The big argument for this cglib is to introduce some type safety to
what is currently string-based bindings. This is a big plus. I
don't think it's worth giving up the "pure Java" aspect that so many
of us obviously like about Wicket.
I'd rather see some IDE support that is smart about flagging bindings
which refer to non-existent bean attributes. That way there's no
magic or performance hit happening at runtime. IntelliJ does
something similar for JPA query Strings. If a query references a non-
existent property it appears as an error, even though they're just
Strings. Obviously things get really sticky when you're dealing with
CompoundPropertyModels, etc.
Best case, of course, would be an addition to the java language that
lets you refer to a property in a type-safe way. Is there a JSR for
this?
-Sam Barnum
360Works
On Feb 7, 2008, at 1:50 AM, Maeder Thomas wrote:
+ 1 for what Igor says. I remember debugging Hibernate code: you debug
as far as your own code goes, and then you just guess. Oh, and yes:
Tapestry anyone?
Thomas
-----Original Message-----
From: Igor Vaynberg [mailto:[EMAIL PROTECTED]
Sent: Donnerstag, 7. Februar 2008 02:34
To: users@wicket.apache.org
Subject: Re: CompoundModel based on proxies
i disagree. i dont think we should be doing more with cglib
in core or any other bytecode magic. have you ever tried to
walk code that uses bytecode generation? its a nightmare. one
of my favorite things about wicket is that it is just java
and its easy as hell to debug. im not really against putting
something like this into extensions, or even having a new
wicket-bytecode/codegen/whatever package that contains things
like these...
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
