On Tue, Feb 09, 2021 at 01:54:33PM +0100, Levente Uzonyi wrote:
> Hi David,
> 
> On Sun, 7 Feb 2021, David T. Lewis wrote:
> 
> >This looks good to me, and it fixes the failing test.
> >
> >Note that none of the historical versions of 
> >SharedQueue>>nextOrNilSuchThat:
> >pass the test, but the issue became visible after CollectionsTests-eem.342
> >last September when the existing test for SharedQueue2 was applied to
> >SharedQueue as well.
> >
> >It would be good if Levente or Nicolas can review this, but +1 from me.
> 
> The fix looks right in the sense that it does what the test case requires.
> But it may affect the way events are processed.
> The only sender of this method is EventSensor which may rely on the 
> current behavior.
> So, I would like someone who understands how EventSensor works confirm 
> that the change is okay before pushing it to the Trunk.
> 

Agreed, although I would say that if no further comments show up after
a couple of days, we should move it to trunk anyway.

The reason I say this is that the only sender is part of the Etoys package,
that that sender is part of SugarLauncher. That means that nobody is
actually using it in Squeak today. The unit test shows an actual error
condition (not a failed test expectation).

Earlier method versions show a test failure (not error) so indeed there
has been some change to the behavior over time. So indeed additional
review would be welcome.

Let's wait a few days and if no objections move it to trunk.

Dave