BYTE_OFFSET=VALUE) or description mode (DEP field_name=value). The /WRITE qualifier is a synonym for /MODIFY
/MOUSE (Default) /NOMOUSE Enable the usage of the mouse in screen (SMG) mode. By default mouse_usage is enabled, but if you want, you can disable it to allow the mouse for selecting text. In the interactive mode the mouse usage can be enabled/disabled via the set mouse command.
Normally DIX will open only one file. If you specify more than one file (via wildcard or list), the program will prompt you which file to use. If you specify /MULTI_FILE, DIX will open all files and lets you switch between them. If you specify DIX/FILE, multi_file is the default.
/[no]OFFSET /OFFSET=DECIMAL Default /OFFSET=HEX /OFFSET=BINARY /OFFSET=DECIMAL If the /NOOFFSET qualifier is given, the program will not display the byte offsets of the data. The default is /OFFSET(=decimal)
If the /OCTAL qualifier is given, the program will display the data in octal. The default is /HEX in raw dump mode. The default is /DECIMAL in interpreted mode. See also the /HEX, /DECIMAL and /BINARY and /RADIX qualifier. /OCTAL is a shortcut for /radix=8
/OUTPUT=Filename Give an alternative output file. This qualifier will only be used if you specify the /FILE qualifier. The default is the terminal. If you specified /WIDE the width will be 132. If you specified /WIDTH=nn, the width will be nn (up to 4095)
/RADIX=nn (nn=2..36) Display/modify data in any radix. This is the general case of /hex, /decimal, /octal and /binary
If DIX can find a description, DIX will use it. If you specify /RAW DIX will display in RAW mode even if it can find a description.
/RFA=(rfaval)
Can be used to select a specific record by RFA
rfaval is either a 3 16bits number value (like dump)
bucketnrl,bucketnrh,offset
or a 32 bits bucketnr, followed by a 16bit offset
Example:
$ DIX FILE.DAT/RFA=(100,12) !Bucket=100, offset=12
$ DIX FILE.DAT/RFA=(100,3,12) !Bucket=3*65536+100, offset=12
/SCREEN If the /SCREEN qualifier is given, the program will display a record, (either interpreted or raw) and lets you specify editing (only in /MODIFY mode) or display commands. You can 1. Alter display formats (raw/interpreted) 2. Modify description files 3. Create description files 4. Modify/insert/delete records (only in modify mode) Press the PF2/HELP key for info in screen mode. The /SCREEN qualifier must be the first on the command line.
/SCRIPT=filename If interactive mode, if you enter /script=file, DIX will take commands from that file (default file type is .DIX), and return to DCL. See also the /define qualifier and the /command qualifier If you are in dcl_compat mode, the default file type is .COM You can also execute a script via the @filename on the command line $ DIX/INT @filename ...
/SELECT=(name[,name...])
In the /FILE mode, DIX will print all fields (of all descriptions) to a file.
If you specify /SELECT, DIX will display only fields with a name that matches
one of the name(s) of the /SELECT. Wildcards (* and %) can be used.
The value is [filemask\][descriptionmask\]fieldmask.
If only one \ is found, it is the descriptionmask, and the filemask is *\ .
If no description mask is given, all description (for the current file) are used.
There is one special name
$LINE : This will display the whole record as a single line, as if a
description with one line "CHARACTER*32767 LINE" has been given.
Examples:
$ DIX/FILE SYSUAF/EQ=SYSTEM ! Will display all the known fields.
$ DIX/FILE SYSUAF/EQ=SYSTEM/SELEC=*PWD* ! Will display only the fields
with PWD in the name.
$ DIX/FILE SYSUAF/EQ=SYSTEM/SELEC=UAF$P* ! Will display only the fields
beginning with UAF$P
$ DIX/FILE SYSUAF/des=(sysuaf,special) -
/EQ=SYSTEM/SELEC=(SYS*\UAF$P*,SPE*\TEMP*)
! Will display only the fields that
match UAF$P* in the description SYSUAF
or TEMP* in the description SPECIAL
$ DIX/FILE SYSUAF/des=(sysuaf,special) -
/EQ=SYSTEM/SELEC=(TEMP*)
! Will display only the fields that
match TEMP* in the description SYSUAF
or TEMP* in the description SPECIAL
$ DIX/FILE SYSUAF /select=$LINE. ! !Display each record with a single
description "CHARACTER*32767"
$ DIX/FILE a*.dat/MULTI/sel=(ax*\*\field1,ay*\*\field2)
! Dump all files a*.dat. For the ax*.dat
files all field1 fields are displayed,
and for the ay*.dat the field2 fields.
All other fields (also in other files)
are NOT displayed.
The select is case insensitive.
/SHARED (Default in DIX mode) /NOSHARED (Default in DCL mode) Open the file share or not shared.
/STARTUP=filename /STARTUP=DIX_INI (Default) Process commands from a startup file. The default filename is SYS$LOGIN:.INI See the topic about the interactive_commands about allowed commands. If the qualifier is defaulted, a missing file (SYS$LOGIN:DIX_INI.INI) will not be signaled. If you specify /startup explicitly, the file must exist.
/STRICT=(mode[,...])
Sets the STRICT mode for DIX for this level. Deeper levels inherit
the strict mode from the outer levels.
Normally DIX is not very strict in the symbol handling.
You can define new symbols by typing
1. a=10 (define the symbol a, type integer, value 10)
and then say
2. a="SKLFG" (redefine a to type character, value "SKLFG"
TYPING : Once a symbol is created, you cannot change the type anymore
In the example above you can do 1. but not 2.
DECLARATION : All symbols have to be declared before usage
This also disallows calling subroutines or @ files
and not having the same number of arguments
given and declared.
LOCAL : Even with /DECLARATION, symbols from an outer level
can be used. With /LOCAL this is not allowed.
/SYMBOL=symbolname
If you specify /SYMBOL and DIX operates in /FILE mode, the
result is printed to the output file, and symbols are defined for
each fieldname.
for example
$ DIX/FILE sysuaf/eq=system/select=*pwd*/symb=test
will display
%Recordsize = 644
340.0|UAF$Q_PWD(1) |22B619B5
344.0|UAF$Q_PWD(2) |E19FAE22
348.0|UAF$Q_PWD2(1) |00000000
352.0|UAF$Q_PWD2(2) |00000000
362.0|UAF$B_PWD_LENGTH |8
372.0|UAF$Q_PWD_LIFETIME |30 00:00:00.00
380.0|UAF$Q_PWD_DATE |17-MAY-2002 18:34:12.00
388.0|UAF$Q_PWD2_DATE |17-NOV-1858 00:00:00.00
But also the following symbols will be set
$ show symbol test*
TESTUAF$B_PWD_LENGTH = "8"
TESTUAF$Q_PWD(1) = "22B619B5"
TESTUAF$Q_PWD(2) = "E19FAE22"
TESTUAF$Q_PWD2(1) = "00000000"
TESTUAF$Q_PWD2(2) = "00000000"
TESTUAF$Q_PWD2_DATE = "17-NOV-1858 00:00:00.00"
TESTUAF$Q_PWD_DATE = "17-MAY-2002 18:34:12.00"
TESTUAF$Q_PWD_LIFETIME = "30 00:00:00.00"
This makes it easy to use DIX in command procedures.
/USER_LIBRARY=usertlbfile /USER_LIBRARY=DIX_DES_USER (Default) If you specify /USER_LIBRARY (Default), DIX will use this .TLB file for extra description records. The default file specification is SYS$LOGIN:.TLB So if you place a DIX_DES_USER.TLB in your login directory, DIX can it. In this file you can store your own descriptions, if you do not have write access to the default DIX_DES.TLB file.
/SYSTEM_LIBRARY=systemtlbfile /SYSTEM_LIBRARY=DIX_DES (Default) If you specify /SYSTEM_LIBRARY (Default), DIX will use this .TLB file for system-wide description records. The default file specification is DIX-image-directory:.TLB If you want to add your own descriptions, you may add them in this file, or (if you do not have write access to this file) in the DIX_DES_USER.TLB file. See the help about [DIX/HELP] "/USER_LIBRARY".
Set the output width (/INFO and /FILE) to 132. This is a shorthand for /WIDTH=132.
/WIDTH=nn Set the output width (/INFO and /FILE) to nn nn will be clipped between minimum 80 and maximum 4095.
If the /WRITE qualifier is specified, you are allowed to modify /delete/insert records, Either in interactive or screen mode, and in raw (DEP BYTE_OFFSET=VALUE) or description mode (DEP field_name=value). The /WRITE qualifier is a synonym for /MODIFY
If you specify /INTERACTIVE the program will enter command line mode.
It will prompt you for commands. Lines can be continued with a -
on the last position of the line (like DCL).
The only exception is the variable-- command, so if you want a continuation
line ending on a -(minus) be sure to add a space between the two -'s
a=a - - (mind the space between the -'s) will be a=a-10
10
versus
a -- will be a--
Blanks are not significant, except in symbol names, verbs and strings.
DIX is by default not case sensitive, except within strings.
DIX can be set to case_sensitive mode, see the help about "set case"
DIX has a powerful set of commands to enable scripting inside a command file.
See the @file_and_commands_in_file help topic.
You can make a one-line command list. To do this you must start the command
with PIPE or a ; and separate the commands with a ;
Example:
DIX> PIPE for k=1,10 ; evalu k ; endfor
will create pseudo-command file containing
for k=1,10
evalu k
endfor
DIX> ; for k=1,10 ; evalu k ; endfor !alternative version
And execute it. In this way you can use the 'file_commands' on the command line
If DCL mode enabled, DIX expects each line to start with a $ (all others are
data lines). To help you in DCL mode DIX will automatically insert a $ before
each line, unless the first line starts with a $. In that case DIX will assume
you understand the need for leading $.
Example (in DCL mode):
DIX> ;show time;show time;show time
DIX will insert a dollar before each line
DIX> ;$type sys$input;line1;line2;$show time
Since the first line starts with a $, DIX assumes you understand the need
for a $ before each command. Line2 and line3 will be regarded as data
for the type command.
If you enter a command without a verb, DIX will try to evaluate the line
as an expression and print the result, so an input like
DIX> 10*10 !will give the answer
100 !just as if you had typed DIX> eval 10*10
If you enable check_symbol DIX will check for default extra settings
for a command. So
DIX> dire !will do a normal directory
$dix_command_directory="/size"
DIX> enable check_symbol
DIX> dire ... !will do a directory/size ...
The symbol is DIX_SYMBOL_COMMANDNAME, where commandname is the fully
specified command (so directory and not dire)
Normally DIX will print the output of the commands to the terminal
(or the filename given with @command/out=filename)
Some commands (the ones with possibly a lot of output) accept the
/OUTPUT=filename qualifier : Output to a file
/PAGE qualifier : Output per page
/SCREEN qualifier : Get all output first and then let
you look at it in a window.
If the output filename is preceded by a # or DCL: or DCL::,
DIX will take the rest of the filename as a symbol name,
and store results in (DCL_) symbols
name,name_1,name_2...name_n.
The symbol name_cnt will be filled with the highest index (might be 0)
In this way you can use DIX to define DCL symbols, e.g. after
$ DIX/COMM/OUT=dcl:abc say 4+5
the DCL symbol ABC will be defined with the value 9
DCL:: will define global symbols, and DCL: local ones
Example: DIX> dir/out=dcl:tmp.tmp ;
DIX> open lun dcl:tmp.tmp ;
DIX> proces data
DIX> close lun [/dispose=delete]
Or
DIX> type mem:tmp.tmp
If the output filename is preceded by MEM:, the output is stored
in a memory table. I.e. DIX> dir/out=mem:tmp.tmp ;
DIX> open lun mem:tmp.tmp ;
DIX> proces data
DIX> close lun [/dispose=delete]
The output files generated via the mem: and dcl: will be treated as
files, so you can also do a directory mem:*. These 'files' can also
be deleted. The names may contain letters, digits and ., but the first
character must be a letter
The 'device'names mem: and dcl: can be shown and changed via the
set/show extra_devices
For all commands you can redirect the output to a (one dimensional)
DIX_symbol via the "symbolname < command" syntax
You may also use the < syntax to get a count of lines/words/chars
the output has generated, if you precede the symbolname with a #
(or specifying only a #), see the examples below
Example:
!default to screen
DIX> SHOW SYMBOL
%OVERFLOW=False
$STATUS=00000001
$SEVERITY=1
$DESCRIPTION=""
$FILE=""
$NRARGS=0
DIX> test<show symbol !Output to symbol testno output
DIX> Show symb/all test !show the symbol (contains 6 lines)
TEST(1)="%OVERFLOW=False"
TEST(2)="$STATUS=00000001"
TEST(3)="$SEVERITY=1"
TEST(4)="$DESCRIPTION="""
TEST(5)="$FILE="""
TEST(6)="$NRARGS=0"
DIX> #<show symbol !example of counting
%OVERFLOW=False !the normal output
$STATUS=00000001
$SEVERITY=1
$DESCRIPTION=""
$FILE=""
$NRARGS=0
TEST= table(6)
7 lines, 7 partial lines, 7 words, 88 characters printed !and the counters
DIX> #test<show symbol !example of counting/symbol output
7 lines, 7 partial lines, 7 words, 88 characters printed
DIX> !no output, except for count line
DIX> show symb test
TEST= table(7) !test is a table of 7 elements
Example for output file specifications:
DIX> show symbols/out=tst.dat !output to the file test.dat
DIX> show symbols/out=#abc !output to ABC, ABC_1, ABS_2...
DIX> show symbols/out=mem:name !output to memory file name
DIX> show symbols/out=dcl:name !output to DCL symbol 'file' name
$ DIX/COMMAND show symbols/out=dcl:name !output to DCL symbol 'file' name
$ show symbol name* !and on DCL level the symbols are there
NAME_1 = "$ARCHITECTURE="ALPHA""
NAME_2 = "$DESCRIPTION="""
NAME_3 = "$DIX_VERSION="8.6""
NAME_4 = "$FILE="""
NAME_5 = "$NRARGS=0"
NAME_6 = "$SEVERITY=1"
NAME_7 = "$STATUS=00000001"
NAME_8 = "%OVERFLOW=False"
NAME_CNT = "8"
Format
@filename[/output=filename][/[no]verb[=readin]][/usage]-
[/dcl_compatible[=value]]-
[/verify] -
[/password=pwd] -
[expression[,expression...]
Take the commands from a command file. The default filespec is
'this_directory'.DIX, or .COM if in DCL compatible mode.
The expressions will be available in the subroutine as local symbols
with the names Pn (As DCL) or any name the ENTRY statement gives them.
There is no limit to the number of arguments.
There is also a symbol $NRARGS that contains the number of arguments.
See also the help about the ENTRY command.
The /[no]verbs decides if verb symbol substitution is ON or OFF for
commands in the file (default is OFF). The OFF mode allows DIX to
precompile all statements in the file, make execution a lot faster.
The ON mode disables the precompile mode, because DIX does not know
if the verb may have been changed (i.e. like dir*ectory = "dir/col=1")
See also the help about compiler hints.
If you specify /verb=readin, the verbs will be translated on reading the
command file, but not again on execution of the statements. In this
way you can use global symbols for your verbs, but you cannot change
the verbs in the command file.
The /usage will display a control-t like line after execution of
each command. If the user had set the statistics to /usage
DIX will output a line for the usage data after execution of the command file
The /[NO]DCL_compatible decides if the command file is executed in DCL mode
If not specified, DIX uses the current setting for DCL_COMPAT
set dcl_compat compatible
set dcl_compat fallback
/dcl_compatible=none No DCL functionality
/dcl_compatible=compatible Support for leading $, replace .xx. by =<>
/dcl_compatible=fallback Compatible + Unknown DIX verbs will be spawned to DCL
The file is read into memory before any statement is executed. At this
point all loop-structures are checked for correct overlap, so
FOR
IF
ENDFOR
ENDIF
is not accepted.
During this phase DIX will pre-compile all statements in the file
(if verb substitution is OFF for command procedures). See the remarks
about /verb above.
Also Verb symbol-substitution for loop control verbs like IF, LOOP,
[END]FOR, [END]WHILE, etc. is NOT supported.
Loop commands can also be entered in terminal input mode (so not in a
command file). DIX will prompt you to enter more command lines until the
END statement is found, and them execute this set of commands.
(This happens as if the commands had been entered in a command procedure
so the command level will be one higher. This may affect symbols.)
Support is for the IF, LOOP, SWITCH, REPEAT, and WHILE statement.
Inside a command file you can use the following extra commands.
If the file is a compiled file (via the compile command) and
contains a password, you can (but maybe should not) specify the
password via the /password= qualifier. If you do not specify it
you will be prompted for the password.
You can give hints to the parser/compiler
Hints may be abbreviated, as long as they are unique
At this moments 2 hints are defined
On starting a command procedure you can specify the /verb qualifier
This qualifier sets the initial setting for the verb substitution.
If verb substitution is ON, DIX cannot precompile command since the
verb symbol may be set somewhere in the procedure.
In the procedure you can switch this behaviour.
##VERB_SUBSTITUTION : From now on verb substitution is possible,
so precompilation is disabled
##NOVERB_SUBSTITUTION : From now on verb substitution is not possible
so precompilation is enabled
This allows DIX to precompile all statements except those between
the ##VERB and the ##NOVERB
Example:
dir1:=dir/date !set a symbol for dir1
k=1 !
gosub werk
set stri/typ/decl
##verb !the next statements will not be precompiled
!since the verb dir1 does not exist
dir1 *.com; !parse at run time to use dir1
##NOverb !From now on statements can be precompiled again
bla: integer a=10
a=20
Call label [expression[,expression...] Call to a subroutine with optional parameters There is no limit to the number of arguments. See help about SUBROUTINE statement The parameters will be local symbols for the SUBROUTINE, see the help about [DIX/HELP] interactive symbols.
DECK [/DOLLARS=string] [/SYMBOL_SUBSTITUTION[=keyword]]
The DECK statement is meant to allow user data to be inside a
command procedure (see the DCL help about DECK)
the /DOLLARS specifies a string used to signal the end of the data
The default is /dollars=EOD. The match IS case-sensitive, so if you
want a lowercase string, include the string within "".
If /SYMBOL_SUBSTITUTION is set, DIX will try to substitute
all strings between ''s. This allows you to use data lines and
still have symbol substitution. See example below
Symbol_substitution is strict or lax (see enable lax_quotes)
The default is the current lax setting (DCL mode automatically enables LAX),
but may be overruled by
/SYMBOL=STRICT Use strict quotes, all quotes must in pairs
/SYMBOL=LAX Use lax quotes (the first non letter/digit stops the symbol)
DIX can operate in 2 modes
DCL_COMPAT : The data is ended by a statement starting with a $.
This may be the EOD command. In this mode, all lines
not starting with a $ will be regarded as data (like DCL,
so the DECK statement is not needed.
Normal : The data is ended by the DOLLARS string
Since DIX cannot distinguish between data and commands
(no leading $), you must use the DECK statement.
Example for DCL mode (leading $ required, so DECK is optional):
$type sys$input ! example with no deck and no EOD
line 1 ! 3 data lines
line 2
line 3
$exit
$type sys$input ! example with deck and eod
$deck ! Using the $DECK
line 1 ! 2 data lines
line 2
$eod ! It stops here
Example for DIX mode, DECK statement required:
type sys$input ! example with explicit /dollars
deck/dollars=test ! the test will be uppercased by DIX
line 1 ! 2 data line
line 2
TEST ! The terminator
exit ! Exit
type sys$input ! example with implicit dollars (EOD)
deck ! so we stop with a line containing EOD
line 1 nodol ! 2 data lines
line 2 nodol
EOD ! The terminator
Example with /symbol_substitution, the examples are for DIX mode:
type sys$input
deck ! without symbol substitution
line 1 nodol 'a' ! 2 data lines
line 2 nodol 'a+1'
EOD ! Terminator
This will return:
line 1 nodol 'a'
line 2 nodol 'a+1'
a=10
type sys$input
deck/symbol ! with symbol substitution
line 1 nodol 'a+2'
line 2 nodol 'a+3'
EOD
This will return:
line 1 nodol 12
line 2 nodol 13
For DCL mode the example is the same, but in this case the $DECK
is necessary, since that is the statement with the /symbol
$a=10
$type sys$input
$deck/symbol ! with symbol substitution
line 1 nodol 'a+2'
line 2 nodol 'a+3'
$EOD
Read the description inline from the command file
OPEN name filename/des=sys$input
DECK
description text
EOD
ENTRY parametername[qualifiers][,parametername[qualifiers]...]
Declare names for parameters on @ed files.
must be the first command in the file
If a call is make to a command file, the (optional) parameters are
evaluated, and the resulting value is placed in the (local) symbol
with the name "parametername" or Pn if there are more parameters
than parameternames.
Qualifiers can be
/REAL
You may specify the size of the real with the
/SIZE=4,8,16
/[U]INTEGER
/CHARACTER
/DATE
/DECIMAL
/LOGICAL The expression given at the @ line must match the
type given to this symbol
/SYMBOL
If the parametername has a /SYMBOL qualifier, the parameter will be
an alias to the original expression parameter. If you use the
/SYMBOL the caller must be a symbol or any subset (except indexed symbols).
If you change the parameter the changes will be done to the
caller symbol. This is a way to let the @file return output values.
If you use /SYMBOL and the parameter is an indexed symbol, you may
only use the whole symbol, not a subset.
If DIX is in STRICT/DECLARATIONS mode, the number of defined and
actual parameters must be equal, and all parameternames must have a type.
Example:
integer x=10
@mysub 10*4,"Text",x,41
and the file mysub.DIX contains the following lines
entry count/integer,name,xparam/symbol
...
...
exit
In the file "mysub" there will be 5 symbols defined
COUNT : Integer with the value 40 (the expression MUST be integer).
NAME : Character with the value "Text"
XPARAM : Integer with the value 10, is an alias for to symbol x.
and modifications are done to the X symbol of the caller
P4 : integer with the value 41
$NRARGS: Integer with the value 4 (4 arguments)
If there are more parameternames than parameters, the parametername
will be defined as a character string with the value "" (empty).
If there are more parameters than parameternames, the symbols
P* are defined with the same type as the parameter.
Both possibilities are not allowed with the STRICT/DECLARATION mode.
Examples:
Example for a subsetted table:
DIX> INTEGER val(10,2)
DIX> @tt: val(1,*)
DIX_1> ENTRY MYVAL/symbol
DIX_1> show symbol myval/fu
MYVAL(Fixed,Integer,Level=2.0,alias->VAL)=table(2)
DIX_1> say myval
[0,0]
DIX_1> myval(2)=12
DIX_1> exit
DIX> say val
[[0,0,0,0,0,0,0,0,0,0][12,0,0,0,0,0,0,0,0,0]]
Example for a structured type:
DIX> load/structure def_main.struct !load the structure
DIX> sh struc/fu def_main !what does it look like
DEF_MAIN, level 1 !
File USER50:[STUBBF.PROGRAMS.DIX.SOURCE]DEF_MAIN.STRUCT;2
INTEGER*4 INDEX
CHARACTER*40 MYSTR
REAL_X*16 REAL16VAL
INTEGER*4 EXTRA_INT
INTEGER*4 EXTRA2(2,3)
DIX> decl/stru=def_main val !declare a symbol of this type
DIX> say val !normal evaluate, only values
[0,,0.0,0,[[0,0][0,0][0,0]]]
DIX> @tt: val.extra2(2,*) !call a subroutine
DIX_1> entry locval/symbol !declare a local value as alias
DIX_1> show symbol locval/fu !is now a table of elements
LOCVAL(Fixed,Integer,Level=2.0,alias->VAL)=table(3) !extra(2,1),(2,2),(2,3)
DIX_1> say locval
[0,0,0]
DIX_1> locval(3)=31 !set the 3 (=(2,3)) to 31
DIX_1> exit
DIX> say val
[0,,0.0,0,[[0,0][0,0][0,31]]] !et voila
Syntax: Goto Label Goto the line with this label. If the label cannot be found, DIX will close this command file (and possibly its parent) until the input comes from a terminal or (in BATCH) exit DIX. The label must be visible from the current context level, so for example "LAB" will not be found, because "LAB" is in the repeat/until structure, and you cannot jump into a loop structure. x=1 repeat lab : x=x+1 until x>5 goto lab
Syntax: GOSUB label Goto the line with this label. If the label cannot be found, DIX will signal an error. Depending on the current ON state DIX may continue or close the command file. The return address will be stacked, and the RETURN command returns you to the next line. The label must be visible from the current context level, so for example "LAB" will not be found, because "LAB" is in the repeat/until structure, and you cannot jump into a loop structure. x=1 repeat lab : x=x+1 until x>5 gosub lab
Each line can be preceded by a label. The label can be any string terminated by a : Label names are not case-sensitive and can have up to 32 characters. All the character must be a-z 0-9 $ or _. Example: This_label12 : statement
On option [then] command
The then is optional
Option can be
Warning
Error
Severe
Control_c
overflow
And can be abbreviated to one letter
For the message levels:
If an error occurs, DIX will execute 'command' for the error level or
a lower level.
Example:
ON error goto label_error.
Any error and severe will jump to label_error. Warnings will be ignored.
on overflow say "got overflow" !say "got Overflow" if it is found
on control_c then debug !Jump to the debugger if you hit a ^c
Will return from a previous GOSUB. Return without an open GOSUB is NOT signaled as an error. DCL does not signal it, and so does DIX.
Format
label: SUBROUTINE [parametername[/qualifiers][,parametername[/qualifiers]...]]
Defines the beginning of a subroutine in a command procedure. The
SUBROUTINE command must be the first executable statement in a
subroutine. The subroutine must be terminated with an ENDSUBROUTINE.
If a call is make to a subroutine, the (optional) parameters are
evaluated, and the resulting value is placed in the (local) symbol
with the name "parametername" or Pn if there are more parameters
than parameternames.
Qualifiers can be
/REAL
You may specify the size of the real with the
/SIZE=4,8,16
/[U]INTEGER
/DATE
/DECIMAL
/CHARACTER
/LOGICAL The expression given at the @ line must match the
type given to this symbol
/SYMBOL
If the parametername has a /SYMBOL qualifier, the parameter will be
an alias to the original expression parameter. If you use the
/SYMBOL the caller must be a symbol or an scalar element of a symbol.
If you change the parameter, the changes will be done to the
caller symbol. This is a way to let the SUBROUTINE return
output values.
If DIX is in STRICT/DECLARATIONS mode, the number of defined and
actual parameters must be equal.
Example:
integer x=10
Call mysub 10*4,"Text",x,4
mysub: subroutine count,name/character,xparam/symbol
...
...
endsubroutine
In the subroutine "mysub" there will be 5 symbols defined
COUNT : Integer with the value 40
NAME : Character with the value "Text" (must be a character expression)
XPARAM : Integer with the value 10, is an alias for the symbol X,
and so modifications are done to the X symbol of the caller
P4 : Integer with the value 4
$NRARGS: Integer with the value 4 (4 arguments)
If there are more parameternames than parameters, the parametername
will be defined as a character string with the value "" (empty).
If there are more parameters than parameternames, the symbols
P* are defined with the same type as the parameter.
Both possibilities are not allowed with the STRICT/DECLARATION mode.
k = 1 !define symbol k value 1
Loop: k = k + 1 !label + increment of symbol
Examine field_'k' !exam with symbol substitution
If (2*(k+10) < 20) goto loop !test for repeat
This following script will increment BYTLM by 1000 for all users with
a name that matches *A*, and have a BYTLM of >=50000 and a JTQUOTA >4000
(maybe not a very useful example, but is shows the possibilities)
nmod = 0
nfnd = 0
on error goto done
loop:
nfnd = nfnd + 1
if (f$matchwild(uaf$t_username,"*A*")) goto next
echo "Found user ''uaf$t_username'"
if (uaf$l_bytlm < 50000 | uaf$l_jtquota <= 4000 ) goto next
deposit uaf$l_bytlm='uaf$l_bytlm + 1000'
nmod = nmod + 1
update/quiet
echo "Updated ''uaf$t_username' to ''uaf$l_bytlm'"
next:
read
goto loop
done:
echo "Found ''nfnd' users, modified ''nmod' users"
DIX has a syntax very similar to DCL. If you want to be able
to use DCL command files, DIX has to be put in dcl_compat mode.
DIX will then remove leading $'s, and if in fallback mode
it will spawn commands that are not recognized by DIX.
You can set/clear dcl_compat mode via the
SET DCL none|compatible|fallback
Or via:
@procedure/dcl[=fallback]
You cannot change the dcl_compat mode within one procedure.
There are 2 levels of dcl_compat mode
compatibility : -leading $'s are removed before parsing
-the DECK statement behaves DCL-like
-.lt., .ge. etc. are replaced by the DIX counterparts < > etc.
fallback : -If a command fails to be parsed by DIX (IVVERB)
DIX will spawn the command line. Be aware that
spawning is not the same as execution under DCL
and that DIX symbols are not DCL symbols,
but maybe the spawn will do what you want.
DIX has a built-in debugger for the debugging of command-scripts
You start this by @procedure/debug
You can set/delete break and watchpoints
You can step [/into] and go
You can also specify most of the interactive DIX commands to
examine/modify symbols/fields etc.
The statement breakpoint in a command procedure will
also activate the debugger
By default the following keys are defined
KP5 : Show calls
PF1-KP5 : Show call /full
KP0 : Step
PF1-KP0 : Step/into
And you may define more via the def/key command.
You can also store commands in a file and execute the file via
@file (default filespec .DIX;), the @commands can be nested 10 deep.
Verb substitution is also active (if you have in enabled, this is the default)
DixDbg> sb:=show break
DixDbg> sb !will show the current breakpoints
Each command procedure has a level
DIX> @first ! level 2 (the DIX> prompt is level 1)
if there is a @ line in first[.dix]
that procedure will be at level 3
Since DIX is an interpreter, DIX does not know anything
about deeper levels until the @ is reached.
So if you are on level 2, you cannot specify breakpoints
on deeper levels
On some commands you can specify lines/labels on lower levels
via the level\line or the level\label syntax
cancel break cancel watch cancel trace
cancel a breakpoint
cancel break/all Cancel all breakpoints
cancel break line [level\]nn Cancel Break on line number
cancel break label [level\]labelname Cancel Break on a label
cancel break call Cancel Break on call
cancel break return Cancel Break on return
cancel break message Cancel Break on error level
cancel a tracepoint
CANCEL TRACE/ALL Cancel all tracepoints
CANCEL TRACE LINE [LEVEL\]NN Cancel trace on line number
CANCEL TRACE LABEL [LEVEL\]LABELNAME Cancel trace on a label
CANCEL TRACE CALL Cancel trace on call
CANCEL TRACE RETURN Cancel trace on return
CANCEL TRACE MESSAGE Cancel trace on error level
cancel watch /all Cancel all watchpoints
cancel watch name Cancel watchpoint "name"
cancel the specified watchpoint
Supported qualifiers
/field Name is a fieldname, if /field is not
specified, name is a symbol
Go [line][label]
(re)start processing until a breakpoint occurs
optionally a line or label may be specified.
DIX will restart execution from that line
line or label must be in the current level
help [token] Help about the debugger
Set some things set break set watch set trace
set watch name
If the symbol/field matching name is touched or modified
display the value (and the new if changed)
Supported qualifiers
/field Name is a fieldname, if /field is not
specified, name is a symbol
/reference If /reference is specified, DIX will also
report references (not changes)
If /reference is not specified, DIX will
only report changes
/break If the symbol/field is touched (with /refer) or
modified DIX will enter the debugger
Watchpoints support the following qualifiers
/when="dix_command"
If the watchpoint is reached, DIX will execute the code specified
The result must be a logical. It the expression evaluates to true
the watchpoint will be taken (but the /after must be satisfied too)
/after=n
Take the watchpoint after the 'n'th time.
If both the /when and the /after are specified, DIX will evaluate
the /when first, and if that succeeds, the /after is tested.
/do="dix_command[;dix_command...]"
When a watchpoint is found (after the /when and the /after are
satisfied, DIX will execute the commands. All commands are
executed, even if one of them is a "go" or a "step"
set break line [level\]nn Break on line number
set break label [level\]labelname Break on a label
set break call Break on call/gosub
set break return Break on return/exit
set break message severe|error|warning Break on error level
set break instruction instructionname Break on a specific instruction
Instruction names are the normal DIX command verbs, the @ command
is named AT, and the assignment (a=10) as ASSIGN.
Both nn and labelname may be prefixed by level\ to
set breaks on lower levels
Supported qualifiers
/when="dix_command"
If the breakpoint is reached, DIX will execute the code specified
The result must be a logical. It the expression evaluates to true
the breakpoint will be taken (but the /after must be satisfied too)
/after=n
Take the breakpoint after the 'n'th time.
If both the /when and the /after are specified, DIX will evaluate
the /when first, and if that succeeds, the /after is tested.
/do="dix_command[;dix_command...]"
When a breakpoint is found (after the /when and the /after are
satisfied, DIX will execute the commands. All commands are
executed, even if one of them is a "go" or a "step"
/temporary
Remove the breakpoint when it is reached
set trace line [level\]nn Trace on line number
set trace label [level\]labelname Trace on a label
set trace call Trace on call/gosub
set trace return Trace on Exit/return
set trace message severe|error|warning Trace on error level
set trace instruction instructionname Trace on a specific instruction
Instruction names are the normal DIX command verbs, the @ command
is named AT, and the assignment (a=10) as ASSIGN.
Both nn and labelname may be prefixed by level\ to
set tracepoints on lower levels
Supported qualifiers
/when="dix_command"
If the tracepoint is reached, DIX will execute the code specified
The result must be a logical. It the expression evaluates to true
the tracepoint will be taken (but the /after must be satisfied too)
/after=n
Take the tracepoint after the 'n'th time.
If both the /when and the /after are specified, DIX will evaluate
the /when first, and if that succeeds, the /after is tested.
/do="dix_command[;dix_command...]"
When a tracepoint is found (after the /when and the /after are
satisfied, DIX will execute the commands. All commands are
executed, even if one of them is a "go" or a "step"
/temporary Remove the breakpoint when it is reached
show break Show breakpoints show trace Show tracepoints show watch Show watchpoints show calls Show call stack
Step[/into] [line][label]
(re)start processing until the next line
optionally a line or label may be specified.
DIX will restart execution from that line
line or label must be in the current level
if the line is a @ line, DIX will not debug the
deeper level unless the /into qualifier is specified
type Type the current line type nn Type line nn type nn:mm Type line nn:mm type * Type all lines type level\.. Type lines from lower levels If you specify /full DIX will also display the level and the command procedure.
ADD DESCRIPTION name[,name...]
ADD VIEW name[,name...]
ADD PARAMETER
ADD VALUE
ADD TYPE
ADD SELECT_FILE
Add new descriptions/views to the current file.
See also the remove command.
Add description mask
Will add the descriptions matching mask to the current file
Parameter/values are a special type of symbol
By default parameter/type/value names are not case sensitive,
unless DIX is put into case sensitive mode, see the help about
"set case"
Parameter, type and values can be defined GLOBAL (/global) or
at a specified level via /LEVEL=n.
n can be 1..current command level
or -1..-n N levels up
The default level is the current command level
ADD PARAMETER name=value Cannot be changed
ADD VALUE=value May be modified
Types are used to specify enumeration types for symbols
See also the help about [DIX/HELP] interactive declaration
ADD TYPE name=value
Example:
DIX> Add type myvals="0=zero,1=one,10=big" !define a type to be used
DIX> a=10 !add a new symbol
DIX> Set symb a/type=#myvalues !add a fieldname list to the
!symbol
DIX> Evaluate a !Evaluate will not use fieldnames
10
DIX> show symb a !show symbol will use the
A=big !fieldnames
DIX> show symb/fu a !show all info about a
A(Integer*4,Level=1.0)=big
Type : #MYVALUE (0=zero,10=big)
syntax : Add select_file name [/qualifiers]
Make a file selection symbol to be used with F$SEARCH, F$SELECT_FILE
or the directory/select=select=name
You can describe which attributes a file must have
Example:
DIX> Add sele big /selec=size=min=1000 !Create a selec mask for files bigger
! than 1000 blocks
DIX> big_f = f$search("*.*",,,,,"big") !return the first file with size>=1000
DIX> say f$select("big_file.dat","big") !return true/false if file big_fil.dat
!is bugger/smaller than 1000 blocks
For more help see also the help about directory/select
Define the pattern in the global symbol table. By default the patterns
is stored at the current command level.
/level=n Define the pattern in the command procedure level n. n can be 1,2,... Absolute level -1,-2 Lower levels
/select=(tag, tag..)
Tag can be
version=(min=nn,max=nn) !check file version number
size=(min=nn,max=nn) !check file size (EOF)
allsize=(min=nn,max=nn) !check file size (Allocated)
headers=(min=nn,max=nn) !check number of file headers
extents=(min=nn,max=nn) !check number of file extents
acl.size=(min=nn,nax=nn) !check ACL size in bytes
acl.count=(min=nn,max=nn) !check ACL count
fileid=(min=nn,max=nn) !check fileid
unused_size=nn !check unused size (between END_OF_FILE and allocated)
by_owner=user !check on username
organization=tag !check file organization
tag can be seq,rel,idx,syml
record_format=tag !check record format
tag can be udf,fix,var,vfc,stream,stream_lf,stream_cr
journal=tag !check journal files
tag can be journal, aij,bij,ru
flag=(tag) !check file flags
tag can be nobackup,locked,nomovefile,wbcache,readverify,
writeverify,directory,badacl,badblock,nocharge,erase
/UIC=UIC Add a UIC selection
/since=date Add a since date. See also the /created, /modified, /backup and /expired
/before=date Add a before date. See also the /created, /modified, /backup and /expired
The /since and/or /before are checked against the creation date
The /since and/or /before are checked against the modification date
The /since and/or /before are checked against the backup date
The /since and/or /before are checked against the expiration date
ANALYZE [/DISPLAY=(option,option..)] [/VERBOSE=n]
option can be
KEY=n : Analyze only key
LEVEL=n : Analyze only key level 'n'
LEVEL="+n" : Analyze only the top 'n' key levels
LEVEL=-n : Analyze only the lowest '-n' key levels
[NO]DATA : Analyze the data buckets too. Since this can take a lot
of time, this is not the default.
[NO]SIDR : Display the SIDR records, this is the default
[NO]KEYBUCKETS : Display the key buckets records, this is the default
DETAIL : The amount of detail 0..4
Be careful, this can generate a lot of output.
OUTPUT=filename : If verbose is >0, the extra output can be redirected
to another file. The normal output will be in the file
specified with /OUTPUT=filename
START=n : Start analysis at bucket 'n' in chain, default=0
(n may be negative, last but 'n' bucket of the chain)
END=n : End analysis at bucket 'n' in chain, default is whole chain
(n may be negative, last but 'n' bucket of the chain)
VERBOSE=0 : Just the normal output
VERBOSE=1 : Print out the buckets read
VERBOSE=2 : Detail 1 and the bucket header info
VERBOSE=3 : Detail 2 and SIDR records and the data records
VERBOSE=4 : Detail 3 and the SIDR pointers
If verbose>0, the output of the detail option can be redirected to another
file. See the OUTPUT= option
For more info see the help about [DIX/HELP] modes analyze
![DIX MODES analyze]DIX modes analyze
symbolname=value fieldname=value You can assign values to symbols or fieldnames Fieldnames can also be assigned via the deposit command There are also four special formats name++ name=name+1 name-- name=name-1 name+=n name=name+n name-=n name=name-n For these formats only [u]integer types (symbols or fields) are allowed In DCL compatible mode the symbol[d1,d2]=expression and :=expression are also supported for strings and integers (like DCL)
symbolname = value
Symbol may be a table or a subset of a table. In this case value
must either be a scalar or have the same shape as the symbol
Examples:
a=10 !scalar assignment
decl a(3,3)/integer !declare an integer table(3,3)
a(*,*) = 0 !all values(9) to 0
a(3,*) = 10 !all values(3,*) (3 of them) to 10
a(3,*) = [1,2,3] !fill a(3,*) with 1,2,3 respectively
For indexed symbols you can also modify an existing entry or add a new
one the indexed symbol
declare/index=char test/int !create an indexed symbol test of type integer
test("a")=1 !add a new element
test("a")=2 !modify element "a"
test("b")=4 !add a new element "b"
test("c")=test("c") + 1 !update element "c"
or create it if not yet present
For description type symbols you may change all individual fields
DIX> load/descr sys$input mydesc
integer*4 testi
real*4 testr
character*20 testc
^Z
%DIX-I-DESADD, Description sys$input:mydesc added
DIX> show des/loaded mydesc
Name : sys$input:mydesc
Status : Fixed
INTEGER*4 TESTI
REAL_F*4 TESTR
CHARACTER*20 TESTC
No variable Data lines
DIX> declare/des=mydesc mysymb
DIX> show symb mysymb/fu/all
MYSYMB(Descr=sys$input:mydesc,Level=1.0)=2 descriptions,2 expanded
MYSYMB.TESTI=0
0 INTEGER*4 TESTI
MYSYMB.TESTR=0.0
4 REAL_F*4 TESTR
MYSYMB.TESTC=
8 CHARACTER*20 TESTC
DIX> mysymb.testi = 12
DIX> mysymb.testc = "abcd"
DIX> show symb mysymb/all
MYSYMB.TESTI=0
MYSYMB.TESTR=0.0
MYSYMB.TESTC="abc"
DIX> show symb mysymb.*i/all
fieldname=value
Fieldname may be the name of one field, or an table
Fieldname may not contain wildcards (Except in the final dimension)
value must have the same shape as fieldname
Examples:
suppose we have a field test.value(2,2) somewhere
test.value(1,1) = 10 !assign to value(1,1)
test.value = 10 !assign 10 to all values (4 fields)
test.value(1,*) = 10 !assign 10 to (1,*) (2 fields)
The following is not supported
Suppose we have a test(3).value(2,2)
test(1).value(1,1) is valid, but
test(*).value(1,1) not
You may also assign values to fieldnames via the DEPOSIT command
The differences are
deposit
allows wildcarding of the name (a text match is done)
the text after the = is interpreted as text, and will be converted
according the fieldname type.
assign
Does not allow wildcarding of the name, but it does support
wildcarding in the lowest dimension
The piece after the = is evaluated as an expression, and must
deliver a result of the same (or compatible) type as the field
Syntax BACK [/LOG] This command allows you to return to the file from which you followed a link to this file. Up to 16 jumps are saved in a history buffer. See also the FOLLOW command
If the break command is found in a command file, DIX jumps to the
DIX debugger.
The command is ignored when issued from the terminal
CD[/log] [string]
A fast way to change directories. You can also use the SET DEFAULT command,
but CD is easier to use.
SDF is a synonym for CD
Syntax:
CD token[.token[.token...]
Token can be
1. The beginning of the directory or a complete device name.
If the directory is not unique, DIX will print a list of
directories, and lets you choose one.
If token ends with a :, DIX will change the current disk
to the token.
2. # : This is the SYS$LOGIN of the current user
3. ~username: The login directory of user 'username'. If username is not
specified, the current user will be used, and it is
equivalent to #. This may need GRPPRV or SYSPRV privilege
4. - Go one level up. If this is the last item on the line you
must add a . to prevent DIX from asking for a continuation
line.
Instead of the . as separator, the / or the \ can also be used.
If the / or the \ is typed as the first character, DIX will jump to
the [000000] directory of the current disk.
Examples:
DIX> CD !display current directories
Current directory:USER50:[STUBBF.PROGRAMS.DIX.SOURCE]
Subdirectories found : DATA T4 TEST
DIX> CD T !go to a directory starting with a T (there are 2)
1:T4 2:TEST !DIX will display all matches and ask you for a choice
Choose 1..2:2
New directory:USER50:[STUBBF.PROGRAMS.DIX.SOURCE.TEST]
DIX> cd -. !go one up again
New directory:USER50:[STUBBF.PROGRAMS.DIX.SOURCE]
DIX> CD ~ !go to my login directory is the same as CD #
New directory:USER50:[STUBBF]
DIX> CD P.D.S !I want to go back to programs.dix.source
1:PCSI 2:PERF 3:PROGRAMS
Choose 1..3:3 !choose programs, and now we again have multiple matches
1:DELETE_INTRUSIONS 2:DISKSTAT 3:DISTRIBUTED 4:DIX
Choose 1..4:4 !Choose DIX, and now there are no more ambiguous directories
New directory:USER50:[STUBBF.PROGRAMS.DIX.SOURCE]
DIX> CD ~OTHERUSER !goto login of otheruser (needs privileges)
New directory:USER100:[OTHERUSER]
CHARACTER[/level=n][/global] name[dimensions][=value] CHARACTER name[=value] Declare a new symbol of type "CHARACTER" and optionally give a value. The value (if present) must be an expression of type "character" Example: CHARACTER mystring="Heading line" See the help about [DIX/HELP] interactive declaration
Compute sha* and md5 checksums of files or strings
checksum/xor=[vms] parameter [/qualifiers] !compute xor hash
checksum/sum parameter [/qualifiers] !compute sum hash
checksum/md5 parameter [/qualifiers] !compute md5 hash
checksum/sha=n parameter [/qualifiers] !compute sha hash
! n=1,224,256,384,512, default=256
DIX will compute the data only from the record contents (if not in block_mode)
If you copy the file to i.e. windows, the checksum may be different. This
has two possible reasons.
a. For text files (ASCII transfer) the windows checksum utility checks
all data in the file including the separating CRLFs.
See the /add qualifier for a possible solution
b. If you checksum a .ZIP file, RMS will read the file as 512 byte records
but the last record may not be completely filled (END_OF_FILE in the middle of
the block). Use the /exact_file_size qualifier for a solution.
The same applies if you checksum the file in block mode.
For the sum and xor checksum you can specify the /byte,/word and /longword
qualifier to select the size of the xor/sum value.
The default size is longword.
If the size in /word or /long, and the record does not contain a multiple
of bytes (2 or 4), the record will be appended by binary 0.
If the user specifies /xor=vms, the last bytes will be added instead
the appending with binary 0. This matches the OpenVMS checksum/algor=xor
method.
Parameter is
a filename to checksum or a string (if /string is specified)
Qualifiers
The parameter is a string and not a filename.
/symbol[=name] Store result in a DIX symbol "name". The default value is CHECKSUM$CHECKSUM If you also specify /dcl, the result will be in a DCL symbol If you are running in command mode (DIX/COMMAND), /DCL is default
If /symbol is specified, the result must be in a DCL symbol If you are running in command mode (DIX/COMMAND), /DCL is default
/logical[=name] Store result in a process logical. The default value is CHECKSUM$CHECKSUM
If /logical is specified, the logical will be stored in a job logical name table
/block_mode
Open the file in blockmode, normally only records are processed
In this case all bytes in the file are used, also the record separators.
See also the /exact_file_size qualifier.
For the sum and xor checksum only. Set the size of the checksum to byte.
For the sum and xor checksum only. Set the size of the checksum to word (2 bytes).
For the sum and xor checksum only. Set the size of the checksum to longword (4 bytes).
/skip=tag
Skip certain characters in input.
Tag can be
LF : Skip LF character
CR : Skip LF character
BOTH : Skip CR and LF character, this is the default /skip
/add=tag
Add certain characters after each record.
Tag can be
LF : Add LF character
CR : Add LF character
BOTH : Add CR and LF character, this is the default /add
For fixed files only. The last record can have less than the fixed record length. RMS will always report the fixed record length. With this switch the checksum will only be computed on the real file size. This may/will be useful for .ZIP files
CHKA Check access to files
Syntax : CHKA filename user /qualifiers
Check if user has access to the file(s).
If no access is possible, DIX will check upper directories until
access is possible.
CHKA will show the ACL or privilege used to access the file
If no ACL or privilege is shown, access will be via ownership/fileprotection
You need privileges (SYSPRV or BYPASS) if you want to query for
anybody but yourself
Parameters
filename : The file to check
User : The user to check for
Qualifiers
Access qualifiers
/read : See if read access is possible
/write : See if write access is possible
/update : See if update access is possible
/delete : See if delete access is possible
the /read is default, and only one access qualifier may be used
Other qualifiers
/defpriv : The default user privileges instead of the authorized
/batch : See is access in batch is allowed
/all_level: Print access of all upper directories
/ids=(id,id..) : Extra identifiers to grant to the user
(so i.e. /batch is the same as /ids=batch)
Examples:
User STUBBF has SYSPRV in authorized, but not in default privileges
DIX> chka sys$system:zic.exe !Read access to zic.exe?
R access to SYS$COMMON:[SYSEXE]ZIC.EXE;1
DIX> chka sys$system:zic.exe /write !write access to zic.exe?
W access to SYS$COMMON:[SYSEXE]ZIC.EXE;1
because priv SYSPRV
DIX> chka sys$system:zic.exe /write/defp !write access to zic.exe with
default privileges?
NO W access to SYS$COMMON:[SYSEXE]ZIC.EXE;1
NO W access to DISK$ALPHASYS:[VMS$COMMON]SYSEXE.DIR;1
E access to DISK$ALPHASYS:[000000]VMS$COMMON.DIR;1
%SYSTEM-F-NOPRIV, insufficient privilege or object protection violation
If ACLS are present like on test.dix. User TEST has no extra privileges.
$ DIR/SEC TEST.DIX
TEST.DIX;13 [STUBBF] (RWED,RWED,RE,)
(IDENTIFIER=[TEST],ACCESS=WRITE)
DIX> chka TEST.dix test !user test has no read access
NO R access to USER50:[STUBBF.PROGRAMS.DIX.SOURCE]TEST.DIX;1
E access to DISK$ALPHASYS:[STUBBF.PROGRAMS.DIX]SOURCE.DIR;1
%SYSTEM-F-NOPRIV, insufficient privilege or object protection violation
DIX> chka test.dix test/write !user test has write access by ACL
W access to USER50:[STUBBF.PROGRAMS.DIX.SOURCE]TEST.DIX;1
Because of ACL (IDENTIFIER=[TEST],ACCESS=WRITE)
CLOSE[/qualifiers] Close the current file
CLOSE[/qualifiers] mask Close file with tag=mask or filename matches mask
Mask can contain wildcards
Will close the file(s) that match 'mask'. If the current file is closed
the first file in the list will become the current file.
/DISPOSITION=what What to with the file when it's closed what can be KEEP : This is the default, do nothing SUBMIT : Submit the file to SYS$BATCH PRINT : Print the file to SYS$PRINT DELETE : Delete the file Only supported for RMS files, no Fast I/O or library module
/ERROR=label If the CLOSE results in an error, DIX will jump to the label. If the /ERROR qualifier is not specified, DIX will look for the ON condition.
/LOG /NOLOG (Default) If you specify /log, DIX will show the file(s) closed. Also the new current file is displayed (or the no files open message)
COMPILE dixfile [/qualifiers]
Compile the DIX file to a compiled file. This may load a little faster,
but runtime performance will be the same, since DIX precompiles all scripts.
However using compiled scripts makes it (nearly) impossible for users to
adapt the script, certainly if you do not specify /include
and even more if you specify /encrypt
Supported qualifiers
/INCLUDE_SOURCE : Include the source of the script. This allows the user
to display the lines during execution if SET VER is on
/ENCRYPT : Encrypt the output file. This makes it very hard to
change the commands.
/OUTPUT=filename: The default filename is 'INPUTFILE'_COMPILED
/LOG : Display logging info
/COMPRESS : Compress the output file. This makes the file smaller.
This is the default
/PASSWORD=string: Add a password for the file. This will also set the
ENCRYPT flag. On execution the user must specify the
password.
/DCL_COMPATIBLE[=compatible|fallback]
: Set the DCL compatible mode
The default is /NODCL, /DCL is /DCL=compatible
/VERB[=none|always|reading]
: Set the verb substitution
The default is /NOVERB, /VERB is /DCL=always
See the help about "intera file_and_commands_in_file"
Example:
DIX> compile script/log !compile script.dix to script.dix_compiled
Compiled DISK:[DIR]SCRIPT.DIX/COMPRES;4
/output=DISK:[DIR]SCRIPT.DIX_COMPILED;2
20 records saved containing 3055 bytes
DIX> @script.dix_compiled !execute compiled script
....
DIX> Compile a.com/DCL !compile a DCL command file
Convert [/qualifiers] inputfile[,inputfile] outputfile Convert an input file to an output file. Just like DCL convert.
Controls whether converted records from an input file are appended to an existing sequential file. Normally DIX would create a new output file, but if you specify /append, DIX will append all records from inputfile to outputfile.
Specifies whether an exceptions file (file type .EXC) is to be
generated during the conversion.
Format
/EXCEPTIONS_FILE [=filespec]
/NOEXCEPTIONS_FILE (Default)
Filespec is by default SYS$OUTPUT
Controls whether the DIX exits when it encounters an exception record. By default, DIX continues processing records when it encounters an exception record.
Specifies whether the DIX uses a fast-loading algorithm for
indexed files.
Format
/FAST_LOAD (Default)
/NOFAST_LOAD
Indicates that an FDL file is to be used in creating the output
file.
Format
/FDL=fdlfile
Specifies the FDL file to be used in creating the output file.
The newly created output file will have the name specified by the
fdlfile command parameter; this name overrides any file name
specified in the FDL file.
The default file type for the FDL file is .FDL.
The FDL file can also be specified inline. If the first character
of the FDL file is a ; DIX will treat the filespec as a (; separated)
string that contains the FDL specifications.
Example:
DIX> convert/fdl=x a b
Convert file a to b using FDL file x.fdl
DIX> convert/fdl="; file ; org seq" a b
Will convert file a to b as if an FDL file
FILE
ORG SEQ
is given.
Controls whether to override the bucket fill percentage parameter
associated with the output file.
Format
/FILL_BUCKETS
/NOFILL_BUCKETS (Default)
If you specify /FILL_BUCKETS, the DIX fills the output file
buckets with as many records as possible.
This option is valid only for indexed output files.
Controls file conversions between files having variable-length
with fixed-length control field (VFC) records and files having
other record formats.
Format
ED_CONTROL
/NOFIXED_CONTROL (Default)
This qualifier applies only to conversions where either the input
or the output file, but not both, uses VFC records. This option
is applicable only to sequential files.
o If you specify /FIXED_CONTROL and the input file uses VFC
records but the output file does not, the fixed-length control
field from the input record is inserted into the output record
as data.
o If you specify /FIXED_CONTROL and the output file has VFC
records but the input file does not, the leading part of the
input record is used to fill the fixed-length control part of
the output record.
o If you specify /NOFIXED_CONTROL and the input file uses VFC
records but the output file does not, the fixed-length control
field from the input record is not included as data in the
output record.
o If you specify /NOFIXED_CONTROL and the output file has
VFC records but the input file does not, the control field
attached to the output record is set to null.
Directs DIX to read records from an indexed file
using a specified key of reference, such as the primary key, the
first alternate key, or the second alternate key.
Format
/KEY=n
Specifies that records are to be inserted into their proper
position in an existing indexed file.
Format
/MERGE
/NOMERGE (Default)
The /MERGE qualifier is useful when your input records are not
sorted and you do not want them to be sorted as they are loaded
into an output file.
If you specify both /MERGE and /CREATE, /MERGE overrides the
/CREATE qualifier.
Determines whether short records are to be padded.
Format
/PAD [=[%b]x]
/NOPAD (Default)
Specifies that the short records are to be padded with either
ASCII characters (A through Z, a through z, or 0 through 9) or
numeric values.
To specify x as a numeric value, you must specify the numeric
base using the percent symbol (%) followed by one of the
following characters:
D Indicates that x is a decimal number.
O Indicates that x is a octal number.
X Indicates that x is a hexadecimal number.
B Indicates that x is a binary number.
Specifies the prolog version number of the output indexed file.
Format
/PROLOG=n
Specifies the prolog number 1, 2, or 3.
Specifies whether each input record is to be read from the file a second time and compared to the record originally read.
Increases DIX's performance by reducing the number of required passes through the input data. This is accomplished by placing alternate key information into the CONVWORK file. Format /SECONDARY=n Qualifier Value n Specifies the number of alternate keys that will be loaded to the CONVWORK file with each pass through the input data. This qualifier is valid when you are fast-loading a file with more than one alternate key. This option allows CONVERT to use more disk space for its work file than would be used by default. The default number of alternate keys written to the CONVWORK file is 1.
Specifies whether the input file is to be opened for sharing with other processes during the conversion.
Specifies whether the input file is to be sorted before being loaded into an indexed file. The sort is done according to the primary key of the output file.
Determines whether statistics about the file conversion are to
be displayed.
Format
/STATISTICS[=keyword]
/NOSTATISTICS (Default)
Keyword Meaning
BRIEF Displays a summary of the file conversion at the
completion of the operation. This is the default.
FULL Displays summary information at the completion of
each key load containing Sort and Load statistics
for the key. A summary of the file conversion is
also displayed at the completion of the operation.
Specifies whether records that exceed the maximum record length for variable-length records, or records that exceed the specified record length for fixed-length records, are to be truncated.
Specifies the number of temporary work files to be used during
the sort process.
Format
/WORK_FILES=n
Specifies the number of work files you want. You can specify 0 or
any value from 1 through 10. The default is 2 work files.
Specifies whether all writes are to be checked by comparing the new disk records with the original records in memory.
Create something CREATE/FILE filename !create new file CREATE/DESCRIPTION new_description !create new description record CREATE/RECORD [tag] !create new record in a file In DIX mode the /RECORD is the default In DCL mode the /FILE is the default
CREATE/RECORD [TAG] Create a new record. The record will be filled with default data - binary null for non-text fields - Spaces for text fields and then allow you to modify this record, and eventually write it out via the Update command. This is the default in DIX mode
CREATE/FILE filename [/fdl=fdlname] [/log] Create a new file. Optionally specify an FDL filename Filename must be specified. If the FDL filename is specified, the file is created using the FDL, otherwise a variable length file is created. This is the default in DIX mode
CREATE/DESCRIPTION name [/qualifiers]
Create a new description. This can be in the system_library (if you have
write access), the user_library or in a file.
Supported qualifiers
/SYSTEM_LIBRARY The new description will be placed in the system_library
/USER_LIBRARY The new description will be placed in the user_library
If you specify nothing, the new description will be in a
file in the current directory.
/TPU Use the TPU editor for editing
/EDT Use the EDT editor for editing
If no editor is specified, use the internal editor
is an alias for CUT
CUT[/qualifiers] [TAG]
COPY is an alias for the word
Save current record values to save area (clipboard) or to a symbol.
TAG is optional. If not specified, the current file is used, otherwise
the file specified by 'TAG' is used.
The PASTE command can be used to restore data from the save area.
See the help about [DIX/HELP] interactive PASTE
Qualifiers can be
/SYMBOL=symbolname
Save the data to a DIX symbol
You can use the PASTE command to restore the record from this save-area
or symbol.
/DCL
The symbol name is a DCL symbol. There are limitations on the length of
DCL symbols, and if you exceed them, you will see an error
/GLOBAL
The DCL symbol is a global.
If the file is a VFC file, and the /symbol is present, 2 symbols are created
1. symbolname : Containing the record data value
2. symbolname_VFC : Containing the VFC data
Skip the rest of the for, repeat, while, loop block and resume execution
at the corresponding ENDxxx statement to start the next iteration.
This command does not exit the current block (as the LEAVE command).
Example:
nrec = 0
REPEAT
read !read next record
if (fieldname=5) cycle !if a fieldname = 5, skip the rest
fieldname = fieldname+1 !update the fieldname
UPDATE !and write out to the file
nrec = nrec + 1 !and count
UNTIL nrec=10 !when 10 processed stop
DATE [/type=enum_type][/level=n][/global] name[dimensions][=value]
DATE /INDEX=indextype name Declare an indexed symbol of type integer
DATE name[=value]
Declare a new symbol of type "DATE" and optionally give a value.
The value must be of type "date"
See the help about [DIX/INT] interact declaration for the enum_type
See the help about [DIX/INT] interact symbols indexed_symbols for /INDEX
Example:
DATE start_date=f$date("1-JAN-2004")
See the help about [DIX/HELP] interact declaration
Deassign logicalname[/qualifiers]
Supported qualifiers are
Qualifier for the table
/cluster
/system
/group
/job
/process
/table=tablename
Qualifiers for the mode of the logical
/kernel
/executive
/supervisor
/user
syntax
decent [/qualifiers] parameter
Decent the DCL procedure or DIX procedure
- Replace all verbs/qualifiers by the full text
- Uppercase/Lowercase all verbs/quals..
- Create a nice indentation around if-then-else-endif
- and lots more, see further help
Parameter can be
- A single command to be decented
- sys$input, sys$command. Dix will ask for a lines to be decented
until the user types a ^Z. The user may also change the settings
with a command starting with a / f.e. /fix=all. /explain is
also supported. Processing a file is supported via @filename
See example below
- dclfile, they may contain wildcards, all files matching will be processed
See also /output and /log and /confirm
Qualifiers are
$dix/command decent sys$input.!no fixes no cases no indent
DCL-Line:sh ti
$SHOW TIME .!all upper
$/explain
Settings after Change
/FIX=NOALL
/ALLOW=NOALL
/FORMAT=NOALL
/WARNING=NOALL
/SPACES=NOALL
/LOWER=NOALL
/UPPER=NOALL
/COMMENT=NOALL
/INDENT=(Initial=1,Before=0,After=0,Labels=0,Continuation=0)
/Twophases/Dcl/Lax_quotes
DCL-Line:/lower=(all,nolabel)/upper=label/indent=initial=8/expl !change setting
Settings after Change !and show
/ALLOW=NOALL
/FORMAT=NOALL
/WARNING=NOALL
/SPACES=NOALL
/LOWER=(Verbs,Qualifiers,Lexicals,Keywords,Data,Strings,Filenames,Logicals,-
Usernames,Input,Symbols)
/UPPER=(Labels)
/COMMENT=NOALL
/INDENT=(Initial=8,Before=0,After=0,Labels=0,Continuation=0)
/Twophases/Dcl/Lax_quotes
DCL-Line:sh ti !try again, now lowercase and at pos 8
$ show time
DCL-Line:label:sh ti !but labels uppercase
$LABEL: show time
DCL-LINE:@proc !will process proc.com in dclmode or proc.dix in dix mode
...
... !output from proc.*
...
DCL-Line:/nodcl.!switch to dix mode
DIX-Line:sh al
SHOW ALL !DIX command (no $ in front)
Line:^Z
$
When you specify /auto, dix will read the first line of the file to see if there is a $ in front. If so, DIX assumes the input file is a DCL command procedure. If the $ is not found it must be a DIX script See also the /dcl and /dix qualifiers. There is no default filetype in this case. Also the /fix=missing_dollar is ignored.
/procedure
/noprocedure
the argument can be either the name of a DCL procedure or a one-line command.
DIX will check p1 for characters / space or =
If any of them is present, the default is /noprocedure so a one-line command
If none is present the default is /procedure, so a procedure name
The qualifier can be used to override this default behaviour
if present
The parameter will be used as the name of DCL procedure, default .COM
if negated
the parameter will be used as a one line DCL command.
/comment=(position=nn, at_end, compress)
position Position in the line for the comments
Comment lines without statements are printed as found
Comment text after statements are by default printed directly after the
command.
If you specify /comment=pos=number, comments will be placed at column 'number'
If you specify /comment=pos=-number, comments will be placed at -'number' after
the command line
If you specify /comment=pos=tab, DIX will insert spaces until a TAB position
This can also be achieved via /spaces=comment=tab, but that will insert
a TAB character, and overrule the comment=pos=tab.
at_end Specifies that comments will only be placed on the
last continuation line
compress Remove leading spaces from comment
keep If the comment is only the comment character !
DIX will remove the ! character also
if you specify comment=keep the ! will be retained
align_last If comment must be printed after the code line,
DIX will normally print the comment at position 2
If you specify align_last, the comment is aligned
with the previous comment
If you specify /confirm, DIX will ask you if you want to process the file. This may be useful if there is a wildcard in the inputfile
The input file is a dcl command procedure. You can either use /dcl or /dix (input file is dix procedure) If neither is specified (or /auto), DIX will look into the file and if the first character is a $, DIX will assume the file is a DCL procedure. If you specify /dcl the default input filetype is .com, If neither /dcl nor /dix is specified there is no default filetype
The input file is a dix command procedure. You can either use /dcl or /dix (input file is dix procedure) If neither is specified (or /auto), DIX will look into the file and if the first character is a $, DIX will assume the file is a DCL procedure. If you specify /dix the default input filetype is .dix If neither /dcl nor /dix is specified there is no default filetype
/explain or / /explain=all or // DIX will print out the various settings. If you specify /explain=all this will be done twice, once after the settings file is processed and once after the command line processed.
If you specify /image and the command is $ MCR IMAGE ..., DIX will look in the image (default filename SYS$SYSTEM:.EXE) for a valid DCL table. If it finds one, that table will be used to check (and expand) the command. This also works if you have a foreign command definition like $ UAF:=$AUTHORIZE or $ UAF:=MCR AUTHORIZE. It even understands extra text in the symbol like $ SYSMANIO:=$SYSMAN IO In the translated command the verb will remain the same. This qualifier is not default.
/indent=(token...)
This sets the values for indentation
Token can be
initial=value Set the initial indentation, value can be tab or a number
before_then=value Set the indentation of the if/then/else/endif command
value is a number of TAB
after_then=value Set the indentation of lines between then[,else] and endif
value is a number of TAB
label=n Set the indentation for labels, value may be negative
or CODE to align with code or 0 for toplevel
continuation=n Set the extra indent for continuation lines, this is
added to the current indentation (value may be negative)
datalines=n Set extra indent for datalines that contain commands in
images. 0=same level as command, >0 deeper, <0 less deep
If you do not specify indent=datalines[=value]
no indentation is done. You can disable indentation
for a specific image in the dix_decent.setup file
comment Align comment lines without commands or labels to the code
The default is init=8,before=1,after=1,labels=0,cont=4
/init=filename
Used to define symbols from a command file.
e.g. LOGIN.COM or SYS$SYLOGIN.
Files may be nested.
Change the output format
/format=(tag...)
tag : keep_lex Do not break lines in the middle of a lexical function
blank_before_label Insert a $! before a label
remove_blank_comment Remove all $! lines
qualifier_on_line Try to break on / (new qualifiers)
dash_in_comment Add a - in the comment if the comment is continued
on the next line. DCL ignores this, but it
signals that the comment is on the next line
semicolon Show a ; between comments on multiple lines in
the original source
quoted_on_line Do not break lines within a "" string
header Insert header line(s) in file, see below
add_defaults Show all default qualifiers/parameters too
identifier=user If in an acl the uic is something like
identifier=[tcpip$aux,tcpip$ftp] DIX will remove
the group info, so the result will be
identifier=tcpip$ftp
identifier=keep Leave the original idname intact so if you
specified id=[1,4] it will not be changed to
id=system
labels Add the name of the subroutine to all labels
found in the subroutine, and also the references
If header is set, the user can select the format of the header in the
dix_decent.setup file via the header command.
DIX will check if the filename occurs in the comment of first 10 lines.
If the name is found, DIX assumes all is okay, and does not insert
extra lines. If the filename is not found DIX will add one or more
lines. See the documentation about the setup file for more info
/fix=(tag...) Fix a number of possible issues with the DCL code tag can be
Add the name of the subroutine in the comment after an EXIT or ENDSUBROUTINE statement.Also the first label of before a return will be inserted in the comment after the return. If the name to be inserted is already present, DIX will not insert it again. The exit of the main level can also be appended if /fix=add_name=all is used. The name added comes from the /name qualifier. See the help there
Replace the assign statement by the define statement. It also
removes the optional : in the second argument of assign
so assign a b: will be define b a
If /QUEUE or /MERGE is present, the change will not be done
If a ' is present, the change will not be done (maybe it
contains /mer or /que) unless you specify /fix=assign=all
if you specify /fix=backslash DIX will remove the \ after the verb this is sometimes used to prevent symbol substitution as in $delete\ x.x;
DIX will remove unnecessary () in keywords and if statements
keywords
direct/sel=(size=(min=10)) will be directory/select=size=min=10
The brackets will not be removed if one of the values contains a '
since that expression may contain a ,
Like
$privs="cmknrl,oper"
$set proc/priv=('privs')
If statements
if (expre) then command to if expre then command
The brackets will not be removed if expr contains ( or )
Assignments
a=(a+b)
DIX will replace the = sign by a : in sort or search /key=(pos=nn,siz=nn) to pos:nn,siz:nn and in protection /prot=rw to /prot:rw
If datalines (in deck part) DIX will look in the image from the previous line to see if that image contains a CLD table. If so DIX will try to expand the datalines Example: $ mcr authorize SHO US * !will be replaced by show user * mod abx/priv=cmk !will be replaced by modify abc/privilege=cmkrnl $ copy sys$input tt: text line 1 !will not be replaced since copy has no CLD table text line 2 !(it is in DCL tables)
Automatically insert $DECK/$EOD around lines that do not start with a $
Some directory names are modified
- If a filename contains an emtpty directory string , DIX will remove it
so disk:[]file will become disk:file
- <> are replaced by []
- 0,0 is replaced by 000000
- [000000.abc] will be replaced by [abc]
Replace operators with no trailing . so .eq will be .eq.
Replace EOJ by EXIT
replace multiple ! with only 1 so command !!!comment will be
command !comment
Add $exit to procedure and not yet seen
the $exit will not be added if the last statement is a return or a
endsubroutine
replace f$fao("abc") by "abc" but not if there is a ! in the string
that might be a directive
/fix=filenames
/fix=filenames=default
/fix=filenames=all
DIX will append missing file type for according to the setting
in the dix_decent.setup file. For example
@filename : default type is .COM
submit file : default type is .COM
call file : default type is .COM
submit xyz/log=file : default type is .LOG
type file : Default type is .LIS
print file : Default type is .LIS
run file : default type is .EXE
convert/fdl : default type is .FDL
create/fdl : default type is .FDL
set command : default type is .CLD
If only the name of the file is present, DIX will not add the
default file type, since the name could be a logical. If any
other part of the filename is present DIX will add the default file type.
If you want to add the file type also if the name is the only part
of the filename, specify /fix=filename=all. Doing this may
make the output file not correct. DIX will check if the
name is a defined logical, if so, DIX will not fix the filename
If you specify /fix=filenames=default, and the command is not one
of the specified in the .setup file, and the cld entity contains a
default name, DIX will use that name as default file type
DIX will make two lines of the THEN statement and ELSE statement This is the the only action taken if you specify /fix without keywords
Force all labels to be printed on a separate line
Remove all leading $! lines from the file
Replace show magtape by show device/full
See if a symbol assignment for a foreign command needs to be replaced by mcr so symb="$image .." replaced by symb="mcr image ..." fix=mcr=never !never replace the $ by mcr fix=mcr=allways !always replace the $ by mcr fix=mcr=system !only replace the $ by mcr if image is in sys$system For /fix=all, the default is fix=mcr=system
If you specify /fix=missing_dollars and the input file does not contain any command with a leading dollar DIX will insert a $ in front of all lines This is ignored unless the /dcl is also present.
multiple $! lines will be replaced by a single one
Replace filenames with [_]nla: and [_]nla0: and [_]nl0: to _nl: on places where the parameter is a filename in the CLD if /fix=nl=all is specified also the define/assign is checked
DIX will replace (some) obsolete commands by the newer commands
lexical function f$logical by f$trnlnm
commands SET ACL by SET SECURITY/ACL
SET FILE/PROT by SET SECURITY/PROT
SET FILE/ACL by SET SECURITY/ACL
/fix=output DIX will process the value of the /output qualifier always as type $outfile Some commands have the type defined as quoted_string (tcptrace f.e.). With this fix you override that type This will not work if the the /output has a userdefined type
Replace RUN SYS$SYSTEM:IMAGE by MCR IMAGE (only if nothing else on the line) and MCR SYS$SYSTEM:IMAGE by MCR IMAGE
Replace the symbol:="string" by symbol="string" So remove the superfluous :
Replace
$a="abc" + -
"xyz"
with $a="abcxyz"
but only if the length of the command will be not too long
symbols All symbol assignments local and global symbols=[no]local Only local symbol assignments symbols=[no]global Only global symbol assignments symbols=both both local and global assignments symbols=[no]all All symbol assignments even with long verbs/qualifiers if you specify /fix=symbols, DIX will try to expand the value in the string assigments so if you have sw:=set ter/wi=132, DIX will expand to sw:=set terminal/width=132 This only works if the asignment is a string assignment (:= or ="") and the string starts with a valid verb If you do not specify symbols=all the found verb must be shorter or equal than the actual verb, so if you have the assignment $dd="Openvms/Wri" If the symbols=all is specified (or /fix=all), this statement will be replaced by "openvms/write", since openvms is recognized as a valid verb (open) If the symbols=all is not specified, the statement is not replaced since Openvms is longer than the actual verb open If you had specified dd:="Openvms/dummy" the statement will never be replaced since this is not a valid statement If you had specified dd:="Ope/Wri" the statements will always be replaced with "open/write"
Do tab expansion in the comment part(replace them by spacs). The tab interval is set via the /tabinterval=nn (default 8)
If you specify /fix=trailing_semicolon (or /fix=all), DIX will remove trailing ; from filenames. This is not done for the delete and the directory command) and also not if a version is specified. so type a.a; will become type a.a
Replace the name UCX by TCPIP
Remove unprintables from source (all except TAB)
/fix=unprintables=(tag,...)
DIX will insert statements like ESC[0,8]=28 and replace the escape
characters by ''esc' or 'esc'
For the unprintable names see DIX> show character/unprintables
Tag can be
top : Insert the definitions only once at the beginning of the outputfile
before : Add the definitions just before the command is printed. This may
add the definition more than once.
prefix=intro : By default only the unprintable name is used, but you may
prefix the name by any string, possibly to prevent
other definitions
So /fix=unprin=prefix=DIX_ will add the DIX_ESC definition
remove=(char...)
: Remove some unprintables from the output file
char is the name of the unprintable char. This name can be found
in DIX> show character/unprintable. Only char codes 0..31 can be
removed. Example /fix=unpr=remove=(nul,ff) will remove the
Null and the Form Feed characters
upper : Normally only characters 00..31 and 128..159 , 127 and 255 will be
changed. If you specify /unprint=upper, also chars 160..254 will
be changed.
With top all definitions are declared only once (at the beginning of the
outputfile). With before, each occurrence of an unprintable will result in
a (new) definition.
If you specify /fix=unprintables=top, /twophase is also set
The default for /fix=unprintables is /fix=unprintables=(before,prefix="")
Allow case changing of unrecognized commands
Remove all trailing $! lines from the file
Replace the filename;0 by filename;
All of the above without the filenames.all and nl.all and assign.all
All means all fixes
/name
/name=name
If you specified /fix=add_name=all, DIX will append a name in the comment
of the exit statements of the main level.
If you do not specify /name, DIX will use #main
If you specify /name without a value, DIX will insert the
name of the procedure (not the type)
If you specify /name=myname, DIX will use myname
/length=n
output qualifiers and keywords with maximal n characters or many as
is needed to be unique. THis can be used to keep the length to a reasonable
value
example
$dix/comm dece def a b /tra=con
Result : $ DEFINE A B/TRANSLATION_ATTRIBUTES=CONCEALED
Normal complete expansion
$dix/comm dece/leng=6 def a b /tra=con
Result : $ DEFINE A B/TRANSL=CONCEALED
expansion to maximal 6 letters
$dix/comm dece/leng=1 def a b /tra=con
Result : $ DEFINE A B/TR=CONCEALED
In this case the minimum length to be unique is 2 since there is also
a /table qualifier
If you specify /log DIX will print the name(s) of the file(s) processed.
Define the place where the output is printed if /output is not present, the output will be to SYS$OUTPUT If /output is present, it will be used the define the output filename the name is defaulted with the current input file The default value for /output=.com_dix So if the input parameter is *.com, DIX will create files with .com_dix
/tabinterval=nn default /tabinterval=8 Set the tabsetting for tab expansion
/width=nn Set the width of the output file. This also determines the place where DIX will wrap the command line. By default DIX will not wrap unless the output file is on a terminal. In that case the width of the terminal will be the width.
/lowercase=(token..) By default all items are unchanged, but you can select what you want to be in lowercase, see also /uppercase token can be [no]verbs Lowercase all verbs [no]qualifiers Lowercase all qualifiers [no]lexicals Lowercase all lexical functions [no]keywords Lowercase all keywords (.ne.,.. goto etc...) [no]labels Lowercase the labels [no]strings Lowercase all possible strings in lexical functions [no]filenames Lowercase filenames in f$search and f$parse [no]logicals Lowercase logical/tables names in f$trnlnm (if no CASE argument) [no]usernames Lowercase usernames in filenames node"user pwd"::... [no]input Lowercase datalines not recognized as commands for an image [no]symbols Lowercase symbolnames in "" strings like abc in "ab''abc'cd" All Lowercase all (except in quoted strings) If you specify /lowercase without any token, all items are lowercased (=all) If you specify a token in /loercase as well as in /uppercase, the text will be changed to lowercase (the uppercase is ignored)
/uppercase=(token..) By default all items are unchanged, but you can select what you want to be in uppercase, see also /lowercase token can be [no]verbs Uppercase all verbs [no]qualifiers Uppercase all qualifiers [no]lexicals Uppercase all lexical functions [no]keywords Uppercase all keywords (.ne.,.. goto etc...) [no]labels Uppercase the labels [no]strings Uppercase all possible strings in lexical functions [no]filenames Uppercase filenames in f$search and f$parse [no]logicals Uppercase logical/tables names in f$trnlnm (if no CASE argument) [no]usernames Uppercase usernames in filenames node"user pwd"::... [no]input Uppercase datalines not recognized as commands for an image [no]symbols Lowercase symbolnames in "" strings like abc in "ab''abc'cd" All Uppercase all (except in quoted strings) Other Uppercase all items not mentioned in /lowercase If you specify /uppercase without any token, all items are uppercased (=all)
/allow=(item,item...) Item can be
if you specify allow short, DIX will allow qualifiers that are longer
then the define ones. For example
In the help about show entry the qualifier /user_name is mentioned.
In the CLD file this qualifier is /USER (so a bug in the docs)
So
$ DECENT/NOPROCEDURE show entry/user_name=system
Will not be parsed, But
$ DECE/NOPROCEDURE/ALLOW show entry/user_name=system
Will translate to SHOW ENTRY/USER=SYSTEM
if DIX find a command with a qualifier in quotes it will not process
the line, since this qualifier could initiate a syntax change and
subsequent qualifiers/parameters could have a different meaning
If you specify /allow=qqualifiers, DIX will continue parsing the
command (and hopes the quoted qualifier does not change the sytax)
f.e.
$dir /'tag'/pa could mean
directory/ftp/passive or directory/paged, the /ftp would change the
syntax so that the /pa would become /passive and not /page
Will allow DIX to decent a statement if quotes are present
Example:
if in a file the following statement is found
set acl/object=devi/acl'p4' 'p5' 'p6'
if /allow=quotes is not present you get
Quote found in ACL'P4' and not allowed via /allow=quotes in: set
acl/object=devi/acl'p4' 'p5' 'p6'
And the line will not be changed
If /allow=quotes is present you will get
SET ACL/OBJECT_TYPE=DEVICE/acl'p4' 'p5' 'p6'
DIX has translated/expanded as much as possible
There is a danger that DIX will expand the wrong text, so this is not the
default, see the example below
$tag = "term"
$show 'tag' /p !will expand to show 'tag'/permanent !terminal
$tag = "dev"
$show 'tag'/p !will expand to show 'tag'/page !device
If you have an image that does the DCL parsing and that has only one verb
like in
define verb getuai
parameter p1
qualifier account
qualifier privs
and in the image you do the DCL parse like
call lib$dcl_parse('getuai '//commandline,..)
you may use this in a foreign command like
$ getuai system/account
But in the DCL command is no verb and DIX will complain about that
and cannot parse the line
If you specify/allow=single_verb, and the image has only one verb
(getuai in this case), DIX will parse the line according to the
getuai verb, and does not signal an error. You can also specify a line
in the dix_decent.setup with image getuai/verb=getuai
In some procedures the symbol assignments for new verbs will be done in a line number after the usage of the changed verb. Example: $ gosub init $ dd /full !usage of dd $exit $init: $ dd:= dire/vers=1 !definition of dd $ return By default, DIX will process the lines in order, and will find the usage (in line 2) before the definition (line 5), and not understand the verb dd. If you specify /TWOPHASES, DIX will read the procedure two times - The first time to remember the redefined verbs - The second time to do the real reformatting. and will process the verb dd (as dire/version=1) correctly.
/setup=file /setup=dix_decent A file containing information about how to process datalines in images and how to append filetypes to filenames. The defaultfilename is dix_decent, and the default filetype is .setup DIX will search for this file in the directory where the DIX executable is. comment is ignored (all things after a !) and blank lines are ignored The following lines can occur in the file
Syntax : header "line1","line2"... DIX can insert headerlines in the files if /format=header is specified In the lines can be control codes supported codes are !AS Insert the filename in uppercase !as Insert the filename in lowercase !UAS Insert the filename in uppercase !LAS Insert the filename in lowercase !XAS Insert the filename as is !%D !n%D !nn%D insert the creation date of the file !%I !n%I !nn%I Insert the onwer name There must be at least a filename control code present If the line does not start with a $!, DIX will insert it If no header statement is found it defaults to header="!AS" that will generate a single line $!name.type Example header "File : !AS","Author: !%I","Date : !20%D","" Will generate a header in the file containing $!File : X.COM $!Author: [STUBBF] $!Date : 12-JAN-2020:12:13:14 $! There must be at least one !AS in the header to prevent DIX from adding a new header if the file is processed twice
syntax : Image imagename [newimagename] [/qualifiers]
Imagename is the name of the image
newimagename is the replacement f.e. latcp does not have the cld
definitions, but sys$share:lat$shr does
so you can use : image latcp sys$share:lat$shr
And DIX can find the cld tables there
Qualifiers can be
/noinput Image does not expect input so datalines after it must be comment
/ignore Ignore all cld tables that may have been found in the image
/lower Convert all datalines to lowercase
/upper Convert all datalines to uppercase
/case COnvert all datalines to the setting of /upper=data or /lower=data
/[no]indent Indent datalines. The /indent=data=nn decides how the indentation
is done.
/[no]decent replace tabs by space and remove all multiple spaces
/[no]dcl Support dcl like input
/dcl and /indent are default if the image contains cld tables, but you
may override with /no...
/decent is default if /dcl is present, but may be overruled with /nodecent
If the image is DCL , imagename is the name of a special dcltable file
If no DCL definition is found, the DCLtables in P1 space are used
/verb=verb Some images take the command line and prefix it with its own verb
f.e. if you specify "dix/command decent f.com", DIX takes the command
line "/command decent f.com", prefixes it with DIX and then parse it.
If you specify image dix/verb=dix, DIX can parse these commands, but
if you do not, you will get a warning "verb is ambiguous"
See also the /allow=single_verb, but that only works if there is only one verb
Replace the found verb with a new one. FOr example: the RMU verb is acually RMUx (in my system RMUJ). If you try to decent the RMU commmand, this will come out as RMUJ command. If you do not wnt that, you can replace the found verb (RMUJ) by RMU. Syntax verb foundverb newverb FOund verb may contain wildcards. So in order to solve the above problem add a VERB RMU* RMU to the setup file.
A way to specify which filetype may be added for images in this specific verb
syntax : file verb/parameter=n type [/qualifiers] Set type for parameters
file verb/qualifier=qual type [/quaifiers] Set type for qualifiers
Type may be specified with and without the ., so lis and .lis are the same
Qualifiers can be
/next_qualifier=qual For parametes folowed by a qualifier like link
link a, b/opt, c/lib See example below
/p1=parameter p1 must be parameter
Examples
file link /parameter=1 obj Parameter p1 is a .obj file
file link /parameter=1/next=option .opt if next_qual=/option it is .opt
file link /parameter=1/next=library .olb
file @/parameter=1 com Parameter p1 is a .com
file @/qualifier=output lis /output is a .lis
file set /parameter=2 /p1=command msg set command , p2 is a .msg
file set /parameter=2 /p1=message exe set message , p2 is a .exe
file fms /parameter=1 flb fms, p1 is a .flb
file fms /parameter=2 frm and p2 is a .frm
For more examples see the dix_decent.setup delivered in the kit
Syntax
decent /qualifier....
You can specify a number of command line qualifier settings.
These setttings set a default, but can be overruled via the command line.
You can have multiple decent verbs in the file
The supported qualifiers are
/FIX=, /FORMAT=, /WARNING=, /INDENT=, /SPACES=,
/UPPERCASE=, /LOWERCASE=, /COMMENT=.
f.e.
Suppose you have the decent/fix=all in the dix_decent.setup file
If you do not specify any /fix on the command line, DIX will apply
all fixes, but you can also specify /fix=noucx to prevent the ucx fix.
You can even specify /fix=(noall,ucx) the and ucx fix will be the only one
If you specify /explain=all, you can see what settings are set after the
.setup file and what after the commandline is processed.
/signal[=token] By default DIX will not complain about commands that it cannot parse. It will just send the original text to the output file If you specify /signal, DIX will print messages to sys$command about the commands that cannot be parsed If you specify /signal=inline, DIX will print the messages in the output file at the point they occur. If you specify /signal=begin, DIX will print the messages before the normal output in the output file If you specify /signal=end, DIX will print the messages after the normal output in the output file. If you specify /signal=file=outfile, DIX will prin t the messages in file "outfile"
/spaces=(tag,tag..)
determines where DIX will insert spaces
[no]assign DIX will surround the [:]=[=] operator with spaces
[no]comment=end DIX will insert a space after the comment character !
[no]comment=begin DIX will insert a space after a ! on a line with no label/command
[no]operators DIX will surround the .eq. etc. + / - operators with spaces
[no]arguments DIX will add a space after a list of values
so write out a,b,c will be write out a, b, c
All All of the above
/spaces without any tag is /spaces=all
And two more
comment=tab DIX will insert a tab before the ! on statement lines
/comment=pos=tab will insert spaces until the next tab
position. This will be overruled by /spaces=comment=tab
comment=one DIX will insert a space before the ! on comment only lines
comment
/table=(tablespec,tablespec...) !a list of spefications
or /table=@filename !a file containing the specifictions
!each line has one specification
!default filename is .lst
You can specify a different CLD table for specific images.
Tablespec has the following syntax
imagename#[cldimagename][#option[#option....]]
if imagename is DCL or not present (and also no #) the
cldimagename is used as an alternate DCL tables for all commands
if imagename is present, the cldimagename is used for commands in the
imagename
Example:
/table=([DCL#]mydcltable,sysgen#sysgentable)
will use the mydcltables for the normal commands and
sysgentable for commands in the image sysgen
You can create a sysgentable via
$ edit sysgentable.cld !and add the verbs/param/quals
$ set command/obj sysgentable.cld
$ link sysgentable
The linker will warn you that no transfer address is specified, but you
can ignore that warning
The default file specification is SYS$SYSTEM:.EXE
Options are
#ignore : Do not look for cld tables in this image
#noinput : This image does not expect input
#decent : convert data lines to lowercase (if /lower=data) and do spacing
#lower : just convert to lower case
#indent : indent datalines to current indentation
/warning=(tag[,tag...])
tag can be
duplicates : DIX will warn you about then same
qualifier being specified more than once.
obsolete : DIX will warn you about specific obsolete commands.
global : DIX will warn you about global assignments
quotes : DIX will warn you about missing ' and "
unprint : DIX will warn you if there are unprintable characters in the file
blank : DIX will warn you about empty lines
dots : DIX will warn you about missing .'s after operators e.g. .EQ
single_quotes : DIX will warn about 'a' in a "" string
double_symbols : DIX will warn you if symbols are defined more than once
unbalenced : DIX will warn you about a () inbalence
long_command : DIX will warn you if the fist 4 character of a verb, but
there are more unmatched characters
dash : DIX will warn you about questionable - useage like .and. -k
data_command: DIX will warn you for datalines that might be dcl-commands
empty_data : DIX will warn you for completely empty data lines
lexicals_in_comment : Warn about lex funtions in comment (except f$verify)
dubious_assign : Warn for possible wrong assignment changes
labels : warn for duplicate label names and for unused and
wrongly used labels (f.e. gosub to SUBROUTINE)
devices : Warn for physical devicenames (disk and tape)
short_file : Warn for filenames with only one character in the name
build_in : warn about $status .eq. instead of .eqs. These symbols are strings
all : All of the above
In order to see the warnings you must specify /signal[=inline] also.
You can also disable a specific flag via f.e. /warning=(all,nolabels)
DECIMAL[/level=n][/global] name[=value]
DECIMAL/INDEX=indextype name Declare an indexed symbol of type date
Declare a new symbol of type "DECIMAL" and optionally give a value.
The value must be of type "decimal"
See the help about [DIX/INT] interact symbols indexed_symbols for /INDEX
Example:
DECIMAL counter=f$decimal("12345e12")
See the help about [DIX/HELP] interactive declaration
Syntax, there are 2 formats
format 1 for a limited number of data types
'type'[/type=enum_type] [/level=n][/global][/overwrite] -
name[['type']*size][dimensions][=value]
[,name[dimensions][=value]...]
where type : REAL[*n], CHARACTER, LOGICAL, INTEGER[*n.m] or DATE, DECIMAL
Format 2, more general
DECLARE[/'what'] [/type=enum_type] [/level=n][/global][/overwrite] -
name[*size][dimensions][=value][,name[dimensions][=value]...]
In this case all 'what' can be any datatype. If you do not
specify 'what', the value has an undefined type
name : The name of the local symbol
size : For integer and reals and dates. Set the size to be fixed.
For integer size is nbyte[.nbits] (0.1 .. 8.0)
For reals size is nbyte (4,8,16)
For dates size is nbyte (4,8)
value : An expression that returns a value of the correct type
dimensions: Up to 3 dimensions (d1[,d2[,d3]])
Each dimension has the format [low:]high[:increment]
low : Upper boundary (default 1)
high : High boundary (must be present)
increment : Increment. If present, low must be present too
example (5:10:2) This results in 5,7,9
Symbols declared in this way cannot be declared again (unless with
the /overwrite qualifier), and cannot be assigned
another value than the declared type. See also the help about
[DIX interactive] SET STRICT
You may declare a symbol at a lower level. If you do not specify /level
the symbol will be declared at the current procedure level.
The value of level can be positive (the absolute level value)
or negative (n levels higher). See the examples
/global is the same as /level=0
For integer data you can set the enumeration_type. This is equal to
the functionality of the "fieldnames" of fields. The 'enum_type' may
be defined before with the ADD TYPE statement, or can be specified
within quotes. If you want the use a previously defined type (add type)
you must prefix the name with a # sign. See below
Example:
INTEGER a=10 Declare an integer a with value 10
INTEGER a=20 Will result in an error "symbol A is already defined."
a=20 Is allowed, change the value to 20 (value must be integer)
a="XYZ" Will result in an error "Symbol A has other type than expression"
DECLARE A/INT=10 Is the same as INTEGER A=10
DECLARE X The value X does not have a fixed type
X=10 Value of x is now 10 (integer)
X="String" Value of X is now "String" (character)
Example with defined types:
DIX> ADD TYPE DAYS="0=sun,1=mon,2=tue,3=wed,4=thu,5=fri,6=sat"
DIX> integer/type=#days weekday
DIX> weekday=wed
DIX> Show symbol weekday
WEEKDAY=wed
DIX> evaluate weekday !the eval will always return the binary
3
Example with direct type:
DIX> integer/type="0=sun,1=mon,2=tue,3=wed,4=thu,5=fri,6=sat" weekday
DIX> weekday=wed
DIX> Show symbol weekday
WEEKDAY=wed
DIX> evaluate weekday !the eval will always return the binary
3
Examples for level/global:
suppose we are at procedure level 3.
Declare /level=1 symbol Will declare a symbol at level 1
declare /level=-2 symbol Will declare a symbol at level 1
Declare /global symbol Will declare a symbol at level 0
Declare new Will declare at level 3
Declare/level=-1 new Will declare at level 2
DECLARE[/WHATTYPE][/type=enum_type][/OVERWRITE] -
[/level=n][/global] NAME[*size][DIMENSIONS]=VALUE
DECLARE/INDEX=indextype/whattype value Declare a 'whattype' symbol with an
index of type 'indextype'
This is the more generic form for INTEGER/REAL/DATE.. declarations.
If you specify WHATTYPE, DIX will define a symbol of the specified
type. You cannot change the 'type' of the symbol later.
If you do not specify the "whattype", DIX will define an "unspecified"
symbol that can change types.
Size can be specified like nbyte.nbit for integers and nbyte for reals
DECLARE /int a*0.4 !declare a 4 bit integer
See the help about [DIX/INT] interact declaration for indextype
Example:
DECLARE/LOG Q=false,A/INTEGER=10,B(10,10)/char="test",K
This will declare
1. A Logical value Q with value false
2. An integer symbol A with value 10
3. A character table (10,10) with all elements initialized to "test"
4. A logical value K (undefined)
Also See the help about declarations for more information
define logicals define keys define screenkeys
Define name logical name equivalence_name[,...] The same as the DCL define Supported qualifiers
The following qualifiers determine the table
/CLUSTER
/SYSTEM
/GROUP
/JOB
/PROCESS
/TABLE=nnn
The following qualifiers determine the mode
/EXECUTIVE
/SUPERVISOR
/USER
/KERNEL
The logical name can have the following attributes
/NAME=CONFINE
/NAME=NO_ALIAS
Each equivalence string can have a the following attributes
/CONCEALED
/TERMINAL
DEFINE/KEY keyname string[/qualifiers] Define keys for the INTERACTIVE mode of DIX. See also the RTL Screen Management (SMG) manual about the possible keys. Also note that the f6..f14 keys are normally reserved for command line editing. You may use those keys when you set the terminal to /noline_edit, or if you prefix the key by a ^V. Also see the OpenVMS help about define/key.
/SET_STATE=name DEFINE/KEY keyname string /SET_STATE=NAME. Sets the key driver to state NAME.
/IF_STATE=name DEFINE/KEY keyname string /IF_STATE=NAME If the user enters "keyname", the string "string" will only be returned if a previous key has put the key driver in the "NAME" state. Name is NOT case sensitive.
After the /TERMINATE, the result will be given to DIX.
Echo the translated strings
DEFINE/KEY keyname string/SET_STATE=NAME/LOCK Normally the state is removed after the typing of the next key. If you define the key (and the SET_STATE) with the /LOCK, the state will remain "NAME" until another state is entered.
DEFINE/KEY skey function [/qualifiers] Define a key combination for the screen mode of DIX. skey : The keyboard key : Almost all keys can be used for keydefinitions. function : The DIX function See the help about [DIX/HELP] key_mapping DIX_functions for allowed functions
/SET_STATE=name The parameter "" must be present but the contents are not used. SMGKEY will set a new state (GOLD and BLUE are used by default, but you may redefine the states if you redefine all keys) State names may not contain a - (hyphen) Example: DEFINE/SKEY PF1/SET=GOLD "" DEFINE/SKEY UPPERCASE_A/IF_STATE=GOLD key_print If the user enters PF1-A, the function key_print is executed. For a list of allowed functions, see the help about [DIX/HELP] KEY_MAPPING
/IF_STATE=NAME Specifies a state that must be in effect for the key definition to work. The state name is an alphanumeric string. States are established with the /SET_STATE qualifier State names may not contain a - (hyphen) DEFINE/SKEY PF1/SET=GOLD "" DEFINE/SKEY F20/IF_STATE=GOLD key_print DEFINE/SKEY F20 key_swap_dis If the user enters PF1-F20, the function key_print is executed. It the user enters F20, the function key_swap_dis is executed. For a list of allowed functions, see the help about [DIX/HELP] KEY_MAPPING
/CONTEXT_SCREEN=Screen_context
Define this key only for a specific screen. The default is that the
key is valid in all screens
Valid screens are
DISPLAY_ALL Display data in all formats
SELKEY Select data for a key value
EDIT Edit description
SELFILES Select one file when multiple files match
EDITFILES Select a file (or add one)
SELDESCRIPTIONS Select a description (or add one)
YESNO Yes/No screen
VIEW View all kinds of data
HELP Help screen(s)
EDITMULTIPLE Edit multiple lines (in show description)
EDITLINE Edit a single line
SELFIELD Select a field
SHOWRAW Show data in raw format
SHOWDES Show data in description format
SELDEMO The select demo screen
DISPLAY_SUBF Display all subfields display
DELETE/KEY name[/state=name]
Delete SMG key definitions for interactive mode
DELETE/SKEY/ALL Delete all (Default) keydefinitions for screen mode
DELETE/SKEY skey [/STATE=NAME]
Delete the key definition for skey (possibly with state NAME)
DELETE/SYMBOL See deeper help
DELETE/SYMBOL/LOG[/all] Mask Delete symbol
DELETE/SYMBOL/LOCAL[/all] Mask Only delete local symbol
DELETE/RECORD[/fast][/log]
Delete the current record (if the file is indexed or
relative and you have opened it with /MODIFY)
This will set the record context to the next record
If you specify /fast, a fast delete is done
DELETE/FUNCTION[/log] mask : Delete functions
DELETE/SEARCH [pattern]/LOG/CONFIRM Delete search patterns
DELETE/VALUE pattern[/log]/global/level=n Delete values
DELETE/TYPE pattern[/log]/global/level=n Delete types
DELETE/PARAMETER pattern[/log]/global/level=n Delete parameters
DELETE/SELECT_file pattern [/log/global/level=n] Delete select_file patterns
If none of the above qualifiers is present, DIX will assume you want to
delete files
DIX> Delete filemask [/confirm][/log][/reverse]
filemask A mask of files to be deleted
/confirm Ask before deleting a file
/log Show the file deleted
/reverse DIX will first make a list of files matching the filemask
and then start deleting these files in reverse order. This
should make a difference on large directories.
A small example on a directory with 1000 files and a size of 115 blocks
DIX> time delete [.abc]*.*.*
ELAPSED: 0 00:00:13.97 CPU: 0:00:03.67 BUFIO: 2002 DIRIO: 4898 FAULTS: 8
DIX> time delete [.abc]*.*.*/reverse
ELAPSED: 0 00:00:11.25 CPU: 0:00:03.82 BUFIO: 2002 DIRIO: 3892 FAULTS:30
DELETE/SYMBOL MASK Delete the symbols matching MASK
If the symbol is an indexed-symbol you can delete elements of the symbol
delete/symbol indsym("A*") Delete all values of indexed symbol indsym
matching A*. If you specify ("*") all values
will be deleted, but the indexed symbol
will not be deleted (it will have 0 values)
See also the /EXACT qualifier below
delete/symbol indsym Delete the whole symbol
Supported qualifiers
/EXACT : If the symbol is an indexed symbol, and you specify the
/exact qualifier, the contents will be used exact, so the
* and % character in the element selection will not be
treated as a wildcard character, so the above example will
delete indsym("A*")
/LOG : Display the symbols deleted
/LEVEL=n : Delete symbols from level n (0=global,1..depth=absolute level
-1..-depth) is relative level
/GLOBAL : Delete global symbols, is the same as /level=0
/LOCAL : Delete local symbols, is the same as /LEVEL=DEPTH
/ALL : Delete all symbols, is the same as MASK= *
This command is only allowed in a command procedure.
The DIX debugger will be started.
You can also use the command "on control_c debug"
If you then type a ^c during execution of a command procedure, the
debugger will be started.
You can use the command "on error/warning debug"
If a command returns an error/warning, the debugger will be started
You can also start a command procedure with /debug
Example:
Suppose the script file (examp.dix) contains
on control_c debug
k=1
10:k=k+1
goto 10
got_l1:say "Seen ^C"
goto 10
DIX> @examp
^C
%DIX-I-DEBUG, Entering debug mode
6:goto 10 !DIX tells you the current statement
DixDbg> eval k !see how we got
172469
DixDbg> go !keep going
^c
%DIX-I-DEBUG, Entering debug mode
6:goto 10 !Again this is where we are
DixDbg> eval k !ask for k again
613507
DixDbg> quit !skip the test, back to interactive mode
DIX> @examp/debug !start in debug mode
DixDbg>
DEPOSIT fieldnamemask=value.
DEPOSIT offset=value
DEPOSIT /symbol symbolname=value
Modify a field (or symbol if /symbol specified) to the given value.
for the DEPOSIT/SYMBOL see the deeper help in the /SYMBOL
DEPOSIT fieldname = value
DEPOSIT/RAW offset[.bit_offset] = value
DEPOSIT/VFC Deposit data in VFC buffer or record
The entered data is first 'unsymbolized' (see substitution)
and then translated according to the field definition, and deposited in the
data-record.
You can modify more than one field in one statement, but all the
used fields must have the same type.
If you modify more than one field, the /CONFIRM may be useful.
DEPOSIT will save the contents of the record before the modification
You can undo the changes for this DEPOSIT with the UNDO/LAST
You may also go back to the original record with the UNDO
The value specified is converted to the binary equivalent. This
will do what you expect, but there are some extensions
See the CONVERSION subtopic.
You may also assign values to fieldnames via the assign (val=expr) command
The differences are
deposit
allows wildcarding of the name (a text match is done)
the text after the = is interpreted as text, and will be converted
according the fieldname type.
assign
Does not allow wildcarding of the name, but it does support
wildcarding in the lowest dimension
The piece after the = is evaluated as an expression, and must
deliver a result of the same (or compatible) type as the field
Examples:
DEPOSIT /word 10=5 !modify the word (16 bits) at offset 10 to 5.
DEPOSIT *COUNTER*=0 !Modify all fields matching *COUNTER* to 0
! you may undo this with UNDO/LAST
DEPOSIT *PRIME*{*DAY*}=false !modify all subfields matching *DAY* of
!all fields matching *PRIME* to false
DEPOSIT UAF$Q_LASTLOGIN=1-JAN-2012:12:12
!set the lastlogin field (type date) to the date
!DIX will convert the string to a date format
Interpret the data in binary this is a shorthand for /radix=2
/BITS=number_of_bits For raw mode you can set any bitsize of 1 up to 64.
Modify data in bytes. Valid only in /RAW mode.
DIX will ask you if the modification should be done. This may be useful
if you specify a wildcard in the fieldname.
If conform is set. DIX will ask before every modification
MODIFY field to value
[y]/q/a/q :
You may answer
Y : Modify this one and keep asking for the next
N : Do not modify this one and keep asking for the next
Q : Do not modify this one and stop processing
A : Modify this one and stop asking (as if confirm was off)
Interpret the data in radix 10. This is only possible for [u]integer fields this is equal to /radix=10
/FORMAT=option
Normally all "character"-like fields will not contain unprintable data.
These field types are CHARACTER,xSTRING.
If the data does contain unprintables (hex 0:1f,7f,80:9f or ff) the
/FORMAT qualifier determines how this data should be displayed.
The default is the current format setting (set/show format)
Option can be
DOT :All unprintables are replaced by a ".". This encoding is not
reversible, after an unprintable value is replaced by a ".",
DIX cannot know what the unprintable value was. This may
be a problem in the screen mode.
HEX :All unprintable data is displayed as %Xdd, a hexadecimal display.
A % in the data will be displayed as %%. This display is reversible,
so DIX can reconstruct the original unprintable value.
If you are in this mode, on input a % must be entered as %%
PASSALL:DIX will not change the unprintable data. The data is displayed as
it is. This mode will not work in screen mode, and maybe poorly
when the output device is a terminal.
DUMP :Unprintable bytes are represented by a 2 or 3 letter mnemonic
like <DEL> or <CR>. A < in the text will be displayed as <<.
This display is reversible, so DIX can reconstruct the original
unprintable value.
If you are in this mode, on input a < must be entered as <<
Modify data in longwords. This is the default in /RAW mode. Valid only in /RAW mode.
/RECOVER (Default) /NORECOVER If you specify /recover, DIX allows you to roll back the changes made in this deposit by using UNDO/LAST If you specify /NORECOVER, DIX will not save the previous context, this may be useful in a script, when you are sure you will never UNDO the change.
Integers are converted as unsigned.
Modify data in words (16 bit values). Valid only in /RAW mode.
Modify data in quadwords. Valid only in /RAW mode.
Interpret the value in hexadecimal. This is a shorthand for /radix=16
Interpret the value in octal. This is a shorthand for /radix=8
/RADIX=nn Interpret the data in a given radix. nn must be between 2 and 36 For /radix=16,8,2 are shorthand defined as /HEX,/OCTAL and /BINARY A radix which is not a power of 2 (so not 2,4,8,16,32) can only be used on [u]integer fields For radix 36 the characters 0..9 and A..Z will be used (so Z=35)
Display data in raw (non description) mode. In this mode the qualifiers /WORD, /BYTE, /LONG, /HEX are significant. The default is the /LONG. Offset can be specified as byte_offset[.bit_offset]
Deposit/symbol[/raw] symbol=value
Allows you to use the deposit for a symbol.
Most of the qualifiers defined for the deposit command work here too,
except /offset, /confirm, /recover, /VFC
If the /raw is not specified, the value will be converted to
binary using the datatype of the symbol
If the /raw is specified, the value will evaluated as an expression
and the result will be stored in the symbol regardless of type.
All bits of the symbol will be overwritten unless the source size is
explicitly specified via /byte,/word,/long,/quad of /bits
See the following example
Somewhere in a file we see the binary value of a date
i.e. in the dump file at location %hex 80 is the dump date
So: $ DUMP/BLOCK=COUNT=1 SYS$SYSTEM:SYSDUMP.DMP
Virtual block number 1 (00000001), 512 (0200) bytes
.....
FFFFFEFC 00000000 FFFFFEFD BF6FC000 .@o?............ 000070
00A9B3FE D141FF1D 00B257CF 0D0A8DF4 t...OW2.....A3). 000080 !Offset 80
.....
DIX> date ddate !make a date type symbol ddate
DIX> deposit/symbol/raw ddate=%x00B257CF0D0A8DF4 !fill the value
DIX> say ddate !and show it
13-DEC-2017 19:14:24.31
Normally the value will be converted to binary, but for some
field-types more options are available. See below
(R)BITS fields
DEPOSIT field=name1,name2,name3 Set bits name1,name2,name3
DEPOSIT field=+name4 Leave the original value, and add
name4
DEPOSIT field=-name1 Leave the original value, and clear
name1
INTEGER*n with n=1,2
Only normal integer conversion is tried
INTEGER*4,*8
First try integer conversion.
Then try %Xnnnn, the %Onnnn, %Bnnnn, %Dnnnn conversions
(hex,octal,binary,decimal)
Then try 'dddd'X, 'dddd'O, 'dddd'B 'ddddd'D conversions
(hex,octal,binary,decimal)
Try to convert as date*4,*8
Try to convert as IDENTIFIER
Try to convert as UIC
Try to convert as 'xxxx ASCII
On VAX INTEGER*8 is only partial supported and is expressed
as INT*4.INT*4
VFC
The format is "DD text"
The DD are 2 hexadecimal chars. Not all combinations are valid.
The text field is ignored.
Only 0..7F : 0.127 linefeeds
80..9F : C0 control codes 00..1f
E0..EF : VFU codes 00..0F
See the OpenVMS manuals about VFC formats.
DEMO [whichdemo [startscreen]]
DEMO/FILE=demofile
DEMO/LIST Show the possible demos
Jump to demo mode. The parameters are
whichdemo : The name of the DEMO, as found in DEMO/LIST (or show demo)
startscreen : The start screen in the demo, default is MAIN
Supported qualifiers are
/SCREENNAME (Default) /NOSCREENNAME Show the name of the screen at the top of the display If you specify /noheader, the screenname will not be visible.
/HEADER (Default) /NOHEADER Display a header line, containing the demo name, and the screen name.
All key input will be uppercased.
DIRECTORY filemask [/qualifiers] [,filemask...] [filemask]
This looks like the normal OpenVMS directory, except that DIX will only
display the highest version, unless you type ;*
Filemask is a normal OpenVMS file specification with wildcards (like directory)
or a library specification: modulemask@librarymask or librarymask(modulemask)
There are 2 special librarynames
DIX_SYS : The system-wide library DIX_DES.TLB
DIX_USR : The user library DIX_USER.TLB
Directory can also do a fast search (using /fast). In this case DIX will
try to open the DEVICE:[000000]INDEXF.SYS (like DFU). You must have read
access to this file.
If you specify /fast or use a library specification, and you do not
specify a /[NO]SORT command, DIX will automatically add a sort over the
whole dev/dir/filename/type/ver[/module]
All type file files/methods can be mixed.
DIRECTORY support up to 8 parameters, each as a list.
DIX will allow you to place a wildcard in the device part. If you do that
DIX will do a scan of all disk (devices of class DC$_DISK) that are
mounted and not members of a shadowed disk (so the DSAdd is searched but
not the members of that shadow set).
Then a match will be done with the fulldevnam, and if that fails the
volume label.
DIX can also show the current open memfiles
DIX> DIR MEM:mask
The following qualifiers are supported
DIX> DIRECTORY *.CLD ! Display all CLD files
Directory: USER50:[STUBBF.PROGRAMS.DIX.SOURCE]
DIX_CLD.CLD;332
Total of 1 file
DIX> DIRECTORY SYS$LIBRARY:*V*.olb,starlet(*ACC*)
Directory: SYS$COMMON:[SYSLIB]
RDBVMSLIB72.OLB;2
Library: STARLET.MLB;1
$ACCDEF $ACCESSDEF $CHECK_ACCESS $CHECK_ACCESSDEF
$CHECK_ACCESS_G $CHECK_ACCESS_S $ICC_ACCEPT $ICC_ACCEPTDEF
$ICC_ACCEPT_G $ICC_ACCEPT_S $MTACCESS $MTACCESSDEF
$MTACCESS_G $MTACCESS_S $SNDACC $SNDACCDEF
$SNDACC_G $SNDACC_S
Total of 18 modules
VAXCCURSE.OLB;1 VAXCRTL.OLB;2 VAXCRTLD.OLB;2 VAXCRTLDX.OLB;2
VAXCRTLT.OLB;2 VAXCRTLTX.OLB;2 VAXCRTLX.OLB;2
VMS$VOLATILE_PRIVATE_INTERFACES.OLB;2
Total of 18 modules, 9 files
DIX> dir sys$sysdevice:*imagelib*.OLB/fast
Directory: SYS$SYSDEVICE:[VMS$COMMON.SYSLIB]
CXML$IMAGELIB_GS.OLB;1 CXML$IMAGELIB_TS.OLB;1
IMAGELIB.OLB;1
Total of 3 files
DIX> dir *@dix_sys !display all modules in the DIX_DES.TLB
Directory: FEKKO$DKA300:[STUBBF.PROGRAMS.DIX]
Library: DIX_DES.TLB;148
$AREA $BUCKET $DEFAULT $KEY $LINE $PROLOGUE
*DAF*.DAT .AUDIT$JOURNAL .BCK|.SAV .CRF_CROSS
.CRF_FILE_NAMES .CRF_MOD_NAMES .DIR ACCOUNTNG ACMSDDF
ACMSUDF ACU_PORTS ACU_USERS BSO_APPLICATIE
BSO_VERSIE DEMO#CREATE_DES DEMO#DIX_GENERAL DEMO#GENERAL_DES
DEMO#GENERAL_RAW DEMO#MODIFY DEMO#SEARCH DOCDB FILECAB
INDEXF KEYS-*.IND KEYS-*.VAL LMF$LURT MNU_USERS
MONITOR NET$PROXY NETPROXY NETWORK PARTITION PENDING
PROFILE QUOTA.SYS RESERVATIONS RIGHTSLIST
SWITCHFILE SYSUAF TCPIP$HOST USERTYPE_MYTYPE
VIEW#.CRF_CROSS VIEW#.DIR VMS$PASSWORD_HISTORY
VMS$SYSTEM_IMAGES.IDX VMSMAIL_PROFILE
Total of 52 modules
DIX> DIR dsa*:[dir]file.type !Search all DSA devices for file [dir]file.type
If some disks have a volume label USERnn,
DIX> DIR USER*:[dir]file.type !Search all devices with
!volume label USER* for file [dir]file.type
If you specify a /before or /since, the date will be checked against the backup date
/before=date Display only files younger than 'date' See the info about /create /expired /modified and /backup date can be a normal OpenVMS-date (including today, yesterday, tomorrow) or one of the keywords created/expired/modified/backup
/BY_OWNER=owner Select only files with this owner
/COLUMNS=n Normally DIX uses the width of the filename to determine the number of columns. If you specify /columns=n, DIX will use this value
/highlight=(colourspec,colourspec..)
You can use colours to highlight certain filetypes
colourspec is filetype#colourname[#modifier[#modifier...]]
or filetype:colourname[:modifier[:modifier...]]
the colours WHITE,RED,GREEN,YELLOW,BLUE,MAGENTA,CYAN,BLACK are supported
modifiers are BOLD, REVERSE, UNDERLINE and BLINKING
Only one colour can be specified, but multiple modifiers are allowed
The first colourname can also be a modifier so exe:underline is valid
You can specify upto 16 filetypes
If you specify /highlight without a value the default is
/highlight=(dir:yellow,exe:red,com:magenta)
If you specify a /before or /since, the date will be checked against the creation date
/DATE=(keyword[,keyword...]) Display the date(s) of the file Keywords are CREATED Creation date MODIFIED Modification date EXPIRED Expiration date BACKUP Backup date ALL All dates
/EXCLUDE=(filespec[,...]) Excludes the specified files. You can use wildcards in the directory, name and type. The device part is ignored.
If you specify a /before or /since, the date will be checked against the expiration date
/extent
/extent=full
Display info about the number of header/extents of the file
in the format "headers\extents" to differentiate with the size
"eof/alloc"
if /extent is given without the full parameter DIX will display the
number of extents in the format \extents
if /extent=full is specified, or if the /FHDRS qualifier is also given
DIX will show both the header count and the number of extents
in the format headers\extents
See also the /FHDRS qualifier
Display the number of file headers in the format headers\ See also the help about /extent
Directory can also do a fast search (using /fast). In this case DIX will try to open the DEVICE:[000000]INDEXF.SYS (like DFU). You must have read access to this file Directory /fast device:[directory]filenamemask By default DIX will sort the resulting files by directory/filename to get a nice display (files per directory). But this may take some time (first all files must be processed before the first output appears). If you specify /nosort DIX will print the files as soon as they are found but the display is set to /nohead/notrail, since files are no longer sorted by directory. Relative files versions also work, but only if you do not specify /nosort
Display the fileid of the file
Display full info of a file
Display a line containing the grand total (over all directories), but no files, and also no directory names
Display the headers (one for each new directory)
If you specify a /before or /since, the date will be checked against the modification date
If the filename is ODS-5 with ^* escapes, DIX will transform the ^* escapes to the real meaning. This transform is done before the possible sorting.
Display the file organization
/output=filename Redirect the output to a file
Display the owner of the file
Display the data one page at the time
Display the record format
Display the data in screen mode
/SELECT=(sel_item[,sel_item...])
Sel_item can be
select_file=name Set the default search pattern from a
select_file definition, see add select
for more help. You can add to or remove from
the select_file by adding additional qualifiers
See the example below
size=min=blocks Select on file size (EOF)
size=max=blocks
allsize=min=blocks Select on file size (Allocated)
allsize=max=blocks
size=unused[=size] Select file where alloc-eof>size
size is default the disk cluster size
version=min=version_number Select on version number
version=max=version_number
organization=org Org=IDX,REL,SEQ,SYMLINK
record_format=fmt fmt=UDF,FIX,VAR,VFC,STREAM,STREAM_LF,STREAM_CR
headers=min=count Select on the number of headers
headers=max=count
extents=min=count Select on the number of mapping extents
extents=max=count
global_buffers=min=count Select on global buffer settings
global_buffers=max=count
aces=count=min=count Select on number of ACEs
aces=count=max=count
aces=size=min=count Select on size of ACEs
aces=size=max=count
aces=ident=(id1[:access][,id2[:access]..)
Select on files that have ACEs with identifier
id1,id2.. Any id will match (or). If you also
specify aces=all, all IDs must be present.
Access can be specified as any combination of
RWDEC. All the letters specified must be in
the ACE and no more. If you append an *, you
select all ACEs that have at least the letters.
aces=all In combination with the above. Match only if
all IDs are present in the ACL
lbn=(start[,end]) Select files that are mapped on LBNs between
start and end. If end is not specified, end=start
journal=(flag,flag..) Select files with journal attributes
Flag can be
journal Select journal files
ai Select files that have AI journaling enabled
bi Select files that have BI journaling enabled
ru Select files that have RU recovery enabled
If you do not specify any flag, all above files will be found
flag=(flag[,flag]..)
Flag can be
nobackup Select files flagged as NOBACKUP
locked Select files flagged as locked
nomovefile Select files flagged as movefile disabled
wbcache Select files flagged as write back cache
readverify Select files flagged as readcheck
writeverify Select files flagged as writecheck
badacl Select files flagged as having a bad ACL
badblock Select files flagged as containing a bad block
nocharge Select files flagged as not charged to quota
erase Select files flagged as erase on delete
directory Select files flagged as directory file
markdelete Select files flagged as marked for delete
file=option (do not) Display parts of the filename
Option can be [no]node
[no]device
[no]directory
[no]name
[no]type
[no]version
Examples for the ACEs:
/SELECT=ACES=(IDENT=(ID1,BATCH),ALL)
Select only files that have an ACE with ID1 AND an ACE with BATCH
/SELECT=ACES=IDENT=(ID1,BATCH)
Select only files that have an ACE with ID1 OR an ACE with BATCH
/SELECT=ACES=COUNT=MIN=10
Select only files that have at least 10 ACEs
/SELECT=ACES=(IDENT=ID1:RWE)
Select all files that have an ACE (ID=ID1,ACCESS=R+W+E)
files that have other access fields enabled are not selected.
but other ACEs may be present
/SELECT=ACES=(IDENT=ID1:R*)
Select all files that have an ACE ID=ID1 with an ACCESS string
that has at least a READ specified, other access are allowed.
Example for select_file:
DIX> Add select_file patt/select=size=(min=10,max=20) !add a select pattern
DIX> direc/select=select=patt !show all files between 10 and 20 blocks
DIX> direc/select=select=patt/by_own=system !add an extra selection
DIX> direc/select=(select=patt,size=nomin) !now show files between 0 and 20
! blocks
/since=date Display only files older than 'date' See the info about /create /expired /modified and /backup date can be a normal OpenVMS-date (including today, yesterday, tomorrow) or one of the keywords created/expired/modified/backup
/SIZE=(keyword[,keyword...])
Support keywords are
all Display used size and allocated size
allocation Display only allocated size
used Display only used size, default if only /SIZE
bytes Display the size not in blocks but in xB (x=K,M,G)
exact Display the size in bytes, exact for sequential files
The normal DCL syntax
/size=units=bytes
/size=units=blocks
/size=units=exact Are also supported
/sort=(field[,field...]])
Sort the files according to the selected items
field can be
ascending The following sorts will be in ascending order
descending The following sorts will be in descending order
case_blind Text compares will case blind
directory Sort on directory
file Sort on the filename only
type Sort on the file type
version Sort on the file version
date[=created|expired|maodified|backup] Sort on one of the dates
created is default
id Sort on fileid
owner Sort on UIC
organization Sort on file organization (SEQ,REL,IDX)
record_format Sort on record format
(DIX,STREAM,STREAM_LF,STREAM_CR,VAR,VFC,UDF)
size[=allocated|used] Sort on size, default is used
headers Sort on the number of headers
extents Sort on the number of extents
For header/extents you need read access to the index file
If the /header or /trailer is active (Default), the sort is done
within each directory. If you specify /noheader/notrailer, the sort
is done over all files (including the directory)
If you have specified /fast, you can also specify /nosort to prevent DIX
from sorting the files (and thus print the files sooner).
DIX uses a special compare routine for the sort operation and this is
incompatible with the Hypersort utility.
If you receive the message %SORT-E-NYI, functionality is not yet implemented
when using directory/sort make sure that the logical SORTSHR does not
exist or points to SYS$SHARE:SORTSHR
Example:
directory/sort=(ascending,type,descending,date=create) *.*
Sort first on file type (ascending) and within equal types sort on
creation date in descending order)