RFR: JMC-6555 Convert JOverflow plugin to SWT

[Arvin Kangcheng Xu]

On Tue, 27 Aug 2019 at 07:53, Mario Torre <neugens at redhat.com> wrote:
>
> On 26/08/2019 17:16, Arvin Kangcheng Xu wrote:
> > I'm unsure if those calls are queued or executed right away. I
> > remembered introducing this guard after I observed multiple updates
> > caused by single user interaction. However, I took a further look at
> > the code and experiment a bit. The guard is indeed unnecessary, at
> > least for the current code.
>
> What I mean is that the event thread should be the one executeing this
> method exclusively, so it cannot be that this code is accessed again
> before the update is finished, it's only one thread.
>
> If you see this call executed multiple times it means it's being
> accessed concurrently. I do not know enough of the Eclipse/SWT internal,
> so I need to defer to Marcus or someone else for this review, however it
> does seem to me that something is calling this method concurrently after
> all, we should see what code does it and possibly fix it, because I
> doubt this is correct.

There is no new thread created once the heap model is loaded, so it cannot
be accessed concurrently. The reason it was called multiple times was that I
had called setSelection on a viewer as part of the updateModel call hierarchy,
which triggers the callback where it handles user clicking and eventually calls
updateModel again. Calls to setSelection were removed later, but not this
guard.

At least that's what I thought what happened. There is a high
likelihood that I'm
wrong. I do agree the code needs extra reviews.

> Cheers,
> Mario
>
> --
> Mario Torre
> Associate Manager, Software Engineering
> Red Hat GmbH <https://www.redhat.com>
> 9704 A60C B4BE A8B8 0F30  9205 5D7E 4952 3F65 7898