Hi,

> I've modified my SCI unit tests to run totally unattended, outputting
> results to a text file. This will allow us to run the unit tests against
> several different interpreters in an automated fashion, and then diff the
> results very easily!

Great, this should be extremely helpful :-)

> I expanded the tests a bit as well to include GetAngle(), GetDistance(),
> and OnControl(). I'll expand it a bit more tomorrow as well and update it
> again.

Thanks-- could you also include results for an earlier interpreter, too,
please?

[...]
> There are some differences in GetAngle and OnControl between FSCI and SSCI
> as well as the priority band differences mentioned before.

OnControl: I haven't had a look at your source yet; what do you do
before checking the control map with this function?

GetAngle: Our current approach to this function is to return the result
it should return, from a mathematical point of view. Sierra uses LUTs to
approximate these results-- trying to emulate their results would only
make sense if we did the same with the various TimesSin, TimesCos etc.
functions, too [if there's any problem with these functions, it's
likely that their inaccuracies are cancelling each other out]. It'd also
require us to build tables for this, which might be more work than it's
worth.
While we may need this for full bug-for-bug compatibility, there's a
good chance that we might end up not having to implement it, so I'll let
this one slip for a while (although I'm not going to stop anyone from
implementing it the Sierra way...)


llap,
 Christoph

Reply via email to