Adam, Thanks for visiting our mailing list. Its always interesting to hear what the EG is up to and what there take is on various things. Oracle folks are always welcome too :-)
The prependId attribute sounds like a step in the right direction. I am not sure if it addresses all of the concerns though. The concept of the naming container controlling whether it prepends id information is an interesting concept though. I hadn't thought of addressing it that way. I guess b/c it's usually one or two components per page where it matters to me instead of all of the components in a form. I'd be interested to hear more of the details. I agree with Heath that subview is a major issue to contend with. There is more than one "naming container" in the API so why would would you not want to address subview, etc. Perhaps this was an oversight but I would urge the EG to reconsider changing only the <h:form> component because this seems like a partial solution. So does the EG agree that the inability to control the id is an issue? If they do not agree I would be happy to provide a long list of reasons why this is important. But if we all agree on that we could save time and discuss possible solutions. Just to reiterrate what Heath said, in the dataTable case, we are appending row information. So if you had <h:inputText id="foo" forceId="true"> you would end up with foo[0] foo[1], etc. This is basically the Struts approach and I find it to be pretty intuitive. Regards, sean On Apr 1, 2005 5:02 PM, Heath Borders <[EMAIL PROTECTED]> wrote: > Sorry, I meant appending, not prepending. > > As far as not allowing turning off all parent NamingContainer > prependings, I'm not sure that the disambiguation from <f:subview> is > acceptable, especially in cases where you are developing a large > application using Tiles. > > In our case, we have our pages nested inside multiple subviews that > wrap various abstractions needed by different pages. However, the > content dropped into those Tiles needn't ever change. If there is no > way to obliterate all prepended ids, it makes it nearly impossible to > develop code inside a page that has no knowledge of its parent subview > structure. > > On Apr 1, 2005 3:47 PM, Adam Winer <[EMAIL PROTECTED]> wrote: > > Prepending a row-number isn't that ideal, since that necessarily > > violates the naming rules for IDs, which must start with letters, > > an underscore, or a colon: > > > > http://www.xml.com/axml/target.html#NT-Name > > > > ... but the decision (and my opinion) was that turning off *all* > > prepending with one flag is simply too blunt, and didn't give > > you a way to say something like "look, I'm sick of that form > > ID getting in the way, but I can accept disambiguation from > > <f:subview>". > > > > -- Adam > > > > On Apr 1, 2005 1:39 PM, Heath Borders <[EMAIL PROTECTED]> wrote: > > > But what about other NamingContainers that are not UIForms, like > > > UIDatas? The way that forceId works is ideal with a UIData (it > > > prepends the row-number onto the id). > > > > > > On Apr 1, 2005 3:35 PM, Adam Winer <[EMAIL PROTECTED]> wrote: > > > > Hi Martin! > > > > > > > > ADF Faces resolves this by virtue of its UIXForm component not > > > > extending UIForm (or even UIComponentBase - we have our own > > > > UIXComponentBase class), and therefore it does not > > > > implement NamingContainer. > > > > > > > > BTW, the subject of "forceId" was raised on the EG a couple > > > > weeks back, and we went a slightly different way to solve this > > > > in 1.2: <h:form> will have a "prependId" flag that can be set > > > > to false. > > > > > > > > -- Adam Winer > > > > > > > > On Apr 1, 2005 11:07 AM, Martin Cooper <[EMAIL PROTECTED]> wrote: > > > > > So now that I know there is at least one ADF Faces person on this > > > > > list (Hi > > > > > Adam! :) ... > > > > > > > > > > I'd be very interested in hearing how ADF Faces resolves the issues > > > > > that > > > > > were discussed here a while ago with respect to JSF generating the > > > > > IDs for > > > > > components, making it difficult to use getElementById(), and > > > > > culminating > > > > > in the addition of the 'forceId' attribute to MyFaces. Is that > > > > > something > > > > > the ADF Faces team might be able to share with us? > > > > > > > > > > TIA! > > > > > > > > > > -- > > > > > Martin Cooper > > > > > > > > > > > > > > > -- > > > -Heath Borders-Wing > > > [EMAIL PROTECTED] > > > > > > > -- > -Heath Borders-Wing > [EMAIL PROTECTED] >
