Intake does not distinguish empty from absent fields

[Rodney Schneider <rl...@arcalink.com>]

Hi all,

One of my fellow developers (not subscribed to this list) sent me this bug
report (or question) regarding Intake...

---------------------------------------------------------------

If a browser submits a form with an empty text field (with, say, the parameter
name="foo"), then the request parameters will come in in the form
"?foo=&bar=..." etc., and both javax.servlet.HttpServletRequest and
org.apache.turbine.util.ParameterParser will report that the parameter "foo"
is present with the value "" (the empty string).  However,
org.apache.turbine.services.intake.model.Field will treat this field as if it
is not present at all: specifically, in the method validate() is the following
code:

            stringValue = pp.getString(getKey());
            if ( stringValue.length() == 0 )
            {
                set_flag = false;
            }

Thus, there is no way to distinguish between an empty field and an absent
field.  This is an issue where the field previously had been non-empty (say
in the corresponding Torque object), and had been explicitly emptied in the
form by the user.  This causes the following problems:

1.) org.apache.turbine.services.intake.model.Group.setProperties(Object obj) 
    will not set this field to empty in obj; this has to be (awkwardly) done
    by hand in the Action class.
2.) If the browser is returned to the form (say, another field was
    invalid), this field will no longer be empty, unless again it is explicitly
    emptied by the Action or Screen class.
 
And if the above explicit coding is done, then:

3.) We will have hard-coded behaviour in the Action redundantly replicating
    information in the intake group, and
4.) This Action cannot be reused by multiple forms: for instance, by
    another form which did not present this field, because the Action
    could not tell whether the field was present and set to empty, or not
    present.

Why has org.apache.turbine.services.intake.model.Field been coded this
way?  And is there another way of achieving what I'm trying to achieve
here?

---------------------------------------------------------------

I could generate a patch that removed the above-mentioned test
(and also the equivalent test for multi-valued fields), but I am
not sure if this will create other problems.

Any help would be greatly appreciated.

Thanks in advance,

-- Rodney

--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>