OK, further experimentation worked for me.
cd /tmp #go to temp directory
mkdir diff #new directory
cd diff #change to it
mkfifo file1 #input file #1
mkfifo file2 #input file #2
mkfifo output #diff output
cp "//'zos.input.file1'" file1 &
cp "//'zos.input.file2'" file2 &
diff file1 file2 >output &
tsocmd "OGET '/tmp/diff/output' VB255.TXT TEXT"
cd /tmp
rm -rf diff #get rid of junk
I actually tried this. I couldn't generalize it because the OGET
command requires a fully resolved path. No use of ~/diff to say
relative to ${HOME}. I preallocated userid.VB255.TXT to VB/255.
On Tue, Jan 8, 2013 at 12:16 PM, Thomas Conley
<[email protected]> wrote:
> On 1/8/2013 12:13 PM, Paul Gilmartin wrote:
>>
>> On Tue, 8 Jan 2013 12:00:49 -0500, Thomas Conley wrote:
>>>
>>>
>>> I've seen this behavior myself, even on smaller files. It stops
>>> recognizing that lines are the same and just has huge blocks as inserts
>>> and deletes, even though the lines are the same. SUPERC support was
>>> unimpressed, I got the WAD.
>>>
>> I'm curious: How does this compare with the behavior of "diff", both
>> on z/OS and on other UNICES. I've not had the bad experience.
>>
>> How small is "smaller"?
>>
>
> Gil,
>
> Smaller is often a 20-50 line PARMLIB or JCL member.
>
> Regards,
> Tom Conley
>
>
> ----------------------------------------------------------------------
> For IBM-MAIN subscribe / signoff / archive access instructions,
> send email to [email protected] with the message: INFO IBM-MAIN
--
Maranatha! <><
John McKown
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN
