Re: [gnugo-devel] Current regression tests

[Gunnar Farneback]

SP Lee wrote:
> Only 89 cases are put in specific categories. The other cases seem to
> involve many areas. I intended to sort the failed cases, but I'm not
> sure if I can really do it.

I'm somewhat doubtful about the usefulness of this. Reasons:
1. Different people would categorize mistakes differently depending on
   their knowledge of various parts of the engine.
2. The categories tend to vary over time. Different versions can make
   completely different mistakes in the same position.

I'm all for adding comments to explain mistakes though. See e.g.
gunnar.tst.

/Gunnar