You wrote:
>
> 1. DBSTAT2 still exists. However, you'll get a System Abort 2559
> when using any of the "global options, such as AUTO or ALL.
> It should give you the information you need when you enter
> one dataset at a time.
System Abort 2559 is "A compatibility mode sudden death has occurred."
Yesterday, Günter Kuhn send me a copy of DBSTATS (which I believe was the
precursor to DBSTAT2), along with the FORTRAN source (about 200 lines of
code - dated 1985). He warned me that IMAGE was a different beast back then
and didn't know how it would react to databases with more than 99 datasets,
or jumbo datasets, or whatever.
I tried it on our problem dataset, and it did an excellent job of
illustrating the problem and led us to a simple solution. I've taken a brief
look at the code, and although my FORTRAN isn't what it used to be, I don't
see anything in the code that could cause a System Abort. It opens the
database in mode 5, does one DBINFO call to get some dataset information and
then does a series of DBGET mode 2's which checks word 5 of the status array
to determine if the record is a primary or a secondary. As far as I can
tell, it should work on current Image databases. It also doesn't have the
global options (AUTO,ALL) that you mention so I'm hoping that it is safe to
run.
The one thing that concerns me is that the compiled program requires
Privileged Mode. From the source, I don't see any reason why it should
require PM (Unless DBINFO used to require PM in olden days).
Do you know if DBSTAT2 (which also requires PM) works the same way, or is it
using PM to do something fancy (and dangerous) with the root file?
The version received yesterday seems so simple and useful that I wonder why
it has never been compiled in native mode, and widely distributed in all the
usual places.
* To join/leave the list, search archives, change list settings, *
* etc., please visit http://raven.utc.edu/archives/hp3000-l.html *
|