/********************************************************************/ /* S A S S A M P L E L I B R A R Y */ /* */ /* NAME: TLKTLOAD */ /* TITLE: TLKTDUMP executable module loading program */ /* PRODUCT: BASE */ /* SYSTEM: ALL */ /* KEYS: */ /* PROCS: PDSCOPY (MVS only) */ /* DATA: NONE */ /* */ /* SUPPORT: Richard D. Langston UPDATE: JUN94 */ /* REF: TLKTLOAD/TLKTDUMP User Documentation */ /* MISC: Companion to the TLKTDUMP program */ /* */ /********************************************************************/ /* TLKTLOAD is a SAS program that reads in a text file generated by */ /* the TLKTDUMP program. It validates the file and ensures that the */ /* file has been tranferred correctly. If all is valid, it creates */ /* the executable(s) dumped by the TLKTDUMP program. */ /* */ /* First %INCLUDE this TLKTLOAD program, then use the %TLKTLOAD */ /* macro to invoke the program: */ /* */ /* %TLKTLOAD(INFILE=dumpfile,OUTFILE=executable); */ /* */ /* The INFILE= option refers to a fileref or a quoted string giving */ /* the name of the input dump file. This file is one that has been */ /* created by the TLKTDUMP program. The INFILE= option must be */ /* given. */ /* */ /* The OUTFILE= option refers to a fileref or a quoted string giving*/ /* the name of the executable module or library. On MVS, this will */ /* be a load module library. All restored modules will be added to */ /* (or replaced in) the library. On CMS, this will be a LOADLIB. */ /* All restored modules will be added to (or replaced in) the */ /* library. On VMS platforms, this will be a specific .EXE file. */ /* It will be replaced. On UNIX platforms, this will be a specific */ /* executable file, with no extension. It will be replaced. On PC */ /* platforms, this will be a specific .DLL file. */ /* */ /* The dump file can consist of several pieces that have been */ /* downloaded separately and "glued" back together, even if they're */ /* not in the right order. As long as each section is separated by */ /* the part delimiter record, any data */ /* between the delimiter records (such as EMAIL address information)*/ /* will be ignored. If any text has been altered, or any data */ /* modified in any way, or if a section appears more than once or */ /* not at all, the TLKTLOAD program will detect it and alert you */ /* in the PRINT file. */ /* */ /* Note that on CMS, it is required that a writeable A disk */ /* be present, since our temporary files are hard-coded to use */ /* a filemode of A. */ /*------------------------------------------------------------------*/ %MACRO TLKTLOAD(INFILE=,OUTFILE=); %LET VERSION = 1.02; /*==================================================================*/ /* In this section, we define the temporary files for the REORDER */ /* and SASCODE filedefs. These files will be used in creating the */ /* intermediate dump file with all parts in the right order and */ /* without any fence information. Also at this point we will try */ /* to determine what the correct DCB info will be for the executable*/ /* if possible. */ /*==================================================================*/ /*------------------------------------------------------------------*/ /* We may need to perform an explicit OS command to delete the */ /* REORDER and SASCODE temporary files used during our execution. */ /* To ensure we can simply reference the macro variable to invoke */ /* the command, we create dummy values for the macro variables that */ /* are simply comments. */ /*------------------------------------------------------------------*/ %LET DELREORD= * No deletion of REORDER necessary; %LET DELSASCD= * No deletion of SASCODE necessary; /*------------------------------------------------------------------*/ /* MVS-type operating systems simply use &&-prefixed data set names */ /* which will be deleted when the SAS job terminates. We defer */ /* setting our file DCB until we determine if the dump represents */ /* an IEBCOPY dump file or a PROC PDSCOPY dump file. */ /*------------------------------------------------------------------*/ %IF "&SYSSCP."="OS" %THEN %DO; %LET REORDER= '&&REORDER'; %LET SASCODE= '&&SASCODE'; %END; /*------------------------------------------------------------------*/ /* CMS uses real file names, and therefore the DELREORD and DELSASCD*/ /* macro variables are set to the proper ERASE commands. Note that */ /* we also set the DCB of the LOADLIB file. */ /*------------------------------------------------------------------*/ %ELSE %IF "&SYSSCP."="CMS" %THEN %DO; %LET DCB=RECFM=V LRECL=32760; %LET REORDER= 'TEMP REORDER A'; %LET SASCODE= 'TEMP SASCODE A'; %LET DELREORD= X "ERASE TEMP REORDER A"; %LET DELSASCD= X "ERASE TEMP SASCODE A"; %END; /*------------------------------------------------------------------*/ /* For VMS, we can place the files in SAS$WORKLIB, which gets */ /* cleared when the SAS job terminates. We can also set the DCB */ /* at this time. */ /*------------------------------------------------------------------*/ %ELSE %IF "&SYSSCP."="VMS" OR "&SYSSCP."="VMS_AXP" %THEN %DO; *-----obtain VMS DCB for EXE file-----*; %LET DCB=RECFM=F LRECL=512; %LET REORDER= 'SAS$WORKLIB:REORDER.DAT'; %LET SASCODE= 'SAS$WORKLIB:SASCODE.SAS'; %END; /*------------------------------------------------------------------*/ /* For PC platforms, we use explicit file names and set the */ /* DEL commands to delete the files. Note that we use OPTIONS */ /* NOXWAIT to ensure that the user doesn't have to terminate the */ /* shell session. */ /*------------------------------------------------------------------*/ %ELSE %IF "&SYSSCP."="OS2" %OR "&SYSSCP."="WIN" %THEN %DO; %LET DCB=RECFM=N; OPTIONS NOXWAIT; %LET REORDER= 'REORDER.TMP'; %LET SASCODE= 'SASCODE.TMP'; %LET DELREORD= X "DEL REORDER.TMP"; %LET DELSASCD= X "DEL SASCODE.TMP"; %END; /*------------------------------------------------------------------*/ /* For UNIX platforms, we use explicit file names in /tmp, but */ /* we also set up rm commands to ensure that the files are cleaned */ /* up. */ /*------------------------------------------------------------------*/ %ELSE %DO; %LET DCB=RECFM=N; %LET REORDER= "/tmp/reorder.tmp.&sysjobid."; %LET SASCODE= "/tmp/sascode.tmp.&sysjobid."; %LET DELREORD= x "rm /tmp/reorder.tmp.&sysjobid."; %LET DELSASCD= x "rm /tmp/sascode.tmp.&sysjobid."; %END; *------print initial information in print file-----------------------*; DATA _NULL_; FILE PRINT; PUT 'TLKTLOAD: A SAS program to create SAS/TOOLKIT(TM)' 'executables from downloaded TLKTDUMP files'; PUT "Version &VERSION."; PUT "Requested INPUT file: &INFILE"; PUT "Requested OUTPUT file: &OUTFILE"; RUN; /*------------------------------------------------------------------*/ /* Since we allow either a fileref or an explicit file name for */ /* both INFILE= and OUTFILE=, we must ensure that we have a fileref */ /* for each, since some code later on may require a fileref and not */ /* permit an explicit file name. We do this by verifying if the */ /* first character of the INFILE and OUTFILE macro variables is a */ /* quote (single or double). If either, then we consider the string */ /* to be an explicit name, in which case a FILENAME statement is */ /* generated in a macro variable (FILEST1 or FILEST2). If the name */ /* is already a fileref, then FILESTx is a comment statement. */ /* We also create the macros INFILEX and OUTFILEX which will always */ /* contain filerefs, either provided by the user via INFILE= and */ /* OUTFILE= or produced via our generated FILENAME statements. */ /*------------------------------------------------------------------*/ DATA _NULL_; LENGTH WHICH $8; WHICH = 'INFILE'; NUMBER=1; LINK DOIT; WHICH = 'OUTFILE'; NUMBER=2; LINK DOIT; RETURN; DOIT:; XFILE=LEFT(SYMGET(WHICH)); IF SUBSTR(XFILE,1,1)='"' OR SUBSTR(XFILE,1,1)="'" THEN DO; FILENAME='RDL'!!PUT(NUMBER,Z5.); STMT='FILENAME '!!TRIM(FILENAME)!!' '!!XFILE; END; ELSE DO; FILENAME=XFILE; STMT='* No FILENAME statement necessary for '!!WHICH; END; CALL SYMPUT(TRIM(WHICH)!!'X',TRIM(FILENAME)); CALL SYMPUT('FILEST'!!PUT(NUMBER,1.),STMT); RETURN; RUN; *-----invoke the possible FILENAME statements------------------------*; &FILEST1; &FILEST2; /*==================================================================*/ /* In this section, we read through the dump file and determine */ /* the start and end locations for each part of the dump. This is */ /* necessary because the parts may not be in the right order, and */ /* also because we permit extraneous records (such as mail headers) */ /* to appear between the parts. With this first pass of the dump */ /* file, the only error checking done is to ensure that each IDENT */ /* indicator is the same and that we encounter and EOF record */ /* somewhere in the dump. */ /* Note that each part is delineated by a header and trailer record */ /* that starts with ****PART in column 1. The part number follows, */ /* followed then by some identifier string. That string can be */ /* supplied by the dump creator, but the string always has the */ /* creation datetime appended to it. */ /*==================================================================*/ DATA PARTS; INFILE &INFILEX LENGTH=L END=EOF; RETAIN MAXPARTN FOUNDEOF 0 SEQUENCE START IDENT_S; LENGTH IDENT $80; *-----read a record of the dump---------------------------------*; INPUT @; INPUT @1 RECORD $VARYING80. L; *-----if the record is a part header, obtain more info----------*; IF RECORD=:'****PART ' THEN DO; *-----get the IDENT info from the record---------------------*; I=INDEX(RECORD,'IDENT = '); IF I THEN DO; I+8; I+(SUBSTR(RECORD,I,1)=':'); RRECORD=REVERSE(RECORD); J=81-VERIFY(RRECORD,'* '); IDENT=SUBSTR(RECORD,I,J-I+1); END; *-----get the part number from the record--------------------*; WHICH=SCAN(SUBSTR(RECORD,13),1,': '); NUM=SCAN(RECORD,2,' '); IF VERIFY(NUM,'0123456789 ')=0 THEN DO; SEQUENCE=INPUT(NUM,BEST12.); MAXPARTN=MAX(MAXPARTN,SEQUENCE); END; *-----part number XXX means we found the last section--------*; ELSE IF NUM='XXX' THEN FOUNDEOF=1; *-----record the record number of start of data--------------*; IF WHICH='START' THEN DO; START=_N_+1; IDENT_S=IDENT; END; *-----record the record number of end of data----------------*; ELSE DO; END=_N_-1; *-----ensure IDENT values match between start and end-----*; IF IDENT_S NE IDENT THEN DO; FILE PRINT; PUT 'ERROR: IDENT labels for start and end do not ' 'match for part ' sequence; ABORT; END; OUTPUT; END; END; *-----if at end of file, ensure we found an EOF record----------*; IF EOF THEN DO; IF NOT FOUNDEOF THEN DO; PUT 'ERROR: An EOF was not found in the dump file, ' 'indicating at least the last section was omitted.'; ABORT; END; *-----save part count for later use--------------------------*; CALL SYMPUT('MAXPARTN',PUT(MAXPARTN,10.)); END; KEEP SEQUENCE START END IDENT; RUN; /*------------------------------------------------------------------*/ /* Verify that all the IDENT values are the same for all parts. If */ /* not, we have an error condition. */ /*------------------------------------------------------------------*/ DATA _NULL_; SET PARTS END=EOF; BY IDENT NOTSORTED; NDIFF+FIRST.IDENT; IF EOF AND NDIFF NE 1; FILE PRINT; PUT 'ERROR: Identifying string is not the same for all parts ' 'of the dump'; ABORT; RUN; /*------------------------------------------------------------------*/ /* Sort the section indicators into the correct sequence. We also */ /* create a set of observations to merge together to ensure that */ /* there is one and only one section per sequence number. */ /*------------------------------------------------------------------*/ PROC SORT DATA=PARTS(DROP=IDENT); BY SEQUENCE; DATA MATCHUP; DO SEQUENCE=0 TO &MAXPARTN; OUTPUT; END; RUN; /*------------------------------------------------------------------*/ /* Now merge the two together and ensure that 1) all sections are */ /* present and 2) no section is repeated. */ /*------------------------------------------------------------------*/ DATA _NULL_; MERGE PARTS(IN=HAVE) MATCHUP END=EOF; BY SEQUENCE; FILE PRINT; RETAIN ERROR 0; IF NOT HAVE THEN DO; PUT 'ERROR: Part ' sequence ' of the dump file is not ' 'present. Processing cannot continue.'; ERROR = 1; END; ELSE IF NOT (FIRST.SEQUENCE AND LAST.SEQUENCE) THEN DO; PUT 'ERROR: Part ' sequence ' of the dump file appears ' 'multiple times. Processing cannot continue.'; ERROR = 1; END; IF EOF AND ERROR THEN ABORT; RUN; /*------------------------------------------------------------------*/ /* At this point, we know that the dump contains all the sections */ /* we need, and we know the locations of each section. We generate */ /* a SAS program into the SASCODE file that will be of the form */ /* */ /* DATA _NULL_; INFILE &INFILEX */ /* LENGTH=L FIRSTOBS=x OBS=y; */ /* FILE REORDER ; */ /* INPUT @; INPUT @1 LINE $VARYING200. L; */ /* IF LINE NE ' ' THEN PUT _INFILE_; */ /* RUN; */ /* */ /* This code will write into the REORDER file only the dump records */ /* we want, eliminating headers/trailers and any other superfluous */ /* text, including blank lines. Note that the MOD option is not */ /* used for the first DATA step. Also, blank lines are preserved */ /* for the first part of the dump, which is textual in nature. */ /*------------------------------------------------------------------*/ *-----ensure we have FILEREFs for REORDER and SASCODe----------------*; FILENAME REORDER &REORDER; FILENAME SASCODE &SASCODE; *-----Generate the SAS code described above--------------------------*; DATA _NULL_; SET PARTS(RENAME=(START=FIRSTOBS END=OBS)); FILE SASCODE; LENGTH MOD $3; RETAIN OBS MOD; IF SEQUENCE=0 THEN CALL SYMPUT('STARTREC',PUT(OBS-FIRSTOBS+2,BEST12.)); PUT "DATA _NULL_; INFILE &INFILEX LENGTH=L " FIRSTOBS= OBS= '; FILE REORDER ' MOD ';'; PUT 'INPUT @; INPUT @1 LINE $VARYING200. L; '; IF SEQUENCE NE 0 THEN PUT "IF LINE NE ' ' THEN "; PUT "PUT _INFILE_;"; PUT 'RUN;'; MOD='MOD'; RUN; *-----Invoke the SAS code that has just been generated---------------*; %INC SASCODE; RUN; *-----Clear out the SASCODE file since it is no longer needed--------*; FILENAME SASCODE CLEAR; &DELSASCD; *------get information from the header portion-----------------------*; DATA _NULL_; INFILE REORDER LENGTH=L END=EOF; *-----get ident, length, version, host-----*; LENGTH LRECL $8 TLKTDMPV $4 HOST IEBCOPY $8 RECFM $8 BLKSIZE $8; INPUT LRECL= HOST= & TLKTDMPV= IEBCOPY= ; FILELREC=LRECL; *-----get IEBCOPY information if necessary-----*; if IEBCOPY='YES' THEN DO; INPUT @19 RECFM= LRECL= BLKSIZE=; CALL SYMPUT('RECFM',TRIM(RECFM)); CALL SYMPUT('IEBLRECL',TRIM(LRECL)); CALL SYMPUT('BLKSIZE',TRIM(BLKSIZE)); END; ELSE IEBCOPY='NO'; *-----make macros out of ident and lrecl for later use-----*; CALL SYMPUT('LRECL',TRIM(FILELREC)); CALL SYMPUT('IEBCOPY',TRIM(IEBCOPY)); *-----verify we are on the right host-----*; IF HOST NE "&SYSSCP." THEN DO; FILE PRINT; PUT 'ERROR: The TLKTDUMP file is meant for the ' HOST ' operating system but this SAS program is being ' "run on the &SYSSCP. operating system."; ABORT; END; *-----warn if the version number does not match-----*; IF TLKTDMPV NE "&VERSION" THEN DO; FILE PRINT; PUT 'WARNING: The TLKTDUMP file was generated with TLKTDUMP' ' Version ' TLKTDMPV ' but this TLKTLOAD is ' " Version &VERSION.. Beware of incompatibilities..."; END; STOP; RUN; /*------------------------------------------------------------------*/ /* Here we have to do some host-specific things. For MVS, we must */ /* now deal with IEBCOPY information. The IEBCOPY utility is very */ /* sensitive about correct DCB information for its unload file. */ /* Therefore, when TLKTDUMP creates a dump file, it records the DCB */ /* of the unload file so that we will be able to recreate the */ /* file with correct DCBs. If instead we're using the PDSCOPY */ /* unload file, we can always use DCB=(RECFM=VB,LRECL=32756, */ /* BLKSIZE=32760). And note that if we are not recreating an */ /* IEBCOPY unload file, we're recreating a PDSCOPY unload file, */ /* which must be a temporary. Therefore, we redefine the output */ /* file to be SEQFILE. Note that we used OUTFILEX earlier to hold */ /* the output fileref. We can redefine OUTFILEX to SEQFILE for */ /* PDSCOPY unload files. For CMS, we have to deal with the */ /* peculiarity that the DATA step does not permit you to use a */ /* LOADLIB file as an output file. This is because it expects you */ /* to treat the LOADLIB file as a library, not as a sequential */ /* file. To overcome this, we use a temporary file TEMP DATA A and */ /* have the fileref RDL00003 refer to it. We reset OUTFILEX */ /* accordingly. Once we populate TEMP DATA A, we'll copy it to the */ /* desired LOADLIB file. */ /*------------------------------------------------------------------*/ %IF "&SYSSCP."="OS" %THEN %DO; %IF "&IEBCOPY." NE "YES" %THEN %DO; %LET DCB=RECFM=VB LRECL=32756 BLKSIZE=32760; FILENAME SEQFILE '&&OUT'; %LET OUTFILEX=SEQFILE; %END; %ELSE %LET DCB=RECFM=&RECFM LRECL=&IEBLRECL BLKSIZE=&BLKSIZE; %END; %IF "&SYSSCP."="CMS" %THEN %DO; FILENAME RDL00003 'TEMP DATA A'; %LET OUTFILEX=RDL00003; %END; /*------------------------------------------------------------------*/ /* Finally, we are at the point where we can read through the dump */ /* file and create our local binary file from the dump. */ /*------------------------------------------------------------------*/ DATA _NULL_; INFILE REORDER FIRSTOBS=&STARTREC END=EOF LENGTH=INRECL; LENGTH STG $160 SRECLEN $20; RETAIN PRECNUM RECNUM PARTNUM 0; *-----not getting hex data (getting length value)-----*; HEX=0; *-----get length of next restore record-----*; LINK GETLEN; IF LENGTH=0 /* EOF */ THEN STOP; *-----read the hex data, computing checksum and making binary---*; CHECKSUM=0; I=1; HEX=1; DO WHILE(I<=LENGTH); *-----extract a piece-----*; LINK GETCOMP; *-----accumulate checksum-----*; DO K=1 TO L; CHECKSUM=MOD(CHECKSUM+INPUT(SUBSTR(STG,K,1),PIB1.)+1,32768); END; *-----write out the piece-----*; LINK PUTDATA; I+L; END; *-----not in hex mode (getting checksum)-----*; HEX=0; *-----flush the binary record-----*; L=0; LINK PUTDATA; *-----absorb comma-----*; L=1; LINK GETDATA; *-----get the checksum and validate against computed one-----*; LINK GETLEN; IF LENGTH NE CHECKSUM THEN DO; FILE PRINT; PUT 'ERROR: For restore file record number ' RECNUM '(input file record ' NRECS ')' ' the provided checksum and the computed checksum ' ' did not match. there is a data integrity problem.'; ABORT; END; *-----on to next restore file record-----*; RETURN; *-----routine to get recnum and length from input file-----*; GETLEN:; J=1; *-----read string of format xxxx:yyyy,-----*; DO UNTIL(STG=','); *-----get a single byte-----*; L=1; LINK GETDATA; *-----concatenate in-----*; SUBSTR(SRECLEN,J,1)=STG; J+1; END; *-----read recnum before colon, length after-----*; RECNUM=INPUT(SCAN(SRECLEN,1,':,'),5.); LENGTH=INPUT(SCAN(SRECLEN,2,':,'),5.); RETURN; *-----routine to generate a stream of uncompressed data-----*; GETCOMP:; *-----read the indicator byte first-----*; L=1; LINK GETDATA; COMPBYTE=INPUT(SUBSTR(STG,1,1),PIB1.); *-----if over 127, this indicates compression-----*; IF COMPBYTE>127 THEN DO; *-----get rid of high-order bit to get length-----*; COMPBYTE=COMPBYTE-128; *-----read the repeating byte-----*; L=1; LINK GETDATA; *-----generate the complete string-----*; STG=REPEAT(SUBSTR(STG,1,1),COMPBYTE-1); L=COMPBYTE; END; *-----otherwise we get the specified number of bytes-----*; ELSE DO; L=COMPBYTE; LINK GETDATA; END; RETURN; *-----read L bytes of binary data from the hex dump-----*; GETDATA:; RETAIN INCOL 1; LENGTH PART $&LRECL; *-----determine actual bytes to read, depending on hex mode-----*; LL=L*(1+HEX); II=1; STG=' '; *-----read that many bytes-----*; DO WHILE(II<=LL); *-----read remaining bytes, or all we need-----*; LLL=MIN(LL-II+1,&LRECL-INCOL+1); INPUT @INCOL PART $VARYING&LRECL.. LLL @@; *-----concatenate that into STG-----*; SUBSTR(STG,II,LLL)=PART; INCOL+LLL; *-----if at end of record, see if next record is a fence-----*; IF INCOL>&LRECL THEN DO; INPUT; NRECS+1; INCOL=1; END; *-----increment data-read count and continue-----*; II+LLL; END; *-----if the data are supposed to be in hex...-----*; IF HEX THEN DO; *-----verify only hex characters used-----*; IF VERIFY(SUBSTR(STG,1,LL),'0123456789ABCDEF') THEN DO; FILE PRINT; PUT 'ERROR: invalid hex characters in string at about ' 'input record ' NRECS ':' STG; ABORT; END; *-----informat the string using $HEX-----*; STG=INPUT(SUBSTR(STG,1,LL),$HEX160.); END; RETURN; *-----write out the resulting string to the output file-----*; PUTDATA:; FILE &OUTFILEX &DCB; IF L=0 THEN PUT; ELSE PUT STG $VARYING80. L @@; FILE LOG; RETURN; RUN; *-----we are done with the dump file and can clear it now------------*; FILENAME REORDER CLEAR; &DELREORD; /*------------------------------------------------------------------*/ /* We have to do some final things on a host-specific basis. */ /* For MVS, if we are dealing with a PDSCOPY unload file, we need */ /* to run PROC PDSCOPY to convert the unload file into a real */ /* load module PDS. For CMS, we must rename the TEMP DATA A file */ /* to the desired LOADLIB file name. This is somewhat tricky on */ /* CMS because if we don't have the fileid, we have to obtain it */ /* by redirecting the SAS log of a DATA step that accesses the */ /* fileref, since the DATA step reports the fileid. */ /* To deal with all this mess, we create a fileref of RDL00004 to */ /* refer to a temporary file RDL00004 RDLLOG A. We run PROC PRINTTO */ /* to redirect the SAS log to that file, then run a DATA step that */ /* references the output file. We then run PROC PRINTTO to stop the */ /* redirection, then read RDL00004 RDLLOG A to get the fileid as */ /* specified by the DATA step. We then can issue a COPY command to */ /* copy TEMP DATA A to the proper fileid, then we ERASE TEMP DATA */ /* A. Note that we can't issue a RENAME because the final LOADLIB */ /* file may reside on a different disk. */ /*------------------------------------------------------------------*/ %IF "&SYSSCP."="OS" & "&IEBCOPY." NE "YES" %THEN %DO; *-----for MVS, convert to a load module library via PDSCOPY----------*; PROC PDSCOPY IN=SEQFILE OUT=&OUTFILE INTAPE; RUN; FILENAME SEQFILE CLEAR; %END; %IF "&SYSSCP."="CMS" %THEN %DO; *-----for CMS, get fileid and issue COPY and ERASE-------------------*; FILENAME RDL00004 'RDL00004 RDLLOG A'; PROC PRINTTO LOG=RDL00004 NEW; RUN; DATA _NULL_; INFILE &OUTFILE; STOP; RUN; PROC PRINTTO; RUN; DATA _NULL_; INFILE RDL00004 LENGTH=L; INPUT @; INPUT @1 LINE $VARYING133. L; LINE=UPCASE(LINE); IF INDEX(LINE,'FILENAME='); FILENAME=SCAN(LINE,2,'=,'); RC = CMS('COPY TEMP DATA A '!!FILENAME!!'(REPLACE'); RC = CMS('ERASE TEMP DATA A'); STOP; RUN; FILENAME RDL00003 CLEAR; FILENAME RDL00004 CLEAR; X 'ERASE RDL00004 RDLLOG A'; %END; *------put trailer with support information-----*; DATA _NULL_; FILE PRINT; PUT 'End of TLKTLOAD processing'; PUT //; PUT 'For questions/comments about this SAS program,'; PUT 'contact:'; PUT // 'Rick Langston'; PUT 'Product Manager, SAS/TOOLKIT Software'; PUT 'SAS Institute Inc.'; PUT 'SAS Campus Drive'; PUT 'Cary, NC 27513'; PUT '919/677-8000 X7613'; PUT 'EMAIL: sasrdl@unx.sas.com'; %MEND TLKTLOAD;