Gerald Richter wrote:
Hi Neil,
I'm attaching the Embperl tar and the source file. You should
be able to duplicate what I am seeing using these... if not
then maybe reality is becoming even more distorted here in
America than I thought!
;-)
Let me know if/what you see...
I found the difference. You really used a "bug" in the 1.3 parser.
You wrote [$ hidden $] and I wrote [$hidden$]. The difference is the space
after the hidden. The 1.3 parser interprets this space as an first empty
argument. You can get the same result with [$ hidden , $] (with or without
spaces) in Embperl 2.
I think the way Embperl 2 behaves is more predictable and clean. I don't
think it makes sense to move this bug over to Embperl 2.
You can simple do a s/hidden/hidden ,/ in your code and things should
work with both versions.
Gerald
Hi Gerald,
Thinking about this some more, I would like to make one more plea to you about making %fdat work as
it did before. The reason is consistency and simplicity. As long as I've used Embperl, I've used
%fdat to handle the form variables being passed in, and to control the values going out in the
generated HTML. When I want to build a populated form, I set variables in %fdat, and these values
are also set up automatically for the script. It's really a very convenient and easy way to handle
this stuff. I believe that it's important to make this mechanism work consistently and in a way that
doesn't introduce needless additional code for application programmers.
For example, if I want to set a value for a form field (not hidden, but a real field) then I can
just set that value in %fdat, somewhere prior to the form definition. Then, later on, Embperl
automatically populates the value of the field accordingly. I believe that this is also the way it
is supposed to work in Embperl 2.0. This should, in my opinion, work in the same way for the
[$hidden$] meta command. It is highly unintuitive to require a comma after the hidden keyword, or to
require a '@ffld = keys %fdat' in order to transfer variables over. You never had to do that before,
so why should it be necessary now? Maybe there was a bug in the prior implementation of [$hidden$],
but that same bug surely did not also apply to the way we set non-hidden form variables. For those,
we simply set the variable value in %fdat. Making it different for hidden variables is just wrong,
to me.
Can't we just make %fdat the one-stop-shop for form variables, as I have always thought it was? I'm
not just saying this because all my code works that way - I'm making an impassioned plea for
simplicity as well as backward compatibility. I realize it's not the way you designed the new
version, but please consider this as being something that might be worth bending.
Of course, if there is a very good reason for it *not* to work this way (apart from ease of internal
implementation) then I'm all ears...
Thanks,
-Neil
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]