Menu
▾
▴
[Springframework-developer] RE: Tags using expression language?? & Petclinic
|
From: Alef A. \(JTeam\) <al...@jt...> - 2003-07-13 17:25:13
|
I'm currently checking out the sourcecode. Somehow I wasn't able to =
during the last couple of days. Well, anyway, I'll have a look into it =
and I'll let you know tomorrow...
Cheers,
Alef
-----Oorspronkelijk bericht-----
Van: Ken Krebs [mailto:kk...@kk...]
Verzonden: Saturday, July 12, 2003 6:22 PM
Aan: j=C3=BCrgen h=C3=B6ller [werk3AT]; al...@jt...
CC: spr...@li...
Onderwerp: Tags using expression language?? & Petclinic
Eliminating repetitive code in the jsp's is something I'd very much like =
to see. Alef: thanks for your contribution, it looks like it will help.
I've been in the process lately of refactoring Petclinic's Clinic=20
implementation along with code tightening and documentation improvements =
and am just about ready to commit the changes. I WILL commit these=20
changes sometime later today. A couple of jsp's have minor changes.
Alef: If you would like to make the changes to the jsp's, please go=20
ahead. If not, I'll do it. Please let me know if you will be doing it.
Regards,
Ken
=20
j=C3=BCrgen h=C3=B6ller [werk3AT] wrote:
>I've introduced support for EL on tag attributes yesterday, already=20
>committed. Thanks for the idea and the prototype, Alef! I hope you are =
satisfied with the integration. You're very welcome to try it out :-)
>=20
>There's a new ExpressionEvaluationsUtils class now in=20
>com.interface21.web.util, taking a String argument value and parsing it =
to Object, String, int, or boolean. It will just attempt EL evaluation =
if the value starts with "${", else it will simply parse the value with =
standard means.
>=20
>All of Spring's tags use this class now for all attributes, simply=20
>delegating to the respective ExpressionEvaluationsUtils method in each =
setter. This means that all attributes still accept normal String values =
as before (like <i21:bind path=3D"person.name">), but also EL =
expressions (like <i21:bind path=3D"${bindPath}">).
>=20
>For EL evaluation, ExpressionEvaluationsUtils depends on Jakarta's JSTL =
>implementation (standard.jar). It doesn't have a runtime dependency if =
just parsing normal String values though, as it delegates EL evaluation =
to an inner class that will just get loaded in case of actual EL =
expressions. So you'll just need the Jakarta JSTL implementation in your =
classpath if you actually use EL attribute values on tags.
>=20
>BTW, our AopProxy uses a similar mechanism to avoid a runtime=20
>dependency on CGLIB. If just proxying interfaces, it will use standard =
J2SE proxies. Only if you try to proxy a class itself, it will invoke =
CGLIB by delegating to a respective inner class.
>=20
>Finally, we should adapt PetClinic to use EL expressions on its bind=20
>tags. Alef has already shown the way in the prototype that he sent. =
Ken, what do you think? Would you like Alef to do this?
>=20
>Juergen
>=20
>=20
>
> -----Urspr=C3=BCngliche Nachricht-----=20
> Von: j=C3=BCrgen h=C3=B6ller [werk3AT]=20
> Gesendet: Fr 11.07.2003 08:13=20
> An: al...@jt...; spr...@li...=20
> Cc:=20
> Betreff: Re: [Springframework-developer] Tags using expression=20
>language??
>=09
>=09
>
> Alef,
>=09
> I've just browsed your code - this looks very interesting! It makes=20
>iterating such repetitive form fields like in the petclinic much =
easier. I'll have a look at the implementation details, and if =
everything works out I'll add a clean room version to the main source =
tree promptly.
>=09
> I wonder if we could make that work with any JSTL implementation, not=20
>just the Jakarta one, although I wouldn't mind a dependency on the =
latter for this feature.
>=09
> Cool stuff :-)
>=09
> Juergen
>=09
>=09
>=09
> -----Urspr=C3=BCngliche Nachricht-----
> Von: Alef Arendsen [mailto:al...@jt...]
> Gesendet: Di 08.07.2003 19:43
> An: j=C3=BCrgen h=C3=B6ller [werk3AT]; =
spr...@li...
> Cc:
> Betreff: RE: [Springframework-developer] Tags using expression =
>language??
> =20
> =20
>=09
> Ok, here we go...
> =20
> Like you said, at first glance, it does not look like the =
i21:bind tag
> needs expressionalization (hmmm, nice huh ;-). However, in the =
petclinic
> demo app I'm seeing a lot of input.jsps in the jsp/fields =
directory.
> These are things I'm hoping to solve using for instance =
EL-based tags...
> Attached you'll find a reworked version of the ownerForm. It =
does not
> use the JSPs from the fields directory anymore, but uses only =
one
> input.jsp that uses the bindstatus object to create the input=20
>field.
> =20
> Ok, it's still all rough and reworking the tags is a little =
bit more
> work than the 15 minutes I spent on it now, but maybe you're =
getting the
> idea.
> =20
> Consequences for adding / reworking the tags:
> =20
> 1. dependency on jakarta-jstl-1.0.3 (jakarta.apache.org/taglib =
-->
> standard).
> 2. dependency on jstl-1.0.3 (java.sun.com)
> 3. In case you want to have both version in there, an =
extending class
> for each tag
> 4. For each tag, a BeanInfo class
> 5. I wasn't able to use the EvalHelper from jakarta, so I =
copied it
> (hmmm... not so nice, is it ;-)
> =20
> Probably I don't have time this week or something to refactor =
them to be
> all EL-based... Just let me know if you'd like it.
> =20
> Well, that's it for now, still discovering really cool =
features and
> already running out (brain)memory to think up everything I =
could do with
> Spring ;-)
> =20
> Cheers,
> =20
> Alef
> =20
>=09
> =
N=18HS^=E9=9A=8A[){([j=1FJ=EB=A2=BAky=C6=AE^=D8=A7j+x:0Z=1Au=DA=95g*)jw`z=
=D6=9F=E7=9B=A20
> Z(~(W(}iT)~{
> +=D7=AFzZ)zXX*kx=1F=C5=A0u=DE=96^X(=1E~zwilq zlX)=DF=A3))~{
> +=D7=AFzZ)
>
>?????????????????????????????????????????=D3=86+=12=17?^?=E9=9A=8A[)?{(?=
?[??=DA=AD?(~?+??=E9=AE=8A=1FY?
>=DA=A6??j?h??^??-?x?????:0?Z=1Aw?jU?l????=DD=81?Z~??n?$?=0C0???j?=1F??(?=
??W???(}?i?_???????????????????????????????????*k?x=1F???=C5=A0??=D7=AFzZ=
)z???X??X??*k?x=1F???=C5=A0??=D7=AFzZ)z???l??.?=C7=9F??=1E?w???i????+-??(=
??=1E~??{?=DE=B7?b????+-?w???k?x=1F???=C5=A0??=D7=AFzZ)
>
>
> =20
>
|