On Wed, Dec 3, 2008 at 8:18 AM, Bruce Johnson <[EMAIL PROTECTED]> wrote:

> On Wed, Dec 3, 2008 at 7:50 AM, <[EMAIL PROTECTED]> wrote:
>
>> To be honest, I wish we would start creating larger .gwt.xml files and
>> make each one that exists inheritable.
>
>
> I agree. It was a rookie decision we made early on to over-emphasize
> fine-grained module reuse, and, like C header files, nobody has enough
> discipline (or incentive) to ensure they are all truly independently
> inheritable.
>

A good question is whether we can just change it. The way things are
currently, you can't inherit any submodule out of user without getting all
of User. You either get errors because you are picking up gunk out of client
that doesn't appear to be translatable or you include one of the few modules
with a circular reference (SplitPanel or Animation). So given that you can't
inherit them now, could we just delete the submodules and inline their
contents into User? The only danger is that someone may've have done this:

<inherits name="com.google.gwt.user.User" />
<inherits name="com.google.gwt.user.DOM" />

or

<inherits name="com.google.gwt.animation.Animation" /> // not realizing that
it loops back to pick up all of User

So removing DOM.gwt.xml would produce an error for those clients.

>
>
>> Doing that would mean that I
>> would get rid of UserAgent.gwt.xml altogether and move its contents into
>> dom.DOM.gwt.xml. (or either create useragent.UserAgent.gwt.xml)
>
>
> The reason not to do this would be if there are other important use cases
> of modules that switch on User Agent but don't use DOM. I can only think of
> one right now: StringBuidler in 1.6 is the first place where the JRE can be
> sensitive to user agent. And you *might* in theory want to do sheer
> computation in GWT without touching the DOM (e.g. in a Gears worker thread).
>
> If we're going to make this kind of change, perhaps we should debate the
> merits of useragent.UserAgent to be darn sure it isn't the better option.
>

This is enough of an argument to make me a supporter of useragent.UserAgent.
Now that you mention it, we have already tried to position it as an
independently inheritable module when we introduced UserAgent JRE
optimizations.

/kel

-- 
If you received this communication by mistake, you are entitled to one free
ice cream cone on me. Simply print out this email including all relevant
SMTP headers and present them at my desk to claim your creamy treat. We'll
have a laugh at my emailing incompetence, and play a game of ping pong.
(offer may not be valid in all States).

--~--~---------~--~----~------------~-------~--~----~
http://groups.google.com/group/Google-Web-Toolkit-Contributors
-~----------~----~----~----~------~----~------~--~---

Reply via email to