My personal feeling is that getValueExpression should NOT return null if explictly called with an empty String. However, this is something for the Unified EL (ie. JSP) specification. If this behaviour is unclear from the spec, it should be explicitly defined there.
--Manfred On 12/21/07, Matthias Wessendorf <[EMAIL PROTECTED]> wrote: > Inside of Trinidad's EditableValueRenderer > (org.apache.myfaces.trinidadinternal.renderkit.core.xhtml.EditableValueRenderer) > > there getReadOnly(...), which is called by all inputXyz Renderers to > check if the component should be rendered "readOnly" or not > That method checks if the component is readOnly AND it also checks if > the underlying EL is readOnly. > > If that method returns TRUE, the component is readOnly (and there for > a inputText isn't editable)...makes sense, so far... > > Here is a "use-case" > > <tr:inputText label="Label a)" value=""/> > (yes, may be stupid, but can happen...) > > Since the getReadOnly() checks if the EL is readOnly... it does this as well: > > ValueExpression ve = getValueExpression(bean); > > > In Jetty (jetty-6.1.2rc2), which uses Sun/Glassfish/RI EL > (com.sun.el.ValueExpressionImpl) > ==> > It returns an object that is readOnly (which is correct) and the > getExpressionString is "" (empty). > > In tomcat 6. which uses this EL-Impl > "org.apache.jasper.el.JspValueExpression", it returns NULL > > > Now, I wonder what the correct EL behavior is. > I tend to think, that Tomcat is right, because there is no > ExpressionString, so not a "real" Expression, > but others could have a different understanding of it. > > What is your take on that ? > > -- > Matthias Wessendorf > > further stuff: > blog: http://matthiaswessendorf.wordpress.com/ > sessions: http://www.slideshare.net/mwessendorf > mail: matzew-at-apache-dot-org > -- http://www.irian.at Your JSF powerhouse - JSF Consulting, Development and Courses in English and German Professional Support for Apache MyFaces
