Mailing-List: contact tapestry-dev-help@jakarta.apache.org; run by ezmlm
Precedence: bulk
Reply-To: "Tapestry development" <tapestry-dev@jakarta.apache.org>
Message-ID: <00ad01c3d3be$892312d0$0900a8c0@mindbridge>
From: "Mindbridge" <mindbridgeweb@yahoo.com>
To: "Tapestry development" <tapestry-dev@jakarta.apache.org>
Subject: Property initializers
Date: Mon, 5 Jan 2004 21:02:58 +0200
MIME-Version: 1.0
Content-Type: multipart/alternative;
	boundary="----=_NextPart_000_00AA_01C3D3CF.4C348FF0"

------=_NextPart_000_00AA_01C3D3CF.4C348FF0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable

Hi,

I am about to check in the docs and examples I was talking about, but =
came across something interesting.=20

The properties defined in the component specifications via =
<property-specification> can be initialized using the 'initial-value' =
parameter. This parameter is an OGNL expression that (at the moment) is =
evaluated once during the creation of the component, the value is =
stored, and is later used as an initialization value whenever necessary.

Shouldn't the 'initial-value' expression be evaluated every time it is =
needed for initialization instead (if not invariant, of course)?=20

This latter approach allows a number of uses that the former one does =
not (e.g. initializing to something that depends on the bindings), and =
perhaps it is more natural to the users.=20

Should this be considered a bug? Should it be changed? (this is backward =
compatible if the initial-value parameter is used as designed at the =
moment).=20

-mb

------=_NextPart_000_00AA_01C3D3CF.4C348FF0--