[gwt-contrib] Re: Wholesale move of the app, requestfactory, and valuestore packages from the bikeshed dir to user... (issue660802)

2010-06-25 Thread fabbott

Sorry to get to this a bit late...

I've got a couple of concerns:
  1. regarding especially the user.compile use of javax validation,
 does that suggest we should embed that library in gwt-dev.jar
 with all of our other runtime dependencies?  (I don't love that
 practice, but it has been our practice...)
  2. bikeshed/scripts/maven_script.sh seems kind of orphaned if the
 rest goes away.  There's nothing really bikeshed-y about it, is
 there?  Should it move up to a top-level scripts directory?

On 2010/06/24 20:29:26, amitmanjhi wrote:




http://gwt-code-reviews.appspot.com/660802/show

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


[gwt-contrib] Re: Wholesale move of the app, requestfactory, and valuestore packages from the bikeshed dir to user... (issue660802)

2010-06-25 Thread Amit Manjhi
Thanks for the feedback. Since I already submitted the patch, I will create
follow-up patches for the issues below.

1. The javax validation is just used in the RequestFactory servlet. It is
currently not used by anything except the expenses sample. So I could get
away with not including it, and still not causing immediate damage.

That said, we do want to include the dependency in the distribution.
However, gwt-dev seems like the wrong place. It would be nice to include it
in gwt-servlet jar instead. What do you think?

2. I agree there is nothing bikeshed-y about the maven_script.sh. However, I
did not want to clutter of a top-level dir which just has a single script.
Move it to tools/scripts?

On Fri, Jun 25, 2010 at 8:03 AM, fabb...@google.com wrote:

 Sorry to get to this a bit late...

 I've got a couple of concerns:
  1. regarding especially the user.compile use of javax validation,
 does that suggest we should embed that library in gwt-dev.jar
 with all of our other runtime dependencies?  (I don't love that
 practice, but it has been our practice...)
  2. bikeshed/scripts/maven_script.sh seems kind of orphaned if the
 rest goes away.  There's nothing really bikeshed-y about it, is
 there?  Should it move up to a top-level scripts directory?


 On 2010/06/24 20:29:26, amitmanjhi wrote:




 http://gwt-code-reviews.appspot.com/660802/show


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

[gwt-contrib] Re: Wholesale move of the app, requestfactory, and valuestore packages from the bikeshed dir to user... (issue660802)

2010-06-25 Thread Freeland Abbott
On Fri, Jun 25, 2010 at 1:23 PM, Amit Manjhi amitman...@google.com wrote:

 Thanks for the feedback. Since I already submitted the patch, I will create
 follow-up patches for the issues below.

 1. The javax validation is just used in the RequestFactory servlet. It is
 currently not used by anything except the expenses sample. So I could get
 away with not including it, and still not causing immediate damage.

 That said, we do want to include the dependency in the distribution.
 However, gwt-dev seems like the wrong place. It would be nice to include it
 in gwt-servlet jar instead. What do you think?


I would be squirmy about being version-married in gwt-servlet.jar; what if
user's server code also used validation?  (We get away with it in gwt-dev,
at least so far, but it's almost surprising to me that we do.)  I'm not sure
how we want to address that.

Also, I just mailed you privately, but I think SampleDataPopulator is also
broken---it pulls classes off gwt-dev.jar (dev.util.Util), which shouldn't
be on the server classpath.



 2. I agree there is nothing bikeshed-y about the maven_script.sh. However,
 I did not want to clutter of a top-level dir which just has a single script.
 Move it to tools/scripts?


Works for me.

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

[gwt-contrib] Re: Wholesale move of the app, requestfactory, and valuestore packages from the bikeshed dir to user... (issue660802)

2010-06-25 Thread Amit Manjhi
Okay, I created a Jira issue (https://jira.springsource.org/browse/ROO-1032)
for figuring out how to ship the javax.validation jars in the M3 timeframe.

Commited the remaining two fixes you pointed out.

On Fri, Jun 25, 2010 at 10:43 AM, Freeland Abbott fabb...@google.comwrote:

 On Fri, Jun 25, 2010 at 1:23 PM, Amit Manjhi amitman...@google.comwrote:

 Thanks for the feedback. Since I already submitted the patch, I will
 create follow-up patches for the issues below.

 1. The javax validation is just used in the RequestFactory servlet. It is
 currently not used by anything except the expenses sample. So I could get
 away with not including it, and still not causing immediate damage.

 That said, we do want to include the dependency in the distribution.
 However, gwt-dev seems like the wrong place. It would be nice to include it
 in gwt-servlet jar instead. What do you think?


 I would be squirmy about being version-married in gwt-servlet.jar; what if
 user's server code also used validation?  (We get away with it in gwt-dev,
 at least so far, but it's almost surprising to me that we do.)  I'm not sure
 how we want to address that.

 Also, I just mailed you privately, but I think SampleDataPopulator is also
 broken---it pulls classes off gwt-dev.jar (dev.util.Util), which shouldn't
 be on the server classpath.



 2. I agree there is nothing bikeshed-y about the maven_script.sh. However,
 I did not want to clutter of a top-level dir which just has a single script.
 Move it to tools/scripts?


 Works for me.



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