A really insidious browse bug...

[GCC Consulting]

> -----Original Message-----
> From: filepro-list-bounces+gccconsulting=comcast.net at lists.celestial.com
>
[mailto:filepro-list-bounces+gccconsulting=comcast.net at lists.celestial.com]
On
> Behalf Of John Esak
> Sent: Wednesday, December 03, 2008 8:25 PM
> To: support at fptech.com
> Cc: 'Lee Machan'; 'glenn.freeman'; 'filePro'; 'Timothy R Barr'
> Subject: Bug: A really insidious browse bug...
> 
> Well, remember I pointed out an obscure bug you could run into if you use
> the old Guru Magazine article I did a hundred years ago about building
> browses in variables and then concatenating the b= var & var & var
> method....
> 
> Well, I found something today that drove me completely nuts... For about
an
> hour.
> 
> If on the lookup line, which looks something like this...
> 
> Lookup alias=file k=(key) i=A -nl b=(b_cfg & b_hdr & b_data)
> 
> You mistakenly do this:
> 
> Lookup alias=file k=(key) i=A -nl b=(b_cfg & b_hdr & b_data)
> *56"
> 
> And I did this because I usually copy the original browse several times
and
> break it into the 3 vars.  It's just an easy way to change a browse that
is
> already written with the browse wizard.  Anyway, the *56"  was hung out
way,
> way past the 80th character on the screen... Meaning OUT OF SITE!!!!  And
> the syntax checker did NOT catch it!!!!!  (Really, it probably shouldn't
be
> looking past a complete lookup... I don't blame the syntax checker.... But
> still, it is something to look for if ever you run into unfigure-outable
> problems.)
> 
> Man, I typed in the lines of the browse over and over again... Thinking I
> had some wired invisible character or who knows.  The symptom was a
skinny,
> minny 1 character wide window... With the highlight bar visible but
nothing
> else of course... Ho much can you show in 1 character? :-)
> 
> So, for you Ray, at FpTech... I'm not sure this could be classified as a
> bug... Because FpTech doesn't suggest or maybe even support breaking the
> browse into variables... But you *have* to do it for those which need to
> take up the whole screen with lots of fields... Otherwise the browse
wizard
> truncates the 2nd line and doubles the quote... Something I have sent to
> tech support several times over 10 years ago and maybe even more recently.
> Obviously, not much you can do, but maybe Ken will want to run his parser
> *past* a completed browse for any "hangouts* which might be past the human
> visible edge of the screen.  Actually, I don't know what made me even look
> over there, I did by accident and ended the hour of torture. :-)
> 
> For the rest of you... Beware of this time waster!
> 
> Everyone have a nice beginning of December! I missed the back from
> Thanksgiving thread... Let's all hope the new year brings back the
economy.
> It's funny, gas is now down to $1.80 a gallon and it's almost unnoticed or
> unmentioned by anyone.... But boy it helps, huh?  Now, if I could only
sell
> an unwanted house!  :-)
> 
> John Esak


If memory serves me correctly, when the browse lookup was first released,
without the wizard, this was one of the techniques suggested.

In any event, )*56 should show an error as this is incorrect syntax for both
a regular lookup or a browse lookup.

I just checked the 5.6.0 manual.  And low and behold, the example for a
browse lookup is just as John used, the browse command broken down into
variables.  So, John, this problem is one which the syntax parser should
catch.

For those of you who don't use this method for a browse, is offers you some
flexibility in how information is displayed. You can change the order in
which fields are displayed or which fields are displayed.  You can even
change the index used based on which display variables are used.


Richard Kreiss
GCC Consulting
rkreiss at gccconsulting.net