[MLton] MinGW signals

[Matthew Fluet]

> The signal numbers in runtime/platform/mingw.h seem a bit odd:
 >
> So, I wonder about the origin of the signal numbers in mingw.h?  Is there
> some specific reason to use those particular signal numbers?

svn blame runtime/platform/mingw.h  reveals that the signal numbers were 
added at revision 3386 on Aug. 23, 2004.  The log for revision 3386 
indicates that Stephen applied a patch by Brent Fulgham to port MLton to 
MinGW.  That patch was announced on the MLton mailing list on Aug. 4, 2004:
   http://mlton.org/pipermail/mlton/2004-August/025740.html
The .diff file still appears at the URL given in the mail, and it shows 
that the signal numbers in mingw.h have been there since that time.

However, that doesn't explain why it maps signal numbers in the 
particular way it does.