[ic] Long-lived co=1 bug

[Josh Lavin]

Quoting Grant (emailgrant at gmail.com):
> >>> > > > There has been a long-lived bug that doesn't allow something like this
> >>> > > > to work properly:
> >>> > > >
> >>> > > > [loop search="fi=orders/ml=999999/st=db/sf=shipment_notification_sent/se=1/op=ne/nu=0"]
> >>> > > > [loop-data orders order_number]
> >>> > > > [/loop]
> >>> > > >
> >>> > > > Even though the shipment_notification_sent field is set to 1 for every
> >>> > > > record in the table, this has a ton of output, maybe all the orders.
> >>> > > > Adding co=1 fixes it, but this isn't a coordinated search.
> >>> > > >
> >>> > > > I've reported this bug before, and someone told me it's a known bug.
> >>> > > > I'm using IC 5.2.  Is there any hope of this being fixed?
> >>> > >
> >>> > > At least it is weird behaviour and I don't like it as well. But I think we
> >>> > > need the expertise of Mike Heins who can tell us what would break if we
> >>> > > fix this somehow :-)
> >>> > >
> >>> > > Bye
> >>> > >         Racke
> >>>
> >>> I'm having a hard time pinning this down so I know when I need to use
> >>> co=1 for a non-coordinated search.  Does anyone know when that is
> >>> necessary?
> >>
> >> I think if the op= parameter is given.
> >>
> >> Bye
> >>        Racke
> >
> > About 5 1/2 years later, I think I've got this more nailed down.
> > Adding co=1 seems to be necessary unless there is only one sf/se block
> > and op=eq is in affect.  Is this really a bug, or am I
> > misunderstanding mv_coordinate?
> >
> > This document:
> >
> > http://docs.icdevgroup.org/cgi-bin/online/search/search_reference.html
> >
> > says:
> >
> > "the so-called "coordinated" search allows for multiple search options
> > to be stacked on top of each other"
> >
> > but it seems to be necessary even when there is only one sf/se block
> > unless using op=eq.
> >
> > - Grant
> 
> The following search returns products in the category "Widgets" and
> products in the category "Big Widgets":
> 
> [loop search="fi=products/st=db/sf=category/se=Widgets/op=eq/nu=0"][/loop]
> 
> If I add co=1, only "Widgets" are returned.  According to the
> documentation, co=1 "allows for multiple search options to be stacked
> on top of each other".  I reported some exceptions to this above, but
> now I'm thinking co=1 should be used in all searches?

Racke is right; you need co=1 if you are using "op":
http://interchange.rtfm.info/icdocs/Search_parameters.html#mv_column_op_op

-- 
Josh Lavin
Perusion -- Expert Interchange Consulting    http://www.perusion.com/