RE: how to create unsigned hexadecimal initial value

Okay.

From Gene's example, I conclude that the value in the type H constant is
copied right adjusted into the variable being initialized - more or less as
though:

DCL DD CONSTANT CHAR(1) INIT(X'0F');
DCL DD VAR      BIN(4);
CPYBRA VAR,CONSTANT;

Of course, if this was true in all elements of its description then Gene's
original example would not have failed.  So there is something special about
Gene's original case copied in here:
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
The following won't compile giving error CPD6416, "Number was either too
large or negative."

DCL DD FLAG AUTO BIN(4) UNSGND INIT(H"80000000");

The following compiles okay but run-time fails with MCH1210, "Receiver
value too small to hold result."

DCL DD FLAG AUTO BIN(4) UNSGND;
CPYNV FLAG, H"80000000";
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
I went back and reread the "create program API description.  It says:
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Integer

Integers define signed and unsigned binary scalar data values. The two forms
of integers are decimal and hexadecimal. The decimal form is a sequence of
digits optionally preceded by a sign. The hexadecimal form is a string of
hexadecimal digits delimited with apostrophes and preceded by an H. Neither
form may exceed the 4-byte limit on binary numbers. When the value of the
integer is between -4095 and +8191, the QPRCRTPG API converts the integer to
an immediate operand where it can.

Example

+123
-1
54788

H'0F0D'
H'0123'
H'5E2D1AB4'
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
I don't think that this description is saying the decimal integers are
signed and hex integers are unsigned.  In fact, I don't see how to create an
unsigned integer constant from this explanation - it appears to me that
using either a type X constant or a type H constant means that you have to
do some work for yourself.

If I got it right from Leif's tests, when a type H integer constant is used
to initialize a numeric constant its hex characters are copied into the
variable right adjusted.  When assigned to a character variable, the type H
integer constant is stripped of its leading zeros then copied into the
character variable left adjusted.  Same thing with the AND operations.

My neck is getting sore from looking at this stuff.  I think I now
understand that type H constants are integer values and that they behave in
interesting but understandable ways when used with numeric versus character
variables.  But I keep looking at all this evidence and I cannot see why the
CPU is throwing a runtime exception when you assign a large negative value
to a bin(4).

Gene, could you try your little demonstration below but make FLAG into an
unsigned number?  I'm wondering about the difference between run time versus
initialization time.

Richard Jackson
mailto:richardjackson@richardjackson.net
http://www.richardjacksonltd.com
Voice: 1 (303) 808-8058
Fax:   1 (303) 663-4325

-----Original Message-----
From: owner-mi400@midrange.com [mailto:owner-mi400@midrange.com]On
Behalf Of Gene_Gaunt/ReviewWorks@reviewworks.com
Sent: Friday, August 25, 2000 6:52 AM
To: MI400@midrange.com
Subject: Re: how to create unsigned hexadecimal initial value


Try this:

DCL DD FLAG(8) AUTO BIN(4) INIT(
   H"F",
   H"FF",
   H"FFF",
   H"FFFF",
   H"FFFFF",
   H"FFFFFF',
   H"FFFFFFF",
   H"FFFFFFFF");
BRK "LOOK";
PEND;

STRDBG
ADDBKP LOOK PGMVAR(FLAG)

then breakpoint displays:

1:   15
2:   255
3:   4095
4:   65535
5:   1048575
6:   16777215
7:   268435455
8:   -1


+---
| This is the MI Programmers Mailing List!
| To submit a new message, send your mail to MI400@midrange.com.
| To subscribe to this list send email to MI400-SUB@midrange.com.
| To unsubscribe from this list send email to MI400-UNSUB@midrange.com.
| Questions should be directed to the list owner/operator: dr2@cssas400.com
+---

+---
| This is the MI Programmers Mailing List!
| To submit a new message, send your mail to MI400@midrange.com.
| To subscribe to this list send email to MI400-SUB@midrange.com.
| To unsubscribe from this list send email to MI400-UNSUB@midrange.com.
| Questions should be directed to the list owner/operator: dr2@cssas400.com
+---