Compile *PRV

Compile *PRV

Post by Melk » Sun, 17 Dec 1995 04:00:00




watson (The TEAMIBM Network) writes:
:>
:>Luis' comments re RPG and CL also apply to COBOL. In order for *PRV to
:>work, you must install the previous release support for the compiler.
:>
:>He's wrong in one aspect, you _will_ be loading another compiler. The
:>*PRV release support is enabled by loading a copy of the "old" compiler
:>which has been recompiled to run on the new release. The CPP decides
:>compiler to invoke based on the release level requested.
:>
:>Jon Paris - AS/400 AD Market Support

The *PRV support for CL installs a set of commands and files from the
previous release(s).  The compiler itself is already present on the system.


...http://www.millcomm.com/~melkor
...Red Rider  -  CBR600F2  -  "A life without risk is no life at all"
...I speak for myself since no one else will let me speak for them

 
 
 

1. ILE dates and TGTRLS(*PRV)

I have an interesting scenario..

My company has two machines: a 530 RISC machine at V3R6, and a F45 at
V3R1.
We currently compile on the 530 and send program objects to the F45. All
of
our objects are compiled with TGTRLS(*PRV). When you look at the
programs
with DSPPGM, they say the earliest release that the program can run on
is
V2R3.

Dates in all our files are in CYYMMDD format. I had thought that *CYMD
support was in in V3R6, but when I attemped to compile the program on
the 530
using *CYMD in factor 1 of a TEST(D) in RPG IV, I got a compiler error.
So I
switched to Plan B:

1. Move the 7-digit date to an 8-digit (numeric) work date.
2. Add 19,000,000 to it.
3. *ISO  TEST(D)    date      95
4. IF NOT *IN95 then move the date to a native work date field, and
depending
on the system's date format (we have a machine where the dates are in
DD/MM/YY format, and so just about every program we have calls an API
we've
written to see what the format is) move to a 6-digit numeric for
printing
with edit code Y.

The stange thing is, that the object on the 530 performed as expected --
but
on the F45, with the SAME OBJECT (converted back to CISC, presumably,
but not
recompiled) the dates show up as zeros. Since I don't get a runtime
error,
I'm guessing that the TEST(D) opcode is setting on *IN95 -- but why only
on
the CISC machine? Anybody have any ideas??

--

--------------------------------------------------------
Opinions are my own. Otherwise, they are someone else's.

2. XHD Super Floppies

3. *PRV and SAVLIB?

4. PODC93 -- Program and Registration

5. Determine which objects could be saved with tgtrls(*prv)

6. unable to get static tunnel and VPN Client 3.6 to work on the same PIX

7. Customising PDM compile options

8. Novice in Need

9. Can't compile C program (CPA include files problem)

10. NEVER MIND READING COMPILE

11. How to compile all program of once

12. CRTBNDRPG parms to compile IFS examples

13. reading the compile