Mikko Ahonen wrote:

Hello,

Small issue I seem to constantly run into while debugging Cells source code:

First, I sometimes get "New as of Cells3: parent must be supplied to make-instance of .." for various reasons.

Hmmm. That is just the family class, yes? If you are not specifying the :fm-parent, yep, that is what happens. Aside: I could probably relax that requirement, I think it just had to do with "kid-slotting", used only in GUI containers. Anyway..

For example now I used c? for cell that did not refer to any other cell. Anyway, this is expected.

Not by me. :) A Cell not using another Cell is fine. I often use a rule just to achieve an initform that can see the instance being initialized. And I don't see how this ties into :fm-parent not being specified.

Probably just a failure of my imagination. Want to post some code?

But even if I run (cells-reset) and then try running it again, I get this kind of error messages below. It is slightly annoying to have to start the environment when this happens.

Current DP 1 not GE pulse 85 of cell (Foo . <vld>)<85:A VISIBLE-NAME/#<ROLE 22356BC7> = Foo>

Is there some simple workaround?

I have to think that although you are doing a cells-reset you are also then continuing work with existing model instances who still remember the last DP. If not, I am stumped and would need to see some code. If so, the simple workaround would be to set up your tests so a new test run starts with cells-reset and then recreates the entire universe of model instances you want to work with.

Other than that, you might be able to hack Cells quickly to respond to a /lower/ DP by simply treating that condition as "out of date" so the recalculation mechanism cuts in, at the end of which the new lower DP would replace the obsolete DP. I doubt we could claim integrity was being preserved because now rules and observers are running for no reason, but I wager there would rarely be an actuall ill effect.

cheers, kenneth \