2-byte RLD entries are correctly handed by the link-editor/binder and
as far as I know they aren’t deprecated. Why would their use by Pl/I (to use
your example) perturb you?
On Sat, Jul 4, 2020 at 7:57 PM -0500, "Seymour J Metz" <[email protected]> wrote:
The default length of a Q-type constant is 4, although early PL/I compilers
used QL2. I don't know whether Enterprise PL/I still uses QL2; I hope not.
--
Shmuel (Seymour J.) Metz
http://mason.gmu.edu/~smetz3
________________________________________
From: IBM Mainframe Assembler List [[email protected]] on behalf
of Keven [[email protected]]
Sent: Saturday, July 4, 2020 8:35 PM
To: [email protected]
Subject: Re: z390 RLD Entries ... Bug or Feature?
Pretty sure that a 2-byte RLD entry results from a Q-Type AdCon and
often represents a CXD or PRV offset. That’s to say, they have their uses
still.Keven
On Sat, Jul 4, 2020 at 7:11 PM -0500, "Seymour J Metz" wrote:
> And of course use of a length-2 RLD is kind of questionable
! constants (external dummies).
> (unless a truncated value is what you are going after).
In the case of PL/I they assumed that there would never be truncation; the code
certainly wouldn't work correctly if the Q-constants were truncated.
--
Shmuel (Seymour J.) Metz
http://mason.gmu.edu/~smetz3
________________________________________
From: IBM Mainframe Assembler List [[email protected]] on behalf
of Peter Relson [[email protected]]
Sent: Saturday, July 4, 2020 9:23 AM
To: [email protected]
Subject: Re: z390 RLD Entries ... Bug or Feature?
I don't know anything about z390, but FWIW z/OS does not support
5/6/7-byte relocatable adcons (let alone use of 5 through 8 on ALn but
they are supported on ADLn). Use of ADL5/6/7 is detected by the binder
with the message:
IEW2353E 243F SECTION TEST CONTAINS INVALID DATA. ERROR CODE IS 250008.
where
250008 - RLD contains an invalid length field. Supported lengths are 2, 3,
4, and 8, where 8 is allowed only for type loader token.
<>
z/OS has never, to my knowledge, supported 1-byte relocatable adcons. And
of course use of a length-2 RLD is kind of questionable (unless a
truncated value is what you are going after).
Peter Relson
z/OS Core Technology Design
