Re: Simple question about GWT.create() signature

Thank you for the reply
I did not know much about GWT-RPC, bay be I should study more before
making questions...
I was aware of the generic call, but looks to me like a cast :)

I was NOT aware of the feature called 'type inference'. I used my IDE
to complete, and noticed that only Object methods were allowed chained
after a GWT.create(), so I thought that I needed a cast to assign to a
variable. Now I readed the java tutorial on Generic Methods and
everything looks much more consistent... I knew about generic methods,
but missing the type inference I always wondered what were they useful
to... :)

On 28 Nov, 10:56, Thomas Broyer <t.bro...@gmail.com> wrote:
> On 26 nov, 21:12, "Alessandro Carraro (JUG Padova)"
>
> <carraro.alessan...@gmail.com> wrote:
> > A simple qestion, sorry if it is a FAQ:
>
> > Why
> >   public static <T> T create(Class<?> classLiteral)
> > and not
> >   public static <T> T create(Class<T> classLiteral)
>
> With GWT-RPC you pass the "synchronous" (extends RemoteService)
> interface as an argument and it returns an implementation of the
> "asynchronous" interface.
> That wouldn't be possible with the Class<T> argument.
>
> > the second one would save me from a lot of unnecessary casts (IMHO).
>
> I hardly ever needed a cast:
>  - either the GWT.create() result is assigned to a variable, and javac/
> GWT compiler is smart enough to infer the generic type argument
>  - or if it's not possible, you can use a "generic call" instead of a
> cast: GWT.<MyClass>create(...) but I admit it adds the same
> boilerplate as a cast...
>
> > What's worse, I tried to write the helper function:
>
> >         @SuppressWarnings("unchecked")
> >         public static <T> T create(Class<T> classLiteral) {
> >                 return (T) GWT.create(classLiteral);
> >         }
>
> > But after a month I discovered that the compiler does not like it at
> > all (it complains that GWT.create must be called only using class
> > literals... could be nice if it could find that MY helper is called
> > with only using  literals...
>
> It's been requested several types in the past (including by people
> that now work at Google on GWT) but I guess it's either low priority
> or it would require too much re-work of the compiler's internals.
>
> > Is it possible to get a hook into the compiler phase to replace all
> > invocations of my helper 'inlining' it?
>
> I don't think so; but is it really worth it? risking bugs in your
> "hook" just to save you from typing a few characters (that, moreover,
> any good IDE is capable of generating for you)

--
You received this message because you are subscribed to the Google Groups "Google Web Toolkit" group.
To post to this group, send email to google-web-toolkit@googlegroups.com.
To unsubscribe from this group, send email to google-web-toolkit+unsubscribe@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/google-web-toolkit?hl=en.