ajax submit and lost styles

[stuartm1971]

If I use any a4j elements to submit a page which navigates to another page, some of the styleClass application on the new page is lost. Any ideas?

eg if I submit using a4j:commandButton and the new page contains a commandButton with a styleClass attribute, the styling doesn't work first time into the page.

[ishabalov]

pls post you usecase with more details
what browser?
how second page look? Is any different between first and second times?

[stuartm1971]

The page I submit is via this control:

<a4j:commandButton id="srchpage_btn" value="Search" styleClass="button" tabindex="17" action="#{searchParameters.validate}">
<a4j:support status="testId" event="onclick" />
</a4j:commandButton>

I also use it to invoke an a4j:status so I can monitor its progress.

On the navigated-to page, I have a button:
<h:commandButton id="selectReceiptBtn" value="Select Receipt"
action="#{resultsData.validate}" styleClass="button" tabindex="0"/>


Using the navigation in the searchParameters.validate action attribute, the style of the h:commandButton is lost and some other is applied.
If I use a normal h:commandButton on the first page this problem doesn't occur.

I'm aware that using a in the appropriate navigation section of faces-config solves this problem but I'm a little reluctant to use this as redirection creates a new request object.

Any ideas why using ajax submission should upset the style application of the new button would be helpful, unless perhaps it's just a browser issue(I'm using IE 6.0.29).

[stuartm1971]

Some missing text from the last post: '....using a redirect in the faces-config....'

[ishabalov]

Do you mean that styleClass is not "button" anymore or you you css is not applyed?
Where is your css referred? Did you try to use style instead of styleClass?

[stuartm1971]

The problem still exists if I use style instead of styleClass. It still appears ad a button bu tthe styling of the button's label changes.

If I use the 'redirect' attribute in faces-config.xml it solves the problem but this starts a new request, which I feel is a bit untidy.