Boost logo

Boost :

From: Mark Rodgers (mark.rodgers_at_[hidden])
Date: 2000-06-22 05:53:25

From: Beman Dawes
> Additionally, the latest draft of the Compiler Status page is there at
> There are currently no
> to that page from the rest of the site to avoid misleading visitors; the
> test results are still highly suspect.

I had a look at the Borland failures.


   The problem is in the test program rather than the library.
   (Well actually the problem is in the compiler, but...)
   The workaround is simple - change runtest from

      void runtest(const std::string & type)


      void runtest(const char *type)


   I believe they should work with the fixes I uploaded previously
   and which Jeremy incorporated into in the vault.


   The problems are a bit deeper. If anyone is interested I've put
   a rational.hpp with workarounds in the vault in

   I don't know if this will break any other compilers. I did test
   it with MSVC 6 and it was still OK, and that was the only one
   on Beman's table that had worked previously. Who knows - it might
   now work with more.


Boost list run by bdawes at, gregod at, cpdaniel at, john at