Address for FP Support

[Ray Scheel]

I wanted to make sure I'd sent a report of a bug or inconsistent behavior
(issue: selection sets formatted to utilize index scan select no records
when invoked with passed parameters) to the right address, since it has been
2 months since I sent the first report and nearly 4 weeks for a resubmission
and I have not heard anything back *AT ALL*.  No an acknowledgement the
letters were received, no requests for more information, nothing.  

Both messages were sent to 'fpsupport at fptech.com', one on May 27 and another
on July 1.  Is dead air the expected response for problems to which
undocumented workarounds exist for undocumented problems with advertised
features?  I used to think the references on the list to fpsupport being a
"black hole" were exaggerations based on the reactions of offense other
members took to those suggestions, but if well documented bug reports are
frequently given the silent treatment, I think I understand why the
frustration level runs pretty high.  

Yes, I figured out how to format the selection set so index scan didn't get
triggered and let the selection set scan through the whole file, and could
write a sort/selection table using "lookup -" and add that to the command
line invocation of this report to speed things up, but that still leaves
this incompatibility totally undocumented, and there is the question of
whether this problem with index scan only impacts passed parameters or all
system maintained variables.

Ray Scheel               ray.scheel at wsdtx.org 
Windham School District     Computer Services
Programmer                       936-291-5206 

> -----Original Message-----
> From: Ray Scheel 
> Sent: Thursday, July 01, 2004 4:21 PM
> To: 'fpsupport at fptech.com'
> Subject: RE: Bug report / undocumented behavior
> 
> 
> I reported this a little over a month ago and never got any 
> sort of reply acknowledging the report, so I was wondering if 
> it was received and the status.  I keep coming across 
> applications where plugging a value into a selection set on 
> the fly would be greatly enhanced if that was compatible with 
> index scan. 
> 
> Ray Scheel               ray.scheel at wsdtx.org 
> Windham School District     Computer Services
> Programmer                       936-291-5206 
> 
> 
> > -----Original Message-----
> > From: Ray Scheel 
> > Sent: Thursday, May 27, 2004 4:21 PM
> > To: 'fpsupport at fptech.com'
> > Subject: Bug report / undocumented behavior
> > 
> > 
> > When using @px in a selection set and PFIDX+ON, I've found 
> > that no records are selected in cases where the fields 
> > compared to a passed parameter match an index, though if you 
> > reverse the comparison (passed parameter to field) for the 
> > same criteria, the appropriate records are selected.  
> > 
> > 
> > With:
> > PFIDK=ON
> > 
> > For this structure:
> > ::::::::::::::
> > att_cla/map
> > ::::::::::::::
> > unitcd ; code of unit:  2:allup:
> > clanum ; number of class on unit:  5:.0:
> > year ; year of record:  4:.0:
> > month ; month of record:  2:allup:
> > clacd ; code of class: 11:allup:
> > syscd ; system code for class:  1:allup:
> > ...
> > 
> > Using this command line call:
> > appl/fp/rreport att_cla -fp yr_to_month -s yr_to_month -m yr 
> > -u -rx 03 -ry 2004
> >  
> > Selects nothing for a selection set configured like this:
> > ::::::::::::::
> > att_cla/sel.yr_to_month
> > ::::::::::::::
> > :4:eqf:@px
> > :3:eqf:@py
> > 
> > But does select records as expected for:
> > ::::::::::::::
> > att_cla/sel.yr_to_month
> > ::::::::::::::
> > :@px:eqf:4
> > :@py:eqf:3
> > 
> > There is an index.D is built on field 3 with a length of 17
> > 
> > Filepro: 5.0.00/5.0.11aR4
> > Linux tyler 2.4.20-6smp #1 SMP Thu Feb 27 09:59:40 EST 2003 
> > i686 i686 i386 GNU/Linux 
> > 
> > I'm not sure if the failure of index scan to utilize a passed 
> > value is intentional but undocumented or a bug.  I have 
> > figured out how to get the expected results, but at the very 
> > least, there needs to be a better explanation of how index 
> > scan works in the online docs with this caveat added.  As it 
> > is, this is a situation where it would be nice if index scan 
> > worked with the passed parameters to speed selections I run 
> > this way out of several large files that have indices built 
> > on the field criteria used for selections.
> > 
> > Thanks,
> > 
> > Ray Scheel