You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@myfaces.apache.org by Simon Kitching <sk...@obsidium.com> on 2005/11/24 01:23:47 UTC
Release method in UIComponentTag (was Reconnecting UIComponentTag
instances ...)
I see that originally _id etc were reset in internalRelease, and they
were moved to the release method by this commit:
------------------------------------------------------------------------
r166747 | manolito | 2004-04-27 00:01:39 +1200 (Tue, 27 Apr 2004) | 2 lines
more reluctant releasing of members for Resin compatibility
So it looks like this is a workaround for a Resin bug. I dont quite
understand that; I can't see what bug Resin might have that would be
fixed by this but wouldn't break many other things. Is "manolito" around
to explain further?
Regards,
Simon
Adam Winer wrote:
> Sounds like it is a bug, though that's a very quick impression
> without serious thought.
>
> Re: what happens if the number of tags change from one request
> to the next: this is why JSF tags inside of c:if are required to have
> an explicitly set ID (think this is stated explicitly in the JSF spec).
>
> This whole mess is why I proposed JspIdConsumer for JSP 2.1,
> which lets you do this reconnecting in a robust manner.
>
> Cheers,
> Adam Winer
>
> On 11/23/05, Travis Reeder <tr...@gmail.com> wrote:
>> Does anyone have any comment on whether this is a bug or not:
>>
>>> As you can see UIComponentTag.doEndTag calls internalRelease(). But
>>> strangely, members _id, _rendered and _binding are cleared only in
>>> release(), not internalRelease(). Maybe that's a bug?
>>>
>
Re: Release method in UIComponentTag (was Reconnecting UIComponentTag instances ...)
Posted by Martin Marinschek <ma...@gmail.com>.
Manolito would be Manfred Geiler.
I suppose that release is called when the tag is given back to the
ServletContainer anyways?
regards,
Martin
On 11/24/05, Simon Kitching <sk...@obsidium.com> wrote:
> I see that originally _id etc were reset in internalRelease, and they
> were moved to the release method by this commit:
>
> ------------------------------------------------------------------------
> r166747 | manolito | 2004-04-27 00:01:39 +1200 (Tue, 27 Apr 2004) | 2 lines
>
> more reluctant releasing of members for Resin compatibility
>
>
> So it looks like this is a workaround for a Resin bug. I dont quite
> understand that; I can't see what bug Resin might have that would be
> fixed by this but wouldn't break many other things. Is "manolito" around
> to explain further?
>
> Regards,
>
> Simon
>
> Adam Winer wrote:
> > Sounds like it is a bug, though that's a very quick impression
> > without serious thought.
> >
> > Re: what happens if the number of tags change from one request
> > to the next: this is why JSF tags inside of c:if are required to have
> > an explicitly set ID (think this is stated explicitly in the JSF spec).
> >
> > This whole mess is why I proposed JspIdConsumer for JSP 2.1,
> > which lets you do this reconnecting in a robust manner.
> >
> > Cheers,
> > Adam Winer
> >
> > On 11/23/05, Travis Reeder <tr...@gmail.com> wrote:
> >> Does anyone have any comment on whether this is a bug or not:
> >>
> >>> As you can see UIComponentTag.doEndTag calls internalRelease(). But
> >>> strangely, members _id, _rendered and _binding are cleared only in
> >>> release(), not internalRelease(). Maybe that's a bug?
> >>>
> >
>
>
--
http://www.irian.at
Your JSF powerhouse -
JSF Consulting, Development and
Courses in English and German
Professional Support for Apache MyFaces