On 1/7/2016 2:37 PM, John Yeung wrote:
On Thu, Jan 7, 2016 at 2:01 PM, Buck Calabro <kc2hiz@xxxxxxxxx> wrote:
On 1/7/2016 1:26 PM, Ken Killian wrote:
Yup, I had to "SCAN" those 72000 characters... <frown>
And I should have said "Close to the end", not the "very end"... <grin>
Ouch. Thanks for sharing that, I appreciate your time. Eyeballing
72000 characters - I don't think that's a skill I could ever master.
I...my eyes just can't go through that much stuff without my attention
wandering.
Well, 72000 characters amounts to 900 lines at 80 characters per line.
Which isn't insurmountable. You know there are folks who will manually
scroll through source code longer than that.
My last post on this, I promise :-)
Imagine this scenario: I've managed to suss out that my code is falling
over on .org addresses, but the first one doesn't occur in the data for
many thousands of rows.
I can't set a conditional breakpoint that will stop when it finds .org
anywhere inside my email address - there's no %scan builtin (green or GUI).
I'm absolutely not going to be setting a breakpoint at the top of the
code and whacking F8 until my eyeball finally sees a candidate email
address.
With the current debugger, I'm going to write some code that will do the
%scan and then set a breakpoint inside that 'dummy' if...endif
structure. Or, I'm going to break when the index is zero, or, or, or.
Anything other than manually eyeball it and hope I catch it as it goes by.
But the issue under discussion is having to scan horizontally. If there
was a variable full of XML data I just know I'd get lost eyeballing it.
One workaround would be to use Windows copy to grab all 72k, paste it
into an editor (like the Eclipse XML editor!) and use search facilities
to look at it. Maybe to alter the data and paste it back into the
program variable to test my hypothesis. But I can't do that because I
can't /see/ all 72k bytes.
And if I had more than one row, I almost certainly would use the same
technique I've always used: insert some debugging stub code into the
program and have the code do the scanning that I wish the debugger could
do. I've been known to condition debugging code like that with a data
area so that I could turn it off and on at will.
I was sort of hoping that Ken had a different take on what to do with
those 72k bytes; something that I in my rut hadn't thought of.
midrange.com
