tapestry-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Howard Lewis Ship <hls...@gmail.com>
Subject Latest changes ...
Date Tue, 08 Mar 2005 22:02:32 GMT
I've been very busy of late.

Tapestry now has improved line-precise exception reporting; the
Exception page now can display the actual file content, highlighting
the line that's in error.

I've also created a bunch of services and such to support "describing"
objects as HTML.  This is mostly used for more Exception page stuff:
describing the HttpServletRequest, etc.

I've been stripping more and more code out of IEngine and
AbstractEngine, moving it elsewhere.

It's been more "positioning" to get things in line for Portlets.  For
Servlet Tapestry, the execution path is:

ApplicationServlet -> ServletRequestServicerPipeline ->
WebRequestServicerPipeline -> Engine -> Engine Service, etc.

The eventual goal is to get as much as possible into the
WebRequestServicerPipeline, so that the code can be shared between
Servlet Tapestry and Portlet Tapestry.

Portlet Tapestry's execution path will be:

ApplicationPortlet -> PortletRequestServicerPipeline ->
WebRequestServicerPipeline -> Engine, etc.

A few service implementations will be specific to Servlet Tapestry or
specific to Portlet Tapestry.  What's going to happen is that the
InitializerPipeline for each (again, pairs) will be responsible for
storing the proper implementations into ApplicationGlobals, and
Infrastructure will be configured from these using the
service-property: object provider.  Everybody else will use
infrastructure: to gain access to the correct flavor.  Lots of fun
indirection!

One thing I'm seeing, especially on my wife's rather anemic laptop, is
that startup time is getting a bit longer than I'd like (about 10
seconds) . I'm going to see if some generous use of HiveMind's
primitive service lifecycle model will help there.  The idea is to cut
down on the number of proxy classes and instances that get created and
get into the execution stack .. that should help with startup and
maybe even execution.

Perhaps the biggest change is that the RequestContext is completely
gone. It was too tightly tied to the Servlet APIs; everything is used
to do has been refactored into new services and pipeline filters. 
Need to update the docs about this.

-- 
Howard M. Lewis Ship
Independent J2EE / Open-Source Java Consultant
Creator, Jakarta Tapestry
Creator, Jakarta HiveMind

Professional Tapestry training, mentoring, support
and project work.  http://howardlewisship.com

---------------------------------------------------------------------
To unsubscribe, e-mail: tapestry-dev-unsubscribe@jakarta.apache.org
For additional commands, e-mail: tapestry-dev-help@jakarta.apache.org


Mime
View raw message