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

Change 20.294 VTS activity reports enhanced and updated. A new

ANAL94 selection parameter, SUBREP= to select desired report:

Dec 17, 2002 ALL, PHY, VIR, TAP.

Dec 30, 2002 Dec 30: cosmetic: "compiler fakers" replaced with RETAIN.

Thanks to Wally Danielson, Airborne, USA.
Change 20.293 CICS Stats "UNKNOWN STID=nnn" messages are printed by MXG

VMAC110 to alert us both that IBM has added a new statistics STID

Dec 17, 2002 segment to the SMF 110 subtype 2 record, one that is not

yet supported by MXG, so that I can begin the process of

finding the documentation from IBM and updating MXG for

that new STID. But when the "UNKNOWN STID=115" message

was printed, it took several iterations of customer's

system programmers time to report the problem and send a

record dump, and hours here trying to find documentation

that doesn't exist, and then reinvolvement of the site's

CICS staff to open a PMR with IBM, and then IBM CICS

support technicians time to finally document that:

The STID=115 for Enterprise Java Beans (DFHEJBDS) was

produced by early levels of CICS/TS 2.1 code but then

withdrawn, with the external mappings and documentation

removed, so these records should not be produced by the

currently supported IBM CICS code. (If you are still

getting these records on CICS TS 2.2, you need to raise

a PMR with your CICS support centre.)

This record type may be reinstated in a future release.

This change in VMAC110 bypasses the UNKNOWN STID= message

for STID=115, and documents the STID number!

02Jan03: APAR PQ69408 was opened 24DEC02.
Thanks to Chris Selley, Zurich Financial Services, ENGLAND.

Thanks to Tony Hurlston, Zurich Financial Services, ENGLAND.

Change 20.292 SAS Technical Support provided updated mapping of SAS SMF

ADOCSASU User record's Procedure Names to the SAS Product Name, so

FORMATS the $MGSASPR format was updated with new procedure names,

Dec 16, 2002 and the table in member ADOCSASU was revised.

Thanks to Tom White, SPRINT, USA.
Change 20.291 Support for APAR OW57121 which documents many new SMF 99

FORMATS Trace Code values; format MG099TC was updated.

Dec 16, 2002
Change 20.290 The Management Class Name variable DMCNAME is padded with

VMACDCOL hex zeros, which are now converted to blanks.

Dec 11, 2002

Thanks to Dave Gibney, Washington State University, USA.

Change 20.289 Support for the Workload Manager's WLM Definitions PDS.

REXXWLM The REXX exec contributed by this user will read the WLM

Dec 11, 2002 PDS to extract the names of tables, keys, and variable

names, and the exec then generates the SAS code to read

the PDS to create a SAS dataset for each WLM table.

Thanks to Sam Bass, McLane Company, USA.

Change 20.288 Support for MQSeries V5.3 (COMPATIBLE) new data:

VMAC115 -SMF 115 Subtype 2 MQMMSGDM Data Set new vars:

VMAC116 QISTDLMM='DELETE*MARKED*MESSAGE*REQUESTS'

Dec 9, 2002 QISTENUM='ENUMERATE*SELECT*REQUESTS'

QISTGETB='GETS THAT*GOT MESSAGE*FROM BP'

QISTGETD='GETS THAT*GOT MESSAGE*OFF DISK'

QISTLOMM='LOCK*MARKED*MESSAGE*REQUESTS'

QISTMBLR='RELEASE*BROWSE*LOCK*REQUESTS'

QISTMCNT='MESSAGE*COUNT*REQUESTS'

QISTRABP='READ*AHEADS*FROM*BUFFER*POOL'

QISTRAIO='READ*AHEADS*DOING*I/O'

QLSTGETL='GET*LOCK*REQUESTS'

QLSTHLDL='TIMES*REQUESTED*LOCK*HELD'

QLSTRELL='RELEASE*LOCK*REQUESTS'

And all of the QSST fields that exist in DB2ACCT:

QSSTABND QSSTCONF QSSTCONT QSSTCONV QSSTCRIT QSSTEXPF

QSSTEXPV QSSTFPLF QSSTFPLV QSSTFREF QSSTFREM QSSTFREV

QSSTGETM QSSTGPLF QSSTGPLV QSSTRCNZ

-SMF 116 MQMACCTQ Data Set new vars:

QWHSRN =' RELEASE*INDICATOR'

QWHSACE ='ACE*ADDRESS'

QWHSSTCK='STORE*CLOCK*VALUE OF*HEADER'

QWHSISEQ='SEQUENCE*NUMBER*FOR*IFCID'

QWHSWSEQ='SEQUENCE*NUMBER*FOR*DESTINATION'

QWHSMTN ='ACTIVE*TRACE*NUMBER*MASK'

WQGETPMS='PERSISTENT*GOT BY*MQGET'

WQPUTPMS='PERSISTENT*GOT BY*MQPUT'

WQPUT1PM='PERSISTENT*GOT BY*MQPUT1'

Change 20.287 PROC DELETE DATA=_Wdddddd; syntax for four z/VM datasets

VMACVMXA were overlooked and are now %VMXGDEL(DDDDDD=dddddd); as

Dec 9, 2002 they should have been. The %VMXGDEL is used only for

datasets that are directly sorted from _W to _Ldddddd,

and prevents deletion when _W is equal to _L dataset.

Thanks to Chris Weston, SAS Institute ITRM, USA

Change 20.286 -Three SMS objects were revised to contain Instance Names;

EXNTSQBP variables SMEXNAME in SMSEXTHR, SMIMNAME in SMSINMEM and

VMACNTSM SMSENAME in dataset SMSSENDR are now supported.

Dec 8, 2002 -Dataset SQLBUFMG has new fields and fields deleted.

-New SQLSERVER:Buffer Partition Object supported.

-INPUT STATEMENT EXCEEDED INPUT with MSExchangeIS object

if it had NRDATA=49; MXG logic corrected.

-Antigen Scan object with only NRDATA=12 supported.

Thanks to Jim Quigley, ConEd, USA.
Change 20.285 Using the _NTIMS macro to tailor Landmark IMS processing

VMACTIMS failed; each line of that MACRO definition was missing

Dec 8, 2002 the work MACRO at the start of each line.

Thanks to Michael Gresham, JCPenny, USA.

Change 20.284 Cosmetic cleanup of first MXG 20.08 only:

VMAC119 VMAC119: PUT at line 1794 was deleted.

VMAC110 VMAC110: PUT at line xxxx was deleted.

VMACNTSM VMACNTSM: Labels for DAPASTRT, DBPAFLRT missing quote.

Dec 8, 2002
Change 20.283 SMF 94 Subtype 2 SMF94SOF/SLN/SON triplet is nonstandard,

VMAC94 so it cannot be used to determine how many pool segments

Dec 5, 2002 need to be output to TYPE942P. SMF94SON should be the

Jan 2, 2002 number of repeated segments, but is always 1. SMF94SLN

should be the length of a segment, but it is 1792, the

total length of all 16 possible pools, even when fewer

pools are used. MXG initially used SON, outputting only

the first pool, but the test site knew they had defined

two pools. Their SMF records had an (undocumented) field

S94PPP='...0....'B in the first 2 segments, and had

S94PPP='...1....'B in the other 14 segments, so that test

was added to VMAC94, to only output if that bit was zero,

while awaiting this answer from IBM:

"We consider all pools to be defined, they just may not

be in current use. Other than assuming that a pool

with no logical volumes in it is 'not defined', there

is no way to make that determination. The bits of

SMF94PPP are now described, but they will not help in

determining which pools are defined."
So the heuristic test was removed, and MXG now always

outputs all 16 pool segments to TYPE942P dataset.

These new variables are created from that S94PPP field:
S94PPPRA='RETURNS*ALLOWED?'

Bit 0:

"Y" if borrowed volumes are to be returned to

the common scratch pool when scratched; blank

if borrowed volumes remain in the pool that

borrowed them when scratched.

S94PPPFM='FIRST*MEDIA*TYPE*TO BORROW'

Bits 2-4 converted to a decimal value.

First media type to borrow, if additional

physical scratch volumes are needed by the

pool.

S94PPPSM='SECOND*MEDIA*TYPE*TO BORROW'

Bits 5-7 converted to a decimal value.

Second media type to borrow, if additional

physical scratch volumes are needed by the

pool, and none of the first media type are

available.
Both are decoded by MG094PP format:

0 - No Borrowing

1 - Borrowing of Media ID 0 is allowed

2 - Borrowing of Media ID 1 is allowed

3 - Borrowing of Both Media is allowed
Change 20.282 The debugging PUT _N_= STARTCOL= .... statement at line

VMACCTLG 580 was removed.

Dec 5, 2002

Thanks to Len Rugen, University of Missouri, USA.

Change 20.281 SMF 90s from OS/390 V2.9 and z/OS 1.3 have VERSN90 as a

VMAC90A binary field in subtype 30 records, although the z/OS 1.4

VMAC90 SMF manual still shows an EBCDIC number. INVALID DATA

Dec 5, 2002 messages were printed, but no impact, as VERSN90 is not

Dec 11, 2002 used. Revised code inputs VERSN90 ?? &NUM.2. and then

tests for missing to re-INPUT as &PIB.2. Subtype 6 have

VERSN90='0002'x, subtype 6s have 'F0F1'x, or EBCDIC 1.

Note: USE TYPE90A, the old TYPE90 was replaced by TYPE90A

and this fix was extended to TYPE90 just for protection.

I have opened a documentation issue with IBM.

Thanks to Peter Webber, Co-operative Insurance Society Limited, UK.
=======Changes thru 20.280 were in MXG 20.08 dated Dec 4, 2002=========
Change 20.280 MXG-constructed variable PLAN in DB2ACCT was incorrect in

VMACDB2 Distributed/Remote records (QWHCATYP=7,8); the statement

Dec 4, 2002 in the DO group for those attachments was changed from

PLAN=SCAN(QWHCCV,1,' .'); to PLAN=SCAN(QWHCPLAN,1,' .');.

Thanks to Warren E. Waid, JCPenney, USA.
Change 20.279 The Local IP address TCBINDIP in TYP119A7 dataset was not

VMAC119 correctly stored; the statement to compress TCBINDIP must

Dec 4, 2002 be: TCBINDIP=COMPRESS(TCBINDIP);

but MXG code had .... TCBINDIC .. causing the error.

Thanks to Thomas Heitlinger, FIDUCIA IT AG, GERMANY.
Change 20.278 The %UPCASE(&SASSWORK) and %UPCASE(&USERWORK) added by

VMXGINIT Change 19.295 were lost, probably by Change 20.019. When

Dec 3, 2002 ITSV set USER=work (lowercase), their absence caused the

VMXGDEL to delete dataset work.ANYTHING because it was

different from dataset WORK.ANYTHING. %UPCASEs restored.

Thanks to Chris Weston, SAS Institute ITRM, USA.

Change 20.277 RMF Channel Activity Report enhanced with two selection

ANALRMFR parmaters, CHP= to select desired CHPID values, and

Dec 3, 2002 CACR= to selected SMF73ACR channel types.

Thanks to Don Isenstadt, Parker Hannifin Corporation, USA.

Change 20.276 Support for numerous NTSMF new objects and new data:

VMACNTSM -Support for MS Exchange 2000 SP3 changes:

VMXGINIT -DATABASE: Variables renamed:

EXNTxxxx TOCAPCHT==>CACHTOHT

Dec 3, 2002 TOCAHIRT==>CACHTORT

TOCAMIRT==>CACHTOMI

Variables removed:

TOOPENRT LGTHWAIT FIOPPEND FIOPRATE LGWRITRT LGRESTRT

CACHSTAL

Variables added:

DBPAFLRT DAPAEVRT DAPASTRT DBCASIZE

-DATABASE ==> INSTANCES: Variables renamed:

TOCAPCHT==>CACHTOHT

TOCAHIRT==>CACHTORT

TOCAMIRT==>CACHTOMI

Variables removed:

CACHFALT CACHSIZE FIBYWRRT FIBYRDRT TOOPENRT FIOPPEND

FIOPRATE LGWRITRT LGRESTRT CACHSTAL

-MSEXCHANGEDSACCESS CACHES: New variables:

CAEXPIRX CAINSRRX TOTENTRX DNENTRX SEAENTRX NDNENTR

NGUENTRX TOTENTMX DNENTMX SEAENTMX NDNENTMX NDUENTM

CAEXPITX CAINSRTX

-MSEXCHANGEIMAP4: New variables:

INVALICR CONNFAIL CONNREJE

-MSEXCHANGEPOP3: New variable:

INVALICR

-MSEXCHANGEIS: New variables:

RPCAVLAT RPCSLOPK

-Support for XP Professional:

-PROCESSOR: New Variables

PCTIDLTM PCTC1TM PCTC2TM PCTC3TM TRNSC1RT TRNSC2RT

TRNSC3RT

-Two new XP Professional Objects supported:

dddddd Dataset Object

NTPSFL PSCHFLOW PSCHED FLOW

NTWMIO WMIOBJCT WMI OBJECTS

-Support for MQSERIES QUEUES: New Instance variable

MSMQINST added.

-Support for new objects: Antigen Scan, Citrix, MSMQ,

and Data Core:

dddddd Dataset Object

NTANGN ANTIGENS ANTIGEN SCAN

NTCIIN CITRIXIN CITRIX IMA NETWORKING

NTCIMF CITRIXMF CITRIX METAFRAME XP

NTDCCA DACOCACH DATACORE CACHE

NTDCDC DACODOMC DATACORE DOMAIN CONTROLLER

NTDCFC DACOFIBC DATACORE FIBRE CHANNEL

NTDCMI DACOMIRR DATACORE MIRRORING

NTDCND DACOSCHD DATACORE SCHEDULING

NTDCNP DACONMVP DATACORE NMV POOLS

NTDCSC DACONMVD DATACORE NMV DISKS

NTMQMQ MSMQUEUE MSMQ QUEUE

NTMQMS MSMQSRVC MSMQ SERVICE
Change 20.275 HSM new variables added to dataset HSMDSRDS:

VMACHSM DSRDRECN='DATA SETS*RECONNECTED'

Dec 2, 2002 DSRBRECN='TRACKS*RECONNECTED*TO TAPE'

DSREXRED='EXTENT*REDUCTIONS'

DSRABACK='ABACKUPS*REQUESTED'

DSRABAKF='ABACKS*FAILED'

DSRABXTR='EXTRA*MOUNTS*RECALL*TAKEAWAY'

HSM now has the ability to "Reconnect" the MCDS record

when migrating a dataset that had been recalled; the

update-bit is tested, and if the dataset was not changed,

HSM "Reconnect"s that MCDS record to the migrated copy

on tape, updates the catalog to MIGRATE, and avoids any

movement of data to tape, counting the tracks that were

not moved to tape!

-There are 13 function types, not 12, so the DO loop was

increased to read all 13 array elements.

Thanks to Frank Cortell, Credit Suisse First Boston, USA.
Change 20.274 Support for RMF APAR OW56739. For RMF 79 subtype 3, adds

VMAC79 4-byte fields at the end of the segment (i.e. COMPATIBLY)

VMAC7072 to replace the existing and now-overflowing 2-byte frame

Dec 2, 2002 count variables. For RMF 72 subtype 3, TYPE72GO dataset,

adds new variable R723PLSC, with the period-level service

class name, which is now used in RMF reports to flag as

'HETEROGENEOUS' if R723PLSC is not identical to R723CLSC

for a certain period across the sysplex.

See Change 20.306 - correction.
Change 20.273 Workload Graph examples now include plots of MSU used by

GRAFWRKX each workload.

Nov 27, 2002

Thanks to Chuck Hopf, MBNA, USA.

Change 20.272 MSU variables that were zero in ASUM70PR/ASUMCEC from old

VMAC7072 S390 data with SMF70CPA=0 are now populated by using the

Nov 26, 2002 CPUTYPE/CPUMODEL variables and the $MG070CP table lookup

to get and store the CECSUSEC value back into SMF70CPA in

the TYPE70PR dataset; the non-zero SMF70CPA is then used

to calculate the MSU values in ASUM70PR/ASUMCEC.

Thanks to Chuck Hopf, MBNA, USA.

Thanks to Alan Sherkow, I/S Management Strategies, USA.

Change 20.271 -VMAC108: DOMPVN now labeled.

several -VMAC42: S42JDDSO now labeled.

Nov 26, 2002 -VMAC60: VVROPIND SMF60FNC VVRAMAT3 VVRTPEXT now labeled.

-VMACTPF: TPFRECNR, GMTOFF, SYSTEM are now labeled, and

temporary variable MZHDVAL is no longer kept.

-VMAC73: SMF73BSY now labeled.

Thanks to Chris Weston, SAS ITRM, USA.
Change 20.270 Enhancement for DB2 Trace IFCIDs 202, 230, and 239; those

READDB2 IFCID values are not SMF 102 subtypes, but are SMF 100

Nov 26, 2002 subtype 2 and 3, and SMF 101 subtype 1. READDB2 will now

create DB2STAT2 (202), DB2GBPAT (230) or DB2ACCTP (239)

datasets when those IFCIDs were requested in your READDB2

invocation.

Thanks to ???, ???, USA.
Change 20.269 Change 20.221 for ASUMUOW created variable TIMESTMP and

JCLUOWP changed the sort order of _SUOWCIC, but the JCL examples

JCLUOWV to use Views (JCLUOWV) or to use Pipes (JCUUOWP) were not

Nov 26, 2002 updated with the revised _SUOWCIC definition, causing the

VARIABLE TIMESTMP NOT FOUND error.

Thanks to Larry Nova, TransAmerica, USA.

Change 20.268 Cosmetic. Syntax of M4=-4; INPUT +M4 ... was replaced

DOC by INPUT +(-4) ... and the Assignment or RETAIN statement

Nov 25, 2002 for variables M1, M2, M4 ... were deleted. There is no

need to create/retain the negative valued variable to

move backwards in an INPUT statement.
Change 20.267 -Type 119 subtype 5 TYP11905 variable TSIPMAXS was not

VMAC119 input, causing the following 8 TSIPxxxx variables to be

Nov 23, 2002 wrong, and the label for TSIPCURS was changed to CURRENT.

-Type 119 subtype 2 TYP11902 variable TTSTATUS was 0 or

16777216 when it should have been 0 or 1; documented as

length 4, it's actually only length 1. Variable TTTOS is

only PIB1, TTXRT is PIB2, and TTINSEG and TTOUSEG are

both PIB8 instead of PIB4; IBM's originaly documentation

was corrected by the z/OS 1.3 online manual.

Thanks to Mark Cohen, DTS, ITALY.

Change 20.266 Most EXdddddd "Data Set Exit" members have one executable

EXDB2ACC statement, OUTPUT _Wdddddd or IF ... THEN OUTPUT ....

EXTY39 which permits instream tailoring syntax that will execute

EXNSPYAP the existing EXdddddd member's statement:

EXIMRACF %LET MACKEEP=

EXDB2ACP %QUOTE(

EXDB2ACB MACRO _EDB2ACC

EXCICJRN IF (your-new-selection-logic-is-true)

EXARBnnn THEN %%%INCLUDE SOURCLIB(EXDB2ACC); %%

Nov 20, 2002 );

However, that logic failed with EXDB2ACC, one of the few

exit members that has more than one executable statement:

IF something THEN DELETE; OUTPUT _WDB2ACC;

DB2 Parallel duplicate must be deleted by MXG, but IBM

techies using MXG inside DB2 Development needed to see

those records, so the normal DELETE was externalized.

Note: That DELETE was removed by Change 22.121.

a. The DELETE statement should not normally be used in

dataset exits, as it deletes the entire record; if

the record has repeated segments, a DELETE will stop

MXG from seeing the rest of that record's segments.

But by adding a DO; END; pair in the exit member code:

DO;

IF DB2PARTY='P' AND QWACPARR='Y' THEN DELETE;

OUTPUT _WDB2ACC; /* DB2ACCT, DEFINED IN IMACDB2 */

END;

there is only one executable statement, so the preceding

logic IF ... THEN %INCLUDE SOURCLIB(EX...) works.

All exit members were examined, and those with more than

one executable statement were wrapped in a DO; END; pair.

HOWEVER: The actual MXG recommended syntax for tailoring

in exit members when overriding the _Edddddd macro, as

shown in examples in member DOCMXG, show that I expected

that YOU would put the DO; ... END; pair in your logic:

%LET MACKEEP= %QUOTE(

MACRO _EDB2ACC

IF selection-critera THEN DO;

%%%INCLUDE SOURCLIB(EXDB2ACC);

END;

%%

);

and I still think that syntax is clearer and safer.
But this is still an worthy enhancement, because, now, no

matter which syntax you use to override _Edddddd macros,

either will work, transparently!

Thanks to Rob D'Andrea, Royal Bank of Scotland, ENGLAND.

Change 20.265 Undocumented fields added by FC4001 cause new SMF 94 data

VMAC94 (added by Change 20.206, using the text of the IBM PTF!)

Nov 19, 2002 to be wrong. IBM Redbook SG24-2229-05 now lists the 128

Nov 25, 2002 byte field we found at offset 400, but an 8-byte field

Nov 26, 2002 between S94OPM_AVG_VDM at offset 534 and S94OPM_DC1 at

offset 536 is not documented; it was located only by this

user's astute validation, knowing what values to expect.

The validation and -05 document made these corrections:

-S94VPSET is now input as PIB2 instead of 1.

-S94PSVC0-S94PSVC3 are now input as PIB2 instead of 1.

-S94DRIC1-8, S94xxWN1-2, S42ADI, S42DTW are recorded in

megabytes and were too small by a factor of 1048076, and

ADI/DTW were not formatted MGBYTES.

-The *1 or *2 text in many labels that didn't belong was

removed (added by erroneous CHANGE X Y ALL edit command).

-Variable SMF94VLC, the 5-byte alphanumeric Library Serial

"number" is now kept in TYPE942 & TYPE942P datasets, and

SMF94VLC should now be used in place of SMF94SNO in your

reports, as the 12-byte SMF94SNO field does not exist in

subtype 2 records. SMF94SNO was replaced by SMF94VLC in

the sort macros for the TYPE94, TYPE942, and TYPE942P

data sets, for consistency and so they could be matched.

SMF94VLC is created from SMF94SNO in (older) subtype 1

records that still have hex zeros in the VLC field.

SMF94VLC is not documented in the SMF 94 subtype 2

record, but MXG creates it as a five-byte character

variable from the 3-byte hex field that was observed to

contain it in the VPS Message Header field. SMF94VLS

is always missing value in TYPE94 when VLC contains a

character, so SMF94VLS is of no use in reporting.

OBSERVATION COUNT CHANGE: TYPE942 fewer obs.

OBSERVATION COUNT CHANGE: TYPE942P more obs.

Thanks to Wally Danielson, Airborne Express, USA.
Change 20.264 Variable MSEXISPU in MSEXCHPU dataset was increased from

VMACNTSM $32 to $64 to hold names like:

Nov 19, 2002 "First Storage Group-Public Folder Store (ISOMAILP1)"

Thanks to Xiaobo Zhang, Insurance Service Office, USA.

Change 20.263 -Support for BMC MVCICS 5.6 and correction to 5.5 CMRDETL

FORMATS support. New data BMC added to file segments in 5.5 was

VMACMVCI optional, the old (5.4) 16 bytes, or an expanded 92 bytes

Nov 16, 2002 but MXG always expected the new 92 byte length, causing

Nov 19, 2002 INPUT STATEMENT EXCEEDED error if you did not turn on the

new optional longer length records. The MXG test to

INPUT the longer length segment was revised to

IF T6ECPRID EQ 'F4'X AND T6EQUAL EQ '1.......'B THEN

using the new T6EQUAL field added in 5.5 to discriminate.
-In MVCICS 5.6, BMC added segment length into a reserved

field; new variable T6EFENLN is created by this Change.

-And MVCICS 5.6 extends the extended information to DB2,

MQSERIES, and DBCTL so the $MGMVCTI format that decodes

the file type variable, TFILI, was updated to identify

those types of files that were accessed from CICS.

Thanks to Franco Carmignani, IntesaBCI, ITALY.
Change 20.262 Support for APAR xxyyyyy added new RMF 99-1 trace codes

FORMATS for WLM dynamic resource group additional measurements,

Nov 16, 2002 that are now decoded by the MG099TC format.
Change 20.261 Preliminary look at Velocity Software ESAMON V3.2 data.

VMACXAM Updated the header from YY to YYYY; received documenation

Nov 15, 2002 of the current version, but yet to validate and compare.

Text of this change will be updated when changes made.

Change 20.260 Support for AS/400 5.2 Collection Services. INCOMPATIBLE

VMACQACS because you must change the LRECL for the files that IBM

Nov 15, 2002 IBM changed; if you upload 5.2 data with the new LRECL,

these data are readable with prior versions of MXG since

IBM added the new data at the end of the record:

Dataset Variable 5.2 LRECL

QAPMSYSL SYLPTB Added at end 3294

But these data were inserted or increased in length, so

you must have this change and correct LRECLs to read:

Dataset Variable 5.2 LRECL

QAPMETH ETMDIF Length increased 266

QAPMETH ETMUPF Added at end 266

Dataset Variable 5.2 LRECL

QAPMDISK DSASPX was DSAASP now NUM 367

QAPMDISK DSASPN Added at end 367

Thanks to Warren Cravey, Fidelity Systems, USA.

Change 20.259 TYPE71 Medium Impact Frames variables ESFRMEAV/MN/MX were

VMAC71 input out of order (AV had Min, MN had Max, MX had Avg)

Nov 15, 2002 this change also affected ESFRHIAV,ESFRHIMN,ESFRHIMX due

May 12, 2003 to subsequent calculations.

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

Thanks to Tony P. Stewart, Royal Mail, ENGLAND.