>>>>> On Fri, 22 Oct 2004 11:16:02 -0700, Adam Winer <adam.winer_at_oracle.com> said:
AW> Ed,
AW> Doesn't this remove support for "onclick" on commandButton?
AW> JSF developers are definitely using that today.
Yes, you are correct.
AW> What happens if there's no CommandLinks on a page? I'd
AW> think that would mean that the commandLink hidden field
AW> wouldn't function, and the commandButton's onclick would
AW> result in a Javascript error.
Actually, in my tests, it does not result in an error, however, you do
have a good point.
AW> Also, getHiddenFieldName() was a reasonable name for
AW> a method on CommandLinkRenderer, but is very confusing
AW> as a name on HtmlBasicRenderer.
Noted.
AW> FWIW, what we (Oracle UIX) did for this category of problem
AW> was to take *all* the "NeededHiddenFields" and explicitly
AW> null them all out in Javascript before continuing with the
AW> code that sets parameters and then submits.
That's what I want to do. However, where do you hang the javascript?
On the form? If so, how do you make it so the nulling out happens
BEFORE the parameter setting code?
>>>>> On Fri, 22 Oct 2004 14:30:53 -0500, Jacob Hookom <jacob_at_hookom.net> said:
JH> One thing I could suggest in for areas where we generate Javascript, if
JH> the attribute already has a value, prepend our script to the existing
JH> value. I'm not sure if the renderkits already works that way?
JH> onClick="alert('Developer debugging');"
JH> becomes
JH> onClick="injectedScript(); alert('Developer debugging');"
That's a good idea to address Adam's initial concern, but it seems like
there would be some cases where we would break user code.
Ed
--
| ed.burns_at_sun.com | {home: 407 294 2468, office: 408 884 9519 OR x31640}
| homepage: | http://javaweb.sfbay.sun.com/~edburns/
| aim: edburns0sunw | iim: ed.burns_at_sun.com
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe_at_javaserverfaces.dev.java.net
For additional commands, e-mail: dev-help_at_javaserverfaces.dev.java.net