> There are no docs on how to do that, because it's not possible with
> run.py. It is strictly designed to make it as easy as possible for our
> regression testers. And hence makes lots of assumptions and assertions.
> I have a recollection that one of the *.cpp status programs is
> maintained to produce library specific status. Not sure which one it is
> though.
>

I believe that is library_status :-)

The problem with compiler_status was that the xml results from some tests were split into 2 xml files.  I just noticed that the xml results have recently changed back to a single xml file although the tests are still reported as failing by compiler_status.  It must be a different problem now.  This is what I meant when I mentioned that it was time consuming maintaining compiler_status, and why we are investigating using the xml/xslt as an alternative.

Chris Cambly
XL Compiler Development