by **Sestir** on Sat Feb 11, 2012 5:38 am

Indeed, this needs to be fixed!

What happens seems to be:

There is an isomorphism that works just fine, with a green bow stretching from the final node of a correct branch (A) to a node midway in another correct variation (B). Now some user tries the JavaScript solver which does not support isomorphism yet. He finds the correct solution (A) but gets a "Wrong" from the solver. He adds one more move to the tree and makes a comment saying "This is correct too, isn't it?" By doing so, he destroys the isomorphism by adding a continuation from the branch which should not have a continuation. Suddenly, the problem is broken for all solvers.

Also, restoring the isomorphism in the editor does not work. It appears to work, but it still does not work after you have saved it. One way to handle this is to make the isomorphism code ignore nodes that are shown with dotted lines. (= = = = =)

Another scary thing about this, that at least I perceive as a bug, is the fact that when you try to look at an old version of the problem, in the edit page, it displays recent user comments within the problem such that they are still abel to break it, even though they were not made until after the next edit.