> I'm afraid, only by examining and hacking source of compiler_status program.
> It is not presently used or maintained by anybody.

We were using it and noticed the change in reporting sometime back in November (between 13th-14th).  We found that maintaining compiler_status for our own internal use is somewhat time consuming since relatively few people use it.  As a result, we started looking at using xml/xsl processing to do a developer summary for internal use.  

Internally, we use these generated results to keep track of Boost for our pre-release compilers for which we cannot publish the results to an external website.  The xml/xsl processing is well maintained.  The pages we generate are similar to the regression testing results that are reported nightly on the boost website but are for our own internal consumption.  

We have had some modest success with this approach.  Although, a few of the nightly regression test processing scripts had to be "modified" in order to achieve the results generation.  I'm not sure if anyone else interested in this as an alternative to compiler_status.
   
Chris Cambly
XL C++ Compiler Development