* copyright (C) 1984-2019 merrill consultants dallas texas usa

with incorrect LRECL value, you got IEC141I 013-20 ABEND.

Thanks to Brian Sanger, Zurich Financial Services, ENGLAND.

Change 18.173 Support for OS/400 Release 4.5.0 (INCOMPAT LRECLs and

VMACQAPM data was inserted in the QAPMETH record). If you don't

Jul 27, 2000 use the QAPMETH record, you can change the LRECLs as

Sep 19, 2000 indicated for QAPMBUS, QAPMJOBS, and QAPMSYS, and the

previous MXG Version will work with 4.5 records.

-QAPMBUS has an undocumented single byte character field

added, changing it's LRECL from 111 to 112. New variable

BUCHAR created, was always 'S' in my sample records.

-QAPMETH had six fields with undocumented length changes.

The suffixes MEXR, MM14, MTR, and MDCN increased from 3

to 6 bytes, and MBRV and MBTR increased from 6 to 8.

The LRECL is now 257 for QAPMETH.

-QAPMJOBS has two undocumented 8-byte fields added at the

end of the record, variables JBUNDOC1/JBUNDOC2. The

LRECL for QAPMJOBS is now 835.

-QAPMSYS has five undocumented 4-byte fields added at the

end of the record, variables; the QAPMSYS LRECL is 3288.

Sep 19, 2000: The undocumented variables are now named,

but I still have no description of their contents:

JOBS: JBEDBC, JBTDBC SYS: SYIFUS, SYIFTE, SYDBC, SYSWC

SYSTEM: SCIFUS SCIFTE SYHFTH SYSDBC SYSSWC

BUS: BUTYPE

Thanks to Stuart R. Burnett, Reynolds Metal Corporation, USA.
Change 18.172 Support for APAR PQ32435 OS/390 IHS WEBSERVER SMF 103

VMAC103 adds variables JOB and ASID to the TYPE1032 dataset.

Jul 27, 2000 The APAR also describes changes in values put into the

"EntityName" and "EntityAddress" fields, and documents

when the interval value is not the expected interval:

- The initial write on startup has an interval of zero.

- The termination write on shutdown or restart has an

interval of the time since last write.

- The initial write on restart has an interval of the

time since the restart termination write.

Change 18.171 MXG's protection logic for TCP SMF 118 records with bad

VMACTCP HFS Filename offsets or lengths was still incomplete and

Jul 27, 2000 permitted INPUT STATEMENT EXCEEDED errors. Logic was

revised and the first two bad records are printed on the

SAS log. This supercedes Change 18.144.

Thanks to Francis Berckman, Astra Zeneca, USA.

Change 18.170 MXG 18.04-18.05 only. Variables ONLINE,CMBINVLD,PARTIAL,

VMAC74 BASE, and VARY were always blank. The five lines setting

TYPE74 those variables blank were not removed after the logic to

Jul 26, 2000 set those variables was moved up thirty lines.

Jul 28, 2000

Thanks to Diane Parker, Bergen Brunswig, USA.

Thanks to Peter Webber, CIS, ENGLAND.
Change 18.169 Variable PERFINDX was incorrect in TRND72GO because the

TRND72GO variables R723CVAL and R723CPCT should have been in the

Jul 26, 2000 ID= list, and not in the SUM= list.

Thanks to Pat Perreca, Bear Stearns & Co., USA.

Change 18.168 Documentation. The SAS option USER=XYZ can be used to

DOCMXG send "//WORK" datasets instead to the "//XYZ" DDname.

Jul 25, 2000 However, for MXG to recognize the USER=option, it needs

Oct 11, 2002 to be set before MXG initialization has executed:

OS/390: put USER=XYZ on the EXEC statement:

//STEPNAME EXEC MXGSAS,OPTIONS='USER=XYZ'

ASCII: put USER=XYZ in the AUTOEXEC.SAS file:

LIBNAME XYZ 'C:\XYZ';

OPTIONS USER=XYZ;

If you must put the USER= option in your //SYSIN stream,

then you must insert a %VMXGINIT; statement after it:

//SYSIN DD *

OPTIONS USER=XYZ;

%VMXGINIT;

which will re-invoke the MXG initialization using your

USER= value (printing a second "WELCOME TO MXG" message).

If the USER= value is put in //SYSIN without the VMXGINIT

MXG will not see your USER= option. If you have USER= on

the // EXEC, and you also have USER= and VMXGINIT in your

//SYSIN, the USER= in SYSIN will be used because of the

second MXG initialization.

Putting USER= on the //EXEC statement is preferred, as it

does not require the extra cost of the second VMXGINIT.

An alternative for re-directing datasets in the //SYSIN

is to use the %LET Wdddddd=XYZ; syntax for the "work"

copy of the unsorted dataset (or use %LET Pdddddd=XYZ;

for the sorted, "PDB" copy). The "dddddd" value for each

MXG dataset is documented in the IMACxxxx member for the

product that creates the dataset.

Oct 11: text was revised for clarity.

Thanks to Mike Delaney, Pershing, USA.
Change 18.167 New variables QBGBGR1N and QBGBGR2N are input as numeric

VMACDB2 variables. The original variables QBGBGR1 and QBGBGR2

Jul 25, 2000 were input as characters (because I didn't realize what

they were), and worse, while input as $EBCDIC6, they were

truncated to a kept length of only $4 because they were

in the $HEX8 format item list! Now the character vars

are kept as full length (by moving to $HEX12 format), and

numeric counterparts now exist in DB2GBPAT dataset for

Global Buffer Pools counts of current or pending

directory entries.

Thanks to Don Deese, (CPExpert), Computer Management Sciences, USA.
Change 18.166 The new R744Cxxx variables in the TYPE74ST Structure data

VMAC74 set were wrong if there were more than one segment, due

Jul 25, 2000 to a typo: the 26 lines with SUM(R744Caaa,R744Caaa) were

changed to SUM(R744Caaa,R744Xaaa). The

RLS and OSAM structures have 20+ and 60+ segments each,

but DB2 group buffer pools have only one segment so their

counts were always correct.

Thanks to Don Deese, (CPExpert), Computer Management Sciences, USA.

Change 18.165 New format MGBYTEN converts bytes to KB/MB/GB/TB/PB like

FORMATS existing format MGBYTES, but MGBYTEN handles negative

Jul 25, 2000 and positive byte values (to track increase/decrease in

DASD memory) so it must be 6 characters wide to have room

for the minus sign, so it had to be a different name than

the 5-character-wide MGBYTES format.

Thanks to Tan York Sin, Singapore Exchange Limited, SINGAPORE.
Change 18.164 Support for Release 321 INCOMPATIBLY changed time stamp

VMACBETA formats for BETABT and BETASTRT. More testing may be

Jul 25, 2000 required if other fields were changed, as only the one

subtype=1 record was received for validation.

Thanks to ???, ???, EUROPE.
Change 18.163 Statement IORATE=0; was removed. The devices with no

ANALCACH cache records needed to have cache-related fields zero,

Jul 25, 2000 but not the IORATE.
Change 18.162 Support for DFSMS/rmm TYPEEDGS/TYPEEDGB was still wrong.

VMACEDGS Zero obs or INPUT STATEMENT EXCEEDED RECORD LENGTH error.

Jul 25, 2000 The "D" and "V" dataset/volume records did change in 1.5

Sep 5, 2000 but the old-format 1.4 records still exist in the catalog

dataset. And the handling of variable length name fields

that worked in 16.16 was somehow lost. In any event, the

change has been tested with TYPEEDGB reading a catalog

with both formats of both records, and works correctly,

using the MD/MVRECLEV variable to identify record format.

There are some 1.4 records that have '40'X in their name

length fields, and have enough record length to contain

a DSN, but they do not contain real DS names; however,

they also do not cause any error.

Sep 5: Variable MVRETDAT was not kept.

Thanks to Richard Fortenberry, Mitsubishi Motor Sales, USA.
Taught classes in Amsterdam and London.
Change 18.161 Y2K support was incorrect, but nobody noticed! Inserted

VMACAXC IF IDTE LT 1900000 THEN IDTE=IDTE+1900000; and changed

Jul 7, 2000 each IDTE,5.),JULIAN5. to IDTE,7.),JULIAN7.

Thanks to Stephen Bell, Informatik Kooperation, GERMANY.

Change 18.160 The %INCLUDE of member IMACKEEP was missing from a number

DOC of MXG TYPExxxx and TYPSxxxx members, which prevented use

Jul 7, 2000 of the new-in-16.04 architecture for instream tailoring.

Most are archaic or not mainstream. Members updated were:

TYPxCRAY TYPxFRYE TYPxHO15 TYPxHPAI TYPxHPCS TYPxHPSU

TYPxHPUX TYPxIMFL TYPxIMS TYPxIMS1 TYPxMWAI TYPxMWSU

TYPxMWTE TYPxMWUX TYPxPW TYPxQTRT TYPxRMF TYPxSAM

TYPxSFS TYPxSNIF TYPxSUPR TYPxTRSN TYPxTUX TYPxVMON

TYPxVMXA TYPxXAM TYPxZARA TYPxZRB TYPEIMSD

Thanks to Bart Decat, KBC Belgium, BELGIUM.

Change 18.159 WARNING: ARGUMENT 3 TO MACRO FUNCTION %SUBSTR IS MISSING

ANALDB2R occurs with SAS Version 8.1, because the length of their

BUIL3001 &SYSVER Version macro was shortened from four to three.

BUIL3005 There is no execution impact, fortunately, except that

BUILD001 the warning now sets a step Condition Code/Return Code of

BUILD005 four instead of zero, and wastes your time in diagnosis!

BUILDPD3

BUILDPDB In Version 8.1, SAS changed the length of their &SYSVER

READDB2 macro from four digits (6.06,6.07,6.08,6.09,8.00) to only

VMXGFOR three digits (8.1), but MXG used %SUBSTR(&SYSVER,1,4):

VMXGINIT to differentiate 6.07 from 6.08, so that options that

VMXGSUM were new in 6.08 would be bypassed under 6.07. While

Jul 6, 2000 the %SUBSTR fails, because the option now exists in

SAS releases, there is no execution error. And the

only purpose was to suppress unneeded SAS messages!

To eliminate this exposure in future SAS Version names,

all MXG references to &SYSVER were revised. Now, all

tests for SAS Version use the MXG-created macro variable

&SASVER, which is set by VMXGINIT from &SYSVER to the SAS

Version number (6, 7, 8, etc.). All of the old 4th-digit

of &SYSVER tests are no longer required:

The tests that set OPTIONS CODEPASS=2 were removed

completely, as that option no longer exists and is no

longer set, and tests around OPTIONS DKROCOND= were

also removed, as that option is now present in all

executable SAS versions.

Thanks to Pat Curren, SuperValu Inc, USA.
Change 18.158 The NTSMF Trending macro variables PTRNTIN and PTRNTLD

VMXGINIT were not defined in VMXGINIT. Three labels in VMACNTSM

VMACNTSM longer than 40 characters, that were not caught by the

Jul 6, 2000 SAS compiler when VMACNTSM ran, but were caught when the

data set was copied.

Thanks to Howard Glasetter, Weyerhaeuser, USA.

Change 18.157 The example REPORT command in the ADOCMWUX member did not

ADOCMWUX create the file that was expected by VMACMWUX, so that

VMACMWUX command was revised to include the new variables that are

Jul 6, 2000 expected. Four variables, TFRSTSEC,TPRMPSEC,TTHNKUSE,

and TPRMPUSE are set missing because I cannot find them

in the current list of MeasureWare fields for HPUX.

Thanks to Roman Gudz, Penske, USA.

Thanks to Bart Decat, KBC Belgium, BELGIUM.

Change 18.156 Variable SMF74CNF was not kept, but not initialized, as

VMAC74 it was INPUT into DEVIND, which was not kept. Now it is.

Jun 30, 2000

Thanks to Freddie Arie, Lone Star Gas, USA.

Change 18.155 MXG 18.05 only. Variables S42DSMXR,S42DSMXS were never

VMAC42 input, having been misspelled as DXMXR and DSMXS in the

Jun 30, 2000 INPUT statement.

Thanks to Bruce Widlund, Merrill Consultants, USA.

======Changes thru 18.154 were in MXG 18.05 dated Jul 1, 2000======
Change 18.154 ANAL4GB member reports on VSAM files approaching 4GB in

ANAL4GB size, but included Extended VSAM files, which show over

Jun 30, 2000 100% utilized space, so IF SMF64EF='Y' THEN DELETE; was

added so those datasets will not be reported.

Thanks to Alfred Holz, Merck & Co., USA.
Change 18.153 Corrections found during QA runs.

VMAC108 -VMAC108. The KEEP list for TYPE108T and TYPE108P did not

VMAC74 keep the variables DOMSPN and DOMSYN, and DOMDBNAM should

WEEKBLD not have been in _BTY108T.

WEEKBLDT -VMAC74. Two variables R745DCID; the "Real CU ID" variable

MONTHBLD was changed to R745DCIR.

VMACEDGS -WEEKBLD/WEEKBLDT/MONTHBLD. The Sort Order for TYPE72SC

VMACMIM was revised to match the VMAC7072 definition, and is now

VMACTMV2 SYSPLEX SYSTEM STARTIME SRVCLASS

VMACSOLV R723SCSN R723SCSR SMFTIME

VMXGCICI -VMACEDGS. Variable MDPDSN repeated in KEEP and LABEL.

Jun 30, 2000 -VMACMIM. Variable DURATM repeated in KEEP list.

-VMACTMV2. Variable WGMSOSCE repeated in KEEP list.

-VMACSOLV. Variable SOLVFLAG had two formats.

-VMXGCICI. Variable STARTIME repeated in FORMAT.

Thanks to Freddie Arie, Lone Star Gas, USA

Thanks to Bruce Widlund, Merrill Consultants, USA
Change 18.152 New _S110ST and _CICSTAT macros are defined in VMAC110 so

VMAC110 that you can control where the CICS Statistics datasets

Jun 29, 2000 are written. With TYPE110 or BUILDPDB, CICS statistics

Aug 29, 2000 datasets are created in the "WORK" DD; BUILDPDB reads 'em

to create the PDB.CICINTRV summary dataset, but those

individual CICS Statistics datasets are not copied or

sorted; they just die at end of step in the //WORK file.

This enhancement lets you control where they written, if

you want to keep any or all of them.

1. You can write them to the 'CICSSTAT' library, sending

all of the CICS statistics datasets (unsorted) to that

DD as the SMF records are read. You would use this:

after your //SYSIN DD statement:

%LET CICSTAT=CICSSTAT;

_CICSTAT;

The _CICSTAT macro invocation uses the value of the

&CICSTAT macro variable as the destination library for

the &WCICddd and &PCICddd CICS statistics data sets.

If you %LET CICSTAT=PDB; before the _CICSTAT invoke,

CICS stats would be written unsorted to the PDB DD.

2. Alternatively, by default, BUILDPDB builds all CICS

Statistics datasets in the //WORK file, and then reads

them to create the PDB.CICINTRV dataset, but then the

CICS Statistics datasets are left in the //WORK DD,

which is temporary and goes away at step end. You can

thus use the member EXPDBOUT to select/sort any of the

CICS Statistics datasets into your PDB. You can use

individual dataset "_Sdddddd" macros for each dataset

that you want to output, or you can use this new

"_S110ST" macro, which will sort only the Statistics

CICS datasets. (The old _S110 macro can't be

used in EXPDBOUT, because it invokes two _Sdddddd sort

macros, _SCICEXC and _SCICSYS, which were already

invoked by BUILDPDB logic.) Just adding _S110ST in

your EXPDBOUT member will sort all CICS statistics

data to the PDB library. You can also invoke the

_CICSTAT macro in member EXPDBOUT after resetting the

default DD name, can could then sort all of the CICS

statistics datasets into the "MYDD" library, using:

%LET CICSTAT=MYDD;

_CICSTAT;

_S110ST

Note: _S110ST did not exist until MXG 18.07, when the

change was revised.

Thanks to Glenn Bowman, Wakefern, USA.
Change 18.151 INPUT STATEMENT EXCEEDED with NETSPY 5.2 subtype "I" and

VMACNSPY MXG 18.03-18.04. In 5.2, the "I" subtype was for TELNET

Jun 29, 2000 sessions, but NETSPY 5.3 deleted the TELNET data, and

reused subtype "I" record for TCP/IP data. MXG 18.03 now

supports the TCP/IP data, but Change 18.069 did not test

for and delete that now-defunct 5.2 "I" record. The code

test after IF NSPYREC='I' THEN DO; was revised to insert

IF NSPYENTL LE 70 THEN DELETE; since the 5.2 "I" record

segment length was 70.

Thanks to Rita Bertolo, Canadian Pacific Railroad, CANADA.

Change 18.150 Negative values for RANDOM PAGES fields in reports. The

ANALDB2R calculation of variable RANDOMP was incorrectly located,

Jun 29, 2000 and was moved to after MEAN=QB&B.TCBA/NUMINTVL;

Thanks to Gary L. Keers, Indianapolis Power & Light, USA.

Change 18.149 New analysis: Who is filling your active SMF VSAM file?

ANALSMJB This program reads the active VSAM SMF file (or any SMF

Jun 29, 2000 file) and counts how many type 14, 15, 30, 42, and 64 SMF

records in that SMF file were created by each JOB, so

that your operators can use this report to cancel a

runaway job. (The program stops reading SMF records when

it sees a timestamp after the start of this program, so

that SAS does not have to read over all the unused CIs

in your VSAM SMF file).
Change 18.148 Documentation only. Change 17.060 showed example syntax

VMXGINIT of %VMXGINIT(MXGWORK=XYZ); to redirect datasets from the

Jun 29, 2000 //WORK default to the "XYZ" libref, but VMXGINIT changes

(VOPTIONS table, resetting MXGWORK based on SASSWORK and

USERWORK) cause that old syntax to no longer work. If you

should ever need to redirect work, now the correct syntax

is to use OPTIONS USER=XYZ; See also Change 18.168

Thanks to MP Welch, SPRINT, USA.

Change 18.147A-SAS V8 no longer requires the MEMSIZE parameter in the

CONFIGV8 CONFIG member, and in some cases it has caused memory

Jun 29, 2000 ABENDs that were eliminated by removing MEMSIZE, so I

have removed it from CONFIGV8 member. Using REGION=0M

on your JOB card is recommended, but you should be aware

that installation exits IEALIMIT and/or IEFUSI can limit

the virtual storage that be allocated.

-I had to reinstate the S=72 and S2=72 options in CONFIGV8

member. It was removed (Change 17.392) to protect from

potential problems if your SOURCLIB was RECFM=VB, but its

absence caused errors if you had numbered lines in your

MXG Tailoring Library (USERID.SOURCLIB), including 180

syntax ABENDs, and APPARENT MACRO &WORDI UNRESOLVED.

Unnumbering your lines eliminated the error, but so does

putting back the S=72 and S2=72 in the CONFIGV8 member,

and you don't have to touch your numbered lines.

Thanks to John Mattson, Espon, USA.
Change 18.147 The data step previously in the _STYMEMx sort macros were

VMACMEMO replaced with PROC SORT and the _BTYMEMx macros now have

Jun 29, 2000 the correct BY list variables.

Thanks to Fred Kuhstoss, Norfolk Southern Corp, USA.

Change 18.146 Variables S42DSMXR and S42DSMXS were added to TYPE42DS

VMAC42 when they were discovered in the SMF manual.

Jun 29, 2000
Change 18.145 -New TRNDNTLD trends NT DISK Space Utilization using the

BLDNTPDB NTSMF dataset LOGLDISK.

NTINTRV -Hardcode dataset names in BLDNTPDB text were replaced

TRNDNTLD with their _Ldddddd symbolic macro name, and NTINTRV was

VMXGINV changed to use the symbolic, and to %INCLUDE the member

Jun 28, 2000 IMACKEEP so the BLDNTPDB builds a fully compatible code

member that accepts IMACKEEP/MACKEEP= tailoring.

If TRENDing is invoked by BLDNTPDB, both TRNDNTSM and the

new TRNDNTLD are invoked.

Thanks to Greg Jackson, National Life of Vermont, USA.

Thanks to Howard Glasetter, Weyerhaeuser Company, USA.
Change 18.144 Protection for bad length/offset for TCP HFS filename

VMACTCP segments was over slightly in error; some records with

Jun 26, 2000 valid HFS lengths of zero were flagged as in error. The

test for IF 0 LT FTPHFS01 LT LENGTH-1 THEN DO; line 440

and IF 0 LT FTPHFS02 LT LENGTH-1 THEN DO; line 472

were changed IF 0 LT FTPHFS01 LE LENGTH-1 THEN DO; 440

and IF 0 LT FTPHFS02 LE LENGTH-1 THEN DO; 472

so only bad segments with non-zero length are detected.

Thanks to Sal Fazzino, First American, USA.
Change 18.143 -Support for new NTSMF Objects in Windows 2000 Server:

EXNTBENG "dddddd" "dataset" vars description

EXNTDHCP NTBENG BEENGINE 131 BE Engine

EXNTDNS NTDHCP DHCP 14 DHCP Server

EXNTFIRC NTDNS DNS 62 DNS

EXNTFIRS NTFIRC FILREPCO 24 FileReplicaConn

EXNTNTPC NTFIRS FILREPSE 91 FileReplicaSet

EXNTNTPS NTNTPC NNTPCMND 44 NNTP Commands

EXNTNTDS NTNTPS NNTPSERV 40 NNTP Server

EXNTNTSH NTNTDS NTDS 140 NTDS

EXNTWMSS NTNTSH NETSHIEL 40 NetShield

EXNTWMUS NTWMSS WINMEDSS 3 Windows Media Station Service

IMACNTSM NTWMUS WINMEDUS 24 Windows Media Unicast Service

VMACNTSM -And the NTSMF 2.2.2 change that puts back into the SYSTEM

VMXGINIT object the variables PCTCPUTM PCTUSRTM PCTPRVTM that

Jun 27, 2000 Microsoft had previously removed from SYSTEM.

-And the NTCONFIG file was corrected; the CPUSPEDn/FAMILYn

MANUFACn/CPUNUMn fields were RETAINed from each 0,0, but

were not initialized, and so could contain data in the

N+1 fields when the value of NRCPUS was only N.

Thanks to Hansueli Vogt, Credit-Suisse, SWITZERLAND.
Change 18.142 CMA SPOOL record, subtype 6, variable SMFT06PC, added by

VMACCMA Change 18.056, was incorrectly decoded, as a two-byte

Jun 26, 2000 reserved field precedes that field. The DSECT shows a

two byte reserved field after SMFT06PC, but it is not

input (to preserve compatibility).

Thanks to Dr. Alexander Raeder, Karstat AG, GERMANY.

Thanks to Harmuth Beckmann, Karstat AG, GERMANY
Change 18.141 Variable DVGSBCNT, the Transfer Byte Count, has always

VMACFTP been wrong; it should have been input as &PD.8., but it

Jun 26, 2000 was documented as binary instead of packed decimal.

Jul 25, 2000 In addition, variable DVGSCFAC should have been &PIB.1.1

but was previously only &PIB.1.

Thanks to John Sheridan, Aer Lingus, IRELAND.

Thanks to Theo Peelen, RABOBANK, THE NETHERLANDS.
Change 18.140 Variables CIOPCT, HITPCT, and RDHITPCT were not being

ASUM42DS recalculated after the summation; the were revised into

Jun 26, 2000 NORM= operands to be properly recalculated as percents.

Thanks to Raff Rushton, Kaiser Permanente, USA.

Change 18.139 IMS Log Version 5.1 cause zero observations in IMSTRAN

TYPEIMSB in STEP4 of JCLIMSL5, because IF _IMSVERS LE 5 THEN DO;

Jun 22, 2000 should have been IF _IMSVERS LE 5.1 THEN DO; in three

places in member TYPEIMSB.

Thanks to Pete Gain, SAS UK, ENGLAND.
Change 18.138 Additional, optional variables are now decoded from the

IMAC6ESS ESS segment in the type 6 record. Comments in IMAC6ESS

VMAC6 describe how to enable the creation of these variables,

Jun 21, 2000 and how to add them to TYPE6, or to PDB.PRINT. The full

list of TYPE6/PDB.PRINT variables that are decoded from

the ESS/IEFDOKEY optional fields are these:

ADDRESS1='FIRST LINE*OF ADDRESS'

ADDRESS2='SECOND LINE*OF ADDRESS'

ADDRESS3='THIRD LINE*OF ADDRESS'

ADDRESS4='FOURTH LINE*OF ADDRESS'

BUILDING='BUILDING'

DEPT ='DEPARTMENT'

DESTNATN='DESTINATION'

ESSBURST='BURST*IN*ESS'

ESSCHARS='CHARS*IN*ESS'

ESS

ESSCOPIE='COPIES*IN*ESS'

ESSFCB ='FCB*IN*ESS'

ESSFMDEF='FORMDEF*IN*ESS'

ESSFORMS='FORMS*IN*ESS'

ESSLINCT='LINECT*IN*ESS'

ESSMODFT='MODIFY TRC*IN*ESS'

ESSMODFY='MODIFY*IN*ESS'

ESSPGDEF='PAGEDEF*IN*ESS'

ESSPRMOD='PRMODE*IN*ESS'