You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@tapestry.apache.org by Howard Lewis Ship <hl...@gmail.com> on 2010/02/19 20:05:07 UTC
Big commit!
Ah, the joy of Git ... that was one big ass commit just now.
You'd be surprised how much of that was coded while sitting sideways
in my seat on the way to and from London, and much of the rest while
sitting in the hospital waiting for my son to be born!
There's still lots of room for tweaking and improving performance, but
creating a ComponentClassTransformWorker that does non-trivial stuff
is now much, much, much easier and there's nothing left in the APIs
that requires Javassist under the covers.
However, many of the existing methods of ClassTransformation have been
deprecated, and some of the deprecated methods are non-functional:
they throw a runtime exception (these are the methods most closely
tied to Javassist).
Basically, where in the past you would specify Javassist script for
part of a method implementation (i.e., to perform some work inside
containingPageDidLoad()) the new approach is to add advice to the
containingPageDidLoad() method to do the desired work.
This will affect performance and memory utilization. Performance may
downgrade some imperceptible amount (there's simply more method
invocations going on). Memory utilization will go up; in fact, much of
the new code creates many more classes in the component class loader.
These new classes exist to provide efficient reflection-style access
to component fields and methods (regardless of visibility).
The flip side, however, is that much more of the meta-programming code
will be in "normal" space, as callbacks from components, rather than
in "javassist" space as runtime-generated bytecode.
--
Howard M. Lewis Ship
Creator of Apache Tapestry
The source for Tapestry training, mentoring and support. Contact me to
learn how I can get you up and productive in Tapestry fast!
(971) 678-5210
http://howardlewisship.com
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@tapestry.apache.org
For additional commands, e-mail: dev-help@tapestry.apache.org
Re: Big commit!
Posted by "Thiago H. de Paula Figueiredo" <th...@gmail.com>.
On Fri, 19 Feb 2010 17:05:07 -0200, Howard Lewis Ship <hl...@gmail.com>
wrote:
> Ah, the joy of Git ... that was one big ass commit just now.
You had problems making Git work with Subversion. Are they solved now?
What was wrong them, so we cannot commit the same erros? :)
> You'd be surprised how much of that was coded while sitting sideways
> in my seat on the way to and from London, and much of the rest while
> sitting in the hospital waiting for my son to be born!
Congratulations for your son! :)
> There's still lots of room for tweaking and improving performance, but
> creating a ComponentClassTransformWorker that does non-trivial stuff
> is now much, much, much easier and there's nothing left in the APIs
> that requires Javassist under the covers.
Nice! Do you have any plans about the proxy implementation in
Tapestry-IoC? It's quite hard to make it support annotations in service
implementations the way it is now.
> Memory utilization will go up; in fact, much of
> the new code creates many more classes in the component class loader.
I guess it won't be significant.
> The flip side, however, is that much more of the meta-programming code
> will be in "normal" space, as callbacks from components, rather than
> in "javassist" space as runtime-generated bytecode.
The simpler, the better. :)
--
Thiago H. de Paula Figueiredo
Independent Java, Apache Tapestry 5 and Hibernate consultant, developer,
and instructor
Owner, software architect and developer, Ars Machina Tecnologia da
Informação Ltda.
http://www.arsmachina.com.br
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@tapestry.apache.org
For additional commands, e-mail: dev-help@tapestry.apache.org
Re: Big commit!
Posted by Howard Lewis Ship <hl...@gmail.com>.
On Fri, Feb 19, 2010 at 1:03 PM, Alfonso Quiroga <al...@gmail.com> wrote:
> Interesting, this means that T5 does NOT deppend on javassist now? thanks
That's a goal, that may take a few iterations to reach. I expect that
the final 5.2 will not depend on Javassist.
>
> On Fri, Feb 19, 2010 at 4:47 PM, Markus Joschko <ma...@gmail.com>wrote:
>
>> > You'd be surprised how much of that was coded while sitting sideways
>> > in my seat on the way to and from London, and much of the rest while
>> > sitting in the hospital waiting for my son to be born!
>>
>> Congratulations, Howard! Wow, your wife must be really supportive.
>> I can't imagine what would have happened to me if I just had brought a
>> notebook with
>> me to the hospital ;-)
>>
>> Good luck,
>> Markus
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: dev-unsubscribe@tapestry.apache.org
>> For additional commands, e-mail: dev-help@tapestry.apache.org
>>
>>
>
--
Howard M. Lewis Ship
Creator of Apache Tapestry
The source for Tapestry training, mentoring and support. Contact me to
learn how I can get you up and productive in Tapestry fast!
(971) 678-5210
http://howardlewisship.com
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@tapestry.apache.org
For additional commands, e-mail: dev-help@tapestry.apache.org
Re: Big commit!
Posted by Alfonso Quiroga <al...@gmail.com>.
Interesting, this means that T5 does NOT deppend on javassist now? thanks
On Fri, Feb 19, 2010 at 4:47 PM, Markus Joschko <ma...@gmail.com>wrote:
> > You'd be surprised how much of that was coded while sitting sideways
> > in my seat on the way to and from London, and much of the rest while
> > sitting in the hospital waiting for my son to be born!
>
> Congratulations, Howard! Wow, your wife must be really supportive.
> I can't imagine what would have happened to me if I just had brought a
> notebook with
> me to the hospital ;-)
>
> Good luck,
> Markus
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@tapestry.apache.org
> For additional commands, e-mail: dev-help@tapestry.apache.org
>
>
Re: Big commit!
Posted by Markus Joschko <ma...@gmail.com>.
> You'd be surprised how much of that was coded while sitting sideways
> in my seat on the way to and from London, and much of the rest while
> sitting in the hospital waiting for my son to be born!
Congratulations, Howard! Wow, your wife must be really supportive.
I can't imagine what would have happened to me if I just had brought a
notebook with
me to the hospital ;-)
Good luck,
Markus
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@tapestry.apache.org
For additional commands, e-mail: dev-help@tapestry.apache.org
Re: Big commit!
Posted by Howard Lewis Ship <hl...@gmail.com>.
We're Jewish, but no bris either (we're not practicing).
On Fri, Feb 19, 2010 at 11:36 AM, Daniel Gredler <dj...@gmail.com> wrote:
> Congratulations on the new Ship!
>
> Have you christened him? ;-)
>
>
>
> On Fri, Feb 19, 2010 at 2:05 PM, Howard Lewis Ship <hl...@gmail.com> wrote:
>>
>> Ah, the joy of Git ... that was one big ass commit just now.
>>
>> You'd be surprised how much of that was coded while sitting sideways
>> in my seat on the way to and from London, and much of the rest while
>> sitting in the hospital waiting for my son to be born!
>>
>> There's still lots of room for tweaking and improving performance, but
>> creating a ComponentClassTransformWorker that does non-trivial stuff
>> is now much, much, much easier and there's nothing left in the APIs
>> that requires Javassist under the covers.
>>
>> However, many of the existing methods of ClassTransformation have been
>> deprecated, and some of the deprecated methods are non-functional:
>> they throw a runtime exception (these are the methods most closely
>> tied to Javassist).
>>
>> Basically, where in the past you would specify Javassist script for
>> part of a method implementation (i.e., to perform some work inside
>> containingPageDidLoad()) the new approach is to add advice to the
>> containingPageDidLoad() method to do the desired work.
>>
>> This will affect performance and memory utilization. Performance may
>> downgrade some imperceptible amount (there's simply more method
>> invocations going on). Memory utilization will go up; in fact, much of
>> the new code creates many more classes in the component class loader.
>> These new classes exist to provide efficient reflection-style access
>> to component fields and methods (regardless of visibility).
>>
>> The flip side, however, is that much more of the meta-programming code
>> will be in "normal" space, as callbacks from components, rather than
>> in "javassist" space as runtime-generated bytecode.
>>
>>
>>
>> --
>> Howard M. Lewis Ship
>>
>> Creator of Apache Tapestry
>>
>> The source for Tapestry training, mentoring and support. Contact me to
>> learn how I can get you up and productive in Tapestry fast!
>>
>> (971) 678-5210
>> http://howardlewisship.com
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: dev-unsubscribe@tapestry.apache.org
>> For additional commands, e-mail: dev-help@tapestry.apache.org
>>
>
>
>
> --
> Daniel Gredler
> http://daniel.gredler.net/
>
--
Howard M. Lewis Ship
Creator of Apache Tapestry
The source for Tapestry training, mentoring and support. Contact me to
learn how I can get you up and productive in Tapestry fast!
(971) 678-5210
http://howardlewisship.com
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@tapestry.apache.org
For additional commands, e-mail: dev-help@tapestry.apache.org