Fw: how to create unsigned hexadecimal initial value

["Leif Svalgaard" <leif@xxxxxxxx>]

I'm reposting this, as I'm not sure my first posting got there.

----- Original Message -----
From: Leif Svalgaard <leif@leif.org>
To: <MI400@midrange.com>
Sent: Saturday, August 26, 2000 7:18 AM
Subject: Re: how to create unsigned hexadecimal initial value


> From: <rjd@us.ibm.com>
>
> > I think even the more recent examples are easy to explain in terms of the
> > definition I gave for H literals.
> > CPYNV  NN, -1;
> > AND(S) NN,H'0003'; /* Equivalent to AND(S) NN, 3; */
> > CVTHC MSG-TEXT(1:4), XX;
> > CALLI SHOW-MESSAGE, *, .SHOW-MESSAGE;    <==== shows  0300    wrong
> >
> > H'0003' represents the value 3 and has length 1 (unlike character
> > constants, the leading zeros are not significant).  Based on the
> > specification for AND(S), the shorter operand is padded on the right with
> > zeros.  So, the expected result of ANDing op1=x'FFFF' with op2=x'03' is
> > x'0300'.
>
> In my rewriting of your definition I said:
> "The length of an H literal is always 4 bytes".
> That does not explain the above behavior. You suggested the asymmetrical
> definition that for non-negative literals, leading zero-valued bytes be
> dropped
> in calculating the length. Although that would explain the behavior, I have
> a semantic problem with it, as it leads to integers of length 1 or 3 which
> are normally not defined (there is no BIN(1)). I also have a problem with
> the value H'00'. Is that a BIN(1) or a BIN(0) value?
>
> Finally, your definition does not explain:
>
> CPYNV  NN, -1;
> AND(S) NN,H'0123'; /* gives 2300, WRONG */
>
>

+---
| 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
+---