Re: Weirdness with EXTDESC, EXTFILE, EXTMBR, and a CPF5152

James,

FWIW, I wrote some tools years ago that worked on source members. I had similar problems with externally defined files -- it seems that not all srcpfs have exactly the same record format. I don't know if this was something that changed over the years or what...?

But, I found that program-described access worked perfectly every time, and now whenever I do stuff like this, I always use program-described access.

And, unlike databases, I haven't found that external definitions help much in a source file... you never add/change/remove fields, et al, so why bother with external definitions??


On 6/19/2014 1:02 PM, James H. H. Lampert wrote:

On 6/16/14 3:00 PM, Barbara Morris wrote:

For the case of writing to source files, using a program-described file
would probably be easier.

chg FMIMEBODY O F 92 DISK USROPN
F EXTFILE('QTEMP/MIMEBODY')
F EXTMBR('DFT')
add D mimeDs ds 92
add D seq 6a inz('000000')
add D date 6a inz('000000')
add D srcdta 80a
...
C EVAL SRCDTA = 'Sch...'
chg C WRITE MIMEBODY mimeDs

I plugged your suggestions into a copy of the source (with the actual
call to MMAIL/EMLMSG disabled) and tried it.

It works beautifully, but I still don't understand why
externally-described mode failed, or especially, why it threw a CPF5152
on a write that actually DID make it into the file.

Regardless, given that satisfying the customer is more important than
understanding the cause, I'm going to plug your changes into the live
version after lunch.

--
JHHL