Meaning of empty from, empty create etc. in expression language - jsf-2

I came across some .xhtml files where for some components the expression language used was like rendered="#{empty from}", rendered="#{empty create}" etc.
I know that empty is an operator in EL, used to check whether a value is null or empty, but I did not understand the meaning of above mentioned ELs.
Can somebody explain to me what above EL's mean?

The rendered attribute is a signal whether JSF should generate HTML for the component or not. If it evaluates false, then it won't generate HTML for the component (nor for its children). The #{empty form} will evaluate false when the #{form} is not null nor empty.
Simple as that. You can find another examples of boolean expressions here: Conditionally displaying JSF components

Related

thymeleaf 3 and th:field syntaxe

I have a form template using thymeleaf 3 and I want to use th:field="*{dueDate}" syntax to bind my model property. I try to do the same as th:value="${#strings.substring(model.dueDate, 0, 19)}" but all my attempts fail.
It is possible to do this into a th:field instead of a th:value ?
EDIT:
I have to substring my date from java.sql.timestamp because it's a little bit different as the format used by my datepicker (2020-01-30 13:00:00.0 instead of 2020-01-30 13:00:00). I understand it isn't a good idea to substring the date string but It is difficult to have a entity method to bind the value as well with th:field. Probably I will use th:value in this case to give me the possibility to create that entity method providing the good format.

Include boolean params with g:include are passed as Strings to the included view

I don't know if the following change is an issue or it is intended.
<g:include view="line.gsp" params="['label':'test', 'progress':false]"/>
Then the expression in line.gsp aways evaluates to true, because the type of the 'progress' param is String, not Boolean.
class is: ${params.progress.getClass()}
<g:if test="${params.progress}">
this should not be displayed
</g:if>
Note that the same is applicable for other types, not just Boolean.
I am using grails 3.3.8
This didn't happen in grails 2.5.2.
I didn't find anything about this online so that's why I am asking here. Thanks.
Edit:
As suggested by Daniel, I've tried with grails 3.3.2 also.
I just created an app with grails create-app and modified the existing index.gsp to include line.gsp, as shown in the code above.
Here is a screenshot:
I realize you have already found a workaround to this, but it was bothering me that your code was not working the same way mine was, so I decided to investigate a bit further in case anyone else runs into this issue.
As other people here have mentioned, params.anything will typically return a String value. This is because params are typically encoded as URI parameters (like ?progress=false) and not autoboxed into any other type. (This is perfectly reasonable; there would be no good way for Grails to know what type they should be.) However, it is possible (and occasionally reasonable) to render your view or template from a controller like render view: whatever, model: [your: model, params: params] where you specifically include params in the model. (Perhaps you have a lot of parameters that you do not need to individually recreate in the model.) In this case, the params map will exist as both URI parameters (?progress=false) and page-scoped model entry (params: [progress: Boolean.FALSE]). Page scope takes priority over URI parameters, and so your types will be preserved.
In my testing, I added your code to an existing page where I was already passing params into the model, and so type was preserved for a newly added parameter as well. (Note that once params are in page scope, they're there for included templates or views as well.) Consequently, I saw progress type of Boolean, while in a basic example it would be type String.
TL/DR: either assume params are strings, or explicitly include params in your page-scoped model when rendering.
Yes, this is true. When you are calling params.something, you are actually accessing GrailsParameterMap which is a Map. So your condition actually evaluates as Map->get() which will be Object->toString() and of course it is true cause it's not null and not empty. So the solution will be the following:
<g:if test="${params.getBoolean("progress")}">
this should not be displayed
</g:if>
You can test this as follows
<g:if test="${params.progress && params.progress == 'false'}">
this should not be displayed
</g:if>

How to null check in EL to show p:dialogs conditionally?

I have two <p:dailog>s and based on the condition of a bean property I want to show one of them. I have used the following code
onclick="#{empty groupBean.selectionGroup?dialog_empty.show():groupDialog.show()}"
But it is not working as it says there is an error in EL expression. I am not sure where the error is. Am I doing it the correct way?
You're treating JavaScript code as part of the EL expression. This would only result in a syntax error because EL cannot find #{dialog_empty} nor #{groupDialog} in the scope. You have to treat JavaScript code as strings by quoting them because they ultimately needs to be written to the HTML response as-is:
onclick="#{empty groupBean.selectionGroup ? 'dialog_empty.show()' : 'groupDialog.show()'}"

What is the preferred way to specify an empty outcome in an EL-conditional expression?

I often want to declare conditional CSS style-classes for components if a certain condition is met. If the condition is not met, then no style should be added. I usually do it like this (which has worked fine so far):
<some:tag styleClass="#{someCondition ? 'SomeClass' : null}" />
Is this OK, or are there any dangerous pitfalls to having null there?
There are no dangerous pitfalls. #{null} is perfectly fine as EL won't emit anything to the response (it evaluates basically to an empty string).
You also sometimes see this style:
<some:tag styleClass="#{someCondition ? 'SomeClass' : ''}" />
This is the style which is usually used in plain Java code as follows:
System.out.println(someCondition ? "SomeClass" : "");
In plain Java code converting null to String would result in "null", not "", which may be undesireable. For ones not so familiar with EL, the above style is actually more self-documenting. Otherwise you get questions like Doesn't it print "null" then?

Accessing a HashMap using Struts 2

I have hashmap that was created on a page using the struts2 <s:set> tag. It looks something like this
<s:set var="mymap" value="#request.mymap"/>
At some point in the page, i need to get a value from the hashmap based upon a key, and i want to do it using OGNL.
The key is generated based upon some logic, which i store using another <s:set> tag. Something like this
<s:set var="mykey" value="1">
I need to get a value from the hashmap using this key. And I need to display it.
How do I simply call the get function on the hashmap?
I tried this
<s:property value="#mymap[#mykey]"/>
and this
<s:property value="#mymap[%{#mykey}]"/>
and this
<s:property value="%{#mymap[%{#mykey}}]"/>
The third one obviously does not work because of the nesting problem.
But the same nesting logic is applicable to the second case as well, due to the manner the value attribute is handled. However none seem to work for me.
The issue here is that my key is unknown. It is a dynamically generated string based upon some pattern. I need to access the object stored in the hashmap using this dynamic key. And due to the inability of nesting ognl, I am in a fix.
I suppose the issue is very simple. I almost feel that I get it, but somehow the solution eludes me.
I suppose I was using a different version of struts wherein using the %{} was required for the expression to be evaluated. I changed the jar files now.
This is what did the job for me:
<s:property value="#mymap.[#mykey2]"/>
My problem was coming because I was trying to use it in a href for a s:a tag. And without the %{} operator, the expression was not being evaluated.
So, i guess, i was right in the beginning itself. Rest of the time, it was just me being silly. :>
Update:
I wrote a blog post on the issue, in case anyone is interested.
http://mycodefixes.blogspot.com/2010/11/struts-2-creating-and-accessing-maps.html

Resources