Oh how stupid of me. Guess which component is probably best suited for
carrying
around important state information? Form :)
The only thing that matters during rewind is making sure the ids generated
are the same as what
were generated when it was first displayed, if they change drastically on
the non rewind cycle it won't
matter. So...Problem solved I think.
On 11/6/06, Jesse Kuhnert <[EMAIL PROTECTED]> wrote:
Hmmm....Maybe we aren't on the same page.
I've solved the getting of client ids before / during (for the most part),
but the after render part has me a little
concerned.
A big reason why tapestry scales so well is in how it handles transient
properties, so I'm not sure
introducing some of these things for the sake of easier use necessarily
makes for a better
solution...I could see changing the _clientId property to be generated on
each render and left "in tact" being
generally ok though. Form does it already right? (In fact, we can probably
eliminate this _name
property from Form altogether now that _clientId and _name generally
always equal for any form
component)
As for "regions" of a page that are being refreshed - this is hopefully
already being handled by
the framework "as is" ? Because of the b0rken way IE handles event
connections we really
don't have a choice in the matter. (Meaning doing node.innerHTML works
just fine in moz without
having to re-wire anything but not at all in IE...and you'll actually end
up with a huge number of phantomed
references lying around in ie that will never get cleaned up if you don't
specifically disconnect them
before doing that...hooray for leaky abstractions. It's good to be
reminded of all the COM running things
to keep us honest.)
On 11/6/06, andyhot <[EMAIL PROTECTED]> wrote:
>
> Jesse Kuhnert wrote:
> > Yeah, the only problem with that is forms are rewound "directly", so
> > other
> > forms
> > won't even be around at the time of rewind.
> >
> > No solution? Of course there is, somewhere.. ;) We don't really have a
>
> > choice I think, seeing
> > as how no one else is going to do it for us. I could always pull out
> my
> > 128bit uuid stuff from
> > old jini days past, but that seems a little brute force for something
> > like
> > this.
>
> ok, let me then define what's the problem i have in mind. Perhaps we're
> really trying to
> solve slightly different things:
> - unique ids in a page
> - after partial refresh all ids in page remain unique
> - elements contained in a refreshed fragment that already existed in the
> initial
> page should retain the same id.
> - ability to get (client-side) id of a tapestry component ( after or
> even before that
> component has rendered).
>
> The first 2 are straightforward.
> The 3rd is needed so that user code that works on a page ( one that
> perhaps does dojo.byId('xxx') )
> continues to work after the refresh.
> The 4th is for allowing users to interact with stuff we/they generate.
>
> >
> > On 11/5/06, andyhot <[EMAIL PROTECTED]> wrote:
> >>
> >> Jesse Kuhnert wrote:
> >> > I think I have a solution in the works to solve this "once and for
> >> > all"..Spent the majority of the day hacking around with things and
> >> > finally
> >> > "get it" from an overarching perspective.
> >> >
> >> > Your namespacing of the form idea is a good one, but I don't think
> it
> >> > will
> >> > be completely reliable either if a few forms are involved in a page
> >> > that sometimes render and sometimes don't.
> >>
> >> Well, the idea was that we'd go on and call renderIdAttribute on all
> >> those
> >> forms even if they're not to be rendered.
> >> Also thought of using the page-level allocator (instead of the one
> >> provided by FormSupport)
> >> but it actually fails worse for similar reasons.
> >>
> >> Anyway, i'm eager to see your solution...
> >>
> >> Hope you can prove me wrong cause I currently think that there's no
> >> clean, transparent and global solution to this.
> >> I feel there's always gonna be a sub-sub-sub-case that will fail.
> >> But perhaps this is not that bad... at least if users are aware of
> what
> >> is happening (and how to deal with it).
> >>
> >>
> >>
> >> >
> >> > I don't know when I'll be done with the changes but hopefully it'll
> be
> >> > today
> >> > / tomorrow- ish.
> >> >
> >> > On 11/5/06, [EMAIL PROTECTED] <[EMAIL PROTECTED]> wrote:
> >> >>
> >> >> Author: andyhot
> >> >> Date: Sun Nov 5 07:54:57 2006
> >> >> New Revision: 471451
> >> >>
> >> >> URL: http://svn.apache.org/viewvc?view=rev&rev=471451
> >> >> Log:
> >> >> namespace the id of each form element with the id of the current
> >> form...
> >> >> resovles TAPESTRY-1131
> >> >>
> >> >> Modified:
> >> >>
> >> >>
> >> >>
> >>
>
tapestry/tapestry4/branches/4.1.1-clientIds/tapestry-framework/src/java/org/apache/tapestry/form/Form.java
> >>
> >> >>
> >> >>
> >> >>
> >> >>
> >>
>
tapestry/tapestry4/branches/4.1.1-clientIds/tapestry-framework/src/java/org/apache/tapestry/form/FormSupportImpl.java
> >>
> >> >>
> >> >>
> >> >> Modified:
> >> >>
> >>
>
tapestry/tapestry4/branches/4.1.1-clientIds/tapestry-framework/src/java/org/apache/tapestry/form/Form.java
> >>
> >> >>
> >> >> URL:
> >> >>
> >>
http://svn.apache.org/viewvc/tapestry/tapestry4/branches/4.1.1-clientIds/tapestry-framework/src/java/org/apache/tapestry/form/Form.java?view=diff&rev=471451&r1=471450&r2=471451
>
> >>
> >> >>
> >> >>
> >> >>
> >>
> ==============================================================================
> >>
> >> >>
> >> >> ---
> >> >>
> >>
>
tapestry/tapestry4/branches/4.1.1-clientIds/tapestry-framework/src/java/org/apache/tapestry/form/Form.java
> >>
> >> >>
> >> >> (original)
> >> >> +++
> >> >>
> >>
>
tapestry/tapestry4/branches/4.1.1-clientIds/tapestry-framework/src/java/org/apache/tapestry/form/Form.java
> >>
> >> >>
> >> >> Sun Nov 5 07:54:57 2006
> >> >> @@ -193,6 +193,9 @@
> >> >>
> >> >> if (isRewinding())
> >> >> {
> >> >> + // even if we're rewinding, make sure we 'train' the
> >> >> idallocator.
> >> >> + renderIdAttribute(writer, cycle);
> >> >> +
> >> >> String submitType = _formSupport.rewind();
> >> >>
> >> >> IActionListener listener = findListener(submitType);
> >> >>
> >> >> Modified:
> >> >>
> >>
>
tapestry/tapestry4/branches/4.1.1-clientIds/tapestry-framework/src/java/org/apache/tapestry/form/FormSupportImpl.java
>
> >>
> >> >>
> >> >> URL:
> >> >>
> >>
>
http://svn.apache.org/viewvc/tapestry/tapestry4/branches/4.1.1-clientIds/tapestry-framework/src/java/org/apache/tapestry/form/FormSupportImpl.java?view=diff&rev=471451&r1=471450&r2=471451
> >>
> >> >>
> >> >>
> >> >>
> >>
> ==============================================================================
> >>
> >> >>
> >> >> ---
> >> >>
> >>
>
tapestry/tapestry4/branches/4.1.1-clientIds/tapestry-framework/src/java/org/apache/tapestry/form/FormSupportImpl.java
>
> >>
> >> >>
> >> >> (original)
> >> >> +++
> >> >>
> >>
>
tapestry/tapestry4/branches/4.1.1-clientIds/tapestry-framework/src/java/org/apache/tapestry/form/FormSupportImpl.java
>
> >>
> >> >>
> >> >> Sun Nov 5 07:54:57 2006
> >> >> @@ -393,7 +393,7 @@
> >> >>
> >> >> String filteredId =
> TapestryUtils.convertTapestryIdToNMToken
> >> >> (baseId);
> >> >>
> >> >> - String result =
> _elementIdAllocator.allocateId(filteredId);
> >> >> + String result = _elementIdAllocator.allocateId(filteredId
> +
> >> >> "_" +
> >> >> _form.getClientId());
> >> >>
> >> >> if (_rewinding)
> >> >> {
> >> >>
> >> >>
> >> >>
> >> >
> >> >
> >>
> >>
> >> --
> >> Andreas Andreou - [EMAIL PROTECTED] - http://andyhot.di.uoa.gr
> >> Tapestry / Tacos developer
> >> Open Source / J2EE Consulting
> >>
> >>
> >> ---------------------------------------------------------------------
> >> To unsubscribe, e-mail: [EMAIL PROTECTED]
> >> For additional commands, e-mail: [EMAIL PROTECTED]
> >>
> >>
> >
> >
>
>
> --
> Andreas Andreou - [EMAIL PROTECTED] - http://andyhot.di.uoa.gr
> Tapestry / Tacos developer
> Open Source / J2EE Consulting
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [EMAIL PROTECTED]
> For additional commands, e-mail: [EMAIL PROTECTED]
>
>
--
Jesse Kuhnert
Tapestry/Dojo/(and a dash of TestNG), team member/developer
Open source based consulting work centered around
dojo/tapestry/tacos/hivemind. http://blog.opencomponentry.com
--
Jesse Kuhnert
Tapestry/Dojo/(and a dash of TestNG), team member/developer
Open source based consulting work centered around
dojo/tapestry/tacos/hivemind. http://blog.opencomponentry.com
