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

Change 11.120 Variable QW0058SN was incorrectly input as $CHAR2,

VMAC102 instead of PIB2., so it was unprintable. Changing to the

Jun 29, 1993 PIB2. will cause the actual statement number to print.

Thanks to Dave Leeker, Southwestern Bell, USA.

Change 11.119 Netview Type 37 "INPUT STATEMENT EXCEEDED RECORD LENGTH"

VMAC37 occurs because APAR OY49717 incorrectly documented the

Jun 24, 1993 format of the "DETT" (Data Network Subfield) text. The

APAR documented an imbedded count of sections BRFDENUM

and an imbedded length of each section, BRFDATLN, yet the

actual data contains neither field; instead, the triplet

LENDETT/NRDETT fields describe the text directly. The

workaround is to insert BRFDENUM=NRDETT; after the

INPUT statement for BRFDENUM, and to insert

BRFDATLN=LENDETT; INPUT @OFFDETT @;

after the INPUT statement for BRFDATLN. The actual MXG

revision eliminates the excess code.

Thanks to Larry Doub, RJReynolds Tobacco Co., USA
Change 11.118 The number of LPAR partitions, NRPRCS, was put back in

ASUM70PR the ASUM70PR data set from TYPE70PR. It had been removed

Jun 24, 1993 by Change 11.003 but turns out useful, so it can now be

Aug 4, 1993 used in reports.

Change 11.117 Support for Top Secret Release 4.3 type 80 SMF record.

VMAC80 Find the test for RACFVRSN following text "Top Secret"

VMAC80A and add OR RACFVRSN=043X to that test in both members

Jun 24, 1993 VMAC80 and VMAC80A to support this release.

Thanks to David Froberg, Software AG, USA.
Change 11.116 APAR OY54370 to NPM allows the Network Transit Time to be

VMAC28 calculated with only user data PIUs that have Definite

Jun 24, 1993 Response (specify TRANSIT=DR in FNMINIT), or now the Time

can include the data flow control PIUs (TRANSIT=DFC).

New MXG variable DFCPIU='Y' if TRANSIT=DFC was specified.

Thanks to Joseph L. Schwartz, CIGNA, USA.

Change 11.115 OMEGAMON V550 User SMF record ("ESRA") fails with INPUT

VMACOMCI STATEMENT EXCEEDED RECORD LENGTH due to MXG logic error.

Jun 24, 1993 Replace @LOC+LENTOTR+1 with @LOC+LENTOTR+1+(-6*(_I_-1))

to correct the alignment.

Thanks to Richard Cooke, Zurich Municipal, ENGLAND.

Thanks to Pete Gain, SAS Institute, ENGLAND.

Change 11.114 APAR OY64585 changes the format of IODFDATE in both type

VMAC73 73 and 74 records to MMDDYY8. (instead of YYMMDD8.), so

VMAC74 both INPUTs were corrected. Variables CYCLE and PARTISHN

Jun 23, 1993 are now kept in TYPE73, and new variable PNCHANBY is now

created in EMIF environments to report the Partition's

use of this CHPID. (PCHANBY is Total CHPID busy, while

PNCHANBY is the Partition's CHPID busy.)
Change 11.113 Support for VM/ESA Release 2.1 Accounting records added

TYPEVM type 0B accounting record for virtual disk in storage,

Jun 23, 1993 which MXG now outputs into dataset VMDEVICE. Variables

VDISKSUB and PAGEUSED in dataset VMDEVICE are non-missing

only for VMID='0B'X virtual disk record.
Change 11.112 Support for VM/ESA Release 2.1 Monitor MONWRITE records

VMACVMXA added several new variables, mostly for virtual disk in

Jun 23, 1993 storage statistics:

VXSYTSHS - QDGSYSLM,QDGUSRLM,QDGSYSCA,QDGLKCNT,QDGDISKS

(size, number, limits on VDISKs). Existing

variable VMSFORE converted to bytes and now

Formatted MGBYTES.

VXUSELOF - New variable VMDVDISK.

VXUSEACT - New variable VMDVDISK.

VXUSEATE - New variable VMDVDISK.

VXIODDEV - New variables VIUSTAMP, VIUSTATE.

(for I/O assist utilization calculations)

Variables now reserved (i.e., have zero value):

VXSTORSP - PLSTRMWT,PLSTRECT,PLSTEFCT,PLSTRDCT,PLSTDFCT,

PLSSYS1 ,PLSSYSE ,PLSSYSP1,PLSSYSPE

New data sets:

VXSTOSHL - 3.15 - NSS/DCSS/SSP Loaded into Storage

VXSTOSHD - 3.16 - NSS/DCSS/SSP Removed from Storage

VXSTOVDK - 3.17 - VDISK Information

Thanks IBM: both data and documentation were provided

for validation!
Change 11.111 VM/ESA dataset VXSYTCPM had incorrect values (DURATM=0,

VMACVMXA PCTCHBSY=wrong), or could ABEND with STOPOVER condition.

Jun 23, 1993 In _CSYTCPM, the ... THEN DO CHPID=0 TO CHPATH; must be

... TO CHPATH-1; Also, SKIP=SKIP-CHPATHLN; must be

replaced with IF SKIP GT 0 THEN INPUT +SKIP @;

Also, after M1=-1; insert SKIP=CHPATHLN-8;

Then in _DSYTCPM, CHPTSTMP+128*4294967296/1E6; must be

changed to CHPTSTMP+128*16777216/1E6 (because CHPTSTMP

is only 3-bytes. Before IF FIRST.BEGINMTR ..., insert

SAMECHAN=DIF(CHPID); DROP SAMECHAN; and add to the

IF FIRST.BEGINMTR OR ... a clauseOR SAMECHAN NE 0 .

Finally, after the END; of DO CHPID=0 TO CHPATH-1; and

before the final END; insert:

SPIN=MRHDRLEN-CALOFFST-CHPATH*CHPATHLN;

IF SKIP GT 0 THEN DO; INPUT +SKIP @; SKIP=0; END;

Thanks to Tom Healy, CHEMNET, USA.

Change 11.110 Support for SAP Releases 4.3.J & 5.0, which incompatibly

FORMATS changed the SAP CICS Journal segment in type 110 records.

IMACICSA Thirty-four new variables are created in the CICSSAP data

VMAC110 set and kept in VMAC110, and format $MGSAPTY was altered

Jun 22, 1993 because STCTASK now contains only a single character.

Note that variables STCIAPPL,STCIDEST,STCIDTYP,STCIMAND,

STCIMODP,and STCIPROG only have value if STC1CPIC='Y',

i.e., for CPIC Communication.

Thanks to Hr. Baehr, Helaba, GERMANY
Change 11.109 MXG 10.10 "Appended" IMS processing members JCLIMSLG and

JCLIMSLG TYPEIMSB were overlaid unexpectedly, and members IMSALOG,

TYPEIMSB IMSFPLOG,IMSTLOG,VIEWS,SUMTAB, and ASMIMSL3 were created

TYPESLRI unexpectedly when the MXG 10.10 tape was unloaded. These

Jun 22, 1993 eight members were supposed to be contained in new member

TYPESLRI as an unloaded IEBUPDTE PDS, but my error was to

fail to change the ./ IEBUPDTE syntax to xy, and thus the

8 members were created in your MXG.SOURCLIB. If you will

delete the last line in TYPEIMSB (a /* in col 1-2), and

ask us to fax you the MXG JCLIMSLG example (116 lines),

you can used the "Appended" IMS processing announced in

Change 10.142. Member TYPESLRI was described in Change

10.290, and now contains xy instead of ./ so that it will

not unload and overlay members again.

Thanks to Kenneth D. Jones, SHL SystemHouse, CANADA.
Change 11.108 DB2 variables CORRNAME and CORRNUM were not decoded for

IMACDB2 IMS connections with QWHCATYP=10. Now, the _DB2CORR IBM

Jun 22, 1993 macro tests for values of 5, 6, 9, or 10 for IMS.

Thanks to Alan Keeble, British Steel, ENGLAND.

Change 11.107 DB2 Trace IFCID=53 and IFCID=58 records can have no data

VMAC102 sections (QWHSNSDA=2 found for CLOSE CURSOR for example),

Jun 22, 1993 but this was not documented by IBM nor expected by MXG,

and those records with QWHSNSDA=2 were not output in

T102S053 or T102S058 data sets. Insert the ELSE OUTPUT

T102S05x; statements as shown:

Change To Read:

OUTPUT T102S053; OUTPUT T102S053;

END; END;

ELSE OUTPUT T102S053;

OUTPUT T102S058; OUTPUT T102S058;

END; END;

ELSE OUTPUT T102S058;

Thanks to John Morgan, Compuzorg B.V., NETHERLANDS.

Change 11.106 Support for DOS/VSE POWER 5.1 has been added. In datasets

TYPEDOS DOSJOBS and DOSREAD, variable TORMOT is now reserved, and

Jun 21, 1993 in DOSSNA variable INVLRESP is now reserved. DOS Network

Account field (8-bytes) was added to DOSLIST, DOSPUNCH,

DOSREAD, DOSXRC, and DOSPOOL, and records inserted or

deleted by OUTEXIT variables were also added in DOSPUNCH,

DOSLIST,DOSXRC, and DOSPOOL datasets. The Execution 'E'

record now documents offsets to either a PUTACCT user

extension or a "VSE/POWER extension area from ACDATE",

I have been unable to find the IBM documentation for that

area. The enhancement code has only been syntax checked.

Thanks to Jean-Claude Burger, Boehringer Ingelheim, FRANCE.

Change 11.105 Cosmetic. The list of members that use %VMXGFOR was out

VMXGFOR of data; now all 95 members are listed in the comments.

Jun 20, 1993

Thanks to Bob Butchco, The Genix Group, USA.

Change 11.104 "VARIABLE SACCT1 NOT FOUND" will result if you DROP all

IMACACCT of the SACCTn variables; BUILDPDB demands that at least

Jun 14, 1993 ACCOUNT1 and SACCT1 exist. You can set the length of

SACCT1 (and ACCOUNT1, if desired) to be one byte and

then DROP SACCT2-SACCT9 ACCOUNT2-ACCOUNT9 if you really

want to minimize the accounting variables that you keep.

Change 11.103 TYPEVVDS can have blank VVRSTGCL, VVRDATCL, or VVRMGTCL

VMACVVDS (Storage, Data, or Management Class) values, because the

Jun 14, 1993 test IF VVRLEN GE 148 ... should have been written as

IF VVRLEN GE 135 ...

Note, however, that there are legitimate TYPEVVDS records

that will still have blank values: Alternate Indexes do

not contain SMS classes, nor will datasets not under SMS!

Thanks to Scott Ashby, First Wachovia Operational Services, USA.

Change 11.102 MXG 11.01 only. Dataset TYPE73 has observations only for

VMAC73 MVS/ESA 4.3 or MVS/ESA 4.2 with APAR OY45991/PTF UY77343

Jun 11, 1993 installed. Change 11.015 was the culprit; VALIDPTH only

exists with the APAR or 4.3. Replace the single line:

IF VALIDPTH='Y' THEN DO;

with these lines:

IF (CFGCHGFL='....1...'B AND VALIDPTH='Y') OR

(CFGCHGFL='....0...'B AND (CHANIND GT 0 OR CHANINDX GT 0 OR

CHANINDY GT 0 OR PCHANBY GT 0 OR

SMF73PBY GT 0 OR SMF73PTI GT 0))

THEN DO;

Thanks to Rich Hawthorne, Nike, USA.

Change 11.101 The MXG Tape Mount Monitor is now the MXG Tape Mount and

ASMTAPES Allocation Monitor. The monitor now creates a subtype 4

EXTMNALO SMF record for each deallocation of a tape drive, with

IMACTMNT the start and end timestamps of how long that drive was

VMACTMNT allocated to that job, so that (finally!) the duration of

Jun 8, 1993 tape drive allocation can be measured, even if dynamic

allocation (eg. HSM, DB2MAST) or dynamic deallocation

(eg., FREE=CLOSE) is used. New MXG dataset TYPETALO is

now created by VMACTMNT (and also by BUILDPDB/3) for each

subtype 4 MXGTMNT record.

The ASM source code for the revised MXGTMNT program is in

ASMTAPES, and installation instructions are in comments;

except for Authorization, it installs exactly as the old

ASMTMNT. Yes, there is a price for this new priceless

data record: the new MXGTMNT load module must now always

be Authorized and stored in an Authorized Library, even

for testing to a flat file. Comments in ASMTAPES discuss

why Authorization is needed.

The allocation monitor was specified by Barry, but the

coding implementation was done for Merrill Consultants

by Bill Fairchild, Royal Software Associates. Chuck Hopf

did a lot of the early testing.

A minor enhancement was also made to the tape mount

monitor: the type of label (SL, NL, NSL, BLP) has been

added to the record and new variable LABEL is added to

TYPETMNT observations.

ORIGINAL TEXT OF CHANGE 11.101:

THIS MONITOR IS STILL IN ALPHA TESTING; IT CREATES VALID

RECORDS, BUT THEN ABENDS WITH AN 0C1, AND WE ARE STILL

DEBUGGING THE NEW MONITOR. DO NOT USE THIS NEW MONITOR!

Text revised Mar 24, 1994, after newsletter TWENTY-FIVE:

See Change 11.357.

Change 11.100 Line 051600 should be INPUT @109+OFFIMS instead of only

VMACCIMS INPUT @109; fortunately, OFFIMS is zero, so there was no

Jun 8, 1993 execution error.

Thanks to Vladimiro Vieno, Assicurazioni Generali, ITALY.

Change 11.099 MXG 11.01 only. A semicolon was missing after MTFLAG=0;

TYPEIMSA

Jun 8, 1993

Thanks to Norbert Korsche, OMV-AG, AUSTRIA.

Change 11.098 The DD name //DASDCOLL in this JCL example should have

JCLDAYDS been //DATASETS, as expected by the (DAILYDSN) program.

Jun 8, 1993

Thanks to David Ehresman, University of Louisville, USA.

Change 11.097 Amdahl format MGAMDDT mapped values 7,8,9 to 6380M1,M2,M3

FORMATS but they should be 6390M1,M2,M3, and have been changed.

Jun 3, 1993

Thanks to Graham West, Amdahl, ENGLAND.

Change 11.096 MXG 11.01 only. Insert an @; immediately before the

VMACDCOL IF LENGTH GE 308 THEN ... statement.

Jun 3, 1993

Thanks to Gary Matney, Twentieth Century Investors, USA.

Change 11.095 Variable STARTIME was added to the KEEP= list for the

VMACTMVS datasets TMVSYSSW and TMVSIV.

Jun 2, 1993

Thanks to Rod Fernandes, Albert Heijn B.V., HOLLAND.

Change 11.094 Several MXG 11.01-only typographical errors:

ASUM70PR ASUM70PR variable LPCTGOV is LABELed twice; delete the

GRAFLPAR second occurrence.

VMXGPRAL GRAFLPAR line 041300 has a missing quote; it should read:

Jun 2, 1993 TITLE2 H=2 F=SWISS 'LPAR % CPU Utilization';

VMXGPRAL utility prints only the first dataset found, and

prints it once for each dataset in the library! These

changes are required, but they also make this utility

work only with SAS Version 6 or later:

Insert COMPRESS=NO after NOBS in the KEEP= list following

the OUT=CONTENTS.

Insert (COMPRESS=NO) after OUT=DATASETS and before NODUP

in the PROC SORT statement.

After DATA _NULL; insert POINT=&I;

Add POINT=POINT to the SET DATASETS statement.

Thanks to Norbert Korsche, OMV-AG, AUSTRIA.

Change 11.093 0C4 ABEND in SASXKERN occurs if you have the IBM supplied

JCLTEST6 blocksize exit IFGOEXOB installed. The //SMFOUT DD in

Jun 2, 1993 JCLTEST6 does not contain a DCB specification, so the

exit picks one, but it conflicts with the BLKSIZE=23440

in the FILE statement in the SMFSMALL program! Adding

DCB=(RECFM=VBS,LRECL=32760,BLKSIZE=23440)

to the //SMFOUT DD avoids this possible abend.

Thanks to John Visser, Delta Motor Corporation, SOUTH AFRICA.

Change 11.092 Omegamon 2.60 Audit Record moved the 8-byte OMSUBSID name

VMACOMAU (specifically, the DB2 subsystem name) to the existing

Jun 1, 1993 reserved OMRSV4 field, but only the first four bytes are

used! Change OMRSV4 $CHAR8. to OMRSVR $CHAR4. +4

and then after the final @;, insert

IF OMSUBSID='0000000000000000'X THEN OMSUBSID=OMRSV4;

Thanks to Wayne Cope, Belk Stores, USA.
Change 11.091 Variable MESSAGE was not decoded in STOPX37 Release 3.5

VMACX37 (because it is not described in their DSECT), but it is

May 28, 1993 located, and @505+OFFSMF MESSAGE $CHAR140. should

be added to the INPUT for Release 3.5.

Thanks to Bob Guthier, Lucky Stores, USA.
Change 11.090 Although ASMVTOC/VMXGVTOF are strongly recommended rather

FMXGUCBL that this archaic FMXGUCBL, sites under MVS/ESA 4.2 will

May 28, 1993 encounter an 0C2 ABEND OUTSIDE THE SAS SYSTEM (because it

required to be Authorized). Rather than putting SAS in

an Authorized Library, the following changes will correct

FMXGUCBL so that it does not require authorization, in

case you are still using this Archaic member.

1. Locate label MVS42AAA DS 0H, and move this line to

after L R2,UCBADDR

2. Locate label MVS42AAB, and change these arguments of

the UCBSCAN macro: Change ADDRESS, to COPY, Replace

UCBPTR=UCBADDR, with UCBAREA=UCBCOPY, Remove NOPIN,

After all of the arguments to the UCBSCAN macro,

and before the L R15,UCBSCNRT, insert

LA R2,UCBCOPY

3. Locate UCBSCNRT DS F, and insert after that line

UCBCOPY DC XL48'00'

4. In the JCL, change C. to ASM. and L. to LKED. and

add a comma after the PARM.ASM field, and remove the

extraneous comma after the PARM.LKED field.

Thanks to Waldemar Schneider, SAS Institute Europe, GERMANY.
Change 11.089 In BUILDPDB, real purge records can go to PDB.NJEPURGE

VMAC26J2 instead of being recognized as real purge records. This

May 26, 1993 can cause jobs to be held in the SPIN library instead of

being output in PDB.JOBS, and when SPINCNT is exceeded

and these jobs are finally written to PDB.JOBS, they will

have ABEND='NOTP' (i.e. did not find a Purge record).

The error was created when IBM replaced PRPRTY with a new

field (SMF26IX2) in ESA 4.3, but when MXG decoded the new

variables DUPJBDLY and OFFLPURG, I failed to decode only

for 4.3+. Variable OFFLPURG is used in BUILDPDB logic to

recognize purge records created by JES2 SPOOL Transfer,

but if the PRPRTY happened to be 4-7 or 12-15, MXG set

OFFLPURG='Y' in VMAC26J2, and then in BUILDPDB that purge

record went to PDB.NJEPURGE! The MXG correction is

to only create DUPJBDLY/OFFLPURG if this is an

MVS/ESA 4.3 record (and, of course, JES2 does not put any

record level indicator in their record, so the only way
to recognize that this is a 4.3 record is to test for the

existence of the accounting triplet that was also added

by 4.3!). In VMAC26J2, add the IF ... DO; and END;

statements as shown:

IF NETFLAG='.......1'B THEN DO;

IF PRPRTY='1.......'B THEN DUPJBDLY='Y';

IF PRPRTY='.1......'B THEN OFFLPURG='Y';

END;

Thanks to Brenda Rabinowitz, Prudential-Bache Securities, USA.
Change 11.088 In line 4800, :"0. Check if the summary table already

TYPESLRI exists;" - the semicolon must be changed to a period, to

May 26, 1993 prevent a syntax error.

Thanks to Steen Pedersen, Danfoss. DENMARK.

Change 11.087 The calculation of LP0MGTTM=LP1UPDTM-LP1UEDTM should have

ASUM70PR obviously been LP0MGTTM=LP0UPDTM-LP0UEDTM, and the

May 24, 1993 variable LP0MGTTM needs to be added to the RETAIN list.

Thanks to Tom Parker, Hogan Systems, USA.

Change 11.086 Variables F20STRT and F20DURA are now FORMATted TIME12.2.

VMACILKA

May 24, 1993

Thanks to Gary Strohminger, AT&T Transtec, USA.

Change 11.085 DB2 Trace dataset T102S145 (Audit Trace) variables

VMAC102 QW0145SC and QW0145LL were not input under DB2 2.3 due to

May 24, 1993 incorrect INPUT logic. Insert an "@;" after the input

of QW0145HO $CHAR4. and insert an "INPUT" before QW0145SC

so the logic ends up looking like this:

IF QWHSRELN GE 2.3 THEN INPUT

QW0145HO $CHAR8. /*HOST OPTIONS*FOR SQL*/

@;

ELSE INPUT

QW0145HO $CHAR4. /*HOST*OPTIONS*/

@;

INPUT QW0145SC $CHAR4. /*SQL*RETURN*CODE*/

QW0145LL PIB2. /*SQL*STATEMENT*LENGTH*/

@;

Thanks to Wilson Kwong, National Australia Bank, AUSTRALIA.

=====Changes thru 11.084 were in MXG PreRelease 11.01, May 20, 1993=====
Change 11.084 ANALNPMR produced only the first report requested when

ANALNPMR PDB=SMF was specified; the same name was used for both

May 19, 1993 input and output, causing no data to be found for the

second and subsequent reports.

Thanks to Victor Chacon, NYNEX Mobil Communications Co., USA.
Change 11.083 Landmark CICS data under DOS/VSE was previously processed

TYPEMOND by MXG member TYPEMOND, but only for Landmark Version 7.

TYPEMONX That archaic member was copied to create TYPEMONX (just

May 19, 1993 in case someone still has Version 7 under DOS/VSE), and

TYPEMOND was completely revised to (hopefully) support

Landmark Versions 8, 9, and 1.0 under DOS/VSE. However,

no test data nor hex dump of a Version 8 record from DOS

was available, so this support has only been syntaxed.

Change 11.082 CICS/ESA journal data produced by the SAP product was not

VMAC110 coded exactly the same as CICS/non-ESA, although the end

May 19, 1993 results were the same. Now, the code segment for both

CICS/ESA and CICS/non-ESA uses the same logic.

Thanks to Isabelle Mejane, SAS Institute, FRANCE.
Change 11.081 DIVIDE BY ZERO error occurs if GEICPUOL is zero. Replace

VMACZRB PCTCPUBY=100*(INTERVAL-(CPUWAIT/GEICPUOL))/INTERVAL;

May 19, 1993 with

IF GEICPUOL GT 0 THEN

PCTCPUBY=100*(INTERVAL-(CPUWAIT/GEICPUOL))/INTERVAL;

ELSE PCTCPUBY=.;

Thanks to Marcel Frechette, Montreal Trust Technology, CANADA
Change 11.080 STARTIME and CPUTCBTM in all CICS/ESA Statistics datasets

VMAC110 (eg., CICINTRV, CICEODRV, CICDS, etc.) are wrong. This

May 19, 1993 error did NOT affect the CICSTRAN dataset.

a.The STARTIME was set to COLLTIME, which is the END of

interval and not the START. Find these lines

COLLTIME=DHMS(SMFSTDAT,CLTHH,CLTMM,CLTSS);

STARTIME=COLLTIME;

IF 0 LE DURHH LE 24 AND 0 LE DURMM LE 59 AND

0 LE DURSS LE 59 AND

SMFSTRQT='INT' THEN DURATM=HMS(DURHH,DURMM,DURSS);

and change them to read:

COLLTIME=DHMS(SMFSTDAT,CLTHH,CLTMM,CLTSS);

IF 0 LE DURHH LE 24 AND 0 LE DURMM LE 59 AND

0 LE DURSS LE 59 AND

SMFSTRQT='INT' THEN DURATM=HMS(DURHH,DURMM,DURSS);

STARTIME=COLLTIME-SUM(0,DURATM);

b.The CPUTCBTM was correct in CICS 3.1.1, because DSGTCT

did contain total TCB time for each TCB, but beginning in

CICS 3.2, IBM reused field name DSGTCT for the DS-Only

TCB time and put the Total TCB time in new variable

DSGACT, but I failed to change the CPUTCBTM equation.

Find the line:

CPUTCBTM=SUM(DSGTCT,DS2TCT,DS3TCT,DS4TCT);

and replace that equation with:

CPUTCBTM=SUM(DSGACT,DS2ACT,DS3ACT,DS4ACT);

c.Also variables DSnTCBFm are now formatted $HEX2., and the

test for DSGNTCB GE 3 was changed to DSGNTCB GE 4, since

CICS 3.3 can have four TCBs.

Thanks to Scott Pearson, King County Medical Blue Shield, USA.
Change 11.079 SAS 6.08 Usage Note 6719 discusses a SAS error in %MACRO

GRAFLPAR compiler which causes GRAFLPAR to fail with an "OUT OF

May 19, 1993 MEMORY" condition (which is NOT the error!). Apparently,

GRAFLPAR broke an internal limit on characters passed as

%MACRO arguments (and blanks at the end of argument lines

are being counted!). We have been able to circumvent the

ABEND by reducing the number of characterspassed by MXG

by reducing four lines of code with two. In three places

find the four lines beginning with

lparnum=lparnum+1;

and move the 2nd and 4th lines to the end of 1st and 3rd!

SAS Institute is still working on a ZAP; the above Usage

Note will be updated when a ZAP exists.

Thanks to Chuck Hopf, Primerica, USA.

Change 11.078 New HSM dataset HSMFSRBO contains both tape and disk data

DIFFHSM set statistics, with several new variables, based on this

EXHSMFSB user contribution. Additionally, VSRTOTKB is now input

IMACHSM as PIB4. instead of PIB2. This change altered IMACHSM,

VMACHSM if you have tailored IMACHSM in your USERID.SOURCLIB, you

May 19, 1993 must retrofit your changes using this new IMACHSM.

Thanks to Neil Campbell, Inland Revenue, ENGLAND.
Change 11.077 The HSMVSRFU dataset previously was OUTPUT only if two of

EXHSMVSF the data variables were non-zero, but now these variables