Two EO questions (dynamic attrs and validation)

Question

Hello all (I really hope Steve M gets to see this),
I am re-visiting an old project (fantasy) of mine to implement spell checking of certain fields in an Entity Object, The concept is that I will implement the spell checking in an overriden validateEntity method. The way I would like this to work is for spelling errors to be reported to the user when they try to commit; if the user commits again without making any changes, I will assume they want to ignore the spelling errors and will allow the Entity to commit. In order to implement this in a really cooI and generic way, Ineed to be able to do a couple of things that have really stumped me.
The first is in relation to entity validation. The doc states that any EO is assumed to be valid after it is read from the database. Is there any way that I can alter this behaviour so that the EO needs to be re-validated even if it has just been read from the database? Extra credit if this can be done in such a fashion as to not cause the EO's data to be re-posted to the database if the user doesn't make any changes. I have tried over-riding lots of things in the EntityImpl and EntityDefImpl, but I cannot find the right place to do this.
The second thing I would like to do is to add dynamic attributes to an EO instance. The reason I want to do this is so that I can add the spell check code in a generic framework extension class and not have to modify each EO for this behaviour. The attribute I want to add is used to track whether spell checking has already been done on the EO row. I cannot use a java class member variable, as the same instance of the java class is not necessarily used across different web requests - in fact, I have found that the same instance is rarely-to-never re-used in my testing. I can make this work by adding a transient attribute to the EO itself, but this defeats my goal of a generic implementation. I have found some methods in EntityImpl (setDynamicAttributeValue and getDynamicAttributeValue), but they don't appear to do what I want. I've also played around a bit with the EntityDefImpl, but no luck there either.
I do promise to post a blog entry detailing how to do this if this pans out. We have looked at some nice JavaScript-based spell checkers that work well (give suggestions in a pop-up window), but we would like multiple fields on a web page to be spell-checked, and this is difficult to do with those spell-checkers (particularly when you have an af:table with varying number rows).
Any assistance is appreciated,
John

Answer 1

Your mechanism to manage css style classes is a good approach; I have used that many times. I do wonder why the style classes were implemented as a list, instead of a set, but there may be good use cases for using a list.
In some cases you can also consider using CSS PsuedoClasses, which were introduced in JavaFX 8. These are a bit easier to use, especially if you have only two options. But a use case might look like:
public class Message {
    public enum Status { NORMAL, WARNING, CRITICAL }
    private final ObjectProperty<Status> status = new SimpleObjectProperty<>(Status.NORMAL);
    private final StringProperty message = new SimpleStringProperty();
    // constructor, getters, setters, and property accessors....
public Label createLabel(Message message) {
    PseudoClass warning = PseudoClass.getPseudoClass("warning");
    PseudoClass critical = PseudoClass.getPseudoClass("critical");
    Label label = new Label();
    label.textProperty().bind(message.messageProperty());
    message.statusProperty().addListener((obs, oldStatus, newStatus) -> {
        label.pseudoClassStateChanged(warning, newStatus == Message.Status.WARNING);
        label.pseudoClassStateChanged(critical, newStatus == Message.Status.CRITICAL);
    return label ;
And then your css looks like
.label:warning {
    -fx-text-fill: orange ;
.label:critical {
    -fx-text-fill: red ;