Hi Chris,
some comments ...

>I like the general idea of providing some tooling to help with Pivot
development, but what about writing tools as Pivot apps, so that they could
also be used as reference applications?
Yes, and up to 1.5.2 there was the Tools subproject, but has been dropped
... and to me seem that some (development only) Tools could be useful to
have.

>Pivot is a UI toolkit so I think it would make more sense to use it to
provide non-CLI tools if they make sense.
Yes, but for example even to execute (as Standalone Applications) already
packaged Demos is not so trivial (but this is a problem in general of Java
deploy, not so friendly), so at least my proposal to have a startup script
to set the right environment, and later have even some utility class to give
some hint.
A similar approach (even for a GUI framework) for example is in Griffon
(another great framework), that a look here:
http://griffon.codehaus.org/Quick+Start

Another very useful feature there (and the same in Grails, Play, etc) is to
give some hint to the user, on parameters for commands, similar commands,
etc. but could be too much for us at the moment.

I agree that this is more for developers than final Users, but things can be
generalized (a little, without spending too much time on it) even for them.


>There is a lot of very useful example code that many people might not be
aware of unless they download and use the Pivot source release, or checkout
from SVN.
Yes, I agree.


>A while ago I thought about creating a simple Pivot app to launch demos and
tutorials so that people could
>
>It would just present the available demos/apps in a searchable way and
allow the user to double click to launch as a desktop (or script)
application.  There would be a link to where the source code could be
found (package & class names/svn directory/website url/whatever).  It could
even show the BXML & resource files, and show a screenshot of the running
app/demo.
Good, but without a script to setup classpath etc it would not so simple to
run.
>
>My initial plan was for the app to scan jar files to find classes that
implement Application, and find BXML files whose root node was a Window
subclass that implements Bindable.  But if that information could also be
extracted as part of the build process and bundled with the release (either
in the same jars as the demos/tutorials, or in its own jar/json/xml file).
This is exactly my proposal for some utility class inside the pivot-shell
(new, dedicated and simple jar), scan candidates (and maybe cache them in a
file under the user profile, or already inside the jar, it's a detail).
And maybe start with an overlay zip to unzip over Pivot binary distribution,
so no change in existing things for the moment. Initial plan could be to
start this as project inside one of our Apache-Extras, and see later, ok ?
And put even this in the infinite TODO list :-) .

As always comments, suggestions, etc are welcome ...

Bye,
Sandro


--
View this message in context: 
http://apache-pivot-developers.417237.n3.nabble.com/Pivot-Shell-tp3313582p3316511.html
Sent from the Apache Pivot - Developers mailing list archive at Nabble.com.

Reply via email to