Reuseable entry screen/processing...

[Scott Walker]

Brian said:

I wouldn't say @wef/@wlf don't work from a call table. I would say they 
don't work as might be expected. Further, I'll say they don't work as I 
expect or desire.

I recently discovered the hard way that if you have a screen command in
a 
call table, and the screen has real fields in it, and there are
@wef/@wlf 
for those fields in the parent process, then when the call table reaches
the 
screen command, the processing jumps back to the parent table to run the

@wef and @wlf, ... And Stays There! The next line of code executed after
the 
first @wlf of the screen field, is the line after the call command in
the 
parent table! This call table did not have any @wef/@wlf in the call
table 
itself by the way. I wasn't surprised that the wrong @w* ran, I was 
surprised that any @w* ran, and even more surprised that my call
processing 
was not proceeding past that screen statement.

The available work-arounds as I see them are to redraw the screen with 
dummies and use the dummies in the call table and ensure that no
@wef/@wlf 
for those dummies exist in the parent. That should stop you from popping
out 
of the call back to the parent before you wanted, but the other half of
the 
battle, I haven't tested if putting @wef/@wlf-dummy in the call table
works 
or not. Or do everything with other forms of input.

Most of my call tables don't happen to collect input via screen fields
so I 
don't have a lot of help, just that gotcha I just happened to hit
yesterday. 
I don't necessarily mean to suggest that this is a bug or even contrary
to 
the documentation. I do feel that this is one of those many little
things 
that aren't spelled out explicitly in the docs, and should be, because
in my 
opinion this isn't something that can be deduced from them as they are,
and 
is something a developer needs to know. I have no strong opinion on
whether 
the behavior should be changed, or if changed, in what way.



Brian,

That sure does sound like a bug to me, or at the very least a design
issue where the implementation of call was not thought all the way
through.  I'm also surprised that it even does the WEF/WLF stuff but it
makes sense if think of the called process sort of like a subroutine.
But I can't come up with a logical explanation for it to be back in the
main prc table as the next step after the WEF/WLF is finished.  

Regards,

Scott