Bug: For some examples, model.Run() returns true when false is expected about wavefunctioncollapse HOT 7 CLOSED

I think this is the intended behavior with the limit property : it actually limits the number of iterations the generation is allowed to do. If the number of iterations reaches this limit without hitting a contradiction condition, the generation is considered a success.

from wavefunctioncollapse.

I included samples with the limit param in samples.xml to show how to generate partially completed results. Because they can be pretty, like the "Circles" one linked from README.

from wavefunctioncollapse.

The same bug occurs with samples.xml

<samples>
    <overlapping name="Village" N="3" symmetry="2" limit="50" periodic="True"/>
</samples>

And seed
int seed = 362313200;

This means that the bug occurs with both the overlapping model and the simple tiled model.

from wavefunctioncollapse.

As I read the source of Run(), you're clearly correct.

But that doesn't seem intuitive to me -- even when providing an iteration limit, the main thing I want to know is whether or not the model now has a usable result. How would I know whether the result was complete or not? So IMO the bug still stands: I expect a return of true when the generation has completed successfully, and false otherwise.

This seems like a case of trying to cram a piece of information with three states (Complete, incomplete, impossible) into a variable with two states.

from wavefunctioncollapse.

I see that the samples.xml in the repo actually have those limit set. I wrongly assumed you added them. The limit properties should probably be removed from the file as they seems to be debug helpers more than anything.

from wavefunctioncollapse.

I edited the issue to indicate that that entry came from the existing samples.xml

from wavefunctioncollapse.

@ejmahler @kchapelier Note sure if people get notifications if I don't mention their username. It seems that everybody understood each other here, closing.

from wavefunctioncollapse.