you're right. That sounds like a great idea. It's a nice one line fix. I did a cheap "grep check" through the source code for the use of UNIQUE_ID_PREFIX (the constant which defines String "_id") and it's always used checking for component ID's starting with "_id".
So unless anyone else knows any other reasons the name spacing order shouldn't be switched...? thanks, -Henrik On 12/28/05, Martin Marinschek <[EMAIL PROTECTED]> wrote: > Wouldn't it be easier to just change the way the encodeNamespace() > method works and first use the _id something, and then the portlet > name? > > Would that break anything somewhere else? > > regards, > > Martin > > On 12/29/05, Henrik Bentel <[EMAIL PROTECTED]> wrote: > > thanks, > > that makes perfect sense. > > > > The natural followup question is: > > Would it makes sense to change HtmlTextRendererBase to check for index of > > "_id" instead? I dug around in the myfaces code and it seems > > PortletExternalContextImpl.encodeNamespace(...) is where the portlet > > name is appended to the generated id. Though indexOf could be an > > expensive operation to do very often. > > > > -Henrik > > > > On 12/28/05, Martin Marinschek <[EMAIL PROTECTED]> wrote: > > > The reasoning is that every id which doesn't start with _id is meant > > > to be a user-set id. If the user sets an id, the assumption is that he > > > wants to do something with it - so this is the reason for rendering > > > out a span, cause if you don't, the id won't be accessible from CSS or > > > javascript. > > > > > > regards, > > > > > > Martin > > > > > > On 12/28/05, Henrik Bentel <[EMAIL PROTECTED]> wrote: > > > > Hi > > > > > > > > This question is related to issue MYFACES-702 > > > > (http://issues.apache.org/jira/browse/MYFACES-702) > > > > > > > > > > > > I've got a couple of JSF Portlet which are running on Liferay 3.6.1. > > > > And I'm also experiencing the issue mentioned in MYFACES-702 except that > > > > pretty much all my page components gets wrapped in a span element. > > > > > > > > I've looked in both Liferay and Myfaces sources and so far I've found > > > > that > > > > Myfaces HtmlTextRendererBase.java is the culprit which appends > > > > the span element. > > > > > > > > The logic in the code is that: > > > > (if the component has an ID) AND (the ID does NOT start with "_id") > > > > then surround the element with a SPAN. > > > > > > > > The problem is that when running in the portal every components id is > > > > "automagically" prefixed with the name of the portlet. So my > > > > compoenent ID's are like this: "_reporting__id0", "_reporting__id1", > > > > and so on where the portlet name is "reporting". The portlet name > > > > matches the <portlet-name> element in portlet.xml. > > > > > > > > The workaround mentioned in the issue tracker works for the most part. > > > > However even verbatim components are surrounded with span tags du to > > > > their ID's automatically being > > > > prefixed with portlet name. And the <f:verbatim> tag doesn't have an > > > > ID attribute so there is no way to avoid the SPAN(spam) elements here. > > > > > > > > I haven't yet tracked down where these ID's are being assigned but I > > > > was curious if anyone knows the reasoning(or requirement) behind > > > > injecting the SPAN element? > > > > > > > > > > > > Thanks, > > > > Henrik Bentel > > > > > > > > > > > > > -- > > > > > > http://www.irian.at > > > > > > Your JSF powerhouse - > > > JSF Consulting, Development and > > > Courses in English and German > > > > > > Professional Support for Apache MyFaces > > > > > > > > > > -- > > http://www.irian.at > > Your JSF powerhouse - > JSF Consulting, Development and > Courses in English and German > > Professional Support for Apache MyFaces >
