Re: [e4-dev] Dynamic Eclipse 4 with Clojure language

[Laurent PETIT <laurent.petit@xxxxxxxxx>]

Hey Paul,

2013/12/2 Paul Webster <pwebster@xxxxxxxxxxxxxxxxxxx>:
> Hi Laurent,
>
> On Mon, Dec 2, 2013 at 8:04 AM, Laurent PETIT <laurent.petit@xxxxxxxxx>
> wrote:
>>
>> I have considered it, not tried yet.
>>
>> So far, I have spotted in past discussions on the subject the
>> following possibilities:
>>
>> 1/ EventBroker + APP_STARTUP_COMPLETE event => when to register to the
>> EventBroker in the context of Eclipse the IDE? By doing it in my
>> current Model Processor hook ?
>
>
> Yes, we often use model processors to register for events (especially
> related to model changes) for the application going forward.  This is the
> Eclipse4 compatible approach.

So I'll use this method, even if, as you pointed out, I will continue
to live in a cross world between 3.x and 4.x APIs, since I won't be
able to migrate all my code at once (and also I depend on so many
things that are still stuck in the 3.x world, I guess ...)

>
>>
>> 2/ Extend org.eclipse.ui.startup ? I have read that I am guaranteed
>> that when my earlyStartup() method is called the Workbench is already
>> totaly operational. But then how to get the MApplication instance?
>> Currently I'm doing
>>
>> PlatformUI.getWorkbench().getService(IEclipseContext.class).get(MApplication.class).
>
>
> If you use org.eclipse.ui.startup, you've picked the right pattern.  This is
> based on 3.x (o.e.ui.startup is not part of Eclipse4) but if you need access
> to things that need the Workbench, then you are tied to 3.x anyway.
>
>>
>> 3/ Just use asyncExec on the display. But how to get the display instance?
>
>
> If you are running in the IDE (which has the Workbench) this is the hardest
> thing to get correct.  Calling Display.getDefault() can instantiate a
> display on the incorrect thread and cause problems.  We use
> org.eclipse.ui.IWorkbench.getDisplay() but that's then tied to 1) running
> against the 3.x Workbench and 2) waiting until the workbench is up, which
> happens after the model processor stuff is called.
>
>
> Later,
> Paul
>
>
>
> --
> Paul Webster
> Hi floor.  Make me a sammich! - GIR
>
> _______________________________________________
> e4-dev mailing list
> e4-dev@xxxxxxxxxxx
> https://dev.eclipse.org/mailman/listinfo/e4-dev
>