[jsr-314-open] [jsf2next] PROJECT_STAGE system property configuration

[Ed Burns]

CH> Still, this points to the fact that the default for the project stage is
CH> probably wrong.

EB> Let's take a step back and look at my motivation for wanting to have the
EB> default be "Development".  One of the biggest classes of gripes about
EB> JSF is the "JSF's error messages suck, if you get them at all" kind of
EB> gripe.  We need to make the out of the box experience address those
EB> kinds of gripes directly.

CH> Just to clarify--everyone seems in agreement that the default should be 
CH> Development, so developers get a good OOBE.

CH> But that's not what the spec says it is. The default is Production 
CH> (http://java.sun.com/javaee/6/docs/api/javax/faces/application/Application.html#getProjectStage()). 

CH> It sounds like you thought the default was Development, in which case 
CH> this would simply be a spec error that can hopefully be corrected quickly.

DA> +1. (It sounds like Ed is arguing with his alter-ego. Hehehe. Just joking)

Gah!  Cay is right, the spec does say the default is Production.  Either
1) I'm not remembering things correctly, 2) the spec is wrong 3) I've
changed my mind since writing the spec.  In any case, this discussion is
apropos.

JD> So, do we want naive users to have imperfect error messages, or a slow 
JD> server?

JD> Because those are the choices...

PM> I like the idea of a container-wide setting - it fits with how
PM> people use container instances (people rarely use the same server
PM> for development and production).

PM> It would also be great to get this pushed into the EE platform - I
PM> know that this has been suggested by the 314 EG, but when I speak to
PM> the EE guys, they also seem keen on this idea...

You weren't here for the JavaOne 2008 Web Tier EG meeting.  I spent a
lot of time there pushing for this in Servlet, but was ultimately
thwarted by that EG and we just kept it in JSF.  

I sent a proposal to Hani Suleiman on 14 May 2008, and he advocated and
forward it to the JSR-315 EG.  Thereafter a 30 message thread commenced,
spanning five months involving such lumaniaries as Joe Walker, Howard
Lewis Ship, Greg Wilkins, and Dhaji Prasanna.  Surely, such great minds
as these would see the value in PROJECT_STAGE.  Right?  Well, the idea
got conflated with JSR-299 development types and thereafter bogged down
and was eventually dropped.

Not content with just the JSR-315 EG, I brought the idea to Sun's EE
architects and met with resistence there.  I even brought it Charles
Nutter, then a Sun employee and stalwart Rails advocate, to promete the
idea.

If it's in JSF, it's in the platform.  I dont think we need to take this
battle any further up the platform.

CH> Absolutely. This is something that I have requested for years. An app
CH> server is a development platform, not just a deployment platform.

Oh do I wish I could share the JSR-315 EG traffic on this topic with
you.  Perhaps you should be on the servlet EG.

DA> We need to make sure this request makes it out of the mailinglist and in
DA> front of the Java EE spec leads. Is there a formal way for one spec to
DA> propose and enhancement to another spec, other than submitting a JSR? We
DA> want to propose it in such a way that they understand that all of us that
DA> support it are pushing/voting for it.

It's already been requested, 19 months ago.  I suggest you get your
Servlet EG rep to advocate for it.

LB> I've been hearing "appserver", but would the "Java EE Umbrella Spec" be
LB> enough to bring this behavior to servlet containers? The last thing we want
LB> is to create another thing that appservers provide, that servlet containers
LB> do not. We still need ProjectStage to make it into the Servlet spec,
LB> explicitly, in order to bridge that gap, right? Is that an accurate
LB> assumption?

As I said, they didn't care too much for the idea.

PM> As I understand it, if the JSF EG feels that this a feature that
PM> should be pushed to another spec, it is the responsibility of the
PM> spec lead to take this up. But rest assured if this doesn't happen,
PM> then we (Red Hat) can push this through other channels such as our
PM> Java EE EG rep :-)

With all due respect: What is it with you Red Hat/JBoss guys always
assuming Ed Burns is standing in the way of your work and not doing
anything on your behalf?  There is a provable pattern of behavior here
stretching back to October 2007 and I'm getting tired of it.  Can you at
least *check* with me before assuming that you need to take it straight
to the highest levels?  On the other hand, I have a provable pattern of
being accomodating and consensus building and continue to expect to run
this EG in that manner.

DA> ...and to add another point, I guess it would be nice if those other
DA> specs had an issue tracker like our where we could record these
DA> wishes.

I have pushed for this too.  For what it's worth, I have forwarded the
JSR-315 discussion of this topic to JSR-314-EG.  I cannot forward it to
JSR-314-OPEN because JSR-315 is not an open EG and I don't think they'd
like me to share their content on an open list.

Ed

-- 
| ed.burns at sun.com  | office: 408 884 9519 OR x31640
| homepage:         | http://ridingthecrest.com/