Small bugfixes

[felixauringer]

We wrote two bugfixes so that the prototype works for our project. We know that our solutions are hacky but we haven't had time to come up with a better solution yet.

• In order to ignore invalid intervals (I described the problem here) and prevent index out of bound exceptions, we only fill the source map until its end.
• Because the root node's parent is nil, the message enclosingRule does not work: Undefined Object does not understand ruleName. We added a nil check to prevent this error from happening.

[coveralls]

Coverage increased (+0.01%) to 85.178% when pulling 8afcaeb on felixauringer:bugfix/interval-and-root-node into 9b85725 on hpi-swa:master.

[codeZeilen]

I will happily merge these as soon as we have the general approach ready! :)

[felixauringer]

I will happily merge these as soon as we have the general approach ready! :)

I think clamping the interval won't be needed anymore because it is just a workaround for interval values that are not expected. And the handling of the root case with its nil-check isn't SWT-compliant yet ;)

[codeZeilen]

hm... true re the clamping.

There is actually a method to cover the logic of the nil-check. I think it is called #isParentNode or something like that.

[felixauringer]

There is actually a method to cover the logic of the nil-check. I think it is called #isParentNode or something like that.

I used #isRootNode, that looks much cleaner.

I think clamping the interval won't be needed anymore because it is just a workaround for interval values that are not expected. And the handling of the root case with its nil-check isn't SWT-compliant yet ;)

Because there are still nodes with an interval outside of the source map's size, the logic is still needed. You actually implemented the same logic in the method OhmSourceMapping>>defaultExpression:.

We think that this could be merged. What do you think?

[felixauringer]

I'll close this and delete my branch as the interval clamping isn't sufficient and the handling of the root node in enclosingRule is on your whitespace-handling-branch