RE: program won't let go of file

[David FOXWELL <David.FOXWELL@xxxxxxxxx>]

> -----Message d'origine-----
> [mailto:rpg400-l-bounces@xxxxxxxxxxxx] De la part de Mark S. Waterbury
>
> DFTACTGRP solution
> =================
> Assuming all of these *PGMs are in fact running in the 
> *DFTACTGRP, then you can just issue:
>
>      RCLRSC    LVL(*)
>
> in PGM1(CLP), right after the CALL to PGM2 returns to PGM1.  
> This will close all open files and end all active programs 
> below the current level in the call stack.
>
> NOTE: if somehow PGM4 or PGM5 were not compiled with CRTBNDRPG ... 
> ACTGRP(*DFTACTGRP) then this will not work. :-o

Thanks Mark. Pgms 4 and 5 are compiled to run in ACTGRP(*CALLER). PGM1 issues a RCLRSC at the end. I tried adding this after the call to PGM2 but it didn't work.

PGM1 does an OVRDBF CLIENTFILE to qtemp/clientfile, then returns to a menu. Another menu option will take me to PGM1b which will also call PGM5 along the line. At this point, although the OVRDBF CLIENTFILE is no longer active, qtemp/clientfile remains open and PGM5 continues to use it instead of CLIENTFILE which is in the *CURLIB.

Any thoughts?

 
> Named Activation Group solution
> =========================
> Another alternative would be to convert all of the called OPM 
> programs to ILE -- CLP to CLLE, and RPGIII to RPGLE, and 
> create PGM2 with a named activation group, e.g. ACTGRP(PGM2) 
> ... and create all the programs it calls with 
> ACTGRP(*CALLER).  Then, when PGM2 returns to PGM1, PGM1 issues:
>
>      RCLACTGRP ACTGRP(PGM2)
>
> This is one of the main benefits of using ILE with named 
> activation groups -- you can control termination of a named 
> activation group far more easily than with RCLRSC in "OPM 
> compatibility mode" (in the *DFTACTGRP).

I'll save this for later. But we need a fix now!