Re: Unicode 9.0 tests

Thanks.

We always have a few test issues in this area, partly because the question of "which Unicode version do we support?" does not have a simple answer: it depends partly on which part of the product we are talking about (the version Saxon uses is not necessarily the same as the ICU library uses), and it also depends on what the Java platform supports: we have one or two tests that pass or fail depending on which version of the JDK we are running on (because of changes in Unicode support in the JDK). And unlike most configuration parameters, we can't choose which version of the JDK to use for each test based on the dependencies in its metadata...

Michael Kay

> On 22 Sep 2016, at 17:02, Abel Braaksma <abel.braaksma@xs4all.nl> wrote:
> 
> I have committed a bunch of Unicode 9.0 tests (there were none). I think it makes sense to retrospectively apply them for 6.0, 7.0 and/or 8.0. Since they are in part generated from the official Unicode docs, I think that can be done without too much pain (we support 9.0, and some characters have been reassigned new categories, hence we fail some of the existing tests of 6.0 or even 3.2).
>  
> This also suggests that the Unicode version is not a forward compatible version. If you support 8.0, you may fail tests in 6.0. Which is odd, but unfortunately the way it is.
>  
> Cheers,
> Abel