RE: OS/390 service discontinued 2004/09/30

As far as syntax is concerned, in place of OS/390, you are using os390. You could use zOS instead of z/OS to replace all of the os390 references.

Dan Phipps 
Infrastructure Management Consultant
ITO Global Service Delivery 
Xerox Business Services, LLC 
427 Cedar Ave
New Smyrna Beach, FL 32169

p 803-223-3770



-----Original Message-----
From: Karl Williamson [mailto:[email protected]] 
Sent: Tuesday, March 24, 2015 12:39 AM
To: Ray Mullins; Perl5 Porters; [email protected]
Subject: Re: OS/390 service discontinued 2004/09/30

On 03/23/2015 06:11 PM, Ray Mullins wrote:
> What would it take to have documentation, etc., updated to reflect the 
> current name of z/OS?

Most of the documentation doesn't mention os390.  z/OS is not a syntactically legal identifier name, so some things need to remain using os390.  But it should be fairly easy to change most of things to z/OS. 
Do you have places you see where there should be changes?
>
> Also, even though there are no longer any users, what is the 
> BS2000/OSD status? It is also EBCDIC.

We generate tables for it, so theoretically it should be quite easy to get this to work.  If there are no users, can we stop generating tables for it?

The only other EBCDIC platforms still
> outstanding are the Unisys OS2200 operating system and IBM's System i 
> (formerly known as AS/400). I'm pretty sure that Perl was never ported 
> to AS/400,

It indeed was purportedly ported to AS/400, and we generate tables for CP 037.  So again, it should be "easy" to get it to work on this.  Do you know of anyone who has access to one, and wants modern Perl on it?

  and I believe it was also never ported to OS2000.

The only EBCDIC tables we generate are for CP 1047, 037, and what was called POSIX BC (for the Siemens).  It should now be trivial to generate tables for any other code page.  There however may be glitches that show up.

One thing that I discovered, and have not committed documentation for is that the official Unicode UTF-EBCDIC is flawed for any other code page besides 1047, wrt Perl anyway.  The reason is that Perl expects the characters that have ASCII equivalents to be the same whether encoded in UTF-EBCDIC or not, and official UTF-EBCDIC is that way only for 1047. 
This caused segfaults on other computers; it doesn't appear from the trouble tickets that people really fully understood the causes and implications of this.  In 5.22, perl will use a variant of UTF-EBCDIC customized for the code page for the system, should there be any system other than z/OS.  To cite an example, Perl wants a '[' character to have the same single-byte representation when it sees it in a UTF-EBCDIC string or just a regular EBCDIC string.

(Unisys MCP
> is ASCII.)
>
> On 2015-03-23 12:22, Karl Nordstrom wrote:
>> /http://www-03.ibm.com/systems/z/os/zos/support/zos_eos_dates.html/
>> /
>> /
>> /OS/390 V2/
>> Program
>> number     VRM     Announced     Available     Marketing
>> withdrawn     Service discontinued
>> 5647-A01     2.10.0     2000/05/16     2000/09/29 2002/12/17 2004/09/30
>> 5647-A01     2.09.0     2000/02/29     2000/03/31 2000/09/14 2003/03/31
>> 5647-A01     2.08.0     1999/02/22     1999/09/24 2000/03/16 2002/09/30
>> 5647-A01     2.07.0     1998/08/18     1999/03/26 1999/09/09 2002/03/31
>> 5647-A01     2.06.0     1998/02/24     1998/09/25 2000/06/15 2002/03/31
>> 5647-A01     2.05.0     1997/09/09     1998/03/27 1998/09/03 2001/03/31
>> 5647-A01     2.04.0     1997/09/09     1997/09/26 1998/03/12 2001/03/31
>>
>>
>>
>
>