Jerry Fochtman wrote:

>Wait a minute, I think you need to re-read my post....

>I did not indicate that I had a problem with the support from the Lab.
>Indeed it was quite the opposite!  It was the lab themselves that were
>interested in understanding more about the issue, such that several
>took time from their busy schedules to look into the issue.

>The roadblock I experienced was one/more engineers 'between' the RC
>and the lab.  These are not, I repeat, the same lab team that did the
>work on large files.  So apparently you do not understand the different
>between the lab, response center engineers and the offline/expert
>center folks and how these engineers inter-relate and their roles.

I do understand the difference.  When the call was open, emails were
exchanged betteen myself, the EC which did not understand, and eventually
the lab, and the lab said they could not reproduce it - end of story.

>Furthermore, the issue with obtaining/writing to mapped pointers
>obtain through means other than HPFOPEN existed prior to large files.
>In fact, it is a re-creatable problem on at least 5.5pp7, 6.0 and 6.0pp1.
>And I also have it from other sources that its been observed on other
>releases as well.  So blaming it all on the work done by the lab to
>provide mapped files is incorrect.  It is quite easy to recreate the
>problem you've also experienced using PM on systems prior to 6.5.

This information would have been most helpful at the time, but it
does nothing to invalidate my point.  It is not a good idea to
provide functionality, which cannot be used.  This is the real
cause of the problem being discussed here.


Ken Graham.