I did write a TAR to unpack some files some twenty years ago.  If I
can find it, I'll put it on the pipelines home page.

   j.

2009/10/22 Bob Cronin <[email protected]>:
> To my knowledge, INMR123 is a PIPE sample/example filter, I don't think it
> is used by SENDFILE, but I admit I haven't looked at the SENDFILE logic in 8
> or 10 years (back when I wrote my *LIST-aware version, SENDLIST), so I
> suppose it might be using Pipelines now. In any event, I use INMR123 (with a
> few private tweaks to handle some degenerate cases that I've seen internally
> that The Piper didn't think were worth updating the sample/example for) in
> our Xagent email gateway (which manages the interchange of all IBM internal
> mainframe-format email with mail clients expecting mail to have been
> composed according to the tenets of RFC2822 and its MIME cousins).
> --
> bc
>
> On Wed, Oct 21, 2009 at 5:50 PM, Paul Gilmartin <[email protected]>wrote:
>
>> On 10/21/09 15:34, Bob Cronin wrote:
>>
>>> See the INMR123 REXX for creating NetData headers. You can set whatever
>>> date/times you like when you create them, then simply RECEIVE them.
>>> --
>>>
>>>  Wow!  so SENDFILE now uses Pipelines.  I'm soooo far behind the
>> times.
>>
>> But I notice that SENDFILE goes to considerable trouble to
>> save and restore the state of 00D, which can never be done
>> perfectly (what if the caller has his punch spooled CONT?)
>>
>> But why bother?  Since the PUNCH stage has the devaddr option,
>> it would be minimally disruptive to CP DEFINE a brand new
>> punch, use that, and CP DETACH it when done.  (I've done
>> this on occasion, long ago.)  Simpler is better.
>>
>> -- gil
>>
>

Reply via email to