CA879-EXIT.<br />EXIT.<br />CA899-EXIT.<br />EXIT.<br />* * * DEFAULT ALGORITHM %SKEYBLD ENDS HERE<br />```** <br />The entire CA800 routine is a Default for the insertion point named %SKEYBLD. Its purpose is to build the key value in (cont.)<br />screen format from the Master key field in the record. It is performed from the NXT Functions. In most cases you would (cont.)<br />not want to touch this routine. In the unlikely event that you need to display the key somehow other than it would be (cont.)<br />built by this routine then you can override the entire CA800 routine with your own.<br />## <br />Service Routines<br />The CA600, CA900, CB200, CB500, CB600, CB900, DA100, DA200, DA300, FA100, FA400, and FA500 routines are not shown here (cont.)<br />because they are used "internally" by the SCN and FND Functions and by the Pattern Editing. You cannot alter them. The (cont.)<br />CC100 routine, also not shown, is used for the VERZUN function which displays statistics for your application. The (cont.)<br />DA400 routine (not shown) is performed to spool report lines. These routines should never be altered by (cont.)<br />you.<br />The CB900-SPOOL routine consists of the insertion point %SPOOL. The Default logic for %SPOOL includes the library (cont.)<br />member named SPOOLIT. You can override this default spooling routine by coding customization for %SPOOL, if you (cont.)<br />wish.<br />| ```<br />**<br />JA100-LOGICAL-JOIN.<br />* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *<br />* LOGICALLY JOIN SECONDARY DATA FILES TO THE PRIMARY FILE *<br />* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *<br />MOVE TWA-DB-REQUEST TO TWA-DB-REQ-SAV.<br />* * * CUSTOM ALGORITHM %JOIN STARTS HERE<br />IF (NOT-FOUND) GO TO JA900-RETURN.<br />MOVE REDKY TO TWA-DB-CMD.<br />MOVE 'SIFK1' TO TWA-DB-KEY-NAME.<br />MOVE 'SIF01 TO TWA-ELT-LIST.<br />MOVE LOW-VALUES<br />TO SIF01-LOVALU.<br />MOVE VAC01-EMPNUM TO SIF01-EMPNUM.<br />MOVE SIF01-MASTER-KEY TO TWA-KEY-VALUE.<br />CALL 'MAGECSET' USING TWA-DB-AREA-A SIF01-ELEMENT.<br />PERFORM AA840-CALL-MAGEC-IO THRU AA899-EXIT.<br />IF (NOT-FOUND)<br />MOVE SPACES TO SIF01-ELEMENT<br />GO TO JA900-RETURN.<br />* * * CUSTOM ALGORITHM %JOIN ENDS HERE<br />JA900-RETURN.<br />MOVE TWA-DB-REQ-SAV TO TWA-DB-REQUEST.<br />MOVE THIS-PGMS-ELEMENTS TO TWA-ELT-LIST.<br />JA999-EXIT.<br />EXIT.<br />```** <br />This is the JA100 routine which is performed from every "read" routine immediately after reading the primary file (cont.)<br />record. It is where you can code the reads for as many other files as you like to construct a "logical record" for the (cont.)<br />MMP to process against.<br />The insertion point named %JOIN is an extremely important one and you will find it very useful. You can not only read (cont.)<br />other files but also compute values into VARIABLE-STORAGE fields which may be specified as "Source/Target" fields to be (cont.)<br />displayed on the screen.<br />Using this insertion point wisely can greatly simplify your MMP's logic and your work. The code shown in this example (cont.)<br />was generated in the Automatic Logical Join process during the development of this application.<br />The coding found in the %JOIN insertion point could be placed there in any of several ways. It could be manually coded (cont.)<br />by the developer as ordinary customization coding, it could be generated via the semi-automated development process (cont.)<br />which requires some programmer input, or it could be generated using the fully-automated development process which (cont.)<br />requires no programmer input. Regardless how it got there, it can be altered or added to online via the standard (cont.)<br />customization processes.<br />| ```<br />**<br />***************************************************<br />* MISCELLANEOUS *<br />* USER-CODED CUSTOM SUBROUTINES WILL BE PLACED *<br />* BELOW. *<br />* *<br />***************************************************<br />* * * DEFAULT ALGORITHM %SUBRTNL STARTS HERE<br />* * * NO STANDARD DEFAULT CODE FOR THIS INSERTION POINT<br />* * * DEFAULT ALGORITHM %SUBRTNL ENDS HERE<br />*<br />* * * DEFAULT ALGORITHM %SUBRTNM STARTS HERE<br />* * * NO STANDARD DEFAULT CODE FOR THIS INSERTION POINT<br />* * * DEFAULT ALGORITHM %SUBRTNM ENDS HERE<br />```** <br />**Above are two insertion points which are provided so that you can insert as much Customization logic as you like to (cont.)<br />be performed from anywhere else in the MMP. Here you might insert a routine to do a special edit on, let us say, dates. (cont.)<br />You might have several date fields on the screen and wish to edit each of them using this subroutine. The code you (cont.)<br />place here may be PERFORMed as many times as you wish from the edit routine (%EDIT) or anywhere else that is needed. (cont.)<br />The insertion point named %SUBRTNL is obsolete and supported for compatibility with older versions of MAGEC. You should (cont.)<br />use %SUBRTNM.<br />| ```<br />* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *<br />* ZZ999 THIS MUST BE THE LAST PARAGRAPH *<br />* IN THIS PROGRAM. IT WILL STOP RUNAWAY LOGIC (FALL-THRU) *<br />* IN THE USER-CODED ROUTINES AND ISSUE AN ERROR MESSAGE. *<br />* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *<br />ZZ999-STOP-RUNAWAY.<br />MOVE FALL-THRU-ERROR TO TWA-ERR-CODES.<br />MOVE F TO FATAL-ERR.<br />GO TO AA900-GOBACK.<br />```** <br />This routine is generated at the physical end of the program. It catches "runaway" logic which might attempt to "fall (cont.)<br />thru" the end of the program. Falling thru the end will always give bad results. In some environments it will execute (cont.)<br />the default GOBACK or STOP RUN which the Cobol compiler puts there - that will usually terminate the whole TP Monitor. (cont.)<br />This "safety net" catches the error before the damage is done.<br />One way you might create such an error would be to PERFORM a subroutine THRU the EXIT and, within the subroutine, GO around the EXIT. With very little effort you can discover other ways.<br />## <br />Appendix A -- Maintenance Mode Setting<br />The logic in the generated MMP which sets the MMP-NEW-TRANSACTION switch is depicted in the following flowchart.This (cont.)<br />switch has significant impact on the operation of the program. It indicates either that the program is in (cont.)<br />NEW-TRANSACTION-MODE or in CONTINUATION-MODE.<br />***<br />## <br />Appendix B -- Browse Mode Setting<br />The setting of the MMP-NEW-TRANSACTION switch in the Browse mainline logic determines whether the program initiates a (cont.)<br />new browse starting with the key value given in SKEY, or continues paging forward or backward from a browse initiated a (cont.)<br />prior transaction.<br /><BR CLEAR=LEFT><br />## <br />Appendix C --Automatic Editing<br />The editing and formatting of screen fields based upon their Edit types is mostly done in MAGEC's Control Program's (cont.)<br />"automatic edit" routine, prior to its passing control to the MMP. The exception to that is the pattern edits (Edit (cont.)<br />types "P" and "#"), which are handled within the MMP in the BB500-SCREEN-EDIT paragraph.<br />The Control Program must make a determination as to whether or not to execute the automatic edit routine based upon the (cont.)<br />function code and whether this will be a new transaction or a contunuation. Refer to "Appendix A" and "Appendix B" for (cont.)<br />a discussion of NEW-TRANSACTION-MODE and CONTINUATION-MODE.<br />Obviously, it would not be necessary to edit and validate screen data for a transaction which is merely an inquiry and (cont.)<br />will not be updating the files or database. The Control Program looks at the SFUNCT and TWA-LAST-FUNCT fields, and the (cont.)<br />SKEY and TWA-LAST-KEY fields before passing control to your MMP. If it determines that it is appropriate to edit the (cont.)<br />screen data, it executes the automatic edit routines which do several things:<br />edit (validate) the contents of each screen field based on the Edit type and associated parameters for each<br />translate lowercase alphabetics to uppercase for type "U" and "T" fields<br />set the error flag (SfieldE field) to space or "E" to indicate whether each field passed or failed the edits<br />fill in the numeric (Sfield-N) value field associated with each screen field having a numeric Edit type<br />normalize the contents of each numeric or date type field into the standard display format for its respective Edit type<br />fill in the DAY-OF-WEEK and JULIAN-DATE fields for all fields having a date Edit type<br />set the appropriate error number(s) into the TWA-ERR fields (occurs 6 times to allow for 6 error messages to be displayed at a time)<br />set the master error flag (FATAL-ERR) to "F" (ERROR-FOUND), or to space<br />fill in the values of all Edit type "W" fields associated with table edited (type "T") fields using the descriptions from the table entry on the TBL file<br />It should also be noted that the contents of the type "W" fields will again be filled in after the MMP exits, just (cont.)<br />before the screen is sent to the operator. This is done so that the MMP can alter the contents of the type "T" field (cont.)<br />and the description will change accordingly.<br />NOTE:<br />**<br />Since the determination of whether or not to execute the automatic editing routine is done prior to entering your MMP (cont.)<br />and is based on the function code (SFUNCT), changing the function code within your MMP's processing will not affect it. (cont.)<br />Thus, if your MMP was entered with a function code which did not indicate the need for automatic editing, but your (cont.)<br />program logic has altered it to another function code which does require editing, the editing will not have been done. (cont.)<br />This means that the numeric values (Sfield-N) will not have been filled in--likewise for all of the other processes (cont.)<br />described above.<br />If you wish to alter the function code and have the editing done, you should do a FTH-FUNCT using the new function code (cont.)<br />so as to give the Control Program a chance to see the new function code and act upon it. Refer to "Appendix B" of the (cont.)<br />Customization* section in your MAGEC Tutorials book. Remember too, the contents of TWA-LAST-FUNCT, SKEY, and (cont.)<br />TWA-LAST-KEY will affect the determination.<br />Normally only the ADD, CHG, and DUP functions require automatic editing. If you wish to indicate that editing is (cont.)<br />required for any other function you can do so on the FCD file definition for that function code. Refer to "Figure 3" in (cont.)<br />the *Security* section of this Programmer's Reference book.<br />Function codes ending in "ADD", "CHG", or "DUP" are considered to require automatic editing in their continuation modes (cont.)<br />regardless what is specified on the FCD definition for them. Any other function codes receive the editing only if you (cont.)<br />have specified "Y" for THIS FUNCTION WILL INVOKE AUTO EDITING (Y OR N): on their respective FCD (cont.)<br />definitions.<br />For purposes of determining whether or not to execute automatic editing for each transaction, the Control Program (cont.)<br />treats any function code having a "Y" specified (as described above) as if it were a "CHG" function. Refer to the prior (cont.)<br />appendices in this section for discussion of the determination logic.<br />Within your MMP updates to the files or database should only occur in CONTINUATION-MODE since no automatic editing will (cont.)<br />have been done for the NEW-TRANSACTION-MODE transaction. Normally NEW-TRANSACTION-MODE involves only the display of a (cont.)<br />screen containing data to be updated (or empty fields in which to type new data to be added). If you chose to add (cont.)<br />customization which does updates to files in NEW-TRANSACTION-MODE you will be responsible for all data validation. (cont.)<br />*Data integrity should always be the paramount consideration!*<br />If you wish to ignore some, or all, of the errors found by the automatic editing routine, you can. The automatic (cont.)<br />editing routine indicates in which fieds it found errors by setting the SfieldE field (in the Mask copybook) to "E", (cont.)<br />and by setting the master error flag and adding the appropriate error number to the list of TWA-ERR fields. You can, if (cont.)<br />you wish, interrogate any or all of these fields to determine if a certain error was found and then unset that error by (cont.)<br />removing (blanking) the error number from the list, resetting the SfieldE field to space, and/or resetting the master (cont.)<br />error flag to space. It would be best to do this early in the program logic, as in %PREINIT, if possible since having (cont.)<br />an error flagged might alter the path of the logic later on. An obvious illustration of this phenomenon is that any (cont.)<br />coding in the %EDIT2 insertion point will be entirely bypassed if the ERROR-FOUND condition is true. Also, the (cont.)<br />ERROR-FOUND condition triggers the Control Program to look up the error numbers indicated in the TWA-ERR's and display (cont.)<br />the associated error messages in SERRMSG before the screen is sent. Of course, the MMP will also bypass the file (cont.)<br />updates if ERROR-FOUND is true.