Re: CCSID - change from 65535 to 37...issues? stories?

On 9/18/2014 3:54 PM, John Yeung wrote:

On Thu, Sep 18, 2014 at 2:28 PM, Gqcy <gmufasa01@xxxxxxxxx> wrote:

about the only problem we had was
that when we changed our Source physical files, it changed the member change
date (like you would see in PDM).
some of our programmers depended on that to denote that a program change was
made...

Well, if you're using PDM, and don't have any version control
software, that date *is* pretty handy.

It would be handy if it could be trusted. After a machine upgrade that
required a save on the old machine and a restore on the new one, the
'last changed' dates for every filt and member were set to the date of
the restore.

source PFs are CCSID 37 but some are 65535. I tried to do CHGSRCPF on
one of the 65535 ones, and it told me the file was changed, didn't
give me any error messages, but the CCSID remained stuck at 65535...

I got very tired of the CCSID(65535) issues, so years ago I set my own
user profile to CCSID(37) as I am in the US. After some time had passed
and the universe didn't implode, I changed the system value during an
IPL. I never noticed any problems that arose from this, but I also
don't recall actively changing any existing files' CCSIDs. The only
thing I changes was the system value.

To try to verify this, I looked at my production QRPGSRC. I myself
always convert to QRPGLESRC, but there are still one or two here who
program in RPG/400 style. This file is marked as being created in 1986,
yet it has a CCSID of 37.

I have no idea how this happened. Was it during a system migration? An
OS upgrade? Beats me. I know there's no way I would have done a
CHGSRCPF on it unless it was en masse, but I have member change dates
(reported by PDM F14) all over the decades - all the way back to 1986.

Beats me. I though I'd be able to help but I'm more puzzled than when I
began to look at this.
--buck